1956-1957 Year Book - Carnegie Institution of Washington

Digitized by the Internet Archive

in 2012 with funding from

LYRASIS Members and Sloan Foundation

http://www.archive.org/details/yearbookcarne56195657carn

YEAR BOOK 56

July 1, 1956— June 30, 1957

r

<*/Ls

p*

t

C

h<

> S

o/yi

4<^

CARNEGIE INSTITUTION OF WASHINGTON

WASHINGTON, D. C.

1957

Library of Congress Catalog Card No. 3-16716
THE LORD BALTIMORE PRESS, INC., BALTIMORE, MARYLAND

CONTENTS

page

OFFICERS AND STAFF v

REPORT OF THE PRESIDENT 1

REPORTS OF DEPARTMENTS AND SPECIAL STUDIES 35

Mount Wilson and Palomar Observatories 37

Committee on Image Tubes for Telescopes 77

Department of Terrestrial Magnetism 81

Geophysical Laboratory 149

Department of Plant Biology 253

Department of Embryology 297

Department of Genetics 357

Department of Archaeology 405

BIBLIOGRAPHY 415

INDEX 417

SUPPLEMENT xiii

Report of the Executive Committee xv

Report of Auditors xvii
Abstract of Minutes of the Fifty-Ninth Meeting of the Board of Trustees xxxiii

Articles of Incorporation xxxv

By-Laws of the Institution xxxix

in

PRESIDENT and TRUSTEES

PRESIDENT
Caryl P. Haskins

BOARD OF TRUSTEES
Walter S. Giflford, Chairman

Barklie McKee Henry, Vice-Chairman

Robert Woods Bliss, Secretary

James F. Bell
Robert Woods Bliss
Lindsay Bradford
Omar N. Bradley
Walter S. Giflord
Crawford H. Greenewalt
Caryl P. Haskins
Barklie McKee Henry
Ernest O. Lawrence
Alfred L. Loomis
Robert A. Lovett
Keith S. McHugh
Margaret Carnegie Miller
Henry S. Morgan
Seeley G. Mudd
William I. Myers
Henning W. Prentis, Jr.
Elihu Root, Jr.
Henry R. Shepley
Charles P. Taft
Juan T. Trippe
James N. White
Robert E. Wilson

TRUSTEES Continued

EXECUTIVE COMMITTEE

Barklie McKee Henry, Chairman
Robert Woods Bliss
Lindsay Bradford
Walter S. GifTord
Caryl P. Haskins

Robert A. Lovett
Henry S. Morgan
Henning W. Prentis, Jr.
Henry R. Shepley

FINANCE COMMITTEE

Lindsay Bradford, Chairman
Walter S. Gififord
Alfred L. Loomis
Henry S. Morgan
Henning W. Prentis, Jr.
James N. White

NOMINATING COMMITTEE

Elihu Root, Jr., Chairman
Walter S. Gifford
Crawford H. Greenewalt
William I. Myers

AUDITING COMMITTEE

Keith S. McHugh, Chairman
Alfred L. Loomis
Juan T. Trippe

RETIREMENT COMMITTEE

Lindsay Bradford, Chairman
Barklie McKee Henry
Henry S. Morgan

COMMITTEE ON
ASTRONOMY

Seeley G. Mudd, Chairman
Crawford H. Greenewalt
Elihu Root, Jr.

COMMITTEE ON
BIOLOGICAL SCIENCES

Alfred L. Loomis, Chairman
Margaret Carnegie Miller
William I. Myers
Charles P. Taft

COMMITTEE ON
TERRESTRIAL SCIENCES

Ernest O. Lawrence, Chairman
Barklie McKee Henry
Henning W. Prentis, Jr.
Robert E. Wilson

COMMITTEE ON
ARCHAEOLOGY

Henry R. Shepley, Chairman
James F. Bell
Robert Woods Bliss
Juan T. Trippe

VI

FORMER PRESIDENTS and TRUSTEES

PRESIDENTS

Daniel Coit Gilman, 1902-1904 Robert Simpson Woodward, 1904-1920

John Campbell Merriam, President 1921-1938; President Emeritus 1939-1945
Vannevar Bush, 1939-1955

TRUSTEES

1904-05

Wayne MacVeagh

1902-07

1925-27

Andrew W. Mellon

1924-37

1934-46

Roswell Miller

1933-55

1902-13

Darius O. Mills

1902-09

1910-29

S. Weir Mitchell

1902-14

1903-14

Andrew J. Montague

1907-35

1929-38

William W. Morrow

1902-29

1916-32

William Church Osborn

1927-34

1925-34

James Parmelee

1917-31

1927-49

Wm. Barclay Parsons

1907-32

1903-23

Stewart Paton

1916-42

1902-03

George W. Pepper

1914-19

1914-24

John J. Pershing

1930-43

1927-52

Henry S. Pritchett

1906-36

1910-14

Gordon S. Rentschler

1946-48

1920-55

David Rockefeller

1952-56

1948-49

Elihu Root

1902-37

1902-15

Julius Rosenwald

1929-31

1902-12

Martin A. Ryerson

1908-28

1924-34

Theobald Smith

1914-34

1924-35

John C. Spooner

1902-07

1902-08

William Benson Storey

1924-39

1902-05

Richard P. Strong

1934-48

1915-29

William H. Taft

1906-15

1902-03

William S. Thayer

1929-32

1902-19

James W. Wadsworth

1932-52

1902-09

Charles D. Walcott

1902-27

1902-02

Frederic C. Walcott

1931-48

1920-49

Henry P. Walcott

1910-24

1903-09

Lewis H. Weed

1935-52

1902-04

William H. Welch

1906-34

1938-44

Andrew D. White

1902-03

1933-49

Edward D. White

1902-03

1904-06

Henry White

1913-27

1934-39

George W. Wickersham

1909-36

1902-09

Robert S. Woodward

1905-24

1914-24

Carroll D. Wright

1902-08

1902-16

Alexander Agassiz
George J. Baldwin
Thomas Barbour
John S. Billings
Robert S. Brookings
John L. Cadwalader
William W. Campbell
John J. Carty
Whitefoord R. Cole
Frederic A. Delano
Cleveland H. Dodge
William E. Dodge
Charles P. Fenner
Homer L. Ferguson
Simon Flexner
W. Cameron Forbes
James Forrestal
William N. Frew
Lyman J. Gage
Cass Gilbert
Frederick H. Gillett
Daniel C. Gilman
John Hay
Myron T. Herrick
Abram S. Hewitt
Henry L. Higginson
Ethan A. Hitchcock
Henry Hitchcock
Herbert Hoover
William Wirt Howe
Charles L. Hutchinson
Walter A. Jessup
Frank B. Jewett
Samuel P. Langley
Charles A. Lindbergh
William Lindsay
Henry Cabot Lodge
Seth Low

Under the original charter, from the date of organization until April 28, 1904, the following were
ex officio members of the Board of Trustees: the President of the United States, the President of the Senate,
the Speaker of the House of Representatives, the Secretary of the Smithsonian Institution, and the President
of the National Academy of Sciences.

vn

STAFF

ASTRONOMY

MOUNT WILSON AND PALOMAR OBSERVATORIES

813 Santa Barbara Street, Pasadena 4, California

Mount Wilson Observatory organized in 1904; George E. Hale, Director 1904-1923, Honorary Director
1923-1936; Walter S. Adams, Director 1924-1945. Unified operation with the Palomar Observatory of
the California Institute of Technology began in 1948.

Ira S. Bowen, Director; Horace W. Babcock, Assistant Director

Halton C. Arp
Walter Baade
William A. Baum
Arthur D. Code
Armin J. Deutsch

Jesse L. Greenstein
Milton L. Humason *
Rudolph L. Minkowski
Guido Munch
Seth B. Nicholson *

Donald E. Osterbrock
Robert S. Richardson
Allan R. Sandage
Olin C. Wilson
Fritz Zwicky

TERRESTRIAL SCIENCES

GEOPHYSICAL LABORATORY

2801 Upton Street, N. W., Washington 8, D. C.

Organized in 1906, opened in 1907; Arthur L. Day, Director 1909-1936; Leason H. Adams, Acting
Director 1936-1937, Director 1938-1952; George W. Morey, Acting Director 1952-1953.

Francis R. Boyd, Jr.
Felix Chayes
Sydney P. Clark, Jr.
Gordon L. Davis
Gabrielle Donnay
Joseph L. England
Hans P. Eugster

Philip H. Abelson, Director

Joseph W. Greig
Gunnar Kullerud
George W. Morey *
J. Frank Schairer
George R. Tilton
Hatten S. Yoder, Jr.

Staff Associate
Gordon J. F. MacDonald

Visiting Investigators
Henry Faul
David B. Stewart

DEPARTMENT OF TERRESTRIAL MAGNETISM

5241 Broad Branch Road, N. W., Washington 15, D. C.

Organized in 1904; Louis A. Bauer, Director 1904-1929; John A. Fleming, Acting Director 1929-1934,
Director 1935-1946.

L. Thomas Aldrich
Ellis T. Bolton
Roy J. Britten
Bernard F. Burke
Dean B. Cowie
John W. Firor

G. N. Cohen
E. H. Creaser ||
W. C. Erickson
K. L. Franklin I

♦Retired June 30, 1957.
% Resigned in 1956.
§ On leave of absence.

Merle A. Tuve, Director

Scott E. Forbush
John W. Graham
Norman P. Heydenburg
Ellis A. Johnson t
Richard B. Roberts

Visiting Investigators

H. Lawrence Heifer

J. J. Leahy

F. T. McClure**

Howard E. Tatel
Georges M. Temmer §
Ernest H. Vestinetl
Harry W. Wells
George W. Wetherill

G. F. Pieper
Irena Z. Roberts
H. Weaver

|| Term of appointment completed in 1956.

j[ Resigned in 1957.

** Term of appointment completed in 1957.

Vlll

STAFF Continued

BIOLOGICAL SCIENCES

DEPARTMENT OF PLANT BIOLOGY

Stanford, California

Desert Laboratory, opened in 1903, became headquarters of Department of Botanical Research in 1905;
name changed to Laboratory for Plant Physiology in 1923; Daniel T. MacDougal, Director 1906-1927.
Reorganized in 1928 as Division of Plant Biology, including Ecology; Herman A. Spoehr, Chairman 1927-
1930 and 1931-1947, Chairman Emeritus 1947-1950. Name changed to Department of Plant Biology
in 1951.

William M. Hiesey
Donald W. Kupke *
Harold W. Milner
Malcolm A. Nobs
James H. C. Smith

C. Stacy French, Director

Visiting Investigators
Per Halldal
Wolf Vishniac

Investigator Engaged in
Post-Retirement Studies
Jens C. Clausen

Research Fellows
F. J. F. Fisher
Paul H. Latimer
Kazuo Shibata

DEPARTMENT OF EMBRYOLOGY

Wolfe and Madison Streets, Baltimore 5, Maryland

Organized in 1914; Franklin P. Mall, Director 1914-1917; George L. Streeter, Director 1918-1940;
George W. Corner, Director 1941-1955.

David W. Bishop
Bent G. Boving
Robert K. Burns
Robert L. DeHaan
Elizabeth M. Ramsey
Royal F. Ruth

James D. Ebert, Director

Consultant
George W. Bartelmez

Special Investigators
Vincent J. De Feo
Seymour Katsh
Malcolm S. Steinberg

Research Associates
Arthur T. Hertig
Chester H. Heuser
Samuel R. M. Reynolds

DEPARTMENT OF GENETICS

Cold Spring Harbor, Long Island, New Yor\

Station for Experimental Evolution opened in 1904; name changed to Department of Experimental Evo-
lution in 1906; combined with Eugenics Record Office in 1921 to form Department of Genetics. Charles B.
Davenport, Director 1904-1934; Albert F. Blakeslee, Director 1935-1941.

Alfred D. Hershey
Berwind P. Kaufmann
Barbara McClintock
Margaret R. McDonald
George Streisinger

Milislav Demerec, Director

Special Investigators
Elizabeth Burgi
Helen Gay
Sheila Howarth X
Etta Kafer §
Andrej W. Kozinski

Ernest L. Lahr
Joseph D. Mandell
Atif Sengiin %
Jun-ichi Tomizawa
Sibergina Wagenaar t

* Resigned September 21, 1956.

t Term of appointment completed during the report year.

§ Resigned during the report year.

IX

STAFF Continued

ARCHAEOLOGY

DEPARTMENT OF ARCHAEOLOGY

10 Frisbie Place, Cambridge 38, Massachusetts

Department of Historical Research organized in 1903; Andrew C. McLaughlin, Director 1903-1905;
J. Franklin Jameson, Director 1905-1928. In 1930 this Department was incorporated as a section of
United States history in a new Division of Historical Research; Alfred V. Kidder, Chairman 1930-1950.
Name changed to Department of Archaeology in 1951.

Tatiana ProskouriakofF
Karl Ruppert *
Anna O. Shepard

Harry E. D. Pollock, Director

Edwin M. Shook t
A. Ledyard Smith
Robert E. Smith

Gustav Stromsvik §
}. Eric S. Thompson

RESEARCH ASSOCIATES

of Carnegie Institution of Washington

William A. Arnold, Oak Ridge National Laboratory

Louis B. Flexner, University of Pennsylvania

Willard F. Libby, University of Chicago

Paul W. Merrill, Mount Wilson Observatory

John von Neumann, || Institute for Advanced Study

Hans Ramberg, University of Chicago

C. E. Tilley, Cambridge University

Evelyn M. Witkin, State University of New York

♦Retired in 1956.

X On leave of absence.

§ Retired in 1957.

|| Died February 8, 1957.

STAFF C ontinued

OFFICE OF ADMINISTRATION

Caryl P. Haskins
President

Paul A. Scherer

Executive Officer

Samuel Callaway

Assistant to the President

Ailene J. Bauer

Director of Publications

Dorothy R. Swift *
Editor

Lucile B. Stryker

Associate Editor

Earle B. Biesecker

Bursar; Secretary-Treasurer, Retirement Trust

James W. Boise

Assistant Bursar; Assistant Treasurer, Retirement Trust

James F. Sullivan

Assistant to the Bursar

Richard F. F. Nichols

Executive Secretary to the Finance Committee

Retired June 30, 1957.

XI

CARNEGIE INSTITUTION OF WASHINGTON

REPORT of

THE PRESIDENT

CARNEGIE INSTITUTION OF WASHINGTON

REPORT OF THE PRESIDENT

It does not matter what a man does; so long as he does it with the attention which
affection engenders, he will come to see his way to something else. After long waiting
he will certainly find one door open, and go through it. He will say to himself that he
can never find another. He has found this, more by luck than cunning, but now he is
done. Yet by and by he will see that there is one more small, unimportant door which
he had overlooked, and he proceeds through this too. . . . Then after years — but probably
not till after a great many — doors will open up all round, so many and so wide that the
difficulty will not be to find a door, but rather to obtain the means of even hurriedly sur-
veying a portion of those that stand invitingly open. — Samuel Butler in
Alps and Sanctuaries of Piedmont and the Canton Ticino

What is a Golden Age ? What echoes of the Age of Pericles, of Renaissance
Italy and the Low Countries and Scandinavia, of Elizabethan England, mark
each as a flood tide in the vast, slow surge of human intellectual development ?
Will such flood tides come again ?

It is interesting to notice, as James Joll has recently done, some of the char-
acteristics that these ages had in common. All of them were times of fervent
intellectual excitement, when major new creations and new experiences and
viewpoints were just coming to wide notice and were on the threshold of gen-
eral acceptance. In all of them one can sense a vigorous address to new ideas —
when indeed opening vistas, half-seen, made of ideas precious coin. All of
them were eras of some physical security and at least some political and or-
ganizational stability. But in all of them, too, stability and security were far
from complete, and there is the flavor of a partnership of disorder and hazard
with vitality and creativeness. None of them, clearly, were especially "com-
fortable" times in which to live, in the sense that static and secure environments
may be comfortable. And yet, as Joll has significantly pointed out, men knew
that they were living in great times. The adventurous in all these periods
would probably have admitted — perhaps bitterly resented — the danger and the
insecurity and the muddled opacity of their days. But if hard pressed probably
no one of them would have admitted a wish to be born in any other era.

Will such times come again? It is hard to imagine that they will not.
Indeed, though we hear our own age criticized as static and as anti-intellectual
often enough, perhaps we ourselves are the restless, insecure, anxious, vital
participants in an era of contemporary intellectual development that other men
sometime, somewhere, may well look back upon as golden too.

If we are in fact witnessing the earlier phases of another era of turbulent
change, when viewpoints shift rapidly and radically, serving as the anvils for
new ideas, we must expect it to differ in many respects from similar periods
in the past. One striking difference will be that we cannot hope to localize it
geographically. The interlocked character of the present world, the growing

3

CARNEGIE INSTITUTION OF WASHINGTON

similarity of all its cultures, the universality of its communication, must make
meaningless any such designation as an Athenian or an Elizabethan age. But
possibly we can characterize it in terms of subject matter, of the loci of ideas
with which it is especially concerned. Prominent among such domains, clearly,
will be the natural sciences.

Such a situation is not new, of course, for ideas in these fields have figured
in the conceptual revolutions of all the Golden Ages. Aristotle and Plato and
Socrates all lived in or close to the times of Periclean Athens; Galileo and
Copernicus, Da Vinci and Vesalius were of Renaissance Italy, Francis Bacon
and William Harvey were of Elizabethan England. But, as any new Golden
Age will be impossible to localize geographically, so will its contributions of
scientific ideas be derived over a wide and sometimes rather inchoate intellec-
tual front. We can already see vivid examples of this development. And if we
compare the current product of the natural sciences over the world for any
single year, not only in volume and diversity of source but in scope of conse-
quences, with the whole product of a Periclean Age, we are all but forced to
conclude that, half-unknowing, half-unrealizing, we are living in proximity
to one of the most astounding Golden Ages of all time.

Surely our age shares many characteristics with the earlier golden times.
There is the relative physical safety and comparative political stability over
much of the face of the globe. There is the wide feeling of insecurity, the
deep-lying anxiety, the sense of confusion, not unlike the earlier times in its
general character even though, to us at least, its causes seem far more complex,
more massive, more intractable. But there is likewise the same intense concern
with new ideas and new concepts, the same eagerness for widened vistas of
understanding. And there is another and an important characteristic of such
times in which our age also seems typical.

The classical Golden Ages were intensely concerned with the problem of
communicating the new ideas that were being born in such profusion. In all
of them there was a preoccupation with the problems of education. All of
them were times for the establishment of special schools of thought and of
great centers of learning, from the Peripatetics to the College of Merton to
Padua to Paris. In this characteristic, too, our age resembles the earlier ones,
even if groping, as yet, toward developments of educational concepts com-
parable to theirs.

In the field of communication in its most general sense, however, our age
confronts a challenge of almost new dimensions, perhaps nowhere more
poignant than in the natural sciences.

Diversity of approach is the very lifeblood of the scientific effort. Science
enlists men of the most unlike temperaments and talents. It unites workers
whose gifts are primarily descriptive with workers whose understanding and

REPORT OF THE PRESIDENT

approaches comprehend symbolism and techniques of the most abstruse and
involved character. Bonded in a common effort are men whose talents are
primarily synthetic with men so keenly and entirely analytical that synthesis
may have little meaning for them. United are investigators of deeply theoretical
bent with investigators of primarily mechanical skills. And since in every in-
vestigation the observer and his "real" world are in some sense in equilibrium,
scientists with divergent gifts and interests, even when concerned with the
same problem, necessarily labor in partly different universes.

It is not only the observers that differ widely in their characteristics, under
the common rubric of scientists. The subject matter diverges even more. Scien-
tific disciplines vary enormously among themselves in their degree of sophisti-
cation and in their intricacy. The attitudes, the modes, the ways of "picking
up the stick," to use Butterfield's expressive phrase, even the underlying atti-
tudes and aspirations of the work, may be almost unrecognizably different in a
mature, well cultivated, highly differentiated discipline on the one hand and in
an exploratory one, still in its primarily descriptive phase, on the other. And
though the newer disciplines must always be in some measure rooted in the
old, and though it is probable that the tested approaches of older fields always
have some relevance for the newer ones, the transfer is far from literal. To
accomplish it successfully requires talent and sophistication in the investigator,
and, above all, that wisdom and sense of proportion that can come only from
broad experience and a flexible viewpoint.

These profound diversities among investigators and within the structure of
science are characteristic and immeasurably precious. But they also harbor all
the dangers of fragmentation and pose the most serious challenges to com-
munication within the very core of the scientific effort. The compartmenting
of subject matter is a constant threat to the unity of science, and many factors
promote it. Mere growth of vocabulary and specialization of terminology in
a given field — to the point where its jargon becomes unintelligible not only to
the layman but even to an investigator working in a nearly adjacent area —
raise practical barriers to understanding, barriers that may be formidable.

But there is a more serious aspect to such failures of communication. Words
are basically the coin of ideas, and to some degree their generators — never
entirely their consequences. So it is not uncommon to find that not only the
words but also some basic concepts governing workers in one field may be
unintelligible to those in another. A particularly vivid historical example of
this situation is presented in the notion, once quite widely held, that the thermo-
dynamic laws underlying life processes must differ in some essential way from
those in force in the nonliving world — an idea whose untenability has only
been generally recognized in rather recent years.

Differences of language and concept tend to be powerfully reinforced by
many of the social factors governing scientific work. The desire of a scientist

CARNEGIE INSTITUTION OF WASHINGTON

to live and talk with those who will understand what he means, the pragmatic
influences that inevitably make him seek professional identification with others
in his immediate field, have their great strengths, both for the investigator and
for his work. For the investigator, such association means immediate identifica-
tion of interest and the satisfaction that only group activity can bring. For the
research, it means the exposure of every man's work to intimate and continuing
criticism by his peers in the same general subject area — the only critical estimate
that can be truly meaningful or can really maintain the standards of the field.
Yet there is a profound debit in this process too. At its worst it can harden an
incipient conventionalism, and can raise the most serious barriers to com-
munication within the body of science, powerfully reinforcing that separation of
fields which, unchecked, leads to unbridled specialism with all its attendant ills.

These challenges to communication within the framework of science are
severe enough. But today a further, and to some extent an intractable, threat of
fragmentation is posed by the very magnitude of the scientific effort itself and
by the tremendous volume of scientific publication that necessarily goes with it.
This is a threat which has been increasing with immense rapidity over the
half -century span of the Carnegie Institution. It is the worse because it is not
only the sheer volume of paper, of titles, of content that must be dealt with.
Some progress has been possible here through modern aids to storing and
sorting information, and intensive research could doubtless carry their effec-
tiveness much further.

But the hard core of the problem remains. It is the basic challenge to com-
munication that lies in all the diversity of the natural sciences. It is the effective
"addition" and the fruitful synthesis of ideas even in one field of work, and
much more generally the transfer of idea-systems from one field into another,
that, successfully met, may lead to major innovations of viewpoint.

Communication of this sort — the counterweight to the forces of fragmenta-
tion in science — can be greatly aided by environments of a very particular kind.
There have been notable examples of them in every scientific age — in the great
universities, and, more recently, in the great research institutes. They have
comprised communities of investigators, working together in a common mode
but in divergent fields, in continuous converse, in sympathy and in rivalry,
without predetermined goal, without overcommitment as a body to any given
sector of nature or to any one approach to the natural world. From such en-
vironments has come a goodly proportion of the real conceptual advances
of science.

As the forces of fragmentation and diversity in science are clearly more
powerful, the barriers to interchange evidently higher and more formidable,
in our own day than in any other age, this kind of communication within sci-
ence is more important now than it has ever been. It may, indeed, be one of

REPORT OF THE PRESIDENT 7

the most important aspects of the whole scientific effort if conceptual advance
is to continue and to accelerate.

The creation of such an environment is a task to which the Carnegie Institu-
tion of Washington is dedicated, and for which it is unusually well equipped.
The scientific community that is the Institution includes among its members
almost the full range of gifts and attitudes that has been described. The scien-
tific fields to which it addresses itself in the various Departments range from
the primarily descriptive to the primarily analytical, from the pioneer to the
more sophisticated. Yet by virtue of the mobility of its organization and its
community of spirit, neither workers nor fields are isolated. Rather the reverse
is true, so that fields of the most divergent character are sometimes included
within the working frame of a single Department and even within the purview
of a single investigator. These circumstances, and the fact that the whole of
the Institution's work is pointed toward the end of uncommitted research, fit
it peculiarly to assist in the major task of scientific synthesis.

The Institution has accomplished much in this direction, and its task in the
future will be yet greater. Notably in the fields of astronomy and physics, and
of physics, chemistry, geology, and biology, syntheses of concept and subject
matter which have been and are being achieved contribute significantly not
only to the breakdown of barriers between those fields but to the creation of
new fields — fields that then lie open to be tilled.

Substantive work of this kind must always remain the most enduring basis
for leadership by the Institution in this task. But there are other avenues too.
Symposia, carefully considered, painstakingly organized, and sensitively timed,
can be exceedingly fruitful in scientific synthesis and in the generation of new
concepts, especially if they conjoin fields that are subtly related and bring
together scientists from America and abroad who would normally foregather
seldom if at all. A number of such symposia have been organized by various
members of the Institution staff, and additional ones are contemplated.

The geographic dispersion of the several Departments of the Institution and
the location of many of them near universities bring further opportunities to
assume the role of a "crossroads" in the scientific effort, through the many
kinds of informal working arrangements that are possible between the staff
of the Institution and of these and other educational establishments. These
potentialities have been considerably explored. They must be developed yet
further.

Finally, the various fellowship programs, augmented now by the first of the
Vannevar Bush Fellowships from the Massachusetts Institute of Technology,
offer splendid opportunities to bring investigators to the Institution for varying
periods of training, of collaborative work, or of independent creative activity.
The newest, and one of the most exciting, of these programs has been designed
to further the work of mature and senior investigators of distinction in the

8 CARNEGIE INSTITUTION OF WASHINGTON

various fields of Institution interest, both at home and abroad. Its initiation
last year was made possible by a generous gift from the Carnegie Corporation
of New York. Guests are currently expected from Holland, Denmark, and
Great Britain, as well as from the United States, to be with the Institution for
varying periods.

Such are some of the challenges to communication presented by the formi-
dable diversities characteristic of the scientific effort. But there is a yet more
important message which must be kept vivid if the promise in our scientific
age is to be wholly realized. It is that of the deeper unities that underlie all
the diversities of the scientific mode — the unities of value, of standard, of goal,
of motivation.

To assume that human communication must always be in words, or even
that it must always take place at the level of consciousness, is an undue restric-
tion of viewpoint. Indeed, there is much to suggest that the kinds of communi-
cation which have had the most profound significance in human affairs have
often been neither wholly conscious nor entirely verbal. They have come in-
stead through the most powerful of all media — the sharing of a common
experience or a common view, simply and grandly symbolized. The sun and
the moon and the stellar firmament, that all men could see and equally know,
must have provided such symbols to innumerable human groups far more
remote in time than the great societies of Peru or Minoa or the Nile. J. Z.
Young has drawn attention to the enormous power of the mountain or the
hilltop, first natural, then man-made as the tumulus or mound or pyramid or
temple, as a towering symbol of communication in ancient societies, and it is
hard to conceive the whole structure and orientation of Renaissance Europe
apart from the glory of its cathedrals, or to reconstruct that society in imagina-
tion without them.

If communication of the most profound sort can thus be nonverbal, it can
also, of course, be largely divorced from material stimuli, a situation well illus-
trated by many highly evolved systems of religious belief. Few bonds of com-
munication can have been stronger, for instance, or can have had a more
important influence on the cohesiveness and the world view of a whole people,
for good and for ill, than that Augustinian concept of knowledge and research
that so dominated early Puritan America. As Perry Miller has described, its
essence was that men must believe in order to know; that the conclusions of
all possible investigations about the world are already given in advance of the
search ; that the most that right reasoning can possibly do is to arrive at them
again by a parallel course and illuminate them in detail ; that since reason is in
any case fallible and likely to fall short of even this secondary goal, it may be
wisest to forego reasoning altogether. No one can deny the power of such a
concept as an instrument of communication, as an integrating and stabilizing

REPORT OF THE PRESIDENT

force — or, in some measure, as a powerful brake to action and to originality —
in the society of its day.

But its very opposite, the approach to nature typified by the sweep of the
scientific revolution of the sixteenth and seventeenth centuries — the deep-lying
belief that all knowledge about the world is not given in advance of the in-
vestigation, that there still are new and profoundly beautiful and exciting
regions of nature and of the relation between nature and the observer that are
accessible to reason and experiment and have not yet been laid bare, that the
effort to explore them may offer spiritual and emotional rewards comparable
to the most exalting of experiences in other fields — this approach has long since
proved itself a mode of communication at least equal in its sweep, more flexible
and perhaps more viable, and, above all, conducive to positive action, to the
growth of ideas, and to human joy. And so the whole orientation of scientific
research can be considered in one sense a powerful symbol, as shining and as
dominating in its way as the simpler symbols of the sun or moon, and in this
context as much nonverbal, as much the sensed epitome of a shared, and ac-
cepted, and dedicated way of life.

The symbol itself is now some three centuries old. It has not dimmed in
those three hundred years, but it has changed extraordinarily in form. From
its very beginning, moreover, it has been in one sense a dual symbol, and this
duality has become emphasized in recent years, especially in our own country.
Neither profound philosophy nor practical experiment was new to the seven-
teenth century, nor, for that matter, to the Greeks. What gave the scientific
revolution its novel character and power — what in fact represented the very
genesis of science — was that for the first time these two strands were effectively
entwined. Science was philosophy joined to practical experiment. And in the
early conjunction there was an implicit realization that the concept must come
first, that experiment must serve as the trap for lines of evidence already
vaguely conjectured, that, in the suggestive simile of H. S. Harrison, experiment
is experience sharpened to a point — less divining rod than digging stick.

Had scientific research not been so eminently successful in a practical sense,
or were our pragmatic genius as a people less, it might not be so important to
make sure, in our day, not only that the symbol of the scientific way stays
bright, but that the strands do not become unwound. In the event, it remains
essential to recall the distinction and the interdependence between the strands —
science as a way of getting things done, and science as a way of life and a view-
point of the world. The first component needs little further emphasis than its
own extraordinary achievements already bring. But the second, and inherently
the more basic, element does require constant reaffirmation among us that it
may retain all the vitality and the allegiance and the comprehension that are
essential to its vigor. The task of such reaffirmation is especially important not
only because this is the more subtle as well as the more fundamental side of

10 CARNEGIE INSTITUTION OF WASHINGTON

science, and therefore less casually appreciated, but also because it is in constant
danger of being overwhelmed by its lusty partner and so lost to view.

This is the task to which, above all, the Carnegie Institution is dedicated.
It is the most essential duty of the Institution to communicate, in virtue of its
own mode and its own being, the essential verities of the scientific way that
are too easily forgotten. On one side lie the joy and the underlying human
values of the road of the investigator, the compelling life challenge that is
offered to the seeker after ideas about the natural world, whoever and wherever
he may be. On the other lie the great unities of approach and of preparation
that bind those dedicated to the scientific path: the requirements of verifi-
ability; the discipline of parsimony; the emphasis on individual effort with its
exacting demands of preparation and dedication, of originality and imagina-
tion, of the maintenance of style. These are not new parameters for the best
in living. They are as old as civilized humanity. But the scientific mode offers
one of the means by which those priceless elements, so often confused or
threatened with destruction in a crowded world, can be assured their proper
and their permanent place. No era which lacks them or to which they have
been lost can be great, whatever may be its other assets. Our time has no more
precious heritage than these qualities — and few tasks more essential than to
defend and reaffirm them.

REPORT OF THE PRESIDENT 11

THE YEAR'S WORK IN REVIEW

As always, selection of the year's researches for inclusion in this review must
be in large measure arbitrary. It cannot imply that those described here are
necessarily more or less important, more or less striking, more or less signifi-
cant in the last analysis, than other programs going forward beside them which
might have been included. They are simply to be considered as representative
examples of the work carried on during the year in the various programs of
the Institution. The reader who is interested in these programs in greater
detail is referred to the fuller individual reviews of the various Departments
that follow.

The Mount Wilson Observatory has pioneered for many years in the study
of stellar magnetic fields. The first definite indication of them was obtained
in the sun by Hale in 1908, using the then newly completed spectrograph of
the 60-foot solar tower on the mountain. Hale's finding that many of the lines
of the spectra of sunspots were split into components which had the charac-
teristic polarization of a Zeeman pattern provided firm evidence of the existence
of such fields. But subsequent search for a general magnetic field of the sun
gave such erratic results that Hale himself was never satisfied with the con-
clusiveness of the work that followed.

Almost forty years later the problem was attacked at the Observatory in a
somewhat different way. The considerable velocity of axial rotation attributed
to nearly all the A-type stars suggested to Horace W. Babcock that these stars
might show relatively large magnetic fields. Early in 1946 he made spectro-
graphs observations of one such star, 78 Virginis, and discovered a general
magnetic field of between one and two thousand gauss. Since that finding,
some hundreds of stars have been under observation with the coude spectro-
graph of both the 100-inch and the 200-inch telescopes on Mount Wilson and
Palomar. Thousands of measurements of stellar magnetic fields have been
made, and these have been collected for publication during the present year.
Eighty-four of the listed stars show definite evidences of magnetic fields, 55
are probably magnetic, and a further 55 show no indications of a coherent
magnetic field.

In 1952, encouraged by these results, Harold D. Babcock and Horace W.
Babcock returned to the older problem of the magnetism of the sun. With a
solar magnetograph constructed to take advantage of improved gratings and
recent advances in photoelectric techniques installed at the Hale Laboratory in
Pasadena, evidence has been obtained of magnetic fields of the order of one
gauss over large areas of the sun's surface. Since this finding, a second improved
magnetograph has been completed for the 150-foot solar tower on Mount
Wilson, and, starting with the end of this report year, a daily record of the
distribution of the magnetic field of the sun's surface is being made. The

12 CARNEGIE INSTITUTION OF WASHINGTON

results should be of great interest, especially in view of the fact that evidence
from many sources increasingly suggests that magnetic fields are of wide
occurrence and probably play a much larger role in astronomical phenomena
than has hitherto been supposed.

Problems of stellar evolution continue to occupy a central position in the
research of the Observatories. One of the important and intriguing aspects of
this field is the question of the evolution of the chemical elements in stellar
systems. G. R. Burbidge and F. Hoyle, with Dr. W. A. Fowler and Dr. E. M.
Burbidge, of the Kellogg Radiation Laboratory of the California Institute of
Technology, have continued their work on the synthesis of elements in the
stars. Eight nuclear processes are found to be necessary to account for the
known abundances of the 327 isotopes recorded in the solar system. The great-
est portion of the energy production of stars is due to the "burning" of hydrogen
(in a nuclear sense), producing helium. When the reaction occurs in a mixture
of hydrogen with other elements, it can result in the building of isotopes of
carbon, nitrogen, oxygen, fluorine, neon, and sodium. In the second process,
the nuclear burning of helium fuel produces C12, and, by further a-particle
addition, O16, Ne20, and perhaps Mg24. In the third pattern, the a process,
through charged-particle interactions, builds the remaining four-structure
nuclei Mg24, Si28, S32, A36, Ca40, and probably Ca44 and Ti48. The fourth pattern,
the e process, builds the elements from vanadium through nickel, comprising
the iron peak on the abundance curve. It takes place at very high temperatures
and densities. The a process and the e process are both thought to take place
shortly before the explosion of a star as a supernova. The fifth process (the s
process) is a slow neutron-capture chain, thought to occur in the interiors of
red giant stars. The r process, a rapid neutron-capture chain, is believed to take
place in supernovae, and to build uranium and thorium, together with a num-
ber of lighter isotopes. The seventh route of synthesis, the p process, is a proton-
capture or photoneutron process which is also thought to occur in some super-
novae. The eighth path of synthesis, the x process, is not as yet fully elaborated.
It may be responsible for building deuterium, lithium, beryllium, and boron,
elements that are unstable in hydrogen burning in stellar interiors. Work on
the nature of the x process is continuing.

One of the most striking features of far outer space is the great clouds of
gas that occupy it. The sources of these clouds may be various. Some of them
are thought to represent the remains of great cosmic explosions, such as those
of supernovae. Two programs have substantially increased our knowledge of
these gaseous nebulae during the past year. In one of them Osterbrock has
taken advantage of the fact that the relative intensity of two forbidden lines
of O II near X3727 varies markedly with density in the range of densities found
in these objects. Observations of these lines have made possible direct deter-

REPORT OF THE PRESIDENT 13

minations of the densities (and of the masses) of the Crab Nebula, the nebula
in Orion, and several planetary nebulae.

In the second program Munch and Wilson, using a multislit technique, have
obtained high-dispersion spectrograms from which the detailed distribution of
velocities through a large object such as the Orion nebula can be measured.
Large and abrupt changes in velocity have been observed, suggesting shock-
wave phenomena rather than a simple turbulence.

In the last report it was mentioned that the year had been an active one in
the comparatively new field of the study of celestial objects as radio sources.
These investigations have been continued in several directions in the current
year, both at the Mount Wilson and Palomar Observatories and in the Depart-
ment of Terrestrial Magnetism. Preliminary results were obtained in the De-
partment last year in the measurement of intensity or flux of radio sources at
frequencies below 30 mc. The greatly increased solar activity of the past year,
which has interfered with the optical redshift measurements of the far distant
galaxies, has also made the observing conditions for these radio sources dim-
cult — so difficult that the preliminary measurements of 1955-1956 cannot, in all
probability, be repeated in the 12 to 15 mc range until, perhaps, 1963-1965.

A review of intensity measurements on discrete sources in the available radio
spectrum, however, has underlined the present inadequate and unsatisfactory
data in this important area of radio astronomy. Accordingly, a program of
flux measurements on a few of the intense radio sources over a wide band of
the useful spectrum has become an important project in radio astronomy in
the Department. Dipole antenna arrays accurately calibrated and with very
precise record scalings have been employed, and measurements have been
made of Cassiopeia A and Virgo A at frequencies below approximately 100 mc.
Cygnus A and Taurus A are also objects of investigation. It is planned to con-
tinue dipole measurements to the limits of sensitivity — possibly several hun-
dred megacycles. Thereafter it is planned to standardize other antennas of
large aperture against dipoles at the transition frequency and to use scaled
versions of the large antenna in subsequent operations. It may thus be possible
to extend the range of absolute measurements to 1000 mc or more.

Measurements of radio emission within our own solar system have also con-
tinued actively. An antenna effective for a detailed examination of the emis-
sion of radio-frequency energy from the surface of the sun presents a number
of special requirements, quite different from those demanded of antennas used
for stellar work. Such an antenna array, specially designed for detailed ex-
amination of solar radio emission, has been built by the Department at the
River Road site near Seneca, Maryland. Preliminary scans of the sun's face
made with this equipment have revealed localized bright sources which move

14 CARNEGIE INSTITUTION OF WASHINGTON

across the disk as the sun rotates. Further work should reveal much of interest
in the finer structure of solar radio emission.

It will be recalled that one of the earliest findings in the Department of
Terrestrial Magnetism in the field of radio astronomy was that the planet
Jupiter was a source of radio noise. Further investigation of this phenomenon
has indicated that there may well be a single center of activity on the planet,
with an approximately uniform rotational period. No visual observations of
Jupiter report surface features having a corresponding rotational period, sug-
gesting that the source of radio radiation lies below the cloud level of the planet
and may well be associated with its actual surface.

Two further developments of note in radio astronomy are under way in the
Department. To obtain a detailed knowledge of the nature of radio sources,
optical identifications must be made. So far, only relatively few celestial objects
have been so identified in a satisfactory way, and the available evidence suggests
that many radio sources may be distant galaxies fainter than the eighteenth
magnitude. To extend the list of optical identifications, precise position de-
terminations for a large number of radio sources should be made. Antennas
constructed for precision position measurements must incorporate very special
features. Such antennas are under experimental design at present. At the
same time, a beginning has been made this year in the actual construction of a
radio telescope to be equipped with a 60-foot dish — a project that has been
under consideration and study in the Institution since 1955.

It is a long way from considerations of distant galaxies and stars — their age,
their evolution, their physical characteristics — to similar investigations within
the confines of our own planet. In some respects, however, the qualities of our
terrestrial environment have an even more vivid and immediate quality for
the imagination. Both the Department of Terrestrial Magnetism and the
Geophysical Laboratory are deeply concerned with the properties of the earth's
crust. In the Department special attention has been focused for several years
on a recognition and understanding of many large-scale physical processes
operating over long periods of geological time which have resulted, among
other things, in the formation of the continents and the ocean depths, moun-
tain ranges and high plateaus. A sine qua non for such understanding is a
comprehensive and quantitative description of the crust of the earth at the
present day.

To this end, seismic and gravity studies continue to form an important pro-
gram of the Department's work. It will be recalled that this program has in
previous years included studies on the Colorado Plateau and, as reported last
year, in Alaska. This year an even more ambitious seismic study has been
undertaken in the Andean highlands, using as indicators the explosions nor-
mally set oft in the operation of large open-pit copper mines in southern Peru

REPORT OF THE PRESIDENT 15

and northern Chile. This program has been undertaken by the Department
in co-operation with the United States National Committee for the Interna-
tional Geophysical Year.

The problem of the age of the various rock deposits of the earth has engaged
the attention of earth scientists since the beginning of geology. Despite wide-
spread and intensive work over the years, however, major problems remain.
Several years ago a program was undertaken jointly by the Geophysical Labora-
tory and the Department of Terrestrial Magnetism to extend the scope of
mineral age measurements involving radiogenic products, originally limited to
rocks containing uranium and thorium — an extension made possible by im-
provements in mass spectrometric techniques and finer developments of chemi-
cal analysis by isotope dilution. Particular emphasis has been given to the
means made available by the natural radioactive decay of potassium to argon,
and of rubidium to strontium.

Agreement of the rubidium-strontium and potassium-argon ages for a sample
of mica is a good indication that the mineral was formed in a closed system,
and such agreement is now considered essential to a satisfactory age determina-
tion. During the year the Geophysical Laboratory-Department of Terrestrial
Magnetism group began an investigation of regional regularities in the ages of
the Precambrian rock exposures. As an early result, it has been found that
over a large part of Arizona, New Mexico, Colorado, and Wyoming rocks
were formed during a period 1300 to 1400 million years ago. In Ontario there
exists a large group of minerals that are apparently much older. Investigation
of these, together with the work of other laboratories, suggests that there is a
long band of rocks, extending from Wyoming through Montana, Minnesota,
Manitoba, and Ontario into Quebec, composed of minerals of approximately
2600 million years of age. Similar areas of very ancient rocks are known in
Africa and Australia.

A striking specific finding in the past year in this area of research has been
the establishment, by concordant rubidium-strontium and potassium-argon
datings, of an age of about 340 million years for the micas in a number of
granitic rock samples collected by Dr. Faul, a guest investigator from the
United States Geological Survey, from the Hercynian Chain of western Europe.
Now the stratigraphic age of these rocks is known to be older than the middle
Carboniferous, and they are usually assigned to the lower Carboniferous. Ac-
cording to the United States Geological Survey version of the Holmes time
scale, the Carboniferous began about 265 million years ago. By this time scale,
then, these Hercynian rocks according to the present findings belong in the
middle Silurian. Thus there is a tremendous difference between the present
age determinations and the commonly accepted stratigraphic assignment. The
discrepancy is so great that it follows either that the time scale is not correct

16 CARNEGIE INSTITUTION OF WASHINGTON

or that the accepted stratigraphic assignment of the formation requires con-
siderable revision.

Phase-equilibrium relations among the major mineral groups continue to be
studied intensively at the Geophysical Laboratory and are making available a
whole series of geological "thermometers" which may be applied to igneous
and metamorphic rocks, supplying means for learning a great deal about the
conditions under which such processes as mountain building occurred. Similar
types of thermometers are being developed for ore minerals. Thus Kullerud
and his associates have performed the basic laboratory work preliminary to
field tests of an additional series of sulfide geological thermometers, including
two based on the stability of pyrite and covellite and a third on the composi-
tion of pyrrhotite formed in the presence of pyrite. The ultimate accumulation
of a number of such thermometers will make possible extensive cross checks
and permit the reliable determination of temperatures prevailing during ore
formation.

For a number of years the Geophysical Laboratory has carried forward a
vigorous program in crystallography, which is becoming perhaps one of the
most vital fields in all of science today. Through the use of neutron diffraction,
nuclear and paramagnetic resonance, and X-ray diffraction the structural chem-
ist is learning how the precise determination of molecular arrangement leads
to real understanding of chemical reactivity. The nature of the chemical bond
as manifested in covalent, metallic, ionic, or intermediate types is related closely
to structure. Interestingly, work on the crystallography of alkali phosphates,
continued this year jointly with Drs. J. W. Gryden and Helen M. Ondik, has
demonstrated the first crystalline ultraphosphate on record.

One of the high points in research for the year in the Institution has been
the work of Chayes in the Geophysical Laboratory. Using an optical analogue,
he has produced diffraction patterns similar to the types obtained from the
interaction of X rays on crystals. By making variations in the analogue he has
been studying the kinds of patterns that might be obtained in various types
of order-disorder in crystals. This method may well represent a major "break-
through" in crystallographic research.

During the year Libby has continued his investigations of simple methods
of absolute counting of 3 radioactivity at the Geophysical Laboratory. He has
evaluated the role of back-scattered radiation with greater accuracy, and has
discovered the important effect of surface roughness in the measurement of
soft 3 rays. One extremely practical consequence of these findings is that they
may now make possible the introduction, into the high school and college
classroom and laboratory, of isotopes of real chemical interest, convenient life-
time, and low enough specific activity to be completely safe, permitting student
use, for example, of the radioactive forms of acetic acid, hydrochloric acid,
sulfuric acid, and the calcium salts.

REPORT OF THE PRESIDENT 17

Yoder and Tilley have continued their studies in the production of various
basalts, with particular attention to the possibility that they can be obtained
naturally by differentiation of a parental magma. Laboratory determinations
of the products formed when a variety of basalts are melted and quenched, as
well as comparisons with minerals obtained from synthetic melts, show that
the basalts must have been formed from different original magmas. Eugster
has determined the effects of oxygen on the stability of the iron mica, annite,
thus demonstrating the feasibility of investigating the important biotite micas.

The Carnegie Institution is deeply concerned with research in fundamental
biology. In five of the seven Departments, investigations are under way in
various aspects of the life sciences. They range from the organization of
molecular units in life processes through questions of the structure and function
of relatively elementary subcellular entities such as the pools for amino acid
concentration and synthesis in single cells, of more grossly defined intracellular
structures such as the chloroplasts of the cells of green plants, and of such ex-
ceedingly complicated and critical cellular organelles as the chromosomes, to
the organization and functioning of cells themselves in the developing meta-
zoan embryo, and finally to questions of speciation and experimental taxonomy,
involving the organization and the interaction of many-celled plants in natural
populations.

In the Geophysical Laboratory, Abelson has continued his pioneering work
on the synthesis of amino acids from simpler components under conditions
simulating those believed to have obtained on the earth in remote geological
periods. Analysis of the available geologic evidence has led to the hypothesis
that the principal constituents of the early terrestrial atmosphere were carbon
monoxide, nitrogen, and lesser amounts of hydrogen, water, and carbon di-
oxide. In a series of laboratory experiments he has simulated the effects of
solar radiation on such an atmosphere and the associated oceans and found that
a considerable production of organic compounds results, including the amino
acids glycine, alanine, and serine, as well as more complicated substances.
These amino acids are important starting points in the building of proteins by
living things.

One of the very interesting consequences of the analysis of probable condi-
tions in the atmosphere and the oceans of the earth at the time when life may
have originated is that carbon dioxide, probably fairly abundant in the early
atmosphere, must have served as a kind of buffer in the oxidation-reduction
system, guaranteeing that the atmosphere could never have been very reducing.
It seems impossible to visualize that any great concentration of such substances
as methane was present together with the carbon dioxide. This picture con-
trasts sharply with some of the earlier ones of the supposed conditions under
which life-like systems might have originated on the earth.

18 CARNEGIE INSTITUTION OF WASHINGTON

For a number of years, the members of the biophysics group in the Depart-
ment of Terrestrial Magnetism have been especially concerned in tracing the
pathways of synthesis of proteins and nucleic acids in microorganisms, concen-
trating their attention especially upon two representative forms, a bacterium
and a yeast. During the earlier years, attention was directed especially to the
synthesis of relatively small molecules. Later the studies were shifted to in-
vestigations of the metabolic pools which are the precursors of the macromole-
cules. Special attention this year has been concentrated on the organization of
structure within the cell in terms of some framework larger than a protein
molecule.

Studies of the metabolic pools have revealed the extremely interesting point
that they are susceptible to osmotic shock, and must therefore be held in some
osmotically sensitive structure. Work conducted this year strongly suggests
that when the osmotic pressure of the medium is suddenly reduced a flow of
water occurs into the sensitive structures, together with a slow loss of solute
from the cell. The stretching of the pool structures that results leads to an
increase of permeability, allowing a faster rate of loss of the solute molecules.
It was found in one set of experiments, for instance, that radioactive S04= and
P04= ions can be taken up after such shock in an amount corresponding to
about 5 per cent of the cell volume at the external concentrations of S04= and
P04= — a situation that cannot occur in the absence of the shock. Stretching of
the pool structures and loss of solute molecules finally lead to a new osmotic
equilibrium in the cell and to the recovery of the normal pool function. These
experiments are the first in which the properties of the pool as a definite,
distensible intracellular structure have been clearly indicated, and they are
correspondingly interesting.

Through the use of amino acid analogues the mechanism in the cell which
selects amino acids for protein synthesis has been investigated by observing the
"mistakes" that a cell can make in protein formation. Through collaboration
with Dr. Georges N. Cohen, of the Institut Pasteur, it has been shown that
selenomethionine, for example, can completely replace methionine and sup-
port exponential growth in a methionine-requiring mutant of the bacterium
Escherichia coli. The experiment, together with related ones conducted at the
Institut Pasteur, demonstrates that the amino acid composition of a protein can
be altered by the presence of such amino acid analogues in the medium.

Perhaps the most dramatic finding of the biophysics group for the year has
been the demonstration that, under appropriate conditions, rather large parti-
cles containing nucleic acids, proteins, and lipids can be made to form "spon-
taneously" from disintegrated cellular material. When cultures of cells of
E. coli suspended in a glucose-mineral salts medium were brought to pH 8,
centrifuged, resuspended, and broken by being forced through a small hole
at high pressure, again centrifuged to remove cell fragments and any unbroken

REPORT OF THE PRESIDENT 19

cells, and the resulting supernatant decanted and once more centrifuged and
diluted, a quite clear fluid remained. When to this fluid MgCl2 and MnCl2
were added (only the manganese being indispensable), the clear solution
became cloudy and a few hours later was found to contain in considerable
density particles which were nearly spherical and ranged in size from about
1 to 5 microns. Since the particles have a definite size and shape, are quite
stable, and are found to contain several of the constituents of bacterial proto-
plasm, they have been named "protomorphs." Further studies of these remark-
able objects are planned.

The Department of Plant Biology has long maintained a central concern
with the nature of the photosynthetic pigment chlorophyll as it occurs in the
living plant. The problem has been attacked by a number of routes. Attempts
to isolate the photochemically active components of disintegrated chloroplasts
made several years ago showed that reaggregation of the finely dispersed frag-
ments restored activity. The size of the chloroplast fragments used at that
time, however, was still too large to permit the isolation of the chlorophyll
complex in pure form. Considerably more success has been achieved in recent
years by the characterization and partial purification of the protochlorophyll
complex from leaves grown in the dark. The chlorophyll-protein complex
formed by illumination of etiolated leaf material, being soluble in water, is
more suitable for certain chemical studies than the natural complex obtained
from dark green plants. The complex in fully developed leaves contains far
more chlorophyll per unit protein, however, so that it is important that it be
studied also. As such complexes are generally water-insoluble, methods of in-
vestigation not based on chemical isolation seem more promising than attempts
at direct chemical characterization.

One evident method for investigation is a study of the chlorophyll absorption
spectrum. Such a study, concerned with the detailed shape of the red absorp-
tion band of chlorophyll in vivo and of the way that its shape may be altered
by various procedures, is under way in the Department at present. The red
band of chlorophyll in living material appears to consist of two components
having a wavelength difference of about 10 millimicrons. These two com-
ponents have not been separated chemically. Their presence, however, may
be indicated by the shape of the derivative of the spectral absorbance curve —
though another interpretation is also possible. The Department is particularly
fitted to undertake an investigation of this kind by virtue of its derivative
spectrophotometer, designed and built in past years by French and his group,
and now in full operation.

Since the evidence for the chemical differences between native and extracted
chlorophyll rests so largely on spectroscopic data, it is especially important to
be sure that the methods for measuring the absorption spectra of pigments in

20 CARNEGIE INSTITUTION OF WASHINGTON

living cells are highly reliable. During the current year Latimer has shown
that large errors can be introduced in determining the peak positions of the
absorption bands of cell pigments by selective scattering of the light by the
pigment molecules — a phenomenon which is sensitively dependent upon the
wavelength of the incident light in the vicinity of a pigment absorption band.
This critical work does not appear to invalidate the conclusion that natural
chlorophyll is spectroscopically different from the extracted material, but it
does raise grave questions of the validity of comparisons of the peak positions
of chlorophyll in living cells spectroscopically measured in different labora-
tories, or even of different samples in the same apparatus. It also raises the
important question whether the double peak of chlorophyll absorption observed
by derivative spectrophotometry is due to two actual components, or whether
one peak may be an optical artifact.

Attempts to isolate the naturally occurring protochlorophyll holochrome
have been continued during the year by Smith, who explored additional meth-
ods of fractionating the leaf extracts. The earlier reported purification of about
75 per cent was not exceeded, but a promising new source of etiolated leaves
has been found in the tropical starch-crop plant taro, which will be interesting
for future work.

Another approach to the study of the natural chlorophyll-protein complex
is to separate it into its protein and chlorophyll moieties and then to recom-
bine them. Dr. Wolf Vishniac, of Yale University, has succeeded in this effort,
and has demonstrated that the recombined "halves" of the complex can again
show photochemical activity. During the year Vishniac spent some time at the
Department, experimenting with and demonstrating his preparations. In the
course of the work he found that purified chlorophyll a can be used in place of
the alcoholic leaf extract with which his earlier experiments had been done,
and which contained substances in addition to chlorophyll.

A more functional kind of investigation of the nature of chlorophyll and its
relation to growth processes has been under way in the Department for some
time. This is a study of the interrelated effects of light intensity and tempera-
ture on the growth of various species of algae, performed by growing the cells
on an agar surface having a gradient of temperature from left to right and a
gradient of light intensity from front to back. The growth pattern produced
on such a plate was found to be very different in different species of algae. In
last year's report the L-shaped pattern of Chlorella pyrenoidosa was mentioned.
It has been found recently that the vertical arm of this pattern is accentuated by
a narrow temperature range around 32° C in which growth is very poor at
high light intensities. On the higher-temperature side of this sharply defined
zone is an equally narrow range of luxuriant growth. At temperatures below
32° C at high light intensity the growth is moderately profuse for a few
degrees, then weakens gradually toward lower temperatures.

REPORT OF THE PRESIDENT 21

It is believed that the narrow temperature range occupied by dark green
cells on plates several days old is artificially sharpened by effects secondary to
the direct influence of intensity and temperature on growth.

Yet another approach to the problem of the role of light in the life of the
plant is a study of the orientation and behavior of freely moving unicellular
algae in the presence of light of varying wavelengths and intensity. Such
studies of phototaxis in the unicellular green algae have been continued at the
Department this year, attention being focused especially on the form Platy-
monas subcordiformis. When this alga exhibited a fixed light response, the
reaction could not be altered immediately by changing either the intensity of
illumination or the absolute or partial pressure of carbon dioxide or oxygen in
the medium. On the other hand, the ions Ca++ and Mg++ were found to be
directly involved, and to be antagonistic in their effects, Ca++ causing a nega-
tive and Mg++ a positive phototaxis. A number of other ions that were tested
produced no reaction. The theory that the driving power in the flagellates is
an "adenosine triphosphate motor" has recently been suggested by Links, in
Holland. Such a hypothesis implies that the mechanism supplying the muscu-
lar energy of animals is similar to that which innervates the flagellar apparatus
of algae. From muscle research it is thought that Ca++ and Mg++ ions may be
involved with antagonistic effects in the ATPase activity. Thus the antago-
nistic effect of Ca++ and Mg++ in these experiments in phototaxis is particularly
suggestive.

Lynch and French last year reported the very interesting finding that the
Hill activity can be restored in ether-extracted chloroplasts by (3-carotene. This
year the work has been continued by Dr. Max Milner, visiting the Department
from Kansas State College, who used somewhat more quantitative techniques.
The surprising effect has been confirmed, and it has been shown that the opti-
mum concentration of (3-carotene required for reactivation is approximately
100 times greater than that used in the earlier work. These and related experi-
ments suggest that one or more fat-soluble factors may be involved in the
reactivation of petroleum ether-extracted chloroplasts in addition to (3-carotene.
Further work is planned to determine the nature and the mode of action of
these components.

Perhaps no aspect of the living cell presents so many facets for investigation
or leads to consequences of wider general interest than the mechanisms through
which heredity is mediated and determined. Concern with the processes of
heredity must underlie biological work throughout the Institution, but, in
both functional and structural aspects, it is the particular province of the De-
partment of Genetics.

On the functional side, the remarkable findings of McClintock concerning
systems of elements in the cell nucleus which control the action of genes in

22 CARNEGIE INSTITUTION OF WASHINGTON

maize have continued to open new vistas. Study has been concentrated on a
system that controls gene actions at two known loci, not directly related to the
Ds-Ac system earlier reported, though the genes affected are themselves ones
which, in other cultures, have come under Ds-Ac control. Particular attention
has been directed to determining the number of recognizable elements com-
posing the system and to studying the transpositions and the changes in modes
of expression that can occur.

Work has also continued in the analysis of a structural modification in maize
chromosome 9, preliminary investigations of which were reported last year.
The substance of the chromosome is distributed between two components, one
member comprising the distal third of the short arm (the "fragment chromo-
some") and the other the proximal two-thirds of the short arm and all of the
long arm (the "deficient chromosome"). The fragment chromosome exhibits
aberrant behavior in somatic cells, sometimes being lost, sometimes undergoing
changes in structural organization or becoming attached to ends or centromeres
of other chromosomes, or being incorporated into another chromosome. The
frequency of occurrence of events leading to such consequences and the time
of their occurrence during the development of a tissue are both known to be
under genetic control. Initial interest in this structurally modified chromosome
was aroused by the aberrant behavior of fragments in somatic cells. It was later
discovered that the fragment could behave unexpectedly in some of the meiotic
cells also, and in plants either heterozygous or homozygous for the structural
modification. It has become clear that the rules supposed to apply to crossing
over in maize are not always followed by the fragment chromosome when it
participates in a crossover event. The genetic significance of these findings may
be very considerable indeed.

For a number of years Demerec and his group have been concerned with the
application of methods of biochemical genetics to investigation of the detailed
constitution of bacterial chromosomes — methods so quantitative as to be ca-
pable of resolving loci concerned with specific amino acid syntheses. In this
connection they have paid much attention to the phenomena of transduction,
in which fragments of a bacterial chromosome can be transferred from one
bacterium to another through the vehicle of a particle of infective bacteriophage,
and of transformation, in which genetic changes can be mediated by nucleic
acid components of a cell. Such studies have been continued this year in the
genetic mechanisms of the bacterium Salmonella typhimurium. Stocks have
been constructed having various combinations of genetic markers representa-
tive of a cystine locus and four tryptophan loci so closely linked that they are
carried together in one transducing fragment. Preliminary experiments with
such stocks have indicated that three separate portions of a fragment may be
incorporated simultaneously into a newly formed bacterial chromosome. If
incorporation is accomplished through a process similar to crossing over in

REPORT OF THE PRESIDENT 23

higher organisms, six simultaneous crossovers would be required to produce
the combinations of markers observed in these experiments. A most interesting
indication has been obtained by Ozeki that a transducing fragment may not
be a randomly selected section of a bacterial chromosome. The evidence sug-
gests that the chromosomes of donor bacteria (infected with bacteriophage)
are partitioned during the lytic process into small sections in some regular way
and at predetermined locations, producing uniform transducing fragments for
any one region.

New evidence that the "chromosome" of bacteriophage T2 consists exclu-
sively of nucleic acid has been obtained by Hershey and his group. They have
found that the chromatin material can multiply in functional form in bacterial
cultures containing chloramphenicol, which inhibits the synthesis of all bac-
teriophage proteins.

The nature of reverse mutations in bacteriophage has been investigated dur-
ing the year by Streisinger, with interesting results. Streisinger's previous work
with bacterial viruses had revealed that mutations at many sites within a certain
genetic region may result in a loss of the capacity of the bacteriophage to attach
to certain strains of bacteria. Such mutations, resulting in a loss of adsorptive
capacity in the phage T2, may take place at any of a large number of genetic
sites. It is now clear that reverse mutations restoring this adsorptive capacity
occur at precisely the sites of the forward mutations from which they are de-
rived. Not all reverse mutants, however, are identical with the original, non-
mutated form.

Work has also gone forward actively during the year on the structural side
of the nature of the hereditary mechanism. Kaufmann, Gay, and their asso-
ciates are combining methods of enzymatic hydrolysis with modern techniques
of electron microscopy. Results obtained after deoxyribonuclease treatment of
salivary-gland chromosomes of Drosophila fixed in osmium tetroxide and
stained by the Feulgen method suggest that the main structural fibers are
immune to nuclease or acid hydrolysis. DNA thus appears to be attached to
the main chromosomal axis. Results obtained with chromosomes of Drosophila
and of plants preserved in other fixatives are somewhat ambiguous ; much work
remains to be done before a generalized conclusion can be formulated about
the location of DNA, RNA, and protein in chromosomes. Inclusion of tritium-
labeled thymidine in the food of Drosophila larvae has afforded some spectacu-
lar radioautographs, prepared in collaboration with Philip Woods, of the
salivary-gland chromosomes at late third instar. Radioactivity was limited to
the banded regions. This finding suggests that thymidine, which is a precursor
of DNA, was not incorporated into the material that occupies the intervals
between the bands.

McDonald has continued her investigations of the intracellular deoxyribo-
nucleases, which appear to be related to the general problem of DNA me-

24 CARNEGIE INSTITUTION OF WASHINGTON

tabolism. These deoxyribonucleases are particularly interesting because of the
unique biological importance of their known substrates as components of
chromosomes. McDonald has concentrated her efforts this year on a survey of
the deoxyribonucleases of various tissues, seeking to determine their precise
relation to cell division and to the synthesis and metabolism of DNA. During
these studies it was found that the deoxyribonuclease content of salmon testes
is especially high, and further investigation suggests that this tissue may be a
potent and extremely useful source of the material.

The organization of cells into tissues, and their modification in the course
of growth and metamorphosis of the many-celled animal, is the peculiar con-
cern of the Department of Embryology. During the year some methods have
been developed and applied which are most noteworthy in their delicacy and
in the extent to which they permit exact inquiry. Perhaps one of the most
striking is illustrated by the current work of Wilt, concerned with a system
which may well permit study of the effects of a hormone system in a meta-
morphosing vertebrate at a molecular level. The bullfrog tadpole is trans-
formed into an adult under the influence of the thyroid hormone. During the
process, an abrupt change takes place in the visual pigment of the retina. The
tadpole's retinal pigment, porphyropsin, is rapidly replaced by rhodopsin, the
pigment of the adult retina. These visual pigments are composed of a protein
moiety (opsin) coupled with a carotenoid. The change at metamorphosis has
been described as a shift from retinene2, the aldehyde of vitamin A2, to retinenei,
the aldehyde of vitamin Ai. Wilt's long-range objective is to determine the
role of the thyroid hormone in this shift. His immediate goal is to explore the
metabolic implications of the predominance of retinene2 in the tadpole and of
retinenei in the adult. He first tested the hypothesis that the tadpole and adult
differ in their content of oxidative enzymes, so that the tadpole can oxidize only
vitamin A2 and the adult only vitamin Ai. His work disclosed that the two
metamorphic stages do not differ qualitatively with respect to retinene reductase,
the enzyme required for the reaction. Both the tadpole and the adult can oxidize
vitamins A2 and Ai. In further experiments Wilt succeeded in preparing and
characterizing highly purified porphyropsin and clearly showed that it differs
from the rhodopsin of the adult, after which he used this porphyropsin in
experiments of great interest. The question has been raised whether the pig-
ment difference in the tadpole and adult was the result of a change in the
ability of the opsin moiety to couple with the retinene. Thus the larval opsin
might couple with retinene2, that of the adult only with retinenei. Wilt has
been able to prove that larval opsin (bleached porphyropsin) will combine
with retinenei, to regenerate rhodopsin, the kinetics of the reaction following
closely those for the regeneration of rhodopsin in the adult retina. He is now

REPORT OF THE PRESIDENT 25

focusing attention on the pathways of synthesis of vitamin A2 and on the
mechanism that confines it to the eye.

In the course of experiments designed to modify the onset of synthesis and
functional activity of acetylcholine in the early chick embryo, DeHaan made
an unexpected observation that has led him to extend his study of spontaneous
contractility to include an analysis of the morphogenetic movements of meso-
derm and endoderm involved in cardiogenesis. It is often unappreciated that
the factors regulating cell movements in the embryo include perhaps the most
obscure phenomena in embryology. Thus it proved of great interest when
DeHaan demonstrated that, in the chick embryo cultivated in vitro, the addition
of a tiny crystal of acetylcholine to the surface of the endoderm resulted in the
formation of two independently beating hearts. On the basis of several lines
of evidence DeHaan has suggested that acetylcholine acts here by the sequestra-
tion of calcium or other divalent cations. Particular interest attached to the
specificity of the effect. Other morphogenetic movements presumably are in
progress at this time, yet only the cardiogenic movements are altered. DeHaan
continues to probe this question, at the same time maintaining his interest in
the onset of spontaneous contractility.

In consultation with Bishop, Dr. Katsh has continued to explore the unusually
interesting observation that the injection into the male guinea pig of testicular
extracts combined with adjuvant (a mixture of oils and certain killed bacteria)
leads to the destruction of the spermatogenic elements in the testes of the recip-
ient. It is suggested that damage to the spermatogenic tissue is the result of an
immune reaction, which may operate in the following manner. The prepara-
tion of the testicular homogenate and its combination with adjuvant result in
a change in its properties sufficient to enable it to act as an antigen in the
homologous species. The antibodies produced by the host, however, are not
sufficiently specific to distinguish normal and slightly altered testicular antigens;
hence the animal's own testis is destroyed. Katsh's studies have disclosed that
sperm-immobilizing, sperm-agglutinating, and complement-fixing antibodies
appear in the sera of the injected animals and not in the sera of animals injected
with adjuvant alone. Yet there are reasons for believing that these circulating
antibodies may not be instrumental in inducing aspermatogenesis. For example,
it has not been possible to demonstrate a correlation between antibody content
and aspermatogenic effectiveness. Moreover, although other animals like the
rabbit are highly effective in producing circulating antibodies of this type,
aspermatogenesis can be induced only with difficulty. These considerations led
to investigations of other types of immune reactions, as a result of which Katsh
has found that the ileum of the sensitized guinea pig injected with homologous
testis or sperm responds in vitro to a challenging dose of guinea-pig sperm by
contracting, a finding that suggests an allergic response of the anaphylactoid
type. While exploring this possibility, Katsh has made progress toward eluci-

26 CARNEGIE INSTITUTION OF WASHINGTON

dating the nature of the antigenic stimulus. He has demonstrated that the
antigens are not species-specific, and that they are shared by brain and testis.
Within the testis, the effective substances are confined to the spermatogenic
elements, a fact established by the failure of homogenates of testes depleted of
spermatogenic tissue to elicit a reaction. A number of thought-provoking ex-
periments suggest, but do not prove, that the antigenic stimulus is provided by
a mixture of bacterial and testicular lipopolysaccharides.

During the past years, Dr. Ramsey's morphological studies have led her to
conclude that the circulation in the maternal placenta of primates is effected by
the vis a tergo of the maternal blood pressure, a hypothesis that contradicts the
traditional belief that the myometrial contractions "squeeze the placenta like a
sponge." Although a few valuable observations have been made in pioneer
studies on human patients, experiments are required to confirm or, if necessary,
modify the hypothesis deduced from morphological studies. During the year,
Ramsey initiated such a program in collaboration with Drs. G. W. Corner, Jr.,
and W. Newton Long, Jr., using experimental material from the Carnegie
monkey colony. Although the first year of the research was devoted primarily
to working out new techniques, the initial results have been encouraging.
Standard procedures were devised for introducing a polyethylene catheter into
the amniotic cavity — the uterus having been exposed at laparotomy — or into
the intervillous space or a uterine vein. Positive results have indicated that
myometrial contractions are reflected in heightened amniotic and intervillous
space pressures, and that variations of absolute pressures in the amniotic and
intervillous spaces correspond. Data have been obtained showing inherent
amplitude-tonus patterns characteristic on the one hand of individual uteri and
on the other of specific developmental periods in animals studied during a
single pregnancy. Although the preliminary findings support Ramsey's hypoth-
esis, the research is just beginning.

At the "highest" level of biological organization, the structure of populations,
the group in experimental taxonomy is continuing its work actively at the
Department of Plant Biology, though this year with a somewhat altered direc-
tion. The research has long been concerned especially with morphological and
genetic studies of widely distributed groups of plants, directed primarily toward
the clarification of the evolutionary relationships between races, species, and
groups of species, and toward their fitness to their natural habitat. Until re-
cently this work was done by comparing the responses of identical population
samples of races and hybrids of plants in contrasting environments, by exten-
sive genetic studies, and by examining chromosome behavior in hybrid and
parental forms of species variously related. Many of these long-range studies
have been completed, and emphasis is now shifting to a study of the compara-
tive physiology of closely related but ecologically distinct races. Research is

REPORT OF THE PRESIDENT 27

being carried on at present with some of the species whose genetic constitution
and evolutionary history have already been considered. Current work is espe-
cially concerned with rates of respiration and photosynthesis of contrasting
climatic races of the monkey flower, Mimulus cardinalis, and with comparative
growth and physiology in many races of the aquatic duckweeds, the Lemnaceae,
and in various groups of the yarrow, Achillea.

Laboratory analyses and studies of data resulting from recent field work by
the Department of Archaeology during the year revealed a number of new and
significant facets of the Maya culture. Studies by Proskouriakoff have developed
some extremely interesting information bearing on the date of the introduction
of metal into Yucatan, suggesting that this event may have occurred later than
has commonly been supposed. Thompson, in a study of effigy incense burners
recovered from Mayapan, was able not only to identify a number of the gods
or personages represented but in addition to show how considerably the religion
of Mayapan was influenced by foreign ideologies. He also succeeded in identi-
fying a number of glyphs representing diseases — a rather striking finding.

The Department of Archaeology has this year given particular attention to
finishing the various studies in hand, preparatory to the completion of its work
in 1958. Effort has been directed especially toward the production of pre-
liminary reports on the field work and to broader studies that will lead to defini-
tive statements covering the results of the more recent program of researches
in Yucatan.

Losses . • •

On June 30, 1957, Dr. George W. Morey retired from the Institution, bringing
to a close an association of forty-five years with the Geophysical Laboratory.
On the following day he joined the Geochemical and Petrology Branch of the
United States Geological Survey to continue his work — a vivid demonstration
of his own conviction that the work of true inquiry is never done.

When Dr. Morey joined the staff of the Geophysical Laboratory in 1912 that
Department was but six years old. Its interest then was concentrated particu-
larly in the study of mineral-forming processes in the interior of the earth. The
new staff member expressed some misgivings about the value of the contribu-
tions a physical chemist could make to such a program. Events, of course, belied
this uncertainty. His initial precise quantitative work on phase equilibria in
silicate systems, in fact, provided the basis for most of his subsequent research.
The label "systems containing water and carbon dioxide," by which he describes
his field of study, covers myriad activities. Intensive experimentation with
silicates and other minerals under extreme conditions of temperature and
pressure, approximating those prevailing deep in the interior of the earth, was
the starting point for new insights into the processes of mineral formation,

28 CARNEGIE INSTITUTION OF WASHINGTON

ore deposition, and the chemistry of magmatic differentiation. It also led to a
new theory of volcanic eruption based on the tremendous pressures developed
when minerals crystallize from an aqueous magma.

Dr. Morey is noted for his skill and ingenuity in devising new apparatus and
techniques. The "Morey bomb" was an important tool in the first production of
synthetic quartz, which, in turn, led to most significant research in the devel-
opment and manufacture of optical glass.

During both world wars, Dr. Morey left the Laboratory to put his theoretical
and experimental work on the chemical constitution of glass to practical appli-
cation in the service of his country. In the first war he manufactured strategic
optical glass for the War Industries Board and was general manager of the
Spencer Lens Company; in the second, he was a member of the optical instru-
ments section of the National Defense Research Committee and, as manager
of a division of the Corning Glass Works, he was responsible for the design,
construction, and operation of the largest optical glass plant ever built. He
invented a new family of glasses with high refractive index and wide disper-
sion especially adaptable for photographic lenses, which have been extensively
used by the Army in aerial cameras. All these contributions to industry, tech-
nology, and military objectives were based on replacing empiricism with pre-
cise knowledge of scientific principles. Dr. Morey is the author of an American
Chemical Society monograph, The Properties of Glass, now in its second edi-
tion, a treatise that stands alone in its field, and of some hundred scientific
papers.

Dr. Milton L. Humason, who came to the Mount Wilson Observatory in
1917, retired on June 30, 1957. Beginning as a janitor at the Observatory, he
soon became a night assistant. Here he showed such gifts of observation that in
1922 he was made a regular member of the Staff of Investigators. His most
important contributions have been in the spectrographic study of very faint
and distant objects. It was largely from the measurements of velocity furnished
by his early spectrograms of distant galaxies that Edwin P. Hubble developed
his famous hypothesis of an expanding universe. In 1928, at the request of
Hubble, Humason began a long series of spectrographic observations of galaxies
designed to investigate the relation between the redshift and the apparent
brightness of the objects. These observations were continued over a period of
twenty-eight years, with the aid of ever newer and more sensitive photographic
plates and of modern spectrographs. After 1950, with the greater light-gathering
power of the Hale telescope, he was able to extend the record of the spectra of
galaxies to distances in space hitherto utterly unattainable. One of his particu-
lar achievements was the development of procedures for locating images ac-
curately on the slit of the spectrograph and holding them there through long
exposures. In the course of his extensive and precise observations he accumu-

REPORT OF THE PRESIDENT 29

lated redshift measurements of 620 separate galactic objects. For this work on
redshift measurements he was awarded the degree of Ph.D. honoris causa by
Lund University in Sweden in 1950.

From 1948 until his retirement Dr. Humason was Secretary of the Mount
Wilson and Palomar Observatories, effectively handling many of the adminis-
trative tasks in addition to his scientific work.

One of the most painstaking investigators of the Institution's astronomical
staff, Dr. Seth B. Nicholson, retired on June 30, 1957, after forty-two years of
service. Dr. Nicholson came to the Institution from Lick Observatory, where
he had already won recognition by the discovery of the ninth satellite of Jupiter
in 1914. Subsequently he detected three other satellites of the planet — the tenth
and eleventh in 1938 and the twelfth in 1951, all with the 100-inch Hooker
telescope on Mount Wilson. Using the 48-inch schmidt camera on Mount
Palomar, he has recently taken a series of photographs of the region around
Jupiter, which he expects to use during the coming year to determine the rela-
tive brightnesses, positions, and magnitudes of all twelve satellites.

When he first came to Mount Wilson, Nicholson collaborated with the Direc-
tor of the Observatory, Dr. George Ellery Hale, in investigations of the sun.
Ever since, his work has been chiefly concerned with solar astronomy, more
particularly with the phenomena of the visible surface of that star. He has car-
ried out extensive studies of the magnetic polarity of sunspots.

Karl Ruppert, a staff member of the Department of Archaeology, retired on
October 1, 1956. He joined the Institution group carrying on research in Middle
American archaeology in 1925 and took an active part in the excavation and
restoration of Maya buildings at Chichen Itza in Yucatan. Subsequently he
made explorations in many other parts of the Yucatan Peninsula, some of which
had never before been entered by archaeologists.

During World War II Ruppert took leave of absence, first for War Depart-
ment activities connected with the American Legation at Guatemala City and
later for ambulance work with the American Field Service in India, Burma,
Italy, and Germany.

In the winter of 1947 he was in charge of a joint Carnegie Institution-United
Fruit Company expedition to Bonampak, Chiapas, Mexico, the archaeological
site that has become famous in recent years for its magnificent Maya mural
paintings. Ruppert has described the location of the site, the history of its dis-
covery, and its architecture in Bonampa\, Chiapas, Mexico, written in col-
laboration with Thompson and Proskouriakoff.

In 1950 the Department of Archaeology began its survey of the ruins of
Mayapan. In this project Ruppert, with A. L. Smith, conducted extensive
excavations of the remains of dwellings, obtaining a rich collection of data the
analysis of which should significantly add to our knowledge of the domestic

30 CARNEGIE INSTITUTION OF WASHINGTON

architecture, the patterns of settlement, and the size and character of the popu-
lation of Mayapan and its vicinity.

Another member of the staff of the Department of Archaeology, Gustav
Stromsvik, retired on June 30, 1957. Stromsvik was first employed in 1926 as a
carpenter at Chichen Itza. He soon acquired such knowledge of that site and
such skill in the engineering aspects of excavating, restoring, and protecting
the ruins as to become one of the most valuable staff members in the field. In
1933 he carried out comprehensive repairs in the Temple of the Phalli at
Chichen Itza, guaranteeing its preservation after excavation. From 1935 to
1942 he was in charge of the joint undertaking of the Government of Honduras
and the Carnegie Institution to excavate, restore, and preserve the highly sig-
nificant ruins at Copan. Here he re-erected the extraordinary monuments of
that city, repaired several of its most important buildings, and restored the great
hieroglyphic stairway. In the course of this program it became necessary to
divert the course of the Copan River to prevent its violent summer floods from
undermining the ruins — a major engineering undertaking.

Stromsvik enlisted in the Royal Norwegian Navy in 1943 and subsequently
took part in the landing of Allied forces on the Normandy beachhead. When
he returned to the Institution in 1945 he resumed his diverse activities of ex-
ploring, mapping, restoring ruins, and collecting artifacts. He participated in
discussions in Honduras leading to the establishment of the Instituto Nacional
de Antropologia e Historia there, for which he served as a Technical Adviser.
He has now returned to Norway, where he plans to continue his archaeological
investigations.

For the past twenty years the Carnegie Institution has been particularly
fortunate in having as its editor Miss Dorothy Swift, who retired on June 30,
1957. The importance of a gifted editor in any publications program cannot
be overemphasized. Miss Swift's contributions to lucidity and exact expression
won the respect and admiration of the staff. Her high standards and her
meticulous attention to perfection in detail are reflected in the publications of
the Institution during the years of her service. Miss Swift plans to continue
her editorial work in Boston.

The death of Dr. John von Neumann, brilliant mathematician and member
of the Atomic Energy Commission, on February 8, 1957, cut short an associa-
tion which the Institution was deeply privileged to enjoy. He had been ap-
pointed a Research Associate on June 1, 1955.

Dr. von Neumann was born in Budapest and received his early training there
and at Zurich. He came to the United States in 1929 as a visiting lecturer at
Princeton University, where the following year he was made professor of

REPORT OF THE PRESIDENT 31

mathematical physics. Three years later he received appointment as one of the
first full professors at the Institute for Advanced Study.

There were few fields of physics or mechanics untouched by von Neumann's
genius. His pioneering investigations of the phenomena of weather are well
known. Perhaps the impact of his gifts was most widely felt in the field of
military armament. His discovery and development of the implosion method
were critical to the development of the first atomic bomb. He was one of the
principal advisers to the United States Air Force and a powerful exponent of
the intercontinental ballistic missile. Von Neumann's contributions to the
design of early fusion weapons were equally significant. He was the recipient
of many honors, among them the Medal of Freedom and the Enrico Fermi
award for his work on the theory and design of computing machines. In
October 1954, he was appointed to the Atomic Energy Commission and served
actively with it, in spite of his illness, almost to the time of his death.

Few losses to contemporary science have been as great as this premature and
tragic ending of the career of a man of extraordinary genius.

On May 14 of this year Dr. Francis G. Benedict, formerly Director of the
Institution's Nutrition Laboratory, died at the age of 86. He began his re-
searches in human nutrition at Wesleyan University under grants of the
Carnegie Institution. In 1907 he became Director of the Institution's Nutrition
Laboratory in Boston and served in this capacity until his retirement in 1937.

The activities of the Laboratory during Dr. Benedict's thirty years as Director
were concentrated largely on establishing standards of basal metabolism of
normal human subjects according to height, weight, age, sex, and race. Special
researches were made on the conditions that may affect basal metabolism, such
as position of the body, temperature, vegetarian diet, athletic activity, environ-
ment, season, and fatigue. Notable contributions were made to the invention
and testing of various types of apparatus for measuring heat production, heat
elimination, respiratory exchange, and surface and internal body temperature.

Henry Norris Russell, who died on February 18, 1957, in his eightieth year,
had long been one of the leading astronomers of this country. A brilliant scien-
tist with an encyclopedic knowledge of astrophysics, a man of inexhaustible
enthusiasms, varied interests, and boundless energy, he exercised an inspiring
influence on all with whom he came in contact.

Although his career was centered in Princeton, he had associations with the
Institution from its earliest days, first as a Research Assistant and later, from
1921 to 1945, as a Research Associate. He worked in England at the Cam-
bridge Observatory from 1903 to 1905. In 1911, at the early age of thirty-four,
he became director of the observatory at Princeton. From 1921 until the late

32 CARNEGIE INSTITUTION OF WASHINGTON

1940's the Institution had the benefit of his presence for a part of practically
every academic year for research and lectures at Mount Wilson.

Dr. Russell's early investigations in the field of stellar constitution and evolu-
tion are of major significance. The Hertzsprung-Russell diagram of the rela-
tion between absolute magnitude and spectral type of stars has played a most
important part in the advance of astrophysics. During the first World War he
contributed to military research in the field of airplane navigation. He was
a pioneer in the analysis of complex laboratory spectra. The role of spectra in
the interpretation of the physical characteristics of stars absorbed him for several
years. His discovery that hydrogen was by far the most abundant element in
the atmosphere of the sun ran contrary to the accepted belief of the day. It
is now a basic fact of cosmology.

Dr. Arthur S. King died on April 25, 1957, at the age of 81. He was one of
the early Research Assistants of the Institution, receiving a grant in 1904 to
investigate emission spectra at high temperatures at the Universities of Bonn
and Berlin. On January 1, 1908, after teaching physics at the University of
California for three years, he was appointed Superintendent of the new Physical
Laboratory of the Mount Wilson Observatory, which was equipped with a
30-foot spectrograph and a large electric furnace designed by him, apparatus
superior to anything then in existence for spectroscopic work.

Over the years until his retirement on February 1, 1943, Dr. King concerned
himself with laboratory investigations of many spectra. His study of the tem-
perature classification of the spectra of chemical elements, as observed in the
arc, spark, and electric furnace, had a far-reaching effect, both on the analyses
of complex spectra and on the interpretation of the solar spectrum, the sunspot
spectrum, and stellar spectra in general. His research also included the study
of the spectra of several of the rare earths (atomic numbers 57 to 71), and he
completed the wavelength measurements and intensity estimates for a great
number of lines in these complex spectra. His laboratory observations, in
collaboration with those of Raymond T. Birge, led to the extremely important
discovery of the carbon isotope of mass 13.

During World War II, Dr. King investigated the velocities of aerial torpedoes
for the Office of Scientific Research and Development at the California Insti-
tute of Technology, and from 1946 to 1954 he served as a mathematician with
the Naval Ordnance Test Station in Pasadena.

• • • and Gains

In consequence of a generous gift from the Carnegie Corporation of New
York, the Institution is initiating this year a series of special fellowships in the
natural sciences, primarily to permit of travel and visits of distinguished scholars
to departments of the Institution carrying on research in their fields of interest.

REPORT OF THE PRESIDENT 33

Such awards are entirely without restriction with regard to their specific use,
but are made in general in subject fields to which the Institution is especially
directing its attention.

Current recipients of the new fellowships are Dr. Evelyn M. Witkin, of the
College of Medicine, State University of New York; Dr. William A. Arnold,
of the Oak Ridge National Laboratory, Oak Ridge, Tennessee; Dr. R. v. d. R.
Woolley, the Astronomer Royal of Great Britain ; and Professor Jan Hendrick
Oort, the Director of the Observatory of Leyden in the Netherlands.

It is a particular pleasure to report that Dr. Vannevar Bush, retired President
of the Institution, was elected Chairman of the Corporation of the Massachu-
setts Institute of Technology on March 4, 1957. On May 21, the New Jersey
Patent Law Association in Newark presented him with the 1957 Jefferson
medal for his notable contributions to the United States patent system.

It gives me great pleasure to announce the following honors that have been
received by directors and members of the staff.

The Catherine Wolfe Bruce gold medal for 1957 for distinguished services
to astronomy was awarded to Dr. Ira S. Bowen, Director of the Mount Wilson
and Palomar Observatories, by the Astronomical Society of the Pacific. Dr.
Jesse L. Greenstein, staff member of the Observatories, was elected to the Na-
tional Academy of Sciences. Dr. Milton L. Humason was elected an associate
of the Royal Astronomical Society. Dr. Harold D. Babcock, retired staff mem-
ber, received the degree of Doctor of Laws honoris causa from the University
of California.

At the Department of Plant Biology, Dr. Jens Clausen, retired staff member,
was awarded the degree of Doctor of Agronomy honoris causa by the Royal
College at Upsala, Sweden, on May 31, 1957, in connection with a celebration
of the 250th anniversary of the birth of Linnaeus.

Dr. M. Demerec, Director of the Department of Genetics, was elected a
member of the Royal Danish Academy of Sciences and Letters on April 6, 1956,
and on February 3, 1957, he received the degree of Doctor of Laws honoris
causa from Hofstra College, Hempstead (Long Island, New York).

The Geophysical Institute of Huancayo, Peru, has been renamed in honor
of the late Dr. John A. Fleming, long associated with the Department of Ter-
restrial Magnetism, as the "Geophysical Observatory John A. Fleming."

Caryl P. Has\ins

REPORTS OF DEPARTMENTS

and SPECIAL STUDIES

MOUNT WILSON AND PALOMAR OBSERVATORIES

COMMITTEE ON IMAGE TUBES FOR TELESCOPES

DEPARTMENT OF TERRESTRIAL MAGNETISM

GEOPHYSICAL LABORATORY

DEPARTMENT OF PLANT BIOLOGY

DEPARTMENT OF EMBRYOLOGY

DEPARTMENT OF GENETICS

DEPARTMENT OF ARCHAEOLOGY

MOUNT WILSON & PALOMAR OBSERVATORIES

Operated by the Carnegie Institution of Washington

and the California Institute of Technology

Pasadena, California

IRA S. BOWEN, Director
HORACE W. BABCOCK, Assistant Director

OBSERVATORY COMMITTEE

Ira S. Bowen, Chairman Robert S. Bacher

Walter Baade Jesse L. Greenstein

Horace W. Babcock Ernest C. Watson

CONTENTS

page

Introduction 39

Observing Conditions 40

Solar Research 40

Solar photography 40

Sunspot activity 40

Magnetic polarities 40

Studies of geomagnetic activity 41

Solar magnetic fields 41

Planets and Satellites 42

Observations of Mars 42

Spectra of Venus 43

Satellites 43

Comets 43

Stellar Spectroscopy and Photometry .... 44

White dwarfs and subdwarfs 44

Late-type stars 46

Variable stars 48

Magnetic stars 49

Globular and galactic clusters and stel-
lar evolution 50

page
Chemical composition of stellar atmos-
pheres 52

Nuclear reactions in stars 54

Gaseous Nebulae 55

Internal motions and radial velocities. 55

Densities of nebulae 57

Galaxies 58

The Andromeda galaxy and other

members of the local group 58

Studies of individual galaxies 59

Catalogues and statistics of galaxies. . 60

Clusters of galaxies 60

Velocities and distances of galaxies. . . 61

Radio Sources 63

Identification of radio sources 63

Instrumentation 64

Guest Investigators 64

Staff and Organization 70

Bibliography 72

Carnegie Institution of Washington Year Boo\ 56, 1956-1957

INTRODUCTION

In 1908, Hale, using the newly completed
spectrograph of the 60-foot solar tower on
Mount Wilson, noticed that many of the
lines of the spectra of sunspots were split
into components which had the character-
istic polarization of a Zeeman pattern.
This observation provided the first definite
evidence for the presence of a magnetic
field in an astronomical body. Later Hale
and his collaborators searched for a gen-
eral magnetic field of the sun, but the field
was so weak that Hale was never satisfied
with the conclusiveness of the evidence.

The large velocity of axial rotation at-
tributed to nearly all A-type stars suggested
to Horace W. Babcock that relatively large
general magnetic fields might exist in these
objects. Early in 1946 he made the neces-
sary spectroscopic observations of one of
these stars, 78 Virginis, with the 100-inch
coude spectrograph, and discovered a gen-
eral magnetic field of between 1000 and
2000 gauss. Since that time some hundreds
of stars have been under observation with
the coude spectrographs of both the 100-
inch and the 200-inch. A few score of the
most interesting objects have been ob-
served many times in order to study the
fluctuations that occur in the magnetic
fields. Many of the variations are irregular,
but in some stars the field reverses with a
regular period of a few days. During the
present year Babcock has collected for
publication all the thousands of measure-
ments of stellar magnetic fields made in
the past 11 years. Eighty-four of the stars
listed show definite evidence for a mag-
netic field; 55 are probably magnetic; an-
other 55 show no evidence for a coherent
magnetic field. One hundred and eleven
of the stars observed had lines too broad
for magnetic measurements.

Encouraged by the discovery of these
general fields of rapidly rotating stars,
Harold D. Babcock and Horace W. Bab-
cock returned to the problem of the gen-
eral magnetic field of the sun in 1952.

Taking advantage of improved gratings
and of current advances in photoelectric
and electronic techniques, they constructed
a solar magnetograph which scans the sur-
face of the sun and records the amount of
the residual polarization at the side of a
certain sensitive spectral line caused by the
Zeeman splitting of the line in the mag-
netic field. Installed in the Hale Solar
Laboratory in Pasadena, the magnetograph
immediately gave definite evidence for
magnetic fields of the order of 1 gauss over
large areas of the solar surface. Further-
more, the pattern of many of these fields
was found to change in the course of a few
days; this finding explains why the results
obtained by Hale and the other early ob-
servers were so erratic and inconclusive.
Recently a second, improved magneto-
graph has been completed for the 150-foot
solar tower on Mount Wilson. Starting at
the end of the report year, a daily record of
the distribution of the magnetic fields over
the sun's surface is being made as a regular
part of the solar program.

Meanwhile indirect evidence has been
found at other observatories for small inter-
stellar magnetic fields extending through-
out large volumes of space. In particular
the discovery in the USSR, by Dombrovsky
and by Vashakidze, of polarization in the
Crab Nebula led to the explanation that
the continuous radiation from this object
was synchrotron radiation caused by very-
high-velocity electrons being accelerated
in a magnetic field. Baade with the aid
of the 200-inch was able to study the de-
tailed structures of the polarization in the
Crab Nebula and also to observe polariza-
tion in the radiation of the jet in M 87
which probably has a similar cause.

All these observations point to the fact
that magnetic fields are very widespread
and play a much larger role in astronomi-
cal phenomena than has hitherto been sup-
posed. They also mean that any attempt

39

40

CARNEGIE INSTITUTION OF WASHINGTON

to explain the structure of stars and nebu-
lae must more and more take into account
the forces exerted on the ionized gas or
"plasma" associated with these fields. Be-
cause of the growing importance of mag-
netic phenomena in astronomy a Sym-
posium on Electromagnetic Phenomena in

Cosmical Physics was held in Stockholm
on August 27-31, 1956, under the auspices
of the International Astronomical Union.
Three representatives from the Mount
Wilson and Palomar Observatories at-
tended, including H. W. Babcock, who was
one of the organizers of the Symposium.

OBSERVING CONDITIONS

For the fifth consecutive year, precipita- ond half of the year. Solar observations

tion was below normal on Mount Wilson, were made on 323 days; observations were

the total for the year being 25.47 inches, made with the 100-inch on 292 nights, and

Nevertheless, the number of cloudy days with the 60-inch on 285 nights,
and nights was large, especially in the sec-

SOLAR RESEARCH

Solar Photography

Solar observations were made by Nichol-
son, Hickox, and Cragg. The numbers
of photographs of various kinds taken be-
tween July 1, 1956, and June 30, 1957, were
as follows :

638
567
936
921
78,000
942

Direct photographs

Ha spectroheliograms, 60-foot focus
Ha spectroheliograms, 18-foot focus
K2 spectroheliograms, 18-foot focus
K2 spectroheliograms, 7-foot focus.
K prominences, 18-foot focus.

Sunspot Activity

The magnetic classification and study of
sunspots and related phenomena have been
continued by Nicholson and Cragg. Co-
operative programs have been carried out
with the U. S. Naval Observatory, the
University of Michigan, the Observatory
of Kodaikanal, the Meudon Observatory,
and the Central Radio Propagation Labo-
ratory of the National Bureau of Stand-
ards. During the calendar year 1956, solar
observations were made on 348 days, on
none of which was the sun without spots.
Only in 1953 were observations made on
more days than in 1956. The total num-
ber of spot groups observed in 1956 was
642, compared with 208 in 1955 and 46 in
1954. The record number of groups ob-

served here in one year is 663 in 1947. The
northern hemisphere was again the more
active, with 53 per cent of the total groups.

An unusual number of spots have ap-
peared in very high latitudes in this cycle.
In the 65 years from 1878 to 1943, during
which time daily solar photographs are
available, only four groups that lasted
longer than a day appeared on the sun
farther than 40° from the equator. In the
last cycle, 1943 to 1954, eight such groups
were observed. In the present cycle thir-
teen have already appeared.

The rate of increase of solar activity from
1955 to 1956 greatly exceeded that from
1945 to 1946. If the number of spots con-
tinues to increase for another year, as it
did at a similar phase last cycle, the com-
ing sunspot maximum will certainly be
the highest ever recorded. Since maximum
activity is expected in 1957 a new record
could be set.

The monthly means of the number of
groups observed daily for the past two and
one-half years are shown in table 1.

Magnetic Polarities

Magnetic polarities in each spot group
have, if possible, been observed at least
once. The classification of groups observed
between July 1, 1956, and June 30, 1957, is
indicated in table 2. "Regular" groups in

MOUNT WILSON AND PALOMAR OBSERVATORIES

41

the northern hemisphere are those in
which the preceding members have N
(north-seeking) polarity; in the southern
hemisphere the polarities are reversed.

TABLE 1

Month

Daily Number of Sunspot Groups

1955

1956

1957

January 2.2

February 2.2

March 0.6

April 1.4

May 2.3

June 2.4

July 3.2

August 3.3

September 4.9

October 6.1

November 7.3

December 6.9

6.4

12.7

9.3

10.9

10.6

13.6

10.8

14.0

9.9

12.7

9.3

15.1

9.3

13.5

15.2

12.9

11.8

13.8

Yearly mean

3.6

10.2

TABLE 2

Hemisphere Regular Irregular Unclassified

North

. 259

7

80

South

312

5

96

Whole sun .

. 571

12

176

Studies of Geomagnetic Activity

The study of the diurnal variation of
geomagnetic activity on very disturbed
days, reported last year, has been extended
by Nicholson and Dr. Oliver R. Wulf, of
the U. S. Weather Bureau, to include the
ten selected quiet days in each month and
the all-except-quiet days. The local-time
component for the quiet days, which is
quite significant, seems to be related to
deviations from the normal diurnal change
in the intensity of the earth's field. This
effect, which changes the phase of the
local-time component for all days, has been
removed from the data. The remaining
local-time component of disturbance re-
flects the known local-time dependence of

magnetic "bays." The additional Universal-
Time component indicates that through-
out the year disturbance is greater when
the sun is over certain longitudes, follow-
ing a pattern like that near the equinoxes,
but that the variations are influenced by
the north-and-south position of the sun as
well. This effect may arise from the chang-
ing position of the ionization cap under
the sun relative to the earth's magnetic
field and to the changing patterns of the
large-scale atmospheric circulation.

Solar Magnetic Fields

In the past year, 160 magnetograms of
the sun have been made by Harold D. Bab-
cock; a total of 786 has been obtained since
the series was begun in 1952. Compilation
of additional data, not photographically
recorded, up to the end of 1956, confirms
published results on signs and intensities
of the high-latitude polar fields of the sun.

Measurements have been made on weak
coherent fields of great area for 17 dates
between December 24, 1953, and March 14,
1954. For 29 of these fields the following
mean values are found: maximum in-
tensity, 1 gauss; total magnetic flux, 1021
maxwells; areas, 0.14 of the sun's disk.

Measurements of magnetic flux, maxi-
mum intensity, and area have been made
for fields of each sign in both sunspot
zones for 21 days between October 8, 1955,
and May 13, 1956. This interval includes
the most notable outburst of activity in the
first 18 months of the present solar cycle.

In 22 selected bipolar magnetic regions,
each dominated by one sunspot, the flux,
maximum intensity, and area have been
measured. In each case, the maximum flux
of the spot umbra has been estimated, for
a comparison between the flux emerging
from the spot and the flux returning
through a photospheric region of opposite
polarity.

A comparison of magnetograms with
charts of coronal activity issued by the
High Altitude Observatory, Boulder, Colo-

42 CARNEGIE INSTITUTION OF WASHINGTON

rado, based on emission of AA5303, 6374,
shows a general correspondence between
magnetic flux and light flux. In every
case, the areas on the sun are much more
extensive in longitude than in latitude.
After putting the new solar magneto-

graph at the 150-foot solar tower on Mount
Wilson into regular operation at a high
level of stability and sensitivity, Vrabec
has used it to map magnetic fields in the
neighborhood of sunspots and to study
their variations.

PLANETS AND SATELLITES

Observations of Mars

Mars was observed at the opposition of
1956 by Richardson on 34 nights for a
total of 110 hours. About twice as much
time was spent on the planet as in 1954,
under generally superior conditions. Only
part of the material secured has been re-
duced as yet.

Four hundred and ten direct photo-
graphs of Mars were taken in blue, orange,
red, and infrared light with a camera and
enlarging lens at the Cassegrain focus of
the 60-inch. The diameter of the images
ranged from 2 to 10 mm. Eastman IV F
emulsion was used throughout the visual
region, and IV N emulsion in the infrared.
Exposure times ranged from J/3 to 20
seconds.

The direct photographs, recording the
appearance of the planet at the time of
observation, can be used for photometry
of the disk. From measures on the posi-
tions of the markings the oblateness of the
planet can be determined, since the period
of rotation is accurately known.

The effective wavelength of the red
images is at 6600 A, at about the position
of the chlorophyll absorption band. The
effective wavelength of the infrared images
is at 8000 A, in the region where chloro-
phyll reflects strongly. Photographs taken
a few minutes apart show the intensity of
the maria relative to the deserts as the
same in both red and infrared. The maria,
therefore, cannot consist of vegetation with
a high chlorophyll content, unless the re-
flectivity of chlorophyll on Mars differs
markedly from that of terrestrial vegeta-
tion.

Spectra of Mars were taken by Richard-
son on 17 nights with the 114-inch camera,

using the grating which gives a dispersion
in the first order of 5.6 A/mm.

Plates in the first order infrared were
taken in a search for the carbon dioxide
bands at A7820 and A8688, first discovered
in the spectrum of Venus. Carbon dioxide
has already been identified in the Martian
atmosphere through the band at 1.6 mi-
crons. The bands at A7820 and A8688 do
not show in the telluric spectrum with a
very low sun. There was no trace of these
bands in the spectrum of Mars. Their ab-
sence should allow us to set an upper limit
to the quantity of carbon dioxide in the
atmosphere of Mars.

The A band of oxygen at A7600 was
photographed in the spectrum of Mars in
July and December 1956, when the velocity
of the planet relative to the earth was —8.9
and +14.9 km/sec, respectively. This
velocity gives a shift of 0.62 A, correspond-
ing to 0.11 mm on the plates. Previous
measures to detect oxygen have been made
on the (1,0) B band at A6900. The (0,0)
A band would seem better suited to this
investigation, owing to the greater intensity
of possible Martian components. Also, the
isotope lines in the band of 016018 arise
from molecules about 0.004 as abundant
as those producing the principal lines, thus
affording a basis of comparison for the
intensity of Martian lines.

The presence of water vapor might be
detected from variations in intensity of the
H2O lines in the maria and polar cap rela-
tive to the adjacent deserts. Inspection of
the water-vapor lines on several spectra in
the region of A8200 showed no variations
in intensity that seemed significant.

Direct photographs of Mars in visual
light show the surface markings distinctly.

MOUNT WILSON AND PALOMAR OBSERVATORIES

43

Those in blue and violet usually show only
a blank disk with a greatly enlarged polar
cap.

Preliminary measures by Richardson on
the spectra from 7500 to 8500 A give a
ratio maria/desert « 0.78. No spectrograms
were taken in the red, but the direct photo-
graphs indicate that this ratio remains
nearly constant from 6000 to 7500 A. Be-
ginning about A5700 the ratio maria/desert
^0.84, and increases gradually until at
about A4400 the maria and deserts become
indistinguishable.

The polar cap is inconspicuous in the
infrared, increases steadily in intensity in
the visual region relative to the deserts, and
is apparently still increasing at the limit
of our spectra at A3600.

Spectra of Venus

Thirteen spectra of Venus were taken
by Richardson for the measurement of
rotation in the A6300 region with the 114-
inch camera in the second order with a
dispersion of 2.8 A/mm. Two spectra of
Mars were taken as controls. Fifteen spec-
tra of Venus have also been taken at the
Snow telescope with a dispersion of 0.75
A/mm. Only a few exploratory measures
have been made on these plates.

Satellites

Perturbations in the motions of the
satellites furnish one of the best means of

determining the oblateness and inner con-
stitution of Mars. A dozen of Richard-
son's best plates which show both Phobos
and Deimos have been sent to the U. S.
Naval Observatory at their request for
measurement and reduction.

Photographs of the satellite field around
Jupiter were taken by Nicholson with the
48-inch schmidt on several nights in April,
May, and June with corresponding ex-
posures on Selected Areas to redetermine
the magnitudes of the faint satellites in a
homogeneous manner. Positions of the
satellites will also be measured on these
plates, each of which covers the entire
satellite field, and their search should re-
veal all of Jupiter's satellites down to the
limiting magnitude of the plates. The
search, measurement, and reduction of
these plates, the last of which were taken
in June, have just been started. A pre-
liminary search of the plates taken in April
located all the known distant satellites with
JXII near the plate limit. No new satellites
were found, but some fifty asteroids were
detected in the search.

Two pairs of plates of Saturn taken in
June by Osterbrock with the 48-inch
schmidt have been searched for satellites
by Nicholson. No new satellites were
found. Among the asteroids detected, one
has the motion characteristic of the Trojan
asteroids, but further observations will be
necessary to determine its orbit.

COMETS

Reduction was completed, under the
supervision of Osterbrock, of a two-year
series of Comet Baade (1954h) and Comet
Haro-Chavira (1954k). These comets were
at a distance of approximately 4 astro-
nomical units from the sun during the
whole period of observation, and the plates
were taken with the 48-inch schmidt in
order to study the comets' tails at these
great distances. The position angle of the
projection of the comet tail on the sky was
measured on each plate; the reductions
show that this angle is not at all strongly
correlated with the position angle of the

projection of the line from the sun to the
comet, but is strongly correlated with the
position angle of the projection of the
comet orbit behind the comet. The reduc-
tions further show that, on the reasonable
assumption that the comet tail is in the
orbital plane of the comet, it lies approxi-
mately halfway between the line from the
sun through the comet and the line along
the orbit behind the comet. This indicates
that, unlike most of the well known bright
comets observed close to the sun, in which
radiation pressure dominates the direction
of the tail, in these more distant comets

44

CARNEGIE INSTITUTION OF WASHINGTON

the pressure of the interplanetary matter
is roughly as important as radiation pres-
sure in its effects on the tail. As the density
of interplanetary material is not believed to
be particularly high at great distances from
the sun, presumably the material in the

tails of these distant comets is relatively
ineffective in absorbing solar radiation.

An extensive series of photographs of
comet Arend-Roland (1956h) were taken
with the 48-inch schmidt camera by
Nicholson, Minkowski, and Schmidt.

STELLAR SPECTROSCOPY AND PHOTOMETRY

Spectroscopic observations have been
regularly carried out with all major tele-
scopes except the 48-inch schmidt during
the moonlit half of each month. Seven
hundred spectrograms were made with the
200-inch telescope, 827 with the 100-inch,
and 978 with the 60-inch during the year.

White Dwarfs and Subdwarfs

The survey of white-dwarf spectra by
Greenstein now includes about 60 objects.
Hydrogen-line profiles were obtained for
25 normal objects of type DA. The use of
the prime-focus spectrograph of the Hale
telescope permits spectra, well widened,
to be obtained at the dispersion of 160
A/mm down to 15.5 mag. Fainter stars
seem difficult during the present sunspot
maximum because of the increased bright-
ness of the airglow, whose emission lines
are now easily detectable on exposures
longer than an hour. Various groups of
peculiar objects have been discovered. The
star HZ 29 has extremely shallow lines
of He I. The star HZ 21 seems to be one
of the hottest probable white dwarfs, of
spectral type dO. It shows very strong
lines of He I and He II. A new type of
yellowish white dwarf, represented by
W 219 and L 879-14, has been found to
show an unidentified broad absorption
near A4670. A recurrent nova, WZ Sagittae
(1913, 1946), has broad hydrogen absorp-
tion lines similar to those in a normal
white dwarf underlying its broad emission
bands. Its luminosity derived from proper
motion is near + 10, so that it seems to be
the expected link between the white dwarf
and the nova stage of evolution.

Most old novae so far observed have rela-
tively narrow emission lines, presumably

arising from post-nova ejection of matter
from the star. Their blue continuous spec-
trum shows no distinct sharp or broad ab-
sorption lines. Nova DQ Herculis has
relatively broad emission lines, with the
higher members of the Balmer series varia-
ble through the 4-hour cycle. A first theo-
retical analysis of the observations of white-
dwarf spectra has been made by Greenstein
with the goal of obtaining a temperature
scale largely based on colors and the central
depth of the hydrogen line. The colors
should be very reliable for the white
dwarfs with nearly continuous spectra.
On the assumption that the interior com-
position is largely helium or heavier ele-
ments, the mean radius deduced for the
white dwarfs is 0.01 solar radius, which
corresponds to about 0.6 solar mass. The
observations of the dependence of line
strength on color demonstrate clearly that
the ratios of abundance of hydrogen to
helium to metals are variable from one
type of white dwarf to another. At ap-
proximately the same color, white dwarfs
can be found with Balmer lines, helium
lines, or essentially continuous spectra.
Possibly another opacity source exists be-
sides hydrogen; otherwise it seems difficult
to understand the hot objects with continu-
ous spectra. A method of observation of
broad features has been developed which
reliably reveals absorptions of 5 per cent
or less in depth, over a wide variety of
temperatures.

A method of analyzing the spectra of
O-type subdwarfs has been developed and
applied by Guido Munch to some of the
stars discovered earlier by him. Essentially,
it is based on the fact that, over a relatively
wide range of effective electron pressure

MOUNT WILSON AND PALOMAR OBSERVATORIES 45

Pe and temperature T, for a given H:He
abundance ratio, the intensity ratios be-
tween lines of He II and H, and between
lines of N III and N II, define the same
curve in a (Pe, T ) plane. An additional
characteristic in this plane independent of
the abundances, as is the Inglis-Teller cri-
terion, then defines both Pe and T. The
analysis of the stars HZ 44 and BD +25°
4655 gives a number abundance ratio H :He
= 1:4, while the relative number abun-
dances of the heavy elements represented
in the spectra with measurable intensi-
ties are N:C:O:Ne:Si = 100:0.5:6:20:6.
The effective temperature is found to be
35,000° K, and the surface gravity around
106-8. No direct information is available
regarding the absolute magnitudes of these
objects, but for assumed masses between
1 O and 15 O, the corresponding visual
absolute magnitudes range from +6.0 to
+ 3.0, respectively. The observational ma-
terial available on the six stars of this type
known at present is now being analyzed.

The subdwarf nature of the blue star
BD +25°2534, found by Dr. A. Slettebak
on objective-prism plates taken at Ham-
burg-Bergedorf, has been verified by G.
Munch on 18 A/mm plates taken at Palo-
mar. Although the star had been observed
before by Munch and Greenstein, it had
not been noticed that, besides the wide and
strong Balmer lines, sharp He I lines and
a shallow diffuse line of He II at A4686
also appear. This observation points to
the existence of a pseudophotosphere with
temperature not as high as would be indi-
cated by the color, where the Balmer lines
originate, and of an underlying hot core,
where the high-temperature color and
spectral characteristics arise. Such a model
was suggested in the last annual report by
G. Munch to explain the spectrum and
color of the blue stars below the horizontal
branch of the H-R diagram of globular
clusters.

Feige has completed the project for the
discovery of faint blue stars on the Na-
tional Geographic Society-Palomar Ob-
servatory Sky Survey plates near the south

galactic pole, and most of the area of the
north galactic pole. Thus far a total of 180
faint blue stars between 11 and 16 mag.
have been found. Some spectra show that
a considerable number of these are the
so-called "halo" blue stars. Others, how-
ever, are helium-rich subdwarfs, and three
definite new white dwarfs are included.

Code has developed a photoelectric scan-
ning spectrograph that has been adapted
for both the 100-inch and 200-inch tele-
scopes and has been used for a detailed
study of the energy distribution in the
spectra of several subdwarfs. All the
energy distributions in the infrared of the
subdwarfs HD 19455, HD 140283, and HD
201891 correspond to lower temperatures
than the temperature inferred from the
spectral type or the B — V color. These
results are to be understood in terms of
weak lines presumably resulting from a
low metal abundance. The infrared, where
line blanketing is very small, provides a
more significant measure of the surface
temperature. Relative to normal main-
sequence stars, the smaller line blanketing
in the subdwarfs leads to a moderate ex-
cess in the B — V color and to the larger
well known U — B excess. Specifically,
HD 19455 has a B — V color corresponding
to an F6 dwarf and an assigned spectral
type of sdF7, while the spectral distribution
in the infrared is that of a GO dwarf. The
subdwarf HD 140283 is similar but some-
what more extreme — a result consistent
with the slightly weaker lines as measured
in its spectrum.

In order to extend the results obtained to
a larger sample of stars a photometer
utilizing a red-sensitive photomultiplier
was adapted by Code for use on the 100-
inch reflector. Several wavelength regions
were isolated by means of glass filters and
with interference filters. It was found that
for stars with the same infrared color index
the weak-line objects averaged about 0.1
mag. bluer in the B — V index and 0.25
mag. bluer in the U — B index.

These results have two important rami-
fications. First, apparently a subdwarf is

46 CARNEGIE INSTITUTION OF WASHINGTON

not as underluminous as was previously year or more, and the velocity amplitudes

thought; second, any interpretation of a rather small, the spectrograms now at hand

color-magnitude array must be made with do not allow any conclusions. On the pos-

due regard for the influence that line sibility that H3 and K3 are produced in

blanketing may have upon the color index, circumstellar envelopes in G and K giants

as well, a few of these stars in binary sys-

Late-Type Stars terns are also being observed on this

Code has applied his photoelectric scan- program,
ning spectrograph to the study of the Coude spectra of the M2 + Ia+ super-
energy distribution of the high-velocity giant VV Cephei, made during the total
giant Arcturus and of two red giants in eclipse of its B-type companion, have re-
M 92. The differences between the ob- vealed both circumstellar and interstellar
served energy distribution and that of nor- absorption lines at H and K. The inter-
mal giants may be interpreted in the same stellar lines have a velocity and strength
manner as the subdwarf results reported appropriate to a distance of about 0.8
in the preceding section. kiloparsec, in good agreement with Kee-

In the spectra of most M-type giants and nan's recent luminosity classification. The

supergiants, the H and K lines of Ca II circumstellar lines are somewhat weaker

show deep absorption reversals (H3 and than was expected by analogy with other

K3) superposed on emission components equally luminous M stars. Possibly the B

(H2 and K2). The velocity from H3 and star has an effect on the envelope that

K3 always indicates that the absorption gas persists, in some degree, even during total

is expanding from the star. Depending on eclipse.

the height at which the lines are formed, The system of Mira is similar in some

the observed expansion velocity may ex- respects. The spectrum of the long-period

ceed the velocity of escape, and the matter variable shows a sharp absorption core at

will be ejected into the interstellar medium. K, with a strength about one-third that

This is true for the M5 II star a Herculis; of K3 in a Herculis A. This feature indi-

it and all other M-type supergiants are cates a variable velocity, which correlates

losing mass at a rate comparable with their with the velocity of the reversing layer but

evolutionary time scale. The question is generally about 2.5 km/sec less. A still

arises whether H3 and K3 in all M giants weaker absorption core at K has been

arise in envelopes so distended that the noted by Joy in the spectrum of Mira B.

observed velocities will suffice for escape. Its velocity, which appears to be constant,

If so, as it was shown last year by Deutsch, is only slightly less than that of the sharp

the observed number of white dwarfs can core in the M star. If both K cores are

be accounted for as the remnants of the produced in a common envelope, then the

massive stars that have condensed into the radius of the envelope must exceed 25

main sequence during the last 5 billion astronomical units,

years. The spectrum of the long-period variable

To investigate the dimensions of the re- Mira has been found to exhibit a number

gions responsible for H3 and K3 in ordi- of remarkable changes from cycle to cycle,

nary M giants, a number of spectroscopic As compared with the spectra of ordinary

binaries have been observed at 10 A/mm. M giants, the absorption features in Mira

If the gas that produces H3 and K3 is really tend to be the most nearly normal at bright

circumstellar, and escaping from the sys- maxima. At most faint maxima, however,

tern, it cannot follow the orbital motion of the atomic absorption lines are strongly

the M star; in this event, H3 and K3 should suppressed, especially in the region of the

be stationary, or nearly so. The periods of TiO bands. As the star fades, the intensi-

M-giant binaries being of the order of a ties of many of these lines change in

MOUNT WILSON AND PALOMAR OBSERVATORIES 47

opposite directions after maxima of these a theoretical derivation of the width-
two kinds. The effects are extremely com- luminosity relationship can be obtained,
plicated, but often very large. Presumably So far, additional spectrograms of the four
they are principally stratification effects, Hyades stars combined with the previous
but as yet they are not well understood. ones yield an average deviation for the

An attempt will be made by Deutsch method of 0.3 mag. Spectrograms, though

this summer to observe the rapid spectro- not of first quality, of the yellow giants in

scopic changes in Mira B found at recent Praesepe give an average deviation of 0.5

minima of the long-period variable. The mag. Further observations of both clusters

companion has been observed visually at should provide information on measuring

each of the last three minima, in roughly error, intrinsic scatter among stars of

the same configuration as at all other re- closely similar luminosities, and the mean

liable observations (p = 0'.'5, 0 = 130°). line width for these eight stars of known

These observations require the abandon- brightness. Even though the calibration

ment of a period near 14 years, as had of the K-line method is still incomplete,

been suggested by Parenago. Evidently it appears already to be adequate to pro-

the motion is very slow, and the mass of vide improved information on the yellow

the system very low. Another long-period and red giants in the solar neighborhood,

variable with a visual companion is X Under consideration are observations in-

Ophiuchi. Recent spectrograms of this tended to extend knowledge of luminosity

companion confirm that it is a subgiant functions, color-magnitude diagrams, and

which is much less luminous than normal masses (through the derivation of im-

K giants. These two systems suggest that proved parallaxes of binaries and physical

long-period variables are likely to be ex- pairs),

tremely old stars. Under the second category, the first

Wilson and M. K. Vainu Bappu have problem is to decide on the nature of the

completed the measurements of the line emitting layers. Wilson favors the view

width of the H and K emission lines in that they are probably optically thin be-

185 late-type stars. As mentioned in the cause of intensity differences between K

last report, the main result is that a plot of and H which can be seen in many stars,

the logarithms of the corrected line widths This evidence is not conclusive, however;

against visual absolute magnitudes is linear the layers may be optically thick. In any

over a 15-mag. range. Another finding is event, if the layers are optically thin, then

that neither surface temperature nor sur- the lines are widened purely by Doppler

face gravity can play much of a role in this effect, the measured widths are approxi-

functional relationship, which appears to mately equal to the Doppler widths, AAd,

be a relation between a chromospheric and the turbulence is large. Under these

phenomenon and the total energy produc- circumstances, AAd a LVq, and the only re-

tion deep within the star. striction on N (particles per square centi-

A number of important possibilities meter) is that it is never so large that the

opened by these observations can be di- optical thickness in the line center exceeds

vided into two categories: use of the rela- approximately 1.

tionship as a tool for the measurement of Dr. L. Goldberg, of the University of

absolute magnitude, and information that Michigan, has pointed out that, with op-

they provide in regard to the physical tically thick layers, AA, the measured

processes operating in the atmospheres of width, is approximately three times the

these stars. Doppler width, AAd. Thus the real turbu-

In considering the first category, the lence is only about one-third as large as

major problem is to refine the calibration, for thin layers. In this case, however,

which is a purely empirical matter unless calculation (unpublished) shows that both

48 CARNEGIE INSTITUTION OF WASHINGTON

AAd and N must vary with luminosity in cepheids in globular clusters and in the
a prescribed fashion in order to agree with galactic field. The line emission, and its
the observations. Thick layers are thus variations, resembles that in W Virginis.
more restrictive on any physical theory In certain globular-cluster cepheids the
seeking to explain the width-luminosity absorption lines double, but in others they
correlation. merely broaden at phases in which W Vir
Emission at He was first noted by Wil- showed line doubling. This material is at
son in the spectrum of Arcturus in 1938. present under discussion for the deter-
It appears on many spectrograms of the mination of the radius changes, spectral-
late-type stars used for the foregoing H type variations, and anomalous spectral
and K investigation. All these plates were behavior of the Population II cepheids.
carefully compared with one of Arcturus The R Coronae Borealis variable, RY
in order to make a preliminary study of Sagittarii, was found by Greenstein to
the intensity of the He emission as related show variable absorption-line doubling
to K emission strength, to spectral type, during small light-variations near its maxi-
and to absolute magnitude. The chief mum, with displaced components up to
results are that He intensity increases on —150 km/sec. Studies by Greenstein and
the average in the direction G— »K— >M, Merrill of the infrared spectrum of R
and that there seems to be a statistical cor- Andromedae showed that this S star had a
relation between emission strengths of K circumstellar envelope, visible in certain
and of He. There is not a one-to-one corre- lines, with about — 20 km/sec relative
lation between He and K, since some stars velocity.

with strong K intensity show no emission Zwicky continued to co-ordinate the
at He. Presumably both the Ca II and search for and the investigation of super-
hydrogen emission are of chromospheric novae conducted at the Lick, Steward,
origin, and they may well supply clues to Berne (Switzerland), and Palomar Ob-
physical properties of the stellar chroma- servatories with the aid of funds from the
spheres. National Science Foundation and the
Variable Stars Swiss Government. During the year two

_ .. . . . bright supernovae were discovered, one in

Radial-velocity measurements were com- NGC 2m . g^-^, ;Q Berne> and the

pleted on spectrograms of the cluster-type othef b M g4 fe Gates a£ palomar> M

variable stars SW Andromedae, DX Del- ' , u n

. . . _T7 TT .. . '_ _ . recent supernovae nave been continually
phini, DY Herculis, and DH Pegasi, , i i i rj . i Vi_ ^i_ aq • L
r , • , i t i i . • , -m a /m? u photographed by Zwicky with the 4o-incn
which had been obtained with the 6U-incn r . * .r , , } .J r ,
telescope during 1955 and 1956 by TifiEt an^OO-inch telescopes in four color ranges
and Bonsack. Radial-velocity and radius- (ultraviolet, blue, green-yellow and red)
variation curves were obtained. The with a vietw to securing four light-curves
velocity curves of DY Her and DH Peg corresponding to these ranges. Efforts by
cannot be described as mirror images of ZwickY to investigate theoretically simple
the published light-curves for these stars, models of supernova outbursts have led to
but the mirror-image relationship does results giving promise that both the light-
hold for SW And and DX Del. curves and the sequences of the spectra in
Tifft and H. J. Smith, of Yale, have supernovae of types I and II may become
completed a study of the three-color light- understandable in the near future,
variation and radial-velocity variation for The supernova in M 84 was observed
the star T Sextantis, an RR Lyrae star of photoelectrically by Baum at the 200-inch
subclass c. telescope. On May 24 its photovisual mag-
Wallerstein has completed his study nitude was 13.32, and its B— V index was
of the velocity curves of Population II +0.80. Multicolor observations were also

MOUNT WILSON AND PALOMAR OBSERVATORIES 49

obtained in order to determine the general
character of its spectral-energy distribution.

Magnetic Stars

The results of an 11-year program de-
voted to the observation of stellar mag-
netic fields have been summed up in a
catalogue of magnetic stars by Horace W.
Babcock. Table 1 of this catalogue is a
compilation of results for 84 stars for which
magnetic fields have been observed by
means of the Zeeman effect. Notes are
given on the spectra, line widths, abnormal
line intensities, and magnetic variations.
Table 2 lists 55 stars that probably but not
definitely have magnetic fields. Table 3 is
a list of 55 stars having sharp lines that
give little or no evidence of the Zeeman
effect; Table 4 is a collection of 111 stars
having lines too broad for magnetic meas-
urements.

The analysis and discussion of the nu-
merous results now available on magnetic
stars lead to the following preliminary
conclusions. Of the magnetic stars in
Table 1 of the catalogue, 1 is a cluster-type
variable, 1 is a subdwarf, 2 are of type S,
3 are M-type giants, 7 are "metallic-line"
stars, and 64 are sharp-line A-type stars.
This variety shows that stellar magnetic
fields are in all probability ubiquitous,
particularly when it is recalled that only
strong coherent fields in sharp-line stars are
susceptible of measurement by available
methods. Another significant result is that
all stellar fields appear to be variable, and
the great majority show irregular varia-
tions. Stellar magnetic phenomena, like
those of the sun, are evidently not simple;
rather, the observations reflect complex
hydromagnetic fluctuations of the photo-
spheric material. It is legitimate to think
of hydromagnetic turbulence in which the
energy of the magnetic field is of the same
order as the kinetic energy of motion of
the gas masses. The statistics of line
widths among the rapidly rotating stars
of type A are in accord with the supposi-
tion that nearly all the sharp-line magnetic

A-type stars are rapid rotators viewed
pole-on. As a working hypothesis, it can
be maintained that the prevalence of rapid
axial rotation plus hydrogen convective
zones is responsible, through a dynamo
process, for the strong fields of the A-type
stars.

Of the magnetic stars found in the
course of this survey, many have peculiar
and variable line profiles, several are spec-
trum variables, and six are new spectro-
scopic binaries. Secular variations in line
intensities and in amplitude of magnetic
variation have been found in certain mag-
netic stars. The variety of significant yet
unexplained phenomena observed is con-
siderable, and this field deserves increased
attention in the future. The interpretation
of the stellar magnetic results will un-
doubtedly be facilitated by the increased
understanding of solar magnetic phenom-
ena, and particularly the solar magnetic
cycle, that is accruing from current studies.

The star HD 125248 reverses its mag-
netic field with a period of 9.3 days, and
shows synchronous changes in line
strength, line width, and radial velocity.
By a kind of harmonic analysis, Deutsch
has shown that it is possible to derive a
rigidly rotating configuration which satis-
fies the observations in any one cycle rea-
sonably well. Superposed on the 9.3-day
cycle, however, is another, much slower,
variation in radial velocity, which has here-
tofore prevented the compilation of ob-
served velocities to form mean curves.

From coude observations accumulated
over a 10-year interval, by Deutsch and
H. W. Babcock, it has now been found
that the secondary velocity variation has a
period of about 1670 days. The amplitude
of this long cycle is about 7 km/sec, or
more than twice that of the 9.3-day cycle.
The long cycle is probably due to Kepler
motion in an eccentric orbit about an in-
visible companion.

When the orbital velocity variation is
removed, the observations may be com-
bined to yield mean velocity curves which
represent the motions due to rigid rotation.

50 CARNEGIE INSTITUTION OF WASHINGTON

One such curve is required for the lines
of Eu II, Gd II, and Ce II; a second for
the lines of Cr I and II; and a third for
the lines of Fe I, Fe II, and Ti II. Since
these mean curves make use of many more
observations, they are much better deter-
mined than the velocity curves previously
used for the harmonic analysis. Accord-
ingly, the harmonic analysis will be re-
peated by Deutsch in an attempt to im-
prove the fidelity of the map that has
been derived.

Spectrograms were obtained by Bonsack
with the 60-inch telescope of the spectrum
variable 56 Arietis. They were measured
to determine the radial velocities of the
individual lines as a test of the rigid mag-
netic rotator model for this star. Analysis
of the wavelength and intensity variations
of the lines showed that the star can be
represented as a rigid rotator of period
0.73 day.

Because of the unusually strong Si II
lines, 21 Aquilae is placed in the group of
"silicon-helium" peculiar A stars although
it is classified as B8 from the He I lines
in its spectrum. Its spectrum lines are un-
usually narrow, so that wavelengths can
be measured with good accuracy. Com-
bining Miss Burd's measurements of plates
taken by H. W. Babcock with data from
a plate taken at the McDonald Observa-
tory, G. R. and E. M. Burbidge have ob-
tained evidence for a systematic shift be-
tween the singlet and triplet lines of He I.
Although some part of this may be due
to Stark effect, comparison with the labora-
tory measures of isotope shifts between
lines of He3 and He4 suggests that there
is a reasonable probability that the major
part of the shift may be due to the pres-
ence of He3 in the star's atmosphere in an
amount comparable with that of He4.

The presence of He3 may be explained
by a theoretical treatment, carried out by
G. R. Burbidge and by W. A. Fowler
and E. M. Burbidge, of the Kellogg Radia-
tion Laboratory, of nuclear reactions on
the surfaces of magnetic stars. It appears
that a flux of neutrons will be produced by

reactions between accelerated protons and
light nuclei, and will be captured by cool
hydrogen to form deuterium. Reactions
between deuterons and protons will either
build He3, or free the neutrons again, so
that eventually they may be captured by
elements in the iron group to build heavy
nuclei.

Globular and Galactic Clusters and
Stellar Evolution

Schmidt has started an investigation of
the color-magnitude diagram for the galac-
tic cluster NGC 6939. The cluster appears
to be relatively old and has an incipient
Hertzsprung gap. More photoelectric
standards are being obtained.

Photographic material has been obtained
for the galactic clusters NGC 2269, 2309,
2311, 2367, 2401, 2453, and 6834. Poor ob-
serving conditions in the winter prevented
the establishment of photoelectric stand-
ards in these clusters.

Schmidt carried out photographic, photo-
electric, and spectroscopic observations on
a small condensation of stars at R.A. 19h
14m 35s, Dec. +15° 137. The diameter is
l'x2', and the tenth-brightest star is
mpg=l65.

Study of color-magnitude diagrams for
a number of star clusters was continued
by Sandage during the report year in con-
nection with the problem of stellar evolu-
tion. Because of the importance of the
ultraviolet excess shown by stars with
B — V— 0.5 in the globular clusters M3 and
NGC 4147 (reported last year), Sandage
and Walker did three-color photometry on
the UBV system for 200 of the brighter
stars in M 92. Photoelectric observations
were made on 8 nights with the 60-inch
telescope, and a special series of photo-
graphic plates were taken with the 100-
inch. The measurements, which have been
completed, indicate that the giant stars in
M 92 also show the A(U — B) excess of
about 0.3 mag. in agreement with the value
found for stars in M 3 and NGC 4147.
Stars on the horizontal branch, however,

MOUNT WILSON AND PALOMAR OBSERVATORIES 51

show an ultraviolet deficiency of nearly
A(U-B)=0.15, again like stars in M 3,
M 13, and NGC 4147. It is therefore be-
coming clear that the energy-distribution
curves for globular-cluster stars differ in a
fundamental way from curves for stars in
the solar neighborhood, and the difference,
which shows up in the three-color pho-
tometry, provides an easy means to search
for globular-cluster-like stars in the gen-
eral field. The differences in the U, B, V
values between globular-cluster stars and
normal field stars are believed to be due
to differences in chemical composition and
the resulting blanketing of the continuum
by the absorption lines in the blue and
ultraviolet region of the spectrum. A spe-
cial study of the blanketing effect was
begun by E. M. Burbidge, G. R. Burbidge,
and Sandage with the globular-cluster
problem in mind. The study is reported
in more detail in another section of this
report.

The globular cluster NGC 5897 was
placed on the three-color program because
a spectrogram by Deutsch showed that this
cluster had very weak absorption lines and
therefore is expected to have a large ultra-
violet excess. Photoelectric observations
with the 100-inch and photographic plates
with the 200-inch were obtained for this
cluster by Sandage. Schmidt is reducing
the photographic material, but the results
are not yet available. Other clusters on the
current program include NGC 6356, which
is one of the W. W. Morgan strong-line
globular clusters near the galactic nucleus;
NGC 6712, a globular cluster situated in
the Scutum star cloud; NGC 7789, an old
galactic cluster probably of the M 67 type;
NGC 7788, which has the three classical
cepheids CF Cassiopeiae, CE a Cas, CE b
Cas associated with it; and NGC 7790,
which forms a double cluster with 7788.
Sandage has obtained photoelectric calibra-
tions and photographic plates for all these
clusters with either the 60-inch or the 100-
inch telescopes. The measurements are
almost complete for most of the clusters,

The work on NGC 7789 is a joint project
with E. M. Burbidge.

The cluster NGC 1866 in the Large
Magellanic Cloud is under study by Arp
and Sandage in collaboration with Dr.
A. D. Thackeray, of the Radcliffe Observa-
tory, Pretoria, South Africa. This cluster
is globular in appearance but has a color-
magnitude diagram that seems at this stage
of the investigation to resemble that of
M 11. Its importance lies in the fact that
9 cepheids with periods ranging from 2.64
to 3.52 days are associated with the cluster.
All these cepheids have the same apparent
magnitude and therefore are at the same
place in their evolutionary history. The
details of the connection of the cepheids
with the nonvariable stars in NGC 1866
promise to clarify our ideas of how cepheid
variables fit into the evolutionary picture.
Photoelectric calibration to V = 19.0 was
completed by Arp while he was in South
Africa; Thackeray has obtained an exten-
sive series of plates with the 74-inch Rad-
cliffe reflector; Sandage has started meas-
urement of the plate material. At the
present writing, NGC 1866 appears to be
one of the most important clusters in the
sky for the evolutionary problem.

Several years ago, E. E. Salpeter, of Cor-
nell, derived a theoretical luminosity func-
tion giving the distribution of stars along
the main sequence at the time of their
formation. This creation function is of
interest because with it can be predicted
the number of stars at any given lumi-
nosity which have been formed in the life-
time of the Galaxy. Because of its signifi-
cance, observational checks on the Salpeter
function are necessary. Sandage showed
that the luminosity functions in each of the
open clusters h Persei, Pleiades, Coma
Berenices, Hyades, and Praesepe agree well
with the Salpeter creation function, and
this agreement supports the argument that
the peculiar form of the van Rhijn lumi-
nosity function for stars in the general field
brighter than Mv— +3.5 is due to stellar
evolution.

Sandage derived a new luminosity func-

52 CARNEGIE INSTITUTION OF WASHINGTON

tion for M 3 which predicts a total mass
for M 3 of 2.45 X 105 MO, a total number
of stars of 5.9 X 105, and a total number of
white dwarfs of 4.8 X 104. By means of the
Salpeter creation function it was estimated
that the stars which are now white dwarfs
in M 3 have shed a mass equal to 1.1 X 105
MO in the form of gas as they evolved
from the main sequence to the white-dwarf
stage. This mass has escaped from the
cluster into the interstellar medium. Simi-
lar considerations for E galaxies show that
the gas formed by the shedding process has
probably not escaped and makes up about
1/200 of the total mass of the E galaxy.
Osterbrock has suggested that this mass
may be the origin of the A3727 [O II]
emission from the E systems.

Semiempirical evolutionary tracks were
computed by Sandage for the stars in M 3
and M 67. M 3 and M 67 are clusters of
about the same age (^5xl09 years), but
the stars in each cluster follow very differ-
ent tracks of evolution in the giant region
even though they are of about the same
mass. Presumably the track differences are
due to differences in the chemical com-
position of the two clusters. The method
of obtaining the tracks of evolution for
individual stars utilizes the information
contained in the observed luminosity func-
tions and color-magnitude diagrams of
clusters. The evolutionary tracks, the time
scale for evolution along these tracks, and
the fraction of the total mass exhausted of
hydrogen at each evolutionary stage are
determined. These semiempirical results
were compared with the predictions of the
Hoyle-Schwarzschild (HS) theoretical
evolving models, and Sandage showed that
the time scale of the HS models gives
nearly the correct luminosity function for
M 3 except at the top of the giant branch,
where the HS time scale is too fast by a
factor of 3. The lifetime of the RR Lyrae
stars is estimated to be 8 X 107 years. This
figure gives an expected rate of change of
period for these variables due to evolution
of A*//=2.40X10-11, which is 0.1 second
of time per century. The final result of the

study shows that stars in M 3 completely
exhaust their energy store of 1.60 XlO52
ergs in their lifetime as they evolve from
the main sequence, through the giant stage,
and finally to the white dwarfs. But stars
in M 67 exhaust only 37 per cent of their
energy store. Since there can be no internal
energy reservoir in the white dwarfs, this
result suggests that mass loss must occur
from stars like those in M 67 before they
become white dwarfs. This conclusion is
supported empirically by Deutsch's study
of the giant a Herculis reported last year.

Chemical Composition of Stellar
Atmospheres

Under the sponsorship of the Physics
Division of the U. S. Air Force, Office of
Scientific Research, a project entitled
"Stellar Composition and Related Nuclear
Processes" has been established at the
Mount Wilson and Palomar Observatories
under the direction of Greenstein. The
goal is to increase the astronomical data
relevant to theories of the origin of the
elements, as well as to bring together a
group of nuclear physicists and theoretical
astrophysicists interested in aspects of this
fundamental question. It is generally ac-
cepted now that certain stars show evi-
dence for the current formation of ele-
ments, for example most obviously the ele-
ment technetium in S stars, originally dis-
covered by Merrill. In addition, the
thermonuclear conversion of hydrogen to
helium, and the burning of helium to car-
bon, in stars has been definitely established.
These processes occur in certain parts of
the evolution of stars, and should be corre-
lated with studies of stellar-interior theory.
The group under Air Force sponsorship
will apply more advanced theoretical tech-
niques to the determination of the abun-
dances in normal and peculiar stars, with
the hope of correlating these results with
nuclear data and theory. The general ques-
tion of the possible evolutionary trend in
the abundance of heavy elements with time
in our own Galaxy will also be studied.

MOUNT WILSON AND PALOMAR OBSERVATORIES 53

Abundance analyses for such elements in
carbon-rich stars, S stars, stars with abun-
dant rare earths, very old and very young
main-sequence and giant stars, and young
and old subgiants will be among the goals
of this research. It is hoped that the study
will be continued for several years.

The two-lined spectroscopic binary
-f74°493, a high-velocity dwarf, has been
studied by Greenstein in collaboration with
Drs. Margherita Hack and Otto Struve,
of the University of California. The
masses seem to be close to those of nor-
mal main-sequence G dwarfs; in other
words, high velocity does not necessarily
mean in the dwarfs a deviation from the
mass-luminosity relation. A spectrophoto-
metric analysis showed that the lines may
be slightly weaker than normal for similar
Population I dwarfs, indicating a slight
underabundance of the metals.

A detailed study of line intensities of the
metals, CN, and CH has been made by
Greenstein and Dr. Philip C. Keenan, of
the Perkins Observatory. Fourteen giants
near spectral type G8 of both high and low
velocities have been studied. The deduced
reduction in the number of metallic atoms
in high-velocity stars is by a factor of about
0.4 to 0.6; the effective number of CN
molecules is reduced to about 0.06 normal,
and CH is hardly affected. Two new ap-
parently carbon-poor stars have been
found. One of the high-velocity stars, u2
Cancri, is shown to be of normal composi-
tion; it is therefore a runaway evolved
Population I star.

Further spectra of the brightest stars in
the globular clusters M 13 and M 92 have
been obtained by Greenstein at 18 A/mm
with exposures running from two to three
nights. The M 13 spectra resemble those
of high-velocity stars in the solar neighbor-
hood; the M 92 spectra show very weak
but numerous metallic lines corresponding
to a quite low excitation temperature. A
quantitative analysis of these stars is
planned.

A survey of early R stars at dispersion
4 and 7 A/mm is being carried out by

Greenstein for the study of the atomic-line
differences between the hydrogen-rich and
hydrogen-poor carbon stars.

Work was begun by Bonsack on an ex-
tensive program to survey a variety of stars
of type K to determine the possible varia-
tions in the abundance of lithium, and a
number of stars of type A to study
beryllium.

The comparison of the spectrum of the
Ba II star HD 46407 with that of the stand-
ard G8 III star k Geminorum has been con-
tinued by G. R. and E. M. Burbidge. The
theoretical curves of growth used were
those by Wrubel for a Milne-Eddington
atmosphere. Conditions in the standard
star k Gem and in HD 46407 were found
to be very similar, as had been anticipated.
There were no detectable differences be-
tween the two stars in Texc, in the degree
of ionization, or in the total velocity (ther-
mal and turbulent). From a preliminary
comparison with the sun, Texc was found
to be 3900° C; with log Pe=-03, Tion
was found to be 4200° C; the velocity
parameter was 4 km/sec.

Relative abundances of the following
elements were determined: sodium, mag-
nesium, aluminum, silicon, calcium, scan-
dium, titanium, vanadium, chromium,
manganese, iron, cobalt, nickel, strontium,
yttrium, zirconium, barium, lanthanum,
cerium, praseodymium, neodymium, and
samarium; also less certain values were
found for copper, zinc, germanium, nio-
bium, molybdenum, ruthenium, europium,
gadolinium, ytterbium, and tungsten. Car-
bon has a slightly greater than normal
abundance in HD 46407. The elements
from sodium through germanium were
found to have the same abundances in the
Ba II stars and in k Gem, apart from barely
significant slightly increased abundances
of aluminum and scandium in HD 46407.
Most of the heavier elements, from stron-
tium onwards, however, were found to
have abundances of the order of 10 times
those in the standard star.

The results have been discussed in the
context of a theory of the stellar origin of

54 CARNEGIE INSTITUTION OF WASHINGTON

the elements in the universe; they are
found to give good support for it. On this
theory certain isotopes of the elements
heavier than iron are built by a slow neu-
tron-capture process in the interiors of red
giant stars. Abundance peaks are produced
at nuclei having a magic number of neu-
trons, e.g. strontium, yttrium, zirconium,
barium, lanthanum, cerium, praseodym-
ium, and neodymium, all of which are
overabundant in HD 46407. Europium, so
prominent in peculiar A stars with mag-
netic fields, is not built predominantly by
neutron capture on the time scale supposed
to occur in red giant stars, and its abun-
dance in HD 46407 is no larger than in
k Gem.

Coude spectrograms have been obtained
by G. R. Burbidge with the 100-inch tele-
scope for a continuation of the program
for the determination of heavy-element
abundances in cool stars. The identifica-
tion by Merrill of Tc I in a carbon star
indicates that element synthesis through a
slow neutron-capture chain occurs in these
stars as well as in S and Ba II stars. A
number of carbon stars have been ob-
served, including "normal" stars with C13
bands, hydrogen-poor stars with no C13,
and CH stars.

Sandage and G. R. and E. M. Burbidge
have started a program for the measure-
ment of the blanketing effect in a number
of standard stars in the range F7 to G2,
and in the Population II star HD 19445.
The latter has a low abundance of most
elements, relative to hydrogen, and conse-
quently weak spectral lines. In a two-color
plot of U — B against B — V, this star has
an ultraviolet excess. It lies close to the line
on which H. L. Johnson and Sandage have
shown the stars in M 3 to lie. The pro-
gram is to see whether the difference in
blanketing between HD 19445 and a nor-
mal star of the same effective temperature
would move the star in the U — B, B — V
plane so as to account for the ultraviolet
excess. It will certainly account for part
of it, but whether there is a remainder
needing some other explanation must be

determined. If blanketing can account for
the whole effect, then we shall be able, in
principle, to use two-color measures to de-
duce whether the elements calcium, iron,
etc., in a distant group of stars are under-
abundant, relative to hydrogen, and by
how much. All the observations, consist-
ing of 10 A/mm spectrograms covering the
range A3300 to A6300, have been obtained
with the 100-inch telescope. Almost all the
tracings have now been completed, with
the Babcock microphotometer, and their
measurement with a planimeter is now
under way.

Nuclear Reactions in Stars

G. R. Burbidge and F. Hoyle, and Drs.
W. A. Fowler and E. M. Burbidge, of the
Kellogg Radiation Laboratory of the Cali-
fornia Institute, have continued their work
on the synthesis of the chemical elements
in the stars. They find that eight processes,
as follows, are necessary to account for the
abundances of all the 327 isotopes found in
the solar system: (1) Hydrogen burning
is responsible for the majority of the en-
ergy production in stars. This process
synthesizes helium, and, when it occurs
in a mixture of hydrogen with other ele-
ments, it builds all those isotopes of carbon,
nitrogen, oxygen, fluorine, neon, and so-
dium that are not built by process 2. It
occurs in main-sequence stars and in shells
around the cores of giants. Its results may
be seen in hydrogen-exhausted hot stars,
some carbon stars, WN and other nitrogen-
rich hot stars, and some white dwarfs.

(2) Helium burning builds C12 from he-
lium and, by further a-particle addition,
O16, Ne20, and perhaps Mg24. Its onset oc-
curs in the cores of red giant stars, and it
presumably continues in stars in later evo-
lutionary stages. Its results may be visible
in some carbon stars and in WC stars.

(3) The a process builds, through charged-
particle interactions, the rest of the four-
structure nuclei Mg24, Si28, S32, A 36, Ca40,
and probably Ca44 and Ti48. (4) The e
process builds the elements comprising the
iron peak in the abundance curve (vana-

MOUNT WILSON AND PALOMAR OBSERVATORIES 55

dium through nickel). It occurs at very
high temperatures and densities, when con-
ditions of statistical equilibrium are set up.
Both it and the a process are thought to
occur shortly before a star explodes as a
supernova. (5) The s process is a slow
neutron-capture chain which builds many
of the isotopes in the ranges 23<^4<46
and 63<A<209. It is thought to take
place in the interiors of red giant stars,
when neutrons are produced, and to be
responsible for the observed anomalies in
S stars, Ba II stars, etc. (6) The r process
is a rapid neutron-capture chain, occurring
on a very short time scale, and is thought
to take place in supernovae. It will pro-
duce a large number of isotopes in the
range 70<A<209, and uranium and tho-
rium. It may also build a few lighter iso-
topes not built by other processes, e.g. S36,
Ca46, Ca48. The decay of radioactive Cf254,
built by this process, is thought to be re-
sponsible for the exponential light-curve
of some Type I supernovae. (7) The p

process is a proton-capture or photoneu-
tron process, also thought to occur in some
supernovae. It builds the remaining iso-
topes in the range 63<A<209 which are
proton-rich, have abundances 0.01 to 0.001
times the near-by normal and neutron-rich
isotopes, and cannot be built by either the
s or the r process. (8) The x process, not
properly worked out yet, must be respon-
sible for building deuterium, lithium,
beryllium, and boron, which are unstable
in hydrogen burning in stellar interiors.
There may be some production in stellar
atmospheres in magnetic stars (including
all stars with flare activity), and deuterium
may be made in some supernovae when a
large flux of neutrons impinges on an ex-
panding, relatively cool envelope in which
hydrogen has not been exhausted.

Further work on the x process is going
on at present. Observations have been
made with Dr. Philip C. Keenan, of the
Perkins Observatory, and theoretical work
is under way with W. A. Fowler.

GASEOUS NEBULAE

Internal Motions and Radial Velocities

In past years, Guido Munch and Wilson
have reported on an extensive series of ob-
servations of the internal motions in the
Orion nebula using a multislit on the 72-
inch camera of the 200-inch coude spectro-
graph. During the current year they have
extended these observations to fainter re-
gions of the nebula by using the multislit
with the 36-inch camera.

All plates have now been measured and
reduced by Miss Flather and Mrs. Coffeen
for radial velocities of the [O II], H, and
[O III] lines. A large number of line pro-
files have also been determined, and the
material is being analyzed. The aspect of
the problem related to the verification in
the nebula of the predictions of the equilib-
rium theory of turbulence at high Reynolds
numbers has been studied. In good agree-
ment with results derived by Dr. S. von
Horner from radial velocities determined
at the Lick Observatory, Munch and Wil-

son find that the mean square difference
between two points in the nebula separated
by a distance d varies nearly as d^, for
values of d between 1" and 60". At vari-
ance with von Horner's conclusions, how-
ever, they find that this statistical relation
cannot be taken as a proof that the Kol-
mogoroff law is satisfied in the nebula, for,
if it were, the line widths predicted would
be 3 times smaller than those observed. If
KolmogorofTs law were satisfied, the aver-
aging effect along the line of sight involved
in the observations should be negligible,
and the nebula would be required to have
the shape of a thin sheet of matter in the
plane of the sky. They find such a con-
figuration inadmissible in the light of di-
rect observational evidence, such as the ap-
pearance of the line 23S-33P of He I in
absorption in the spectra of the illuminat-
ing stars and also the reddening of the
Trapezium cluster. The failure of the
Kolmogoroff law to describe the state of

56 CARNEGIE INSTITUTION OF WASHINGTON

motion of the nebula is due to the com- supports Oort's suggestion that the Cygnus

pressibility effects. The spectra show many loop is a shell originally ejected at a high

areas in the nebula where the lines appear velocity and decelerated by collisions with

distinctly as double, suggesting the ex- interstellar clouds. In order to form an

istence of discontinuities in the flow, pro- object like the Cygnus loop in a region of

duced by shock waves. Munch and Wilson average interstellar density (A/h = 0.01 to

visualize these shock waves as a result of 1), the shell must have a momentum of the

the interaction of the expanding H II order of 1043 cm. g. sec-1. This value is too

region with the surrounding cold material, high for shells ejected by ordinary novae

An investigation of the radial velocities and even by supernovae of Type I, such as

in the Cygnus loop has been completed by the Crab Nebula, but it appears possible

Minkowski. The average picture found is that such a value pertains to the shells of

that of an expanding incomplete thick supernovae of Type II. The Cygnus loop

shell. The velocity of expansion is 65 and other objects of this type thus may be

km/sec at the inner boundary with a di- the hitherto unobserved remnants of these

ameter of about 80' and 115 km/sec at the supernovae.

outer boundary with a diameter of about A preliminary discussion of the observa-

170'. This picture resembles in many ways tions of the radial velocities of the bright

the appearance of IC 443, which is obvi- filaments in the Crab Nebula has been pre-

ously an object of the same type as the pared by G. Munch. It was found that

Cygnus loop. If the velocity of expansion many of those filaments that by their

at the outer border is combined with the radial velocities can be considered as a unit

outward motion of 0703 per year found by are oriented at right angles to the direction

Hubble, a distance of 770 parsecs is ob- of the magnetic field, as determined by

tained. The diameter of the main part of Hiltner from polarization observations of

the nebula is then 40 parsecs; faint matter the synchrotron radiation. This observa-

in the south extends to a distance of 35 tion suggests that the filaments move in a

parsecs from the center. "force-free" field, in which the electric

The absence of an exciting star suggests currents are parallel to the magnetic field,
that the excitation of the emission spec- The radial velocities of the brighter central
trum in the Cygnus loop is collisional. The filaments also show that their motion is
observational decision on the type of exci- not one of uniform expansion from the
tation depends essentially on the ratio of star supposed to be the remnant of the
the intensities of Ha relative to H3 for supernova. Study of the large-scale geom-
which conflicting results had been obtained etry of the filamentary mass disclosed that
by various observers. New measures by the direction of the major axis of the
Minkowski confirm the result by Pikelner nearly elliptical outer boundary is that of
that the average ratio (Ha+[N II]) :H|3 the galactic equator, to the degree of pre-
has the value 6.1. The [N II] lines, how- cision with which such major axis can be
ever, have considerable strength, and the determined. The importance of the role
average value of the ratio Ha:H(3 is 3.4, that the prevalent galactic magnetic field
with local variations up to 5.4. These values plays in determining the present structure
are consistent with the interpretation that of the shell thus becomes apparent. The
the excitation is collisional. The lower outer galactic magnetic field being re-
values of the ratio conform to electron tern- sponsible for the ellipticity of the shell, it
peratures of the order of 100,000°, which follows that the expansion velocity of the
is in general agreement with the tempera- shell at present may have values along the
tures found from the relative intensities of lines of force differing from values at right
the [O III] lines A4363 and A5007/4959. angles to them. The distance of the Crab

The fact that the excitation is collisional Nebula, determined on the assumption of

MOUNT WILSON AND PALOMAR OBSERVATORIES 57

the equality of the cross motions with the
radial velocities, may have to be revised
and increased upward. Further observa-
tions of the fainter filaments are being
planned to elucidate this point.

Because of the renewed interest in the
intensity changes in the continuum of the
Crab Nebula, the continuum was regularly
photographed by Baade in the range A5300
to A6400. No moving wisps were observed
during the fall of 1956.

In continuation of observations going
back to the late 1930's, the shells of Nova
T Aurigae (1891), Nova Cygni (1920),
and Nova Herculis (1936) were photo-
graphed by Baade at the prime focus of
the 200-inch. The progressive expansion of
all three shells is very conspicuous in the
blink comparator.

Radial velocities of 101 faint planetary
nebulae have been determined by Minkow-
ski; almost all of them are for objects
which are within 10° from the position of
the galactic center. In this region the
velocities range between about —300 and
+ 300 km/sec, showing the presence of a
large velocity dispersion in the central part
of the Galaxy. The observations are still
being continued, but the available results
are already adequate for a general discus-
sion of the motions of planetaries, which
Schmidt has started.

Densities of Nebulae

Osterbroek continued measurements of
the intensity ratio of the two components
of the [O II] A3727 doublet in order to
determine the electron densities in gaseous
nebulae. Measurements of this ratio show
that the density in the brightest filaments
of the Crab Nebula is of the order of 1000
electrons/cm3. The total volume of such
filaments, estimated from Baade's direct
plates of the nebula, is approximately
8 X 1051 cm3, and the resulting mass of the
system of brightest filaments is about 0.02
solar mass. Rough estimates of the densi-
ties and volumes of the fainter filaments
can also be made, leading to a total mass
of the whole filamentary system of the

Crab Nebula in the range 0.05 to 0.1 solar
mass.

Plates were also obtained for measure-
ment of the A3727 intensity ratio in a
number of planetary nebulae, particularly
IC 418, NGC 6720, and NGC 7293. These
observations were planned to study the
spatial-density variation in the various
planetaries, for the electron density gives
a good measure of the mass density in
these objects. Reductions of the NGC 6720
observations show that in this well known
ring planetary all the regions that emit
A3727 have essentially the same density,
about 1000 electrons (or hydrogen ions)
per cm3. Additional observations of the
A3727 ratio in the Orion nebula were ob-
tained, to complete the study of the large-
scale structure of this object.

Several diffuse nebulae in which a dark
cloud is ionized by an O star outside the
cloud were studied by Osterbroek. The
ionized material, seen in emission, lies
mostly between the star and the dark
cloud, and is sharply bounded on the side
toward the cloud by a bright rim, while
on the other side it fades out gradually.
In each of these nebulae, long-exposure
schmidt photographs taken in Ha show
that there is parallel filamentary structure,
approximately perpendicular to the bright
rim, in the emission nebulosity (these stri-
ations were observed some years ago in
IC 434, the brightest member of this group,
by Duncan, with the 100-inch telescope).
These parallel bright filaments must result
from regions of high density in the origi-
nal cloud, which are drawn out by expan-
sion without being mixed appreciably.
There are two possible interpretations for
the fact that the filaments are long but
remain unmixed: either there is a mag-
netic field that prevents motions perpen-
dicular to its own direction in all these
objects, or else the mechanism that gen-
erates turbulence in most diffuse nebulae
cannot operate in these objects, perhaps
because they are expanding into regions
of very low density.

58 CARNEGIE INSTITUTION OF WASHINGTON

GALAXIES

The Andromeda Galaxy (Messier 31) and photovisual measurements for the color-

Other Members of the Local Group magnitude diagram of this system. The

During the past year the photovisual ob- results can bf Pref nted in final £orm as

servations of the cepheids in the outer field soon as thf Photoelectric sequence in Mes-

of the Andromeda galaxy, 96' south pre- sier 13> fhlch has been measured by Dr.

ceding the nucleus, were concluded by HL. Johnson, of the Lowell Observatory,

Baade. Combined with the earlier photo- and whlch has, been transferred to the

graphic series of plates they should furnish Drac° system, becomes available. Since

the necessary data about color excesses and enouSh Plates rhave been, obtained for the

absorption in this field. Still under way is investigation of the variables in the Leo II

a survey of this same field for faint red system> only the Ursa Minor and Leo I

variables which were below the plate limit systems remain under observation,

of the 30-minute photographic series at the The studY o£ the stellar contents of M 33,

200-inch but show up in remarkably large reported last year, has continued. Huma-

numbers on photovisual and red plates of son and Sandage began a search for the

long exposures. These faint red variables red supergiants which are expected from

are clearly members of the Population II evolutionary considerations to accompany

which pervades the whole disk of the the blue O and B type stars in the spiral

Andromeda galaxy (the "interarm" popu- arms- °ver 100° red supergiants were

lation). They deserve further investigation found in M 33 by Humason by blinking

because they should throw much needed Pairs of blue (103a-O + GG 13) and yellow

light on the characteristics of the interarm (103a-D + GG 11) plates taken with the

population of our own Galaxy. 200-inch. A special field south preceding

Most of Baade's observing time during the nucleus was chosen to study these stars,
the past year was devoted to the dwarf E and Plates taken sporadically during the
galaxies of the local group which as typical season indicated that many of these red
representatives of the pure Population II stars are irregular variables. They may be
are of special interest. Four of them (the the extragalactic analogues of the M super-
Sculptor, Ursa Minor, Draco, and Leo II Siants found in h and X Persei, for exam-
systems) are close enough to reach the Ple* Photometric measurements on the
cluster-type variables with the large mod- UBV sYstem are in progress with a stand-
ern telescopes. Although all four are E ard photoelectric sequence set up by Dr.
galaxies of very low luminosity, the cluster- H- L- Johnson in M 33 several years ago.
type variables appear in them in large plates in blue and yellow wavelengths
numbers (on the average more than 200 for a number of other galaxies (NGC
cluster-type variables per system). Type II 6822, Sextans Dwarf, Leo Dwarf, W-L-M
cepheids with periods longer than a day Dwarf, NGC 2403, M 101) have been ob-
are rare in these dwarf systems. The Draco tained for a similar study of red super-
system, for instance, contains only two, giants. Blink surveys show that the red
and the same number was found by the supergiants are indeed present in these
Pretoria observers in the Sculptor system, galaxies as well as in M 33.
But with increasing stellar content their Schmidt has computed a model of the
number increases rapidly; examples are the distribution of mass in M 31. It is based
Fornax system and the Leo I system north on recent results obtained by Dutch ob-
of Regulus. servers from observations at 21-cm wave-
Miss Swope has finished the investigation length, and gives a total mass of 3.4 X 1011
of the variables in the Draco system. She solar masses. About 50 per cent of the mass
also completed the photographic and has a mass-luminosity ratio of around 27;

MOUNT WILSON AND PALOMAR OBSERVATORIES 59

the ratio for the whole system has about
the same value.

Studies of Individual Galaxies

Osterbrock began obtaining spectra at a
dispersion of 66 A/mm of those elliptical
galaxies already known from the observa-
tions of Humason to have [O II] A3727 in
emission. Only a few galaxies have been
observed to date, but several interesting
preliminary results have appeared. In all
the galaxies studied, except a single EO, the
material emitting A3727 is in rapid rotation.
One EO and one El, both very nearly
round objects, as well as several more-
elongated objects, show this rotation of
the interstellar matter quite definitely.
That the A3727 lines are resolved into two
components in only one object indicates
quite high turbulent velocities as well as
systematic rotational velocities for the in-
terstellar matter near the nuclei of these
elliptical galaxies. Among the ellipticals
known to have A3727 in emission, there are
wide variations in the strength of the line,
but it is invariably so strong at the 66
A/mm dispersion that possibly all ellipti-
cals will have detectable A3727 at this dis-
persion.

The blue galaxies recently discovered at
Tonantzintla by Dr. G. Haro are being
observed both spectroscopically and directly
by Guido Munch. Among those already
observed, an appreciable fraction have been
found to have regular geometry, resem-
bling that of early and intermediate spirals.
When spiral arms can be seen, as in NGC
263 and NGC 2415, they appear thicker
and with brighter condensations than in
typical spirals. Their spectra show the
emission lines characteristic of planetary
nebulae, superposed on a continuum corre-
sponding to an early spectral type. Strong
Balmer lines in absorption are prominent,
and in one case the line He A4471 has been
measured in absorption. The emission
lines and early-type spectrum extend
through the entire body of the systems,
with increased relative strengths of the
emission lines in the nuclei and in the con-

densations. The components of the [O II]
doublet are clearly resolved, indicating
that, unlike the condition in the normal
ellipticals, the random motions of the gas
producing them are small. The relative
intensities of the [O II] lines indicate num-
ber densities of free electrons of the order
of 103 cm"3.

Code has made scans of several elliptical
galaxies in the Virgo cluster with his
photoelectric scanning spectrograph. His
curves should provide further information
on the energy distribution in the spectrum
of elliptical galaxies and on their possible
stellar content.

TifTt has begun the photoelectric pho-
tometry of the brighter galaxies in four
colors in the range A3400 to A6000 A. This
work will be extended to the red and
infrared this coming year. Field galaxies
of nearly all types, with special emphasis
on those for which W. W. Morgan has de-
rived spectral types, are being obtained.
The Coma and Virgo clusters are included
for comparison. The hope is that the
photometric system will be calibrated in
absolute energy units. The goal is a study
of the variations of color between galaxies
and within galaxies, as well as an attempt
to synthesize the observed colors in terms
of the stellar populations present.

Zwicky has continued the investigation
of pairs and groups of galaxies that are in-
terconnected by luminous intergalactic fila-
ments. All types of galaxies have been
found to be thus linked by luminous mat-
ter, usually blue, suggesting clouds of sub-
luminous blue stars whose absolute photo-
graphic magnitude is greater than Mp = 0.
Most of the double nebulae photographed
originally by Pease with the 60-inch reflec-
tor were rephotographed and found to
show faint countertides on one or on both
components. The absence of countertides
would lead the investigator to suspect
strongly that he was dealing with optical
doubles whose differences in radial veloci-
ties should not be used for a determination
of the masses involved.

60

CARNEGIE INSTITUTION OF WASHINGTON

Catalogues and Statistics of Galaxies

Work on the catalogue of galaxies
brighter than the apparent photographic
magnitude +15.5 has been continued by
Herzog and Zwicky with a grant from the
Office of Naval Research. Charts are being
prepared for publication of all results in
the strip from a = 7h to 17h between
h— —10° and +20° with all the measured
galaxies indicated by various symbols giv-
ing the photographic magnitudes in ranges
of 1 mag. Each chart is on the scale of the
48-inch photograph, and for the conveni-
ence of the users is centered near the center
of a 48-inch Sky Survey chart, 1950 co-
ordinates being used. The catalogue, when
completed, is expected to give the magni-
tudes and positions as well as certain other
data for 35,000 galaxies.

In preparation for a study of the statistics
of galaxies and clusters of galaxies the strip
between a = 12h and 14h 40m from 8= +5°
to +15° was completely covered by Zwicky
with 27 fields at the 48-inch, each field hav-
ing been photographed on 103a-O with
exposures of 3.5 and 10 minutes and on
103a-D behind yellow Plexiglass filter with
exposures of 5 and 15 minutes. Extensive
counts covering large numbers of galaxies
in a large cap around the north galactic
pole have confirmed the previous result
that fields containing many near-by galax-
ies (large near-by clusters) contain rela-
tively few distant galaxies. If m is the
number of galaxies brighter than + 15.5 on
one of the charts of 36 square degrees of
our new catalogue, and n/ is the number
of galaxies in the approximate range from
+ 15.5 to +19.0 counted on good red 48-
inch schmidt plates, we have approxi-
mately mXnf = B. The value of B de-
creases rapidly when we come near the
belt of interstellar obscuration, but B is ap-
proximately a constant =5x 106 in a large
cap of 40° radius around the north galactic
pole. The lowest values of «/ are obtained
near the centers of the near-by clusters in
Virgo and Coma, suggesting that inter-
galactic dust producing obscuration to the

amount of 0.25 to 0.5 mag. is locally con-
centrated in the central regions of these
clusters.

The nonuniformities in the distribution
of galaxies are exceedingly pronounced,
both locally and integrally. Zwicky be-
lieves that this can be understood only on
the basis of the assumption that clustering
of galaxies is a universal phenomenon and
that the apparent distribution of galaxies
is affected in an intricate way by the effects
of both interstellar and intergalactic ob-
scuration. These circumstances make it
difficult to arrive at any definite conclu-
sions regarding the real distribution of
galaxies throughout cosmic space.

Clusters of Galaxies

About two dozen rich globular clusters
of galaxies were chosen by Zwicky for fur-
ther detailed investigation concerning total
population, radial distribution, structural
index and distribution index, distance,
symbolic velocity of recession, and internal
velocity dispersion.

The faintest among these clusters, barely
recognizable on red 48-inch schmidt plates,
are expected to have a symbolic velocity
of recession Vs = cX AA/A of approximately
150,000 km/sec. Each cluster is photo-
graphed with the 200-inch three times on
103a-O plates (exposures 3, 9, 27 minutes)
and three times on 103a-D behind GG 11
filter (exposures 5, 15, 45 minutes). The
main results obtained so far are: (1) The
total population of these clusters in the first
three magnitude ranges mmax to Wmax + 3
is very closely the same regardless of the
value of ramax. (2) The structural indices
and the distribution indices are closely the
same for clusters at all distances. (3) The
segregation of bright and of faint galaxies
within the clusters is the same regardless
of their distance. (4) The total population
of the clusters in a first approximation is
proportional to y — Yo, where y is the angu-
lar diameter, and y0 is a constant. (5) The
number nc of member galaxies per limiting
square degree in the center of the clusters

MOUNT WILSON AND PALOMAR OBSERVATORIES

61

in a first approximation is inversely propor-
tional to the angular diameter y> provided
that no interstellar or intergalactic absorp-
tion interferes. These results allow a dis-
tance determination from counts of cluster
galaxies alone, the distance being inversely
proportional to the apparent structural
index.

During the summer of 1956, Abell com-
pleted the compilation of a catalogue of
2712 rich clusters of galaxies discovered on
the National Geographic Society-Palomar
Observatory Sky Survey. From these he
selected a homogeneous sample of 1682
clusters each of which contained at least
50 members within 2 magnitudes of the
third brightest member. A statistical analy-
sis of this sample indicates: (1) The dis-
tribution function of clusters according to
richness increases rapidly as the population
per cluster decreases. (2) The data allow
no significant decision that the spatial
density of cluster centers varies with dis-
tance. (3) Galactic obscuration of the
order of a few tenths of a magnitude
(photo-red) exists at high northern galac-
tic latitudes around longitude 300°. (4)
There is a highly significant nonrandom
distribution of clusters in direction in the
sky, both when clusters at all distances and
when clusters at various distances are
considered.

In conjunction with the new catalogue
of galaxies, a catalogue of the richest clus-
ters of galaxies is being constructed by
Zwicky, including the 1950 positions of the
centers, the populations, characters, ap-
parent diameters, and estimated distances
of all clusters. Contours of all clusters, that
is isopleths for which the population per
square degree is about twice that found in
the surrounding field, are plotted on the
same charts containing the galaxies of the
catalogue mentioned.

Clusters of galaxies are the only objects
that can be used for distance indicators or
for the study of such problems as the
velocity-distance relation out to the extreme
range of the large telescopes. Because of
the importance of locating as distant clus-

ters as possible, Baum made a series of
experimental exposures with the 48-inch
telescope in a study of the most effective
techniques for locating clusters at the ex-
treme limit of the instrument. Experi-
ments were made with Eastman 103-U
plates in combination with a Chance OR-1
filter to obtain a range of response from
roughly 6500 to 7500 A. This range was
selected for two reasons: it lies within a
spectral region unusually free of radiation
from the night airglow; it particularly
favors galaxies of the type sought, namely,
those having redshifts equivalent to about
half the velocity of light. Although the ad-
vantage over the red-sensitive plates ordi-
narily used amounts to only a fraction of
a magnitude in threshold detection, about
a dozen suspected clusters were observed
in this manner, and four of them were
rephotographed with the 200-inch tele-
scope. Since all these clusters were found
to have relatively small membership, it
will be worth some further searching,
possibly with the additional aid of finer-
grained emulsions.

Velocities and Distances of Galaxies

During the past three years, observations
have been made by Humason at the prime
focus of the 200-inch for the purpose of
obtaining larger redshifts than had hereto-
fore been measured. The results of this
investigation have been negative in that it
has not been possible to identify known
spectral features with the certainty re-
quired for the measurement of redshifts.
Inability to obtain usable spectra is due to
the following reasons. As the present sun-
spot maximum develops, intensity of the
airglow spectrum has increased to the point
where it almost obliterates the spectrum of
a faint galaxy. Magnitudes of the brightest
members in the clusters observed are
fainter than 20.0. The spectra are ex-
tremely narrow because the diameters of
distant galaxies are small. Redshifts of
very distant clusters are large enough to
displace such well known features as the

62 CARNEGIE INSTITUTION OF WASHINGTON

G band, H and K, and the emission at
A3727, beyond the long-wavelength limit
of fast blue emulsions. All this has necessi-
tated the use of slower panchromatic emul-
sions, which increases the exposure times
by a factor of 10 or more.

The negative results obtained for this
particular investigation are reported on
here both because the observations have
consumed much valuable 200-inch time
and because the data given below will lo-
cate and identify the clusters for other ob-
servers who may intend to continue this
type of research. All the clusters listed in
table 3 were first found on plates of the
National Geographic Society-Palomar Ob-
servatory Sky Survey and later rephoto-
graphed by Sandage with the 200-inch.

TABLE 3. Clusters

(1950)

R.A

Dec

1.

Possible

Redshif t

0h 24m

0s

+ 16°

53'

103,000 <

Dr 167,000

10

44

11

+ 9

20

63,000

13

32

10

+ 28

28

74,000

14

47

33

+ 26

22

118,000

Co-ordinates are for 1950, and refer to the
center of each cluster. In the last column
appear possible values of the red displace-
ments. It should be stated that they are
unmeasured, and their reality is uncer-
tain. They were obtained from one or
more spectral features which seemed to
be dimly visible but could not be positively
identified.

Two uncertain displacements are given
for the first cluster. If either is real it is
most probably the smaller value. The most
certain value is that for the second cluster
in the list, +63,000 km/sec.

During the report year, Baum has con-
tinued the photoelectric program for de-
termining both the redshifts and the mag-
nitudes of remote galaxies by multicolor
photometry. It consists in measuring their
relative luminosities in a number of differ-
ent colors ranging from 3800 A in the
ultraviolet to 10,000 A in the infrared. The
data for each object yield a curve of radi-
ated energy E as a function of wavelength

A. For a galaxy that is shifted to the red,
the whole E(X) curve is displaced toward
longer wavelengths, and the amount of the
displacement yields the "velocity" of re-
cession. The result is the same as a spec-
trographic measurement of the amount by
which individual spectrum lines are dis-
placed, but multicolor photometry has the
advantage of being able to reach galaxies
considerably fainter and more distant than
those within spectrographic reach. The
E(X) curves also provide bolometric mag-
nitudes directly without the need for K
conditions.

This multicolor photoelectric procedure
rests on two conditions: the E(X) curves
must be on a true scale of energy per unit
wavelength; the galaxies of one cluster
being compared with those of another
cluster must be intrinsically similar, that
is the differences between their E(X)
curves must be due largely to their red-
shifts and not to other effects. Although
this second condition cannot be guaranteed
for unlimited distances, neither Whitford
nor Baum now finds any clear evidence
for other reddening effects within the
range of distances reachable spectroscopi-
cally. There is, however, the possibility
that effects due to evolution or obscuration
may influence the magnitudes more
strongly than the E(X) redshifts. This pos-
sibility can be checked by comparing mag-
nitudes with apparent angular diameters
estimated from photoelectrically measured
brightness profiles, and efforts were con-
tinued during the report year to obtain the
profile observations required.

The photoelectric redshift-magnitude
data are being accumulated as rapidly as
telescope time permits, and some interest-
ing results are beginning to emerge. Ob-
servations obtained thus far extend from
the near-by cluster of galaxies in Virgo
(about 10 megaparsecs distant) to one of
the remotest clusters of galaxies detectable
on the 48-inch Sky Survey plates. The
E(X) redshif t of this latter cluster was
found to be of the order of 120,000 km/sec
(or OAc), which is roughly twice that of
the present spectroscopic limit.

MOUNT WILSON AND PALOMAR OBSERVATORIES 63

RADIO SOURCES

Identification of Radio Sources able that most radio sources are distant
In a sample area between 0b and 6h, gala*ies fainter than 18th magnitude.
-9° to +3°, a detailed study is now being Two regions, previously found by Mat-
carried out by Minkowski of those objects thews t0 have an excess of 21-cm radiation,
that may be of interest in connection with have been identified by him with features
the problems of the identification of radio appearing on the 48-inch survey. The first
sources. For this area the records and re- o£ these regions is in the OB aggregate I
suits by Mills and the results of the Cam- Camelopardalis, where the schmidt plates
bridge Survey were compared with the show a region of very high absorption
prints of the National Geographic Society- about 4° across with a few spots of weak
Palomar Observatory Sky Survey during Ha emission around the edges. About
Minkowski's stay at the Division of Radio- three-quarters of the stars in the OB aggre-
physics of the Commonwealth Scientific gate are located less than 1°5 from the ab-
and Research Organization, Sydney, Aus- sorbing region. Their proximity, together
tralia. The aim of the study is not pri- with the observed 21-cm radial velocity of
marily to identify a few more radio sources, the feature, suggest that the distance to the
but to obtain a clear picture of the limita- absorbing cloud is 1 kiloparsec.
tions to which attempts at identification The 21-cm observations show that the
are subjected, in particular for extragalactic feature is 6° across, which gives a diameter
sources. Since the intrinsic strength of of 105 parsecs. The total mass of neutral
sources has a very large dispersion, the hydrogen in the region is 7.8 X 104 times
situation differs for strong and for weak the mass of the sun, giving a mean density
sources. Intrinsically very strong extra- of 5.5 H atoms/cm3. This mass is the same
galactic sources are expected to be optically as the mass of neutral hydrogen in the
of low apparent brightness, since Cygnus Orion region given by T. K. Menon. The
A, the prototype of this class, is already of 21-cm profiles show that the expansion of
18th magnitude. The accuracy of the posi- the region, if any, is less than 3 km/sec.
tions in the present radio surveys does not Photoelectric measures of selected stars in
permit the identification of objects fainter the OB aggregate and in an open cluster
than this, and additional identifications of situated within the area of the absorbing
intrinsically strong radio sources will be cloud are in progress,
possible only when more precise positions The second region under study is at
become available. Intrinsically weak extra- galactic longitude 100°, latitude +11°.
galactic sources may be optically relatively The 48-inch survey shows some absorbing
bright, such as NGC 5128 or NGC 1275. clouds to be present. Deep Ha photo-
Such objects can easily be photographed graphs, taken with the 48-inch schmidt,
and recognized as peculiar at distances at show the presence of faint Ha emission in
which they are beyond the sensitivity limit the same region that has the excess 21-cm
of present radio telescopes. At present, radiation. No star or stars, down to the
only radio sources of intermediate intrinsic 10th magnitude, are known either within
intensity can be expected to be identifiable, the region or near by which could provide
The results obtained in the still-incom- the ionizing radiation to produce the Ha
pleted study show already that at the pres- emission. Thus collisional excitation is
ent stage of technical development not probably present. Further observations are
more than a small percentage of all sources in progress, and calculations will be made
can be identified with galaxies. Since the to check this hypothesis,
available evidence suggests strongly that The 32-foot-diameter radio telescope on
stars are not radio sources, it seems prob- Palomar Mountain is being used by Mat-

64 CARNEGIE INSTITUTION OF WASHINGTON

thews to make a 21-cm survey of the Milky
Way region between galactic longitudes
190° and 250°. The coverage in galactic
latitude extends up to 16° from the plane
of the Galaxy. Preliminary results on the
distribution of neutral hydrogen in the
galactic plane show a good agreement with
the Dutch and Australian results. The
hydrogen belonging to the Orion arms
shows the presence of systematic velocities
in varying amounts up to 10 km/sec be-
tween longitudes 200° and 240°. A survey
of the hydrogen distribution near selected
galactic radio sources is also in progress.
In order to provide independent evidence
on the distance of the Cassiopeia radio
source, a search for faint B stars in that
field was made by Luis Munch on Ha
plates taken with the Tonantzintla pris-
matic camera and a red filter. Three stars
tentatively classified as B stars were found
at distances of 1'6, 2'2, and 53 from the

center of the source as given by Baade and
Minkowski. Photoelectric colors in the
UBV system of these three stars, deter-
mined by Luis and Guido Munch at the
60-inch telescope, have shown that indeed
these stars are of type B, with color ex-
cesses E(B — V) around 1.0 mag. Al-
though no structure in the interstellar lines
of one of these stars has been observed,
the measured K-line radial velocity sug-
gests that the star is in the Perseus spiral
arm. Since the reddening of the source
estimated by Baade and Minkowski is
larger than that of this star, it would seem
that the distance of the source is larger
than 2 kiloparsecs, in agreement with the
distance determination from the 21-cm ab-
sorption lines. Spectroscopic observations
are planned for the two other fainter stars,
which are at closer distance to the center
of the source, in order to make the evi-
dence somewhat stronger.

INSTRUMENTATION

The idea of the pneumatic mirror,
formed by stretching a thin solid film over
an optically finished ring, and aluminizing
it, has been developed by Horace W. Bab-
cock, and experimental work on such a
mirror up to 10 inches in size is being
carried out in the laboratory. Such a mir-
ror, when provided with a backing plate
a short distance behind the film, has pneu-
matic stiffness against fluctuations in at-
mospheric pressure. It also has the inter-
esting property that, in the flat form, its
optical figure is independent of the tem-
perature of the film. The film can readily
be cooled from behind, as is desirable for
mirrors of solar telescopes. In principle, if
a differential pressure is applied to the
opposite sides of a uniform film, a parabo-
loidal figure is obtained.

Swanson has ruled 12 large gratings dur-

ing the year, and has devoted much time
to tests and improvements of various parts
of the ruling machine. The most notable
accomplishment was the ruling of an ex-
cellent grating having the exceptional
width of 10 inches and a groove length of
6 inches. It is not only the largest high-
precision grating ruled here, but it is fully
the equal in quality with any of the earlier
and smaller gratings. Its resolving power,
though not yet quantitatively measured, is
distinctly the best yet seen, showing the
anticipated improvement over 8-inch grat-
ings. This accomplishment proves that the
ruling engine is completely successful up
to the limits of its dimensional capacity,
and that a ruling diamond can, on occa-
sion, produce more than 14 miles of uni-
form precision grooves on a single plate
without appreciable wear.

GUEST INVESTIGATORS

The Observatories have invited a num- required by the programs of the regular
ber of guest investigators to make use of staff. The following studies have been
such observational facilities as were not carried out by these investigators.

MOUNT WILSON AND PALOMAR OBSERVATORIES 65

Dr. George O. Abell, of the Department
of Astronomy at the University of Cali-
fornia at Los Angeles, investigated the
bright end of the luminosity function of
galaxies in rich clusters with the 48-inch
schmidt telescope. For the investigation
about 30 clusters have been chosen from
the catalogue of clusters compiled from the
National Geographic Society-Palomar Ob-
servatory Sky Survey photographs. Magni-
tudes of galaxies (to within about 0.1
mag.) are obtained by extrafocal photo-
graphic photometry, extrafocal images of
galaxies being compared with extrafocal
images of stars which are calibrated with
photographic transfers from Selected
Areas. In addition to the luminosity func-
tion, information is obtained about the
spatial distribution within clusters of gal-
axies of various luminosities and (pre-
sumably) masses. The observational phase
of the program is now about 30 per cent
complete.

Observations of 26 planetary nebulae
have been secured with a photoelectric
scanner attached to the 60- and 100-inch
telescopes by Dr. Lawrence H. Aller and
Dr. William Liller, of the University of
Michigan. Disperson is provided by a 600
line/mm reflection grating, with identical
//5 Newtonian systems used as the colli-
mating and focusing units. The spectrum
is scanned with either a blue-sensitive or
a red- and infrared-sensitive photomulti-
plier permitting a study of the spectrum
from A3200 to A12,000 A. The total nebular
brightnesses were measured photoelectri-
cally at Mount Wilson in 1954. By com-
bining both the direct and spectral observa-
tions with previously obtained photo-
graphic data for the fainter lines and the
isophotic contours, Drs. Aller and Liller
expect to obtain a better assessment of
stratification effects together with im-
proved ionic concentrations and electron
temperatures. The variability of the spec-
trum of IC 4997 has been established.

Dr. Dinsmore Alter, of the Griffith Ob-
servatory, has continued his photographic
observations of the moon with the 60-inch

telescope on 11 nights of above-average
seeing. Photographs were taken in blue-
violet and in infrared light, special atten-
tion being given to the regions of Ptole-
maeus, Alphonsus, Arzachel, and Atlas
under a low setting sun in an attempt to
obtain evidence for scattering by traces of
escaping gas.

Dr. James Cuffey, of Indiana University,
obtained photoelectric measures of colors
and magnitudes of faint stars in the globu-
lar clusters M 53 and NGC 5466 with the
60- and 100-inch reflectors in April 1957.
The photoelectric standards are being used
to calibrate photographic observations of
the color-magnitude relations as faint as
the 19th magnitude.

The solar furnace located on the roof of
Robinson Hall at the California Institute
has been reconditioned and used for high-
temperature materials studies by the Stan-
ford Research Institute in collaboration
with Dr. Paul Duwez, of the Department
of Engineering of the California Institute.
The great advantage of a solar furnace in
materials research is the high concentra-
tion of radiant heat over a small area (l/2
inch in diameter in the present apparatus) .
Because of this highly localized heat flux,
melting of a portion of a solid sample may
be achieved so that the material under
study serves as its own crucible. Refrac-
tory substances that react with any known
crucible can therefore be melted without
being contaminated. Compounds involv-
ing two of the most refractory oxides,
namely, thorium oxide (3200° C±100° C)
and zirconium oxide (2750° C), were suc-
cessfully melted in the furnace. The struc-
ture of these compounds was investigated
by X-ray diffraction methods, and the re-
sults of these studies led to a better under-
standing of one of the most heat-resisting
solid materials of engineering interest. At
present, the solar furnace is being used for
the study of compounds of uranium di-
oxide and zirconium dioxide, which are of
great potential interest in the development
of fuel elements for high-temperature nu-
clear reactors.

66 CARNEGIE INSTITUTION OF WASHINGTON

Dr. Carlos Jaschek, of the Observatorio cases are found it will be possible to apply

Astronomic© de la Universidad Nacional the usual methods, involving intensities

de La Plata, obtained an extensive series of and displacements of the interstellar lines,

spectra of metallic-line stars in order to to derive mean luminosities,
ascertain the existence of families among Studies of the air currents and "seeing"

these stars in analogy to the existing fami- at the 60-foot solar tower have been carried

lies among peculiar stars. Spectra of 35 out by Dr. R. B. Leighton, of the Physics

stars of this type were taken with the 60- Department of the California Institute,

inch telescope at a dispersion of 21 A/mm. These investigations have shown that it is

Five spectra of peculiar A-type stars were possible to reduce the effects of local

taken with the 100-inch coude spectrograph thermal air currents by appropriate treat-

for a detailed analysis of the atmospheres ment of exposed surfaces near the optical

of these objects. path through the tower, and there appears

Spectrograms of T Coronae Borealis at to be considerable hope of prolonging the

10 A/mm were obtained by Dr. Robert P. period of early-morning good seeing to an

Kraft, of Indiana University, in an attempt hour or more.

to classify the motions of the components Three hundred feet of 16-mm Koda-

of this binary star. Dr. Kraft also observed chrome film were exposed by Dr. Leighton

the spectra of several of the fainter mem- in a photographic study of Mars during

bers of the open cluster NGC 6664 in order the period August 6 to October 15, 1956.

to obtain radial velocities and spectral Many photographs of excellent quality

types. were obtained. Study of these photographs

Spectrograms of 39 long-period variable is still in progress,
stars with types MO to M5 were taken with Visual observations of double stars were
the 4-inch camera on the X spectrograph made with the 60-inch reflector at the
of the 60-inch by Dr. Philip C. Keenan, of Cassegrain focus on portions of the nights
the Perkins Observatory. This group of of August 16 to 19 and all night at the
variables was selected because they are Newtonian focus on August 20-21 by Dr.
made up, at least in part, of Population II William Markowitz, of the U. S. Naval
stars which include some with luminosity Observatory. The seeing was good enough
high enough to make them valuable as on all nights to permit measurement of
distance indicators for moderately distant close doubles. The Airy disks were seen
stellar systems. The spectra give evidence on all nights. The powers generally used
of a considerable spread in luminosity were 1300 at the Cassegrain and, with a
within the group. RT Cygni and Z Ophiu- Barlow lens, 2000 at the Newtonian,
chi, for example, are probably supergiants Images obtained in the second manner
at least as bright as a Ononis, while the were superior to those obtained at the
least luminous stars in the group are more Cassegrain without a Barlow lens. Thirty-
comparable to ordinary red giants. More one measures of 23 pairs were made. Most
data will be needed to fix the scale of of the pairs had separations from 0'.'09 to
absolute magnitudes of these variables, but 0'.'19. L726-8 was measured. The results
one means of doing that was found when of these tests indicate that the 60-inch may
two of the stars, R Trianguli and X Mono- be used to a limited extent for the meas-
cerotis, were photographed in the red urement of very close pairs and in search-
region with the 100-inch coude spectro- ing for duplicity in suspected faint stars,
graph. The large velocity shifts in these The co-operative program with the Mc-
spectra made it possible to see interstellar Math-Hulbert Observatory was continued
components of the sodium D line, and throughout the year as in the past 6 years,
these components turned out to be quite The observer for the project stationed on
strong in both stars. When more such Mount Wilson was Mr. Thomas K. Jones.

MOUNT WILSON AND PALOMAR OBSERVATORIES

67

Dr. Robert R. McMath spent a number of
days in February inspecting the Snow
telescope instrument and reviewing the
various aspects of the program. The pro-
gram for the Snow telescope for the cur-
rent report year has been the following:
(1) continuation of the systematic obser-
vation of the infrared helium line (10830
A) at the limbs of the sun and in the plage
regions; (2) systematic observation of the
central structure of the K line (3934 A)
at the limbs of the sun and in plage re-
gions; (3) observations of a number of se-
lected lines in the region 7500 to 12000 A
that should be good indicators of physical
conditions on the sun; (4) observations of
a number of lines that are badly blended
with water-vapor lines (such as Hot) for
comparison with McMath-Hulbert tracings
made under conditions of high water-
vapor content in the earth's atmosphere;
(5) continuation of infrared-sunspot trac-
ings as suitable spots have developed on
the solar disk; (6) repetition of tracings
for wavelength measurement in the 3 to
5 u region with the Lallemand PbTe cell.

Observations were made on 120 days;
760 tracings were produced. These totals
are considerably less than in previous years
because instrumental troubles with the
spectrometer made necessary an almost
complete overhaul with considerable re-
building of worn parts of the mechanism.

The observations on dry days on Mount
Wilson are proving to be very valuable for
eliminating the effect of water-vapor lines
from tracings made with higher dispersion
at the McMath-Hulbert Observatory with
the McMath-Hulbert vacuum spectro-
graph.

The programs of observation of the Ca+
K lines and the 10830 A line of helium are
now approaching a critical stage as the sun-
spot activity increases. The continuation
of the recording throughout the present
maximum will complete the observational
history of the variation of these lines
throughout a sunspot cycle.

With the present Babcock gratings and
the reconditioned spectrometer, measure-

ments of improved precision are being
made in the 3 to 5 |j region. The new
measurements are being made on tracings
obtained with a Lallemand PbTe cell,
whose increased sensitivity with respect to
earlier cells also makes increased accuracy
possible.

Spectroscopic observations of the eclips-
ing systems RZ Scuti and U Cephei were
made by Dr. D. H. McNamara, of Brig-
ham Young University. The Ha line in
the spectrum of RZ Set is of great interest.
The absorption line is bordered by emis-
sion that is particularly pronounced during
eclipse but is also faintly present outside of
eclipse. The Ha absorption line undergoes
a change in width as well as in intensity
during the primary eclipse. When the
brighter component emerges from eclipse,
the line is 1.3 times wider than at other
phases. Asymmetries can be detected in
the Ha line several days before and after
eclipse. For U Cephei a new velocity
curve has been derived from plates taken
with the 60-inch telescope. Of particular
interest is the discovery of a new distortion
in the velocity curve in the form of a rapid
rise and decline in the velocity of the order
of 25 km/sec occurring near phase 0.8 day.

Dr. \Y. W. Morgan, of the Yerkes Ob-
servatory, continued his work on the spec-
tral classification of galaxies which he had
started earlier in collaboration with Dr.
N. U. Mayall, of the Lick Observatory.
While at the Mount Wilson and Palomar
Observatories he studied the files of spec-
trograms obtained by Humason and the
direct photographs of Hubble, Humason,
Sandage, and Baade. A period of five
nights with the 100-inch nebular spectro-
graph was used to obtain new spectra of
several elliptical galaxies; these made possi-
ble the determination of spectral types of
the systems in the ultraviolet. A discussion
of the above material, together with spec-
trograms obtained by Mayall at Lick, has
led to the following conclusions: The
great majority of the galaxies can be classi-
fied by their spectra into about five classes;
if these are denoted by the spectral types

CARNEGIE INSTITUTION OF WASHINGTON

as determined in the region of 4000 A, the
spectroscopic groups range from A to K.
There is a marked correlation between
the degree of central condensation and the
spectral type of the system as a whole;
those of type A show little or no central
condensation; on the other hand, those of
type K consist of the ellipticals and the
Sa and Sb systems having pronounced
central condensations.

From a study of objective-prism plates
taken with the schmidt camera of the
Tonantzintla Observatory a number of
hitherto unclassified OB stars in longitude
82° to 92° have been selected in order to
determine to what extent the OB associa-
tions I, II, and III Cassiopeiae are separated
from one another in space. Fifty stars have
been observed with the 4-inch camera of
the X spectrograph of the 60-inch telescope,
and their spectral types on the Yerkes sys-
tem have been determined by Luis Munch,
of the Tonantzintla Observatory. It is
planned to reobserve spectroscopically as
many of these as possible to derive radial
velocities. The photoelectric colors and
magnitudes of these stars are also being
measured to obtain spectroscopic parallaxes.

A number of stars classified on objective-
prism plates as peculiar A stars have been
observed with the X spectrograph in order
to relate the criteria of peculiarity with
those established in the Yerkes system.
The stars BD -4°1644, -1°1414, and
— 3° 1665 were also found to be spectrum
variables, and their variations are being
studied. A period of 4.1 days has been de-
rived for the variation of Ca II K in the
spectrum of BD +46° 1913. Photoelectric
observations of these same objects are being
carried out with the 20-inch telescope of
Palomar Mountain.

The radial velocities of the most distant
early-type stars known in galactic longi-
tudes 325° to 10° are being determined
from plates taken with the 8-inch camera
of the X spectrograph by Luis Munch in
collaboration with Guido Munch. The
purpose of this program is to reanalyze
the problem of the discrepancy between

the rotational velocity of the inner parts of
the galactic system as determined from
stellar radial velocities and 21-cm-line radio
observations.

Dr. L. Plaut, of the Kapteyn Astronomi-
cal Observatory, Groningen, has completed
the observational part of his investigation
of the large-scale structures of the halo of
the galactic system. Nearly 500 plates were
taken with the 48-inch schmidt camera.
They are now being searched in Gronin-
gen for variable stars. Four areas are being
investigated, centered at the following ga-
lactic co-ordinates: /=327°5, £=+28°;
l = 327°5, £=-12°; /=331?0, £=+12°;
and/=147°5, £=+15°.

Dr. Daniel M. Popper, of the University
of California at Los Angeles, has continued
his investigation of the orbits and masses
of eclipsing binaries. He has recently re-
discussed the eclipsing binary Z Vulpeculae
on the basis of new photoelectric observa-
tions at Palomar as well as of spectroscopic
observations. The principal interest in Z
Vul is that it contains an A-type star of
luminosity class III, the first star of this
type to have reliable determinations of
mass and radius. During the current re-
port year, one star, V 477 Cygni, has been
added to the list of systems in need of re-
vision of masses and radii, and two stars,
RY Persei and RS Vulpeculae, have been
added to the list of systems with com-
ponents above the main sequence for
which masses will eventually be deter-
mined. Observations on RX Herculis and
RS Canum Venaticorum have been nearly
completed. The mass of the former system
is found to be 30 per cent larger, that of
the latter about 40 per cent smaller, than
the previously published values. Clearly,
higher dispersion than heretofore must be
applied to many systems. Fair progress
has been made on the difficult problem of
determining the masses of £ Aurigae from
spectrograms. Although final results are
not available, the published value for the
mass of the cool supergiant appears to need
serious revision downward.

Drs. Otto Struve and Jorge Sahade, of

MOUNT WILSON AND PALOMAR OBSERVATORIES 69

the University of California at Berkeley,
have continued their spectroscopic observa-
tions throughout the year. Among the re-
sults obtained from these observations are :

1. Emission lines (strong at Ha) were
discovered in the spectrum of Algol. The
emission is observed at quadratures, and is
different in character from the emission
previously observed in eclipsing systems.

2. Ha emission was also discovered in
3 Cephei, U Coronae Borealis, and HD
47129.

3. Individual cycles were found to differ
in velocity amplitude in h Delphini, the
total range varying from 1-2 km/sec to
4-5 km/sec.

4. Extensive studies were made of the
spectrum of e Aurigae, which emerged
from its last eclipse in May 1957. All the
radial velocities have been measured, and
the results indicate large changes of an ir-
regular nature which are not caused by
the geometrical properties of the double-
star system. A spectrophotometric study of
the best plates during and outside of
eclipse, made in Berkeley with the co-
operation of Dr. Margherita Hack, has
indicated that the general character of the
eclipsing body is very much like that previ-
ously described by Struve. The invisible
nucleus in the center of the large eclipsing
shell may be a subluminous B-type star.

5. The spectrograms of 3 Lyrae have
been enlarged and assembled in the form
of an atlas. The radial velocities of 3
Lyrae have been used to determine the
orbital elements with the help of the IBM
701 in Berkeley. The precision of this orbit
is of a high order, and will form the basis
for future studies of the perturbations in
the system.

6. Observations have been made of the
following additional stars: W Serpentis,
AZ Cassiopeiae, h Capricorni, 12 Lacertae,
and 17 Leporis.

Professor A. Unsold, of the University
of Kiel, obtained a series of spectrograms
for a detailed study of the atmospheres of
several stellar types. They included spectra
of medium-type subdwarfs, red and infra-

red spectra of normal bright stars, spectra
of later-type main-sequence stars, and spec-
tra of M dwarfs of emission and nonemis-
sion types.

Photoelectric observations of the short-
period eclipsing binary Nova DQ Herculis
(1934) were obtained on five nights during
July and August with the 100-inch reflector
by Dr. Merle Walker, of the Warner and
Swasey Observatory. From these observa-
tions, the following improved elements
have been derived:

Min = Hel JD 2434954.94475 + 0<?19362060E

An analysis shows that the form of the
eclipse curve has changed since 1954, pre-
sumably owing to the presence of gas
streams or of detached material in the sys-
tem. It is not possible to obtain reliable
photometric elements from the 1956 obser-
vations, and the change in the light-curve
throws considerable doubt on the validity
of the photometric elements derived from
the 1954 observations. The observations
were made in yellow light to avoid the
effect of the nebulosity surrounding the
system. A more detailed study of the 1-
minute periodic variations has been made
from the 1956 observations. It has been
possible to derive the period of these varia-
tions over a 2-day interval. The elements
are

Max = Hel JD 2435660.71 13 + 0<?000822528E

for the observations made in July; for those
in August, the epoch is

Hel JD 2435695.7499

The change in the visibility and amplitude
of these oscillations with the phase in the
4-hour period suggests that they originate
on the hemisphere of the Nova facing the
secondary star. There are complications,
however.

Simultaneous photometric and spectro-
scopic observations were obtained of AE
Aquarii by Dr. Walker in collaboration
with Deutsch. The photometric observa-
tions, made with the 60-inch reflector, were
used to direct the spectroscopic observer at

70 CARNEGIE INSTITUTION OF WASHINGTON

the 100-inch to place the star in one of two
regions of the slit, depending upon whether
it was momentarily bright or faint. In
this way two spectra were eventually built
up, corresponding to the maxima and
minima of the rapid variations in light
displayed by the star. In the red, the
equivalent widths of all absorption and
emission lines, including Hoc, appear to be
the same in the two spectra. In the photo-
graphic region, the continuous spectrum
is much bluer at maximum light, and all
emission and absorption lines have de-
cidedly smaller equivalent widths, includ-
ing the Balmer lines.

The infrared spectra of bright stars were
examined by Dr. A. E. Whitford, of Wash-
burn Observatory, University of Wiscon-
sin, using a germanium photodiode at the
exit slit of the scanning spectrograph. The
new detector shows considerable promise
in the range from 10000 to 16000 A, having
a detection limit about 100 times better

than lead sulfide photoconductive cells. An
exit slit 20 A wide gave a resolving power
6 to 10 times better than had previously
been achieved on stars beyond the photo-
graphic limit. This resolution was ade-
quate to show the stronger atomic absorp-
tion lines, such as the Paschen series of
hydrogen, and molecular absorption bands.
The detector sensitivity would permit go-
ing to a resolution of 2 to 4 A, which
would bring out subordinate metallic lines;
a different optical arrangement would be
necessary, however.

Over the more conventional photomulti-
plier range of 3400 to 10000 A, Dr. Whit-
ford scanned the spectrum of NGC 4374,
one of the bright elliptical galaxies in the
Virgo cluster, in order to compare its spec-
tral energy distribution with that previ-
ously found for M 32 by Code. Selected
pairs of reddened and unreddened stars
were also observed with the scanner as a
test of the law of reddening.

STAFF AND ORGANIZATION

Dr. Milton L. Humason and Dr. Seth B.
Nicholson retired from the staff of the
Observatories on June 30, 1957. Dr. Huma-
son joined the Mount Wilson Observatory
in 1917, first as janitor and then as night
assistant. In the latter position he displayed
such skill as an observer that he was made
a member of the Staff of Investigators in
1922. He first assisted Dr. Merrill in a
survey for early-type stars with bright
hydrogen lines. Later he collaborated with
Drs. Adams and Joy and Miss Brayton in
their very extensive study of stellar abso-
lute magnitudes and spectroscopic paral-
laxes, which resulted in the publication of
the magnitudes and parallaxes of 4179 stars
in 1935.

In the course of these studies Humason
developed a very unusual proficiency in the
photography of spectra of very faint ob-
jects. After the discovery by Hubble in the
1920's of the major role played by the
galaxies in the structure of the universe,
Humason turned his attention to the spec-
tra of these objects and soon accumulated

spectra of a substantial number of these
galaxies spread over a wide range of dis-
tances. A study of the relationship be-
tween the velocities as measured on these
spectrograms and the distances of these
galaxies led Hubble to the concept of the
expanding universe. For the next quarter
century Humason devoted most of his at-
tention to this problem. The introduction
of extremely fast photographic plates, the
development of new and very rapid spec-
trographs, and the completion of the 200-
inch Hale telescope enabled Humason to
push his observations to fainter and fainter
and therefore more and more distant
galaxies. These techniques now permit
photographing the spectra of galaxies far
too faint to be seen visually with the tele-
scope used to collect the light. Humason
therefore had to develop elaborate offset
procedures that ensure locating invisible
images accurately on the slit of the spectro-
graph and holding them there during long
exposures. His studies culminated in the
publication in 1956, in collaboration with

MOUNT WILSON AND PALOMAR OBSERVATORIES 71

Dr. N. Mayall, of the Lick Observatory,
and with Dr. Sandage, of the velocities of
over 900 galaxies. Some of these velocities
are as high as one-fifth that of light.

In 1948 Dr. Humason was appointed
Secretary of the Observatories. As Secre-
tary he has ably handled the correspond-
ence and public relations as well as many
of the other administrative problems of the
Observatories.

Dr. Nicholson joined the staff of the
Mount Wilson Observatory in 1915. Dur-
ing the first few years he investigated the
orbits of several of Jupiter's satellites, the
ninth of which he had discovered at Lick
Observatory in 1914. In collaboration with
Dr. Pettit he developed a very sensitive
vacuum thermocouple. This they used to
measure the total radiation and surface
temperature of stars, the planets, and the
moon. Their data on the cool long-period
variables were of special value. Studies of
the rates of cooling of the moon's surface
during an eclipse gave a measure of the
thermal conductivity of the surface rocks
and provided information on their physical
characteristics. In the late 1930's and again
in the early 1950's Dr. Nicholson returned
to the observation of Jupiter's satellites,
discovering the tenth, eleventh, and twelfth
of these objects and determining the posi-
tions necessary to fix their orbits.

Throughout Dr. Nicholson's 42 years at
the Observatories a large part of his efforts
has been devoted to solar observations, at
first in collaboration with Dr. Hale. He
has developed an extraordinarily detailed
knowledge of the complex phenomena of
the sun's visible surface. He has supervised
the systematic collection of data on sun-
spots, including the polarity and strength
of their magnetic fields. In collaboration
with Dr. Oliver Wulf, of the U. S.
Weather Bureau, he has made detailed in-
vestigations of the correlation between
solar and terrestrial phenomena.

Mr. Edgar C. Nichols retired as Chief
Designer and Superintendent of the Instru-
ment Shop on February 28, 1957, after 46
years of service at the Observatories. Many

of the instruments on Mount Wilson owe
much of their efficiency and ease of opera-
tion to Mr. Nichols' skill as a designer.

Dr. Horace W. Babcock was appointed
Assistant Director of the Observatories
effective January 1, 1957. Dr. Arthur D.
Code became a member of the Staff of the
Observatories on September 1, 1956, and
Dr. Halton C. Arp on July 1, 1957.

Research Division
Staff Members

Halton C. Arp
Walter Baade

Horace W. Babcock, Assistant Director
William A. Baum
Ira S. Bo wen, Director
Arthur D. Code
Armin J. Deutsch
Jesse L. Greenstein

Milton L. Humason, Secretary of the Ob-
servatory *
Rudolph L. Minkowski
Guido Munch
Seth B. Nicholson x
Donald E. Osterbrock
Robert S. Richardson
Allan R. Sandage
Olin C. Wilson
Fritz Zwicky

Carnegie Research Fellows
Geoffrey R. Burbidge
Thomas A. Matthews
Maarten Schmidt

Research Assistants

Sylvia Burd
Mary F. Coffeen
Thomas A. Cragg
Dorothy S. Deutsch
Edith Flather
Emil Herzog
Joseph O. Hickox
A. Louise Lowen
Mildred Matthews
Carol Nordquist
Henrietta H. Swope

Student Observers
George O. Abell
Walter K. Bonsack
Jacques Feige

1 Retired June 30, 1957.

72 CARNEGIE INSTITUTION OF WASHINGTON

William G. Tifift
Dale Vrabec
George Wallerstein

Editor and Librarian
Alexander Pogo

Photographer

William C. Miller

Instrument Design and Construction

Lawrence E. Blakee, Electronic Technician

Floyd E. Day, Optician

Kenneth E. DeHuff, Machinist

Robert D. Georgen, Machinist

Don O. Hendrix, Superintendent, Optical

Shop
Melvin W. Johnson, Optician
Edgar C. Nichols, Chief Designer, and

Superintendent, Instrument Shop 2
Bruce Rule, Project Engineer
Oscar Swanson, Instrument Maker

Maintenance and Operation

Mount Wilson Observatory and Offices
Audrey A. Acrea, Stewardess
Paul F. Barnhart, Truck Driver
Ashel N. Beebe, Superintendent of Con-
struction

2 Retired February 28, 1957.

Wilma J. Berkebile, Secretary
Ernest V. Cherry, Janitor
Hugh T. Couch, Carpenter
Eugene L. Hancock, Night Assistant
Emerson W. Hartong, Truck Driver
Anne McConnell, Administrative Assistant
Leah M. Mutschler, Stenographer and Tele-
phone Operator
Bula H. Nation, Stewardess
Alfred H. Olmstead, Night Assistant
Arnold T. Ratzlaff, Night Assistant
Clyde Sanger, Gardener
John E. Shirey, Janitor and Relief Engineer
Benjamin B. Traxler, Superintendent

Palomar Observatory and Robinson Labora-
tory

Fred Anderson, Machinist

Dorothea Davis, Secretary

Eleanor G. Ellison, Secretary and Librarian

Ferd Feryan, Mechanic

Arlis Grant, Stewardess

Leslie S. Grant, Relief Night Assistant and
Mechanic

Byron Hill, Superintendent

Charles E. Kearns, Night Assistant

Harley C. Marshall, Office Manager

George W. Pettit, Janitor

Robert E. Sears, Night Assistant

William C. Van Hook, Electrician and As-
sistant Superintendent

Gus Weber, Assistant Mechanic

BIBLIOGRAPHY

Abell, G. Ogden. The distribution of rich clus-
ters of galaxies. Read at 1956 Berkeley
meeting of Am. Astron. Soc; (abstract)
Astron. J., 62, 2 (1957).

Adams, Walter S., and Paul W. Merrill. Mount
Wilson spectrograms of P Cygni. Astrophys.
J., 125, 102-106 (1957).

Aller, L. H., and R. Minkowski. The interpre-
tation of the spectrum of NGC 7027. Astro-
phys. ]., 124, 110-115 (1956).

Aller, L. H. See also Minkowski, Rudolph.

Arp, Halton C. Three-color photometry of
Cepheids W Virginis, M5 Nos. 42 and 84,
and M10 Nos. 2 and 3. Astron. /., 62,
129-136 (1957).

Arp, Halton C., and George Wallerstein. Ceph-
eids in M2. Astron. J., 61, 272 (1956).

Baade, Walter. Wieweit sind Milchstrasse und
Andromeda-Nebel im Aufbau vergleichbar?
Mitt. Astron. Ges., 1955, 51-59 (1956).

Baade, Walter, G. R. Burbidge, F. Hoyle, E. M.
Burbidge, R. F. Christy, and W. A. Fowler.
Supernovae and californium 254. Pubs.
Astron. Soc. Pacific, 68, 296-300 (1956).

Babcock, Horace W. The magnetic variable
HD 71866. Astrophys. J., 124, 489-498
(1956).

Babcock, Horace W. Review of The history of
the telescope, by H. C. King. /. Opt. Soc.
Am., 47, 191 (1957).

Baker, Robert H., and Hildred V. Baker. Ultra-
violet light-curves of selected variable stars
in M3. Astron. J., 61, 283-289 (1956).

Bappu, M. K. Vainu. See Wilson, O. C.

Baum, William A. Globular clusters observed
through a crystal ball. Smithsonian Contribs.
to Astrophys., 1, 165-176 (1956).

Baum, William A. Electronic photography of
stars. Astron. Soc. Pacific Leaflet 326, 8 pp.
(1956).

Baum, William A. The detection of faint im-
ages against the sky background. Trans.
Intern. Astron. Union, 9, 681-687 (1957).

Baum, William A. The polarization of some
selected patches in the Crab Nebula. Pubs.
Astron. Soc. Pacific, 69, 231-232 (1957).

Baum, William A. Photoelectric determination
of redshifts beyond 0.2c. Read at 1956

MOUNT WILSON AND PALOMAR OBSERVATORIES 73

Berkeley meeting of Am. Astron. Soc.; (ab-
stract) Astron. ]., 62, 6-7 (1957).

Bonsack, Walter K., and others. The interpre-
tation of photoelectric colors for stars of
types B-F. Astrophys. ]., 125, 139-152
(1957).

Bowen, Ira S. Optics. Smithsonian Contribs. to
Astrophys., 1, 1-3 (1956).

Bowen, Ira S. Mount Wilson and Palomar Ob-
servatories. (Reports of observatories, 1955-
1956.) Astron. J., 61, 336-341 (1956).

Burbidge, E. Margaret, and G. R. Burbidge.
The chemical composition of five stars which
show some of the characteristics of Popula-
tion II. Astrophys. J., 124, 116-129 (1956).

Burbidge, E. Margaret, and G. R. Burbidge. On
the possible presence of He3 in the magnetic
star 21 Aquilae. Astrophys. ]., 124, 655-659
(1956).

Burbidge, E. Margaret. See also Baade, Walter;
Burbidge, Geoffrey R.; Hoyle, Fred.

Burbidge, Geoffrey R. On cluster-type variables
and magnetic fields. Astrophys. }., 124, 412—
415 (1956).

Burbidge, Geoffrey R. On synchrotron radia-
tion from Messier 87. Astrophys. /., 124,
416-429 (1956).

Burbidge, Geoffrey R. Galactic radio emission
and the energy released in nuclear collisions
of primary cosmic-ray protons. Phys. Rev.,
103, 264-265 (1956); 104, 1808 (1956).

Burbidge, Geoffrey R. The Crab Nebula: a cos-
mic synchrotron. Astron. Soc. Pacific Leaf-
let 332, 8 pp. (1957).

Burbidge, Geoffrey R., and E. M. Burbidge.
Anomalous abundances of manganese, stron-
tium, and europium in HD 151199. As-
trophys. ]., 124, 130-133 (1956).

Burbidge, Geoffrey R., and E. M. Burbidge.
The sources of radio emission in NGC 5128
and NGC 1316. Astrophys. J., 125, 1-8
(1957).

Burbidge, Geoffrey R., and F. Hoyle. Matter
and anti-matter. // Nuovo Cimento, 4, 558-
564 (1956).

Burbidge, Geoffrey R., and F. Hoyle. Matter
and anti-matter. Read at 1956 Berkeley
meeting of Am. Astron. Soc; (abstract)
Astron. ]., 62, 9 (1957).

Burbidge, Geoffrey R., F. Hoyle, E. M. Burbidge,
R. F. Christy, and W. A. Fowler. Cali-
fornium-254 and supernovae. Phys. Rev.,
103, 1145-1149 (1956).

Burbidge, Geoffrey R. See also Baade, Walter;
Burbidge, E. Margaret; Hoyle, Fred.

Christy, R. F. See Baade, Walter; Burbidge,
Geoffrey R.

Cragg, Thomas A. The eruptive prominence
of May 18, 1956. Pubs. Astron. Soc. Pacific,
69, 268-270 (1957).

Cragg, Thomas A. Solar activity in 1956. Pubs.
Astron. Soc. Pacific, 69, 166-168 (1957).

Deutsch, Armin J. The peculiar A stars. In
Vistas in astronomy, ed. by A. Beer, 2, 1421-
1427 (1956).

Deutsch, Armin J. The dead stars of Popula-
tion I. Pubs. Astron. Soc. Pacific, 68, 308-
317 (1956).

Deutsch, Armin J. See also Miczaika, G. R.

Flather, Edith. See Munch, Guido.

Fowler, William A. See Baade, Walter; Bur-
bidge, Geoffrey R.; Hoyle, Fred.

Franklin, F. A. See Miczaika, G. R.

Greenstein, Jesse L. The technique and possi-
bilities of astronomical spectrophotometry.
In Vistas in astronomy, ed. by A. Beer, 2,
1299-1308 (1956).

Greenstein, Jesse L. The spectra and other
properties of stars lying below the normal
main sequence. Proc. 3d Berkeley Sym-
posium Math. Statistics and Probability , 3,
11-29 (1956).

Greenstein, Jesse L. Comments on Hsue-shen
Tsien's paper on thermonuclear power
plants. Jet Propulsion, 26, 564, 575 (1956).

Greenstein, Jesse L. Evolution and the white
dwarfs. Read at 1956 Berkeley meeting of
Am. Astron. Soc; (abstract) Pubs. Astron.
Soc. Pacific, 68, 501-503 (1956).

Greenstein, Jesse L. Evidence for instability
among sub-luminous stars. In Intern. As-
tron. Union Symposium 3, ed. by G. H.
Herbig, p. 41. Cambridge, University Press,
1957.

Greenstein, Jesse L. Report, Commission 29, Stel-
lar spectra. Trans. Intern. Astron. Union,
9, 386-416 (1957).

Greenstein, Jessse L. Some high-velocity Popu-
lation I stars. Read at 1956 Berkeley meet-
ing of Am. Astron. Soc; (abstract) Astron.
]., 62, 16 (1957).

Greenstein, Jesse L. See also Bonsack, Walter
K.; Miczaika, G. R.; Sanford, Roscoe F.

Hoyle, Fred, and A. Sandage. The second-
order term in the redshift-magnitude rela-
tion. Pubs. Astron. Soc. Pacific, 68, 301-307
(1956).

Hoyle, Fred, W. A. Fowler, G. R. Burbidge, and
E. M. Burbidge. Origin of the elements in
stars. Science, 124, 611-614 (1956).

Hoyle, Fred. See also Baade, Walter; Burbidge,
Geoffrey R.

Humason, Milton, L. Red-shifts in the spectra
of extragalactic nebulae. In Vistas in as-
tronomy, ed. by A. Beer, 2, 1620-1628 (1956).

Johnson, H. L., A. R. Sandage, and H. D. Wahl-
quist. The galactic cluster Mil. Astrophys.
]., 124, 81-89 (1956).

Johnson, H. L., and A. R. Sandage. Three-color
photometry in the globular cluster M3. As-
trophys. ]., 124, 379-389 (1956).

Joy, Alfred H. Radial-velocity measures of SS
Cygni at minimum light. Astrophys. ]., 124,
317-320 (1956).

74 CARNEGIE INSTITUTION OF WASHINGTON

Joy, Alfred H. Walter Sydney Adams, 1876—
1956. Pubs. Astron. Soc. Pacific, 68, 285-
295 (1956).

Kraft, Robert P. The Ca II emission in classical
Cepheid variables. Read at the 1956 Berke-
ley meeting of Am. Astron. Soc; (abstract)
Astron. J, 62, 21 (1957).

Mathis, J. S. Statistical equilibrium of triplet
levels of neutral helium. Astrophys. J., 125,
318-327 (1957).

Mathis, J. S. The ratio of helium to hydrogen
in the Orion Nebula. Astrophys. ]., 125,
328-336 (1957).

Mathis, J. S. See also Bonsack, Walter K.

Matthews, Thomas A. Observations of neutral
hydrogen in the 21 -cm line at various lati-
tudes near galactic longitude 100°. Read at
1956 Berkeley meeting of Am. Astron. Soc;
(abstract) Astron. J., 62, 25-26 (1957).

Merrill, Paul W. Spectroscopic classification of
red stars. Pubs. Astron. Soc. Pacific, 68,
356-358 (1956).

Merrill, Paul W. Oxygen in the stars and nebu-
lae. Astron. Soc. Pacific Leaflet 328, 8 pp.
(1956).

Merrill, Paul W. Stars with expanding atmos-
pheres. In Vistas in astronomy , ed. by A.
Beer, 2, 1375-1380 (1956).

Merrill, Paul W. A spectrogram of R Hydrae
near minimum light. Pubs. Astron. Soc.
Pacific, 69, 77-7S (1957).

Merrill, Paul W. See also Adams, Walter S.

Miczaika, G. R., F. A. Franklin, A. }. Deutsch,
and J. L. Greenstein. A spectrophotometric
analysis of two metallic-line stars. Astrophys.
J., 124, 134-154 (1956).

Miller, William C. Part-time explorer. Eng.
and Sci., 20 (no. 4), 23-25 (1957).

Minkowski, Rudolph. Colliding galaxies. Sci.
American, 195, 124-134 (1956).

Minkowski, Rudolph, and L. H. Aller. Spectro-
photometry of planetary nebulae. Astrophys.
J., 124, 93-109 (1956).

Minkowski, Rudolph. See also Aller, L. H.

Munch, Guido. Stochastic processes of astro-
nomical interest. In Applied probability,
Proc. 7th Symposium Applied Math., Am.
Math. Soc, pp. 51-66. New York, McGraw-
Hill, 1957.

Munch, Guido. The age of early B-type super-
giants. Pubs. Astron. Soc. Pacific, 68, 351-
353 (1956).

Munch, Guido. Interstellar absorption lines in dis-
tant stars. I. Northern Milky Way. Astro-
phys. J., 125, 42-65 (1957).

Munch, Guido. The mass of M 81. Read at
1956 Berkeley meeting of Am. Astron. Soc;
(abstract) Astron. ]., 62, 28 (1957).

Munch, Guido, and Edith Flather. The radial
velocity of 53 Arietis. Pubs. Astron. Soc.
Pacific, 69, 142-146 (1957).

Nicholson, Seth B. A sunspot group in excep-
tionally high latitude. Pubs. Astron. Soc.
Pacific, 69, 80-82 (1957).

Nicholson, Seth B. Solar observations. Eng.
and Sci., 20 (no. 9), 14-15 (1957).

Nicholson, Seth B. Comets of recent years.
Astron. Soc. Pacific Leaflet 338, 8 pp. (1957).

Nicholson, Seth B. Review of Vistas in astron-
omy, 1, ed. by A. Beer. Pubs. Astron Soc.
Pacific, 69, 188 (1957).

Nicholson, Seth B., and others. Summary of
Mount Wilson magnetic observations of sun-
spots for March, 1956, to February, 1957.
Pubs. Astron. Soc. Pacific, 68, 365-368, 460-
463, 550-552 (1956); 69, 86-90, 180-183,
270-273 (1957).

Osterbrock, Donald E. Comet-tail structures in
emission nebulae. Astrophys. J., 125, 622—
635 (1957).

Osterbrock, Donald E. Electron densities in the
filaments of the Crab Nebula. Pubs. Astron.
Soc. Pacific, 69, 227-230 (1957).

Osterbrock, Donald E. See also Seaton, M. J.

Parker, James M. Line contours and wave
lengths in the flash spectrum without eclipse.
Read at 1956 Berkeley meeting of Am.
Astron. Soc; (abstract) Astron. ]., 62, 29
(1957).

Pettit, Edison. The achromatic refractor. As-
tron. Soc. Pacific Leaflet 330, 8 pp. (1956).

Pettit, Edison. The reflector. Astron. Soc. Pacific
Leaflet 331, 8 pp. (1956).

Reaves, Gibson. See Walker, Merle F.

Richardson, Robert S. Preliminary report on
observations of Mars made at Mount Wilson
in 1956. Read at 1956 New York meeting
of Am. Astron. Soc; Pubs. Astron. Soc. Pa-
cific, 69, 23-30 (1957).

Richardson, Robert S. Some observations of
Mars made at Mount Wilson in 1956. As-
tron. Soc. Pacific Leaflet 333, 8 pp. (1957).

Rogerson, John B., Jr. Photoelectric observations
of solar-line profiles. Astrophys. J., 125,
275-284 (1957).

Sandage, Allan R. The systematics of color-
magnitude diagrams and stellar evolution.
Read at 1956 Berkeley meeting of Am.
Astron. Soc; (abstract) Pubs. Astron. Soc.
Pacific, 68, 498-500 (1956).

Sandage, Allan R. The red-shift. Sci. Ameri-
can, 195 (no. 3), 171-182 (1956); (no. 5),
14-16 (1956).

Sandage, Allan R. Observational approach to
evolution. I. Luminosity functions. As-
trophys. ]., 125, 422-434 (1957).

Sandage, Allan R. Observational approach to
evolution. II. A computed luminosity func-
tion for K0-K2 stars from Mv = +5 to Mv
= —4.5. Astrophys. J., 125, 435-444 (1957) .

Sandage, Allan R. The birth and death of a
star. Eng. and Sci., 20 (no. 4), 17-21 (1957).

MOUNT WILSON AND PALOMAR OBSERVATORIES 75

Sandage, Allan R. See also Hoyle, Fred; John-
son, H. L.

Sanford, Roscoe F., and J. L. Greenstein. The
absolute magnitude of Nova Puppis 1942.
Pubs. Astron. Soc. Pacific, 69, 75-77 (1957).

Savedoff, Malcolm P. A photographic color-
magnitude array of the globular cluster
M 13. Astron. ]., 61, 254-261 (1956).

Schmidt, Maarten. Exploring our Galaxy with
21-centimeter radio waves. Astron. Soc. Pa-
cific Leaflet 335, 8 pp. (1957); Danish
versions: Nordisl^ Astron. Tidsshjift, no. 2,
pp. 51-56 (1957), and Urania, 14, 41-45
(1957).

Seaton, M. J., and D. E. Osterbrock. Relative
[O II] intensities in gaseous nebulae. As-
trophys. ]., 125, 66-83 (1957).

Tifft, W. G. See Bonsack, Walter K.

Wahlquist, H. D. See Bonsack, Walter K.;
Johnson, H. L.

Walker, Merle F. Studies of extremely young
clusters. I. NGC 2264. Astrophys. J., Suppl.
Ser., 2, 365-387 (1956).

Walker, Merle F., and G. Reaves. Photoelectric
photometry of the lunar eclipse of July 26,
1953. Pubs. Astron. Soc. Pacific, 69, 153-
157 (1957).

Walker, Merle F. See also Wilson, O. C.

Wallerstein, George. The absolute magnitude
of U Sagittarii and its membership in M 25.
Pubs. Astron. Soc. Pacific, 69, 172-175 (1957).

Wallerstein, George. See also Arp, Halton C.;
Bonsack, Walter K.

Wilson, O. C. The new Cassegrain spectro-
graph for the Mount Wilson 60-inch tele-
scope. Pubs. Astron. Soc. Pacific, 68, 346—
350 (1956).

Wilson, O. C. The analysis of the chromosphere
of Zeta Aurigae. /. Roy. Astron Soc. Can.,
51, 70-74 (1957).

Wilson, O. C., and M. F. Walker. Simultaneous
spectrographic and photometric observations
of the short-period variables SX Phoenicis
and CC Andromedae. Astrophys. ]., 124,
325-341 (1956).

Wilson, O. C., and M. K. Vainu Bappu. H and
K emission in late-type stars: dependence
of line width on luminosity and related
topics. Astrophys. J., 125, 661-683 (1957).

Zwicky, F. Relative populations of rich spheri-
cal clusters of galaxies. Pubs. Astron. Soc.
Pacific, 68, 331-338 (1956).

Zwicky, F. Statistics of clusters of galaxies.
Proc. 3d Berkeley Symposium Math. Statis-
tics and Probability, 3, 113-144 (1956).

Zwicky, F. Morphologic Industrielle Organi-
sation, 25, 425-428 (1956).

Zwicky, F. Morphological astronomy. 299 pp.
Berlin-Gottingen-Heidelberg, Springer- Ver-
lag, 1957.

COMMITTEE ON IMAGE TUBES
FOR TELESCOPES

Mount Wilson and Talomar Observatories, Department of Terrestrial Magnetism,
National Bureau of Standards, United States Naval Observatory, and California

Institute of Technology

W. A. BAUM, Mount Wilson and Palomar Observatories

JOHN S. HALL, United States Naval Observatory

L. L. MARTON, National Bureau of Standards

MERLE A. TUVE, Chairman

Carnegie Institution of Washington Year Boo\ 56, 1956-1957

During the past year the Committee has
continued to encourage or foster several
exploratory projects for the development
of image tubes useful for increasing the
range of telescopes. Several of these proj-
ects were outlined in the last report.

Although the secondary-emission image-
multiplier tube still appears to offer the
most attractive over-all solution to the
image-tube problem, serious difficulties
were encountered in attempts to make a
pilot model.

Work both here and abroad has been
carried on in an effort to make an image
converter containing one or more stages
of image intensification. Each stage con-
sists of a very thin membrane with a
phosphorescent screen on one side and a
photocathode on the other. A gain of 10
or more per stage has been reported. These
projects have thus far had very limited
success, because of technical difficulties,
dark emission, and loss of resolution.

During the past year no important ad-
vance, useful to astronomers, which has
to do with the development of an elec-
tronic storage tube with a built-in electri-
cal read-off system, has come to the atten-
tion of the Committee.

The main effort of the Committee was
directed toward the development of thin-
film image converters. Dr. W. Kent Ford,
Jr., continued his important work, with
the help of Committee funds, on the prob-
lem of making suitable thin metal films
at the University of Virginia. The films
transmit electrons but protect the photo-
cathodes of the tubes from the molecules
exuded from the emulsions of the nuclear-
track plates which record the images of
stars. The manufacturer mounted the films
in place of the phosphorescent screens
used in commercial tubes for viewing
images directly. A special glass envelope
was also attached to the rear end of the
tube by the manufacturer to protect the

film from rupture as the tube was evacu-
ated and sensitized. After the finished
tube is installed in a special chamber
mounted on the tail end of the telescope
and a suitable vacuum around the glass
cap is achieved, the cap is removed by
cracking the glass along a circular groove
by means of an electrically heated wire.
The plate is next admitted through an
air lock, and its emulsion is placed about
0.3 mm behind the thin film. Exposures
are made by application of high voltages
to the converter.

Dr. Ford's most successful films consist
of an aluminum coating on Formvar and
have a total thickness of about 0.15 of a
wavelength of visible light. A large per-
centage of these films, mounted on Kovar
rings, withstand baking for several hours
at 300° C and are strong enough to be
shipped through the mails.

One highly sensitive converter, having
a thin film made by Ford and an Sll
photocathode, was tested on the Naval Ob-
servatory's 40-inch telescope at Flagstaff,
Arizona, by Drs. W. A. Baum and J. S.
Hall. Unfortunately, the thin film was
broken as the glass cap was removed just
a few seconds before the first exposure
was made. The contamination of the
cathode was so rapid that no stellar image
appeared. This rupture of the thin film
was doubtless caused by gas pressure gen-
erated in the cap just as the glass was
heated by the wire during the breakoff
procedure. This phenomenon was not
experienced during a test reported a year
ago; in that test, however, the thin film
contained pinholes, and only one useful
exposure could be made.

In August 1957, after the close of this
report period, a tube of low sensitivity was
successfully used to record the images of
several bright stars with the 40-inch tele-
scope at the United States Naval Ob-
servatory.

79

DEPARTMENT OF TERRESTRIAL MAGNETISM

Washington, District of Columbia MERLE A. TUVE, Director

RICHARD B. ROBERTS, Acting Director, July-November 1956

CONTENTS

page

Introduction 83

Experimental Geophysics 86

Radio astronomy 86

Radio emission from the sun 86

Absolute intensity measurements of

discrete radio sources 87

Precise position apparatus 89

Radio emission from Jupiter 90

Search for Venus 90

Radio hydrogen 91

The upper atmosphere 93

Winds and radio star scintillations . . 93

The earth's crust 94

Seismic studies 94

Rock magnetism 97

Mineral age measurements 100

Theoretical and Statistical Geophysics . . 108

Equatorial electrojet 108

Variability of geomagnetic diurnal
variation and of ionospheric

winds 109

Cosmic-ray investigations 110

page

Laboratory Physics Ill

Nuclear physics Ill

Coulomb excitation studies Ill

Investigation of the mechanism of

nuclear reactions 114

Biophysics 118

Amino acid pools in E. coli 119

Amino acid pools in yeast 125

Protoplasts 131

Fractionation of cell juices on ion-
exchange columns 133

Incorporation of amino acid ana-
logues into proteins 136

Particles 138

Protein synthesis in mouse tissues. . 142

Visitors 142

Operations and Staff 143

Co-operative work of the Department. 143

Administration and operation 144

Bibliography 144

Personnel 146

Plates 2-5, following page 148

Carnegie Institution of Washington Year Boo\ 56, 1956-1957

bo

Si

<3

V-l

>^ 9

h a J* £ ^ a b©

^ G f g

o y <u £ o °P

rt ° - ^<3 <u <u

G r" • "3 r-1 to

■3 =" -G S O

>h r- 2 > ,G

j-i J-1 G q 75 4-j

5T1 i.« '*-' <D l-i

O bJO <*> 4-j s-c

3

W><u

CO

3 u .55

t; c

as n!

G

^ -5 G w -^ :

G O U3 VM --H

IO U „. £j N •

a M « y «

bc^? rt bO

ai

rt O

<U OJ

OS

G >

. u <u

> 4-j G <L> rj

<u u a-> o H

00 O G H

G ,±J „ <-G

§ fc J3 ^ .y ?

^ 2 >^ rtH G s

O u, o >» n? • bO ^

; tj a •> g s M -5 °

g a|SrSS-s %&

*j u « u i, oi u t) ^

bo

n TO ^-

St! ~ * r

.,• • -< G rt 4-

7 Oh M P rt^ 3
<l> G oj ^ u S

1.a.§e*.5--s

2 4-> W3 G g U T-j

5 a^ t! u 3 a

r r'.y vh 5 « u „

— — -- ~

- 3 'g

OJ 4-1

^ q ^

J ^

O O «J

- tJ ■*-' ^ o ^

bO >^ 3 <u o G _o

rS^^ G rtT3 rtl3

<u G G ni -G -G o

4_) ^G ,—^

2 <u OJ G t> Q

"^ _, O H u

^^ o

2 2

G rt

O

°^

'as oj o d "5b "5
o

,-- w u .rr

y o ^

"3 uj 0>+4 !2 Q h U

y d o 'G a -G in

o-d o3^-q g ^

rt bo 5 M

c q u

<S1

<

c a^

. •

! , * i

MMBhtJ

CvJ

b q

INTRODUCTION

Basic research is usually described as
simply the active expression of a curious
or inquiring mind, whereas applied re-
search is goal-directed and motivated by
the hope of useful (and profitable) ap-
plication.

There is a considerable area of overlap,
however, as much of the activity of an in-
vestigator in "pure science" is goal-directed
and practical to a high degree. This is
because his deep interest in certain prob-
lems, or his "love for a subject," drives the
investigator to find usable solutions to dif-
ficulties which he discovers must be over-
come before he can examine or understand
certain processes or conditions he has en-
countered in his studies. His actual work
is nearly always "goal-directed," but its
usefulness he measures in terms of the
extent to which it enlarges his understand-
ing and sharpens his concepts, and the
profit lies in the corresponding enrichment
of his own enjoyment of his subject. The
subtle effect of his free-ranging curiosity
and his corresponding willingness to be
deflected from his immediate aims for a
time, in order to inquire into some un-
expected feature which he observes in the
course of his systematic efforts, is another
facet of difference between the man in
basic research and his former classmate in
the applied field. The principal distinc-
tion, however, probably lies in the different
motivation of the individual and in the
scale of values applied by the man and by
his colleagues to the different types of con-
tributions each man makes as time goes on.

This special kind of motivation of the
individual investigator in "pure science"
is of course strongly reflected in the char-
acter of the encouragement and support
which the members of such a research
staff are given. No fully effective research
man is ever "dispassionate, detached, and
disinterested" — on the contrary, he is pas-
sionately interested and very much im-
mersed in his work. His intense emotional

investment simply makes greater demands
on the homely qualities of honesty and in-
tegrity which are always so indispensable.
This report illustrates some of the "goal-
directed" activities of a group of physicists
who have been encouraged for some years
to study and investigate those problems
which have seemed to them of the most
compelling interest, and each man has thus
been invited to expand and deepen his
love for his subject. The detailed problems
range from hydrogen clouds among the
stars to biochemical complexes, but this
whole range of continually shared enthusi-
asms in the group finally adds up to a
vigorous expression of the scope of
"natural philosophy" and the creative use
of experimental laboratory methods under
intelligently analytical scrutiny.

In our studies of the radio waves ar-
riving from outside the earth's atmosphere
it is becoming increasingly apparent that
optical identifications of the radio sources
are necessary for progress in understanding
just how large clouds of gas can radiate
in the radio region. Even in studying that
relatively near-by object, the sun, it seems
that the position and intensity of the radio
emission must be related to optically ob-
served features on the sun before a satis-
factory theory can be evolved. We have
been led by this necessity to build, for
scanning the sun, an antenna with angular
resolution sufficient to separate two sources
one-sixth of a solar diameter apart. This
antenna takes the form of a line, 2000 feet
long, of helical receiving elements, all con-
nected in phase. With this antenna, and a
receiver recording the power received at
335 mc/sec, the sun is slowly scanned by
its own motion several times each day near
local noon.

Observations of radio sources with test
arrays at 400 mc/sec and with the 300
mc/sec helix array this year have provided
interesting information on the possibility

83

84 CARNEGIE INSTITUTION OF WASHINGTON

of constructing relatively simple arrays
capable of resolution of the order of one
or two minutes of arc. Such accuracy is
required to aid in the important problem
of reliably identifying radio sources with
optical objects.

In the last annual report, a discussion
was given of some evidence suggesting
that directed stresses occurring in the crust
might have a profound influence on the
magnetizations of various rocks. This pos-
sibility was examined by direct experiment
this year with new equipment constructed
specifically for the purpose; the results in-
dicate that many conclusions that have
been offered by various authors in recent
scientific journals on the basis of rock mag-
netism data, relating to polar wandering,
continental drift, secular variation and re-
versal of the earth's magnetic field, are
subject to serious doubt because of the
demonstrable sensitivity to stress of the
magnetizations of many rocks.

By a combination of the rubidium-stron-
tium, potassium-argon age methods it is
now possible to date reliably the time of
crystallization of igneous rocks and the
time of formation of metamorphic rocks.
As a result of such measurements made
here and by workers elsewhere, it has been
found that within geographically large
regions all the igneous and metamorphic
rocks were formed about the same time.
In the Grenville subprovince of Ontario
the almost uniform occurrence of rocks
approximately 1000 million years old has
been known for some time. Similarly,
in the Appalachians all the rocks that have
been measured give an age of 300 ±100
million years. Data obtained here and by
workers in other institutions show the
presence of a large belt of 2500-million-
year-old rocks extending from Wyoming
through Montana, Minnesota, Manitoba,
Ontario, and Quebec. In addition, a large
region of 1350-million-year-old igneous and
metamorphic rocks has been found in
southwestern United States as a result of
measurements made in this laboratory.

Similarly, areas of 1000- and 2600-million-
year-old-rocks have been found in Africa,
and a group of 2700-million-year-old rocks
has been found in Western Australia.

These measurements support the older
idea that earth history is characterized by
orogenic episodes during which a large
belt of the earth's crust is deformed, up-
lifted, intruded by igneous magmas, and
subjected to regional metamorphism. An
investigation is being made in order to see
whether there are any regularities in the
geographic distribution of successive oro-
genic belts of this kind. Measurements
are also being made to find out whether
these orogenic episodes of approximately
200 million years' duration consist of a
series of short episodes or whether the
formation of igneous metamorphic rocks
is essentially continuous.

Soon after the Huancayo Magnetic Ob-
servatory was established by the Depart-
ment, in 1922, the records of geomagnetic
variations obtained there showed that the
amplitude of the quiet-day diurnal varia-
tion, Sq, in the horizontal magnetic com-
ponent, H, is abnormally large. This ab-
normally large diurnal variation in H is
due to the existence, during midday, of a
band of concentrated electric current flow-
ing eastward in the ionosphere over Huan-
cayo. This current, known as the equa-
torial electrojet, is superposed on the cur-
rent system responsible for the normal
quiet-day diurnal variation, Sq.

To determine the height, intensity, and
the pattern of the current flow into and
out of the electrojet it is necessary to de-
termine the variation, with latitude, near
the magnetic equator, of the amplitude
of the diurnal variation in the three com-
ponents of the geomagnetic field. As its
contribution to the U. S. International
Geophysical Year effort, the Department
carried out, from March to May 1957, with
the co-operation of the Instituto Geofisico
de Huancayo, a survey on the west coast
of Peru, to obtain data for answering these
Questions. In addition, locations were

DEPARTMENT OF TERRESTRIAL MAGNETISM

85

chosen for recording continuously the geo-
magnetic variations during the IGY. These
data will indicate whether the same, Sq,
electro jet is responsible for the large lu-
nar diurnal variations at Huancayo, and
whether "electro jet" effects exist for the
"sudden commencements" of magnetic
storms and for other magnetic variations.

In nuclear physics a study has been
made of the angular correlation between
the directions of emission of the proton
and y ray in an a-particle-induced reaction
on fluorine. The results show that there
is indeed a definite correlation, that some-
how the residual neon nucleus remembers
the direction in which the proton was
emitted, for a sufficiently long time to
allow the y ray to be emitted in a definite
direction with respect to the proton di-
rection even though the natural period of
nucleon motion within a nucleus is very
much shorter than the decay time for the
Y ray. The actual form of the correlation
shows in several cases a surprising agree-
ment with the predictions of a simple di-
rect or surface interaction mechanism.

In the work of the biophysics section it
is increasingly apparent that it is now
timely to attempt to interpret the chemical
activities of the cells in terms of cellular
structures. Kinetic models of the bacterial
cell which take into account the known
structures of the cell have been considered
in detail. One model which fits the present
data suggests that there may be a kinetic
relationship among the various classes of
particles found in bacterial cells, that one
type of particle grows and becomes a dif-
ferent type. Experiments to test this hy-
pothesis have been started. The data have
not yet had sufficient refinement to test
the detailed predictions of the model, but

they do show marked differences in the
rates at which radioactive materials appear
in different cellular structures. Another
feature of one of the models is the assump-
tion that the particles are arranged within
the cell in a definite spatial array. Al-
though this feature is difficult to test ex-
perimentally it has received some support
from an unexpected direction. Clear solu-
tions of bacterial juices were observed to
give rise to cell-like bodies containing pro-
tein and nucleic acid. Those forms that
"reconstitute" from the much smaller par-
ticles of the fluid may quite possibly be a
different expression of the forces that or-
ganize the material of the cells.

That the mechanism which inserts
amino acids into the peptide chains of
protein operates with a high degree of
accuracy is shown by studies of amino
acid sequences. Recent work at the In-
stitut Pasteur showed, however, that the
mechanism was not perfect, and that cer-
tain analogues of amino acids could be
incorporated. The resulting proteins were
sufficiently altered so that normal growth
was not possible. During the year, studies
carried out here in collaboration with
Dr. G. N. Cohen, of the Institut Pasteur,
showed that methionine could be com-
pletely replaced by its selenium analogue.
In this case the resulting altered proteins
still have a sufficient enzymic activity to
permit continued growth. These results
show that some errors in the formation of
proteins are acceptable, and the mechanism
of amino acid selection does not have to
be perfect. Studies along these lines are
being continued to determine whether the
degree of substitution is the same in all
proteins or whether it varies from one pro-
tein to another.

86 CARNEGIE INSTITUTION OF WASHINGTON

EXPERIMENTAL GEOPHYSICS

RADIO ASTRONOMY tion and the behavior of localized bright

B. F. Bur\e, W. C. Ericsson, /. W. Firor, H. L. areas on the sun is different from the prob-

Heljer, H. E. Tatel, M. A. Tuve, and lem of preparing to search for new radio

T T TT7 Tjr 7 7 1~ L CD

n. w . weus stars or to measure accurately the position

RADIO emission FROM THE SUN Gr intensity of radio stars. When searching

It has been the expectation that our for stars the investigator must be able to

understanding of the solar atmosphere scan large parts of the sky to see many

will be increased by a search for relation- sources, to detect weak ones, and to give a

ships between the radio emission from the single position in the sky for each. In

sun and features of the sun observed op- the solar case a single strong object is being

tically. Almost always the attempt has studied, and arrangements to make meas-

involved measurements of one feature of urements on this one object continuously,

the radio emission and some selected fea- or at least repeatedly, during the day are

ture of the optical observations: for ex- necessary. On the other hand, advantage

ample, measurements of the total emission may be taken of the fact that there is only

from the whole sun at one radio frequency one source like the sun in the sky, and the

and the sunspot number or the sunspot antenna pattern may have considerable

area or some weighted average of the two. positional ambiguity— the measurements

Although the relation between the spots may be consistent with many different

and the radio emission was never found to positions in the sky as long as only one

be close, the matter has been pursued to of them falls near the known position of

the point of trying to determine, statis- tne sun-

tically, from which spot on the disk the An antenna designed especially for solar

radio emission comes or, by extrapolation studies was used by Christiansen in Sydney

of statistical data, what would be the level for several years to determine the bright-

of radiation in the absence of spots. Most ness distribution of the quiet sun. His

of the conclusions reached in this manner arrangement was an array of 32 small

have later been found to be incorrect; and, paraboloids in a line 1000 wavelengths long

further hindsight shows, many mistakes and operated at a wavelength of about

would have been avoided by having in 20 cm. All the elements were connected

hand a more complete radio description together in phase. The resulting recep-

of the phenomenon before the optical con- tivity pattern in the sky was a number of

nection was attempted. In the examples narrow parallel lobes, about 3' of arc wide

mentioned, a measurement of the position and spaced about 2°. These fan-shaped

on the solar disk of the source of radio beams remained fixed in the sky, and the

radiation would have guarded against the sun, moving across the sky in its diurnal

incorrect conclusions. path, was scanned by each beam (lobe)

Although many characteristics of the in turn. As used by the Australian group,

solar radio emission that could be measured this antenna permitted bright areas on the

would add to the radio description of the sun to be recognized and allowed for in

sun, the measurement of the positions on trying to deduce the distribution of bright-

the disk of the sources of the radiation, or ness of the quiet sun.

in other words the distribution of bright- Such an antenna arrangement can clearly

ness on the solar disk, seemed to us to be be utilized for studying bright areas as

the one that would lead most directly to well as for eliminating their influence,

fruitful optical comparisons. Each scan of the sun by one of the fan

The problem of designing an antenna beams gives a position line across the sun

for studying the solar brightness distribu- for any bright areas as well as the inten-

DEPARTMENT OF TERRESTRIAL MAGNETISM

87

sity of the region. Later scans then reveal
changes in position or intensity.

The selection of the wavelength depends
on the end in view. For studying the
quiet-sun radiation, almost any wavelength
in the radio-astronomy range is of interest,
for only when the results from a wide
range of wavelengths are available will it
be possible to derive a radio picture of
the chromosphere and the corona. A sim-
ilar statement could be made for the ac-
tive radiation from the sun — the localized
bright areas and the bursts. The active
radiation is very complex, however, not
only in its variation with wavelength, but
also in its polarization, variation with time,
and change in position, so that it is advan-
tageous to have a number of different types
of measurements made on a single active
event at similar wavelengths. For this
reason a wavelength near to that used by
other solar investigators in the same hemi-
sphere is desirable.

An antenna meeting most of these re-
quirements has been built at our River
Road site near Seneca, Maryland. The
array is modeled after the Australian one,
but employs a longer wavelength (90 cm;
328 mc/sec) and different receiving ele-
ments. The individual elements are pairs
of 10-turn helices mounted on a common
ground screen. One helix is mounted with
a half-turn rotation with respect to the
other and so is out of phase. The two
helices can thus be connected to the two
sides of a balanced transmission line. (See
fig. 1, pi. 2, facing p. 148.)

At present the array has 30 of these
elements and is a little over 600 wave-
lengths long. The measured beamwidth
of one of the fan beams is 4.8' of arc to
half-power points, and the fans are spaced
2J/2°« A comparison-type receiver com-
pares the power received by the antenna
with that from a room-temperature re-
sistor at a 1000-cycle rate.

Four scans of the sun taken near noon
on successive days are shown in figure 2.
Two bright regions can be seen moving
across the disk as the sun rotates. One of

them is variable in time — changes take
place during the few seconds required for
the beam to scan it. The other region is
not only steady during the scan but also
is much the same day after day. For the
steady regions the apparent motion of the
source across the solar disk can be derived
from a series of these scans, and a height
above the photosphere can be found for
the source. For the more changeable re-
gions, connections will be sought with the

Fig. 2. Scans of the sun on four successive
days taken with the helix array. Two regions
of enhanced radiation are seen which move from
left to right as the sun rotates. One of the
regions (near the center) is relatively stable; the
other (near the right, or west, limb) changes in
intensity in the few seconds required by the nar-
row antenna beam to scan it.

optically observed features near the correct
position line given by the scan. (/. W. F.,
W. C. E., B. F. B.) '

ABSOLUTE INTENSITY MEASUREMENTS OF
DISCRETE RADIO SOURCES

The report for 1955-1956 described pre-
liminary results of intensity or flux meas-
urements on radio stars at frequencies be-
low 30 mc. Greatly increased solar activity
of the past year has resulted in an iono-
sphere which usually is not sufficiently

CARNEGIE INSTITUTION OF WASHINGTON

transparent for additional radio star meas-
urements at the lower frequencies. In all
probability, it will be impossible to repeat
our observations in the 12 to 15 mc range
until perhaps 1963-1965.

A review of intensity measurements of
discrete stars in the available radio spec-
trum has emphasized the very unsatis-
factory and inadequate nature of measure-
ments in this important area of radio as-
tronomy. As yet, there is not one accepted
standard of reference at any frequency.
Even in measuring the same star in the
same frequency range there is very little
agreement among different investigators.
Few have reported any flux measurements
at more than some isolated frequencies,
and those may lack consistent observa-
tional procedures. No one has assumed
the task of making a careful, systematic
series of "absolute" measurements over a
wide range of frequencies.

Accordingly, one of our objectives in
radio astronomy has evolved into a pro-
gram of flux measurements on a few of
the intense radio stars over a wide band of
the useful spectrum. For the purpose of
minimizing all errors, both relative and
absolute, we are undertaking a series of
measurements by systematic and consistent
procedures over the frequency range. In
theory a flux measurement is simple. The
intensity, S, is obtained from

S = \T/A

where S is in watts per square meter per
cycle per second, T is the equivalent tem-
perature of the signal at the antenna, \
is Boltzmann's constant, and A is the
effective area of the antenna in square
meters. In practice, however, there are
several potential sources of error. T de-
pends on (1) precise knowledge of all
transmission line losses between antenna
and receiver, (2) accuracy of the calibrat-
ing system, and (3) the precision of the
record scalings. The effective antenna area,
A, is calculated from knowledge of its
polar diagram, the efficiency of the ground
screen being taken into account. Experi-

ence indicates that the most likely sources
of error may lie in the calibration method
and in the determination of antenna aper-
ture. Such errors are now minimized by
calibrating noise sources against heated
resistors at each frequency, applying cali-
bration signals while antennas are also
connected to the receiver, and using the
simplest of antennas (dipoles) for all pri-
mary observations.

The use of dipole antennas limits the
primary observations that can be made at
this latitude to one or two relatively iso-
lated radio sources. Even with one primary
measurement at any frequency, however,
higher gain arrays with adequate resolu-
tion are then employed to obtain relative
intensity measurements from which an
accurate flux may be determined for other
radio stars.

At frequencies below approximately 100
mc we have made dipole (primary) meas-
urements of both Cassiopeia A and Virgo
A. We plan to continue dipole measure-
ments to the limit of sensitivity, which
may be several hundred megacycles. For
observations at higher frequencies it will
be necessary to standardize other antennas
of larger aperture — perhaps horns — against
dipoles at the transition frequency and to
use scaled versions of the large antenna in
subsequent operations. In this manner, no
difficulty is anticipated in extending the
series of absolute measurements to 1000
mc or more.

The present series of measurements in-
cludes 18.5, 27, 50, 87, and 108 mc. The
sources of principal interest are Cassiopeia
A, Cygnus A, Taurus A, and Virgo A.
Some of the final primary observations on
Cassiopeia at 87 and 108 mc are scheduled
for July 1957, at which time the source
will be transiting just before sunrise. The
next recording frequency is 207 mc. Ad-
vance preparations are nearing completion,
and observations are scheduled to start in
late June or early July 1957. The next step
will be to about 300 mc or higher, depend-
ing on the 207-mc results.

When our measurements are combined

DEPARTMENT OF TERRESTRIAL MAGNETISM 89

with all other known values, several in- of the gas or matter emitting the radio

teresting characteristics promptly become waves are known. Some sources have steep

apparent: (1) there appears to be a "knee" density gradients to high pressures, as in

or change in slope of Cassiopeia A below the solar atmosphere; others appear to

50 mc; (2) there is so much scatter of be extremely tenuous and extended gas

points in the range 200 to 600 mc that clouds; some are relatively much cooler

earlier curves showing a "flat" spectrum than others; all appear to be highly turbu-

may be seriously questioned; (3) at fre- lent. Until much larger numbers of the

quencies up to approximately 1000 mc the various types are known and studied by

ratio of Cassiopeia A to Cygnus A is de- optical methods we can hardly consider

creasing (Cygnus A is getting relatively radio observations to have contributed

stronger at the higher frequencies), but much to astronomical knowledge. More-

the ratio of Virgo A to Cygnus A is essen- over, faint but easily observable radio

tially unchanging over the same range. In sources may lie beyond the radius of the

connection with the ratio measurements it optically observable universe. Therefore,

is perhaps significant to note that the ratio the statistics of faint radio sources may

of two extragalactic sources (Virgo A and yield information about the large-scale

Cygnus A) is steady but the ratio of Cas- structure of the universe,

siopeia A, which is a galactic source, to In order to extend the list of optical

Cygnus A is changing in a manner to identifications, precision position determi-

indicate that the very distant source in nations for a large number of radio sources

Cygnus A is getting relatively weaker at must be made. The few identifications so

the lower frequencies. Absorption in inter- far have proved to be interesting objects,

stellar space — a rough distance scale — is but if meaningful identification with ob-

one of several possible explanations of the jects as faint as 18th magnitude are to be

trend. (H. W. W.) made, radio positions should be measured

to the order of several square minutes of

precise position apparatus arc? a precision attained so far only for

The optical radiation from an astro- the most intense sources. The strongly

nomical radio source is capable of yielding felt need for a greater number of precise

a great wealth of information that the positions has led us, therefore, to investi-

radio radiation cannot give. Therefore, to gate the problem. An antenna constructed

obtain detailed knowledge of the nature for precise position measurements must

of radio sources, optical identification must possess not only the necessary mechanical

be made. Up to this time, however, a dis- and electrical stability but also sufficient

couragingly small number of radio sources angular resolution to guarantee that most

have been unequivocally identified with of the sources measured are truly discrete,

optically observed objects. These sources not simply blends of several sources (often

represent a great variety of objects, rang- referred to as "confusion"),

ing from the sun and planets, through A long array, to give a narrow "line of

supernova remnants, and peculiar galactic position" in the sky, or an interferometer

emission nebulosities, to external galaxies, composed of a pair of long arrays along

Only a few objects of each class have been one line, appears promising, particularly

identified. Before definitive conclusions in view of the success of the 328 mc/sec

about the general nature of radio sources helix array we have in use on the sun.

can be drawn, many sources of each class The arrays must be oriented in several

must be identified. It is impossible to carry different azimuths, since a single orienta-

out any extensive astrophysical studies us- tion gives high precision for only one co-

ing radio information unless the state of ordinate. At least three different orienta-

aggregation and other physical properties tions of the arrays appear desirable, in

90

CARNEGIE INSTITUTION OF WASHINGTON

order to overdetermine the position and
guard against confusion errors.

Several small arrays and a receiver were
constructed for use at 400 mc/sec to aid
in design of the large arrays with the aim
of achieving position accuracy within 4
square minutes of arc. These arrays are
expected to be made in small sections, each
about 10 feet long to facilitate remounting
along different lines of azimuth. The most
suitable element investigated so far is a
60° V reflector excited by a line of full-
wave dipoles. The test arrays used open-
wire feeder lines which can be constructed
to give low losses and good phase stabil-
ity. Trial interferometric measurements of
right ascension were made with small ar-
rays and a short baseline (50 meters) on
the strong sources in Cygnus and Cassi-
opeia, the results agreeing with the known
positions to within a minute of arc. Prep-
arations have been completed, and the
construction of a large array is in prog-
ress. (B. F. B., /. W. F., W. C. E.)

RADIO EMISSION FROM JUPITER

Regular observations of the planet Jupi-
ter were not continued, but data taken in
previous years at this laboratory were re-
examined in the course of preparing a sum-
mary of the present status of this phase of
radio astronomy. In the previous report
it was mentioned that the region of the
planet which appeared to be the most per-
sistent source of noise during 1955-1956
was certainly not the same as that reported
by Australian workers from their predis-
covery records taken from August to Sep-
tember 1951. An analysis of the time de-
pendence of all obtainable observations,
including the Australian observations in
late 1950 and early 1951, the August-Sep-
tember 1951 Australian series, the pre-
discovery observations at this laboratory
during June 1954, the Mills Cross series
from January to May 1955, and the 1955-
1956 Carnegie observations, revealed that
the most active region could always be
reconciled with a single center of activity

on the planet, having an approximately
uniform rotational period of 9h 55m 28.5s.
Interestingly, none of the visual observa-
tions report any surface features exhibit-
ing this rotational period; the source of
radio radiation therefore probably lies be-
low the cloud level, and might well be
associated with the surface of the planet
itself. (B. F. B.)

SEARCH FOR VENUS

During the fall of 1956, when Venus
was at elongation, two interferometers
were placed in operation to check on the
validity of the reported low-frequency non-
thermal radiation from this planet. These
interferometers operated at frequencies of
22 and 26.75 mc/sec. The 22 mc/sec array
consisted of two elements, each contain-
ing eight half-wave dipoles, which were
phased in such a direction that Venus rose
through the beam about 2 hours before
sunrise each morning. Each element of
the 26.75 mc/sec array consisted of four
half-wave dipoles phased in a similar man-
ner. The observations were attempted be-
fore sunrise, since after sunrise strong in-
terfering signals of terrestrial origin make
identification far more difficult. Observa-
tions were made continuously from Sep-
tember 19 to October 25, 1956, but no ra-
diation of Venusian origin was found.
If during this period Venus had been a
source of equal intensity to the Crab Neb-
ula, it would have easily been observed.

During these observations, one interest-
ing effect was noted. Approximately 30
minutes before sunrise, just as the inter-
fering signal strengths were rising, a few
lobes of an interference pattern were often
found at both frequencies. They could not
be associated with radiation originating at
Venus, since they disagreed both in phase
and in period with radiation coming from
the direction of that planet. The effect
may be due to the build-up and movement
of ionized regions in the atmosphere as
they are illuminated by sunlight, some of
these regions being responsible for the re-
flection of the interfering radiation, which

DEPARTMENT OF TERRESTRIAL MAGNETISM 91

may be of terrestrial or solar origin, to the
observing site. (B. F. B., W. C. E.)

Other interferometer records of the sum-
mer of 1956 using simple dipoles at 18.5
and 26.75 mc were scanned in the search
for Venus. No events at either frequency
could be uniquely identified as having
origin in Venus, although several interest-
ing borderline occurrences were noted on
the 18.5-mc instrument. These tests indi-
cate that no pronounced signals from Ve-
nus were received, but do not conclusively
prove the absence of some weak emissions
at random intervals. (H. W. W.)

RADIO HYDROGEN

During the past year a large part of our
effort in the 21-cm hydrogen-line program
has been devoted to the improvement of
our observing equipment. We have de-
signed an equatorial telescope mount suit-
able for large dishes, installed a multi-
channel spectrometer, and improved the
over-all stability of our recording equip-
ment. In addition to this emphasis on in-
strument improvement, however, we have
continued our measurements at a modest
rate. Several more meridian galactic sur-
veys have been finished, and a thorough
survey of the Pleiades and the II Persei
clusters is nearing completion. One find-
ing is that a large region of the sky in the
direction of these clusters is covered by a
homogeneous hydrogen gas cloud. Pre-
liminary examination of the Doppler-shift
cross section of this cloud indicates that a
Gaussian velocity distribution represents
the major portion of the observed inten-
sity. The velocity distribution parameter
is 5 km/sec, corresponding to 600° K,
whereas the "spin temperature" is assumed
to be about 150° K. It should be noted
that the optical absorption of interstellar
lines indicates that this mass of gas is
probably 150 to 350 parsecs distant.

The general problem of the design of
large parabolic antennas was resolved into
two principles: astronomical uses make it
desirable to have an equatorial mount;
and the structure should be designed with

rigidity commensurate with the precision
of the reflector. Hence the main drive
gears and major structural units should be
large, so that machining tolerances can
also be large. Numerical studies of our
resulting initial design showed it to be
adequate and economical for supporting
reflectors as large as 85 feet. The National
Science Foundation is planning to build a
large "national facility for radio astron-
omy" at Greenbank, West Virginia. Hence
we carried out some careful studies of the
features required for larger parabolic re-
flectors and mounts. A 200-foot bar of
steel first compressed and then stretched
by its own weight will elongate about 0.4
cm. Since any major structural unit must
support more than its own weight, its
elongation will be greater. If it supports
2l/z times its own weight the elongation
is 1 cm. In a reflector the depth is much
less than the breadth, so that major struc-
tures have even larger loads than simple
trusses of equal width and breadth — per-
haps twice as much. If the deflection be-
comes 5 cm it reaches (exceeds) the limit
for a reflector intended for 21-cm waves.
This very rough calculation, which neg-
lects some features such as "hoop strength,"
indicates that a simple cantilever structure
of steel cannot be expected to be useful as
a steerable astronomical instrument if it
is much longer than about 200 feet. For
larger structures other features of design
must be devised to surmount the problem
of the elasticity of steel. The economic
problem is also considerable. Devices of
this magnitude constructed with present
techniques are expected to have costs in
the range of perhaps 10 million dollars.
Our hydrogen-line radiometer has been
converted to a multichannel device. It is
now used to record in 54 channels 10
km/sec wide, spaced at frequencies equiva-
lent to a Doppler shift of 4 km/sec between
channels. The basic system is a Dicke-
Ewen comparison receiver with two local
oscillators alternately switched on and off.
The oscillators are tuned separately, so that
the two receiving frequencies and their

92

CARNEGIE INSTITUTION OF WASHINGTON

differences can be adjusted with precision.
The received signal, detected in a crystal,
is doubly converted to a band extending
from 1.5 to 2.5 mc/sec, at an output level
of 2 to 3 volts rms. A common output
amplifier drives 54 separately tuned fil-
ters. At the output of each filter is an in-
dividual amplifier with ample feedback
stabilization.

The signal detected at the output of
each of these radio-frequency amplifiers
contains the switching frequency (about
450 cps). It is amplified in a feedback-
stabilized audio amplifier and detected in
a phase-sensitive (450 cps) detector. The
output of the final amplifier on each chan-
nel is 2 volts rms. Between this output
and the final detector is a potentiometer
gain control for each individual channel.
The phase detectors charge l-|jf condensers
through 400 or 4000 megohms, depending
on the desired output time constant. The
voltage on the storage condensers is read
out by means of an electrometer tube and
recording potentiometer. Read-out time is
2 minutes for the 54 channels. This sys-
tem has very great advantages over the
single channel unit with frequency scan,
but there are several features that we wish
to modify. The output diodes have a sta-
bility of about l/2 millivolt per day, which
is equivalent to a signal of 1° K at the
antenna, so that the diodes need adjust-
ment every two or three days. We there-
fore expect later to increase the audio gain
by about a factor of 8. The gains of the
feedback amplifiers are difficult to moni-
tor; we calibrate the over-all system with
a modulated noise diode at the input, and
statistical fluctuations during the calibra-
tions are important. Our wide-band ampli-
fier (1.5 to 2.5 mc/sec) does not have a
sufficiently low impedance to drive the
high capacity of the existing cables feed-
ing the 54 filters, so that we are forced to
operate at a low output diode level. These
cables must be replaced by a low-capacity
feed.

The whole system is flexible enough to
use in many problems. The oscillators can

be tuned over a wide range. Channels
can be connected in parallel so as to have
greater bandwidth and smaller fluctua-
tions. We are now in the process of de-
veloping the observing procedures for the
best use of this powerfully analytical in-
strument.

Another problem of great difficulty and
even greater importance is the zero sta-
bility of the detection system. Many in-
teresting problems require the detection of
minute signals of the order of 1° K an-
tenna temperature, or less. Such measure-
ments require a system with zero drifts
and changes not exceeding 0.5° or 03° K
over periods of several hours. Our detec-
tion system last year had a variable stability
with occasional fluctuations even as high
as 10° K over a run of several hours. After
much testing and many changes, including
the physical separation of the local oscil-
lators from the receiver units and im-
provement in the shielding of transmission
cables, we have been able to reduce the
zero changes to a drift of less than 1° K
per week on the dummy antenna. On the
sky antenna the system is not as free of
drifts; at times it may shift as much as
±1° K in several hours. For the narrow-
band short time constant (7 minutes) this
amount of drift is not important, but in
looking for signals of a few tenths of a
degree Kelvin it is still a limiting factor.

Using this new multichannel recording
instrument we have in progress a repeti-
tion and extension of our survey of the
hydrogen clouds within ±20° of the galac-
tic plane, for a great many points spaced
from 2° to 10° apart along the galactic
equator as visible in Washington. Resid-
ual hydrogen is also found at all points
examined to date at higher galactic lati-
tudes. (H. L. H., H. E. T., M. A. T.)

Tidal distortion of the galaxy. The
Leiden and Australian groups have shown
that, if the Lund galactic pole is shifted,
the major concentration of galactic hydro-
gen lies close to the plane, with small
deviations. The deviation for the outer-
most spiral structure is noticeably greater,

DEPARTMENT OF TERRESTRIAL MAGNETISM 93

and the 21-cm meridional surveys taken at 45°. Such large-scale drifts are predomi-
this laboratory were examined to see nantly downward and into the east. Often
whether the effect was systematic. Sur- there is no significant change either in
veys were available here, at galactic longi- velocity or in direction as the disturbance
tudes 50°, 60°, 80°, 90°, 110°, 180°, 200°, progresses downward through the outer
and 210°, which showed that in general atmosphere. Effects such as described could
the center of mass of the outermost hydro- be caused by an inclined wave front mov-
gen lay consistently below the plane on the ing horizontally, an inclined wave front
side closest to the Large Magellanic Cloud moving vertically, or combinations of both
(LMC), but above the plane on the more motions. Predominant velocities are be-
distant side, except for longitude 110°. Al- tween 100 and 200 m/sec.
though the effect is qualitatively in the In another review of ionospheric winds
proper direction to be due to tidal effects from other radio methods of measurement,
of the LMC, the observed deviation ap- an intercomparison was made of results
pears to be much too large if conventional from the "meteor-Doppler" and the "fad-
values for the masses and distance scales of ing" techniques. The outcome was the
the galaxy and LMC are assumed. At interesting fact that the two methods inde-
1=210°, the deviation is about 300 parsecs, pendently establish similar characteristics
which is approximately 20 times larger of EAaycr winds. The outstanding features
than would be expected from gravitational are (1) velocities in the range 50 to 100
effects, if Schmidt's galactic model is used, m/sec, (2) large semidiurnal components
together with a mass of 4 X 109 solar masses with clockwise rotation of wind directions
for the LMC. The effect is sensitive to the in the northern hemisphere, and (3) coun-
assumed force perpendicular to the galactic terclockwise rotation in the southern
plane, but it appears that both a larger hemisphere.

mass for the LMC and a larger distance Radio star scintillations have been ob-
scale for our galaxy are required if the served at all operating frequencies between
effect is to be explained using gravitational 18.5 and 108 mc. Although the scintilla-
forces only. (B. F. B.) tions at 108 mc, which is the IGY satellite

frequency, are not a normal occurrence, it

THE UPPER ATMOSPHERE [s c}ear ^at varying refraction in the iono-

H. W. Wells sphere causes the apparent "radio" position
WINDS AND RADIO STAR SCINTILLATIONS of an object to change by significant
A paper, "Large scale movements of the amounts. At 50 mc, scintillations were
layers," was presented to the AGARD prevalent during the period of observation.
(Advisory Group for Aeronautical Re- At 27 mc, preparations have been made
search and Development) at Oslo, Nor- for the operation of a three-station network
way, in July 1956. The activities of this to measure the drift characteristics of the
Department and others were reviewed in ionized clouds producing the scintillations,
the light of additional analyses. Sweep- To explore the possibility of long-
frequency observations of the ionosphere at distance propagation of electromagnetic
networks of stations spaced 20 to 200 miles waves from power lines some brief experi-
reveal the nature of apparent large-scale ments were conducted at the Derwood
movements and permit a three-dimen- Field Station. It was assumed that any
sional interpretation. Analyses clearly peak in radiation at 50 cycles would indi-
show for the first time that traveling dis- cate foreign origin. Filters were incorpo-
turbances have vertical as well as hori- rated so that the strong 60-cycle radiation
zontal components of motion. The dis- from domestic sources would not interfere,
turbance wave front is often inclined about The recordings showed the presence of

94 CARNEGIE INSTITUTION OF WASHINGTON

very substantial energy between 45 and 55
cycles without any peak in the 50-cycle
region. Random atmospheric noise and
other disturbances left a high residual noise
level, however. A few isolated cases of ac-
tivity during thunderstorms clearly estab-
lished the fact that electromagnetic radia-
tion of these extremely long wavelengths is
generated at such times.

Preparations for the Xllth General As-
sembly of the International Scientific
Radio Union at Boulder in August-Sep-
tember 1957, and attention to many fea-
tures of the program of the International
Geophysical Year, especially in the areas
of ionospheric physics, have occupied some
members of the DTM staff during a con-
siderable part of the current report year.

THE EARTH'S CRUST

L. T. Aldrich, J. W. Graham, H. E. Tatel,
M. A. Tuve, and G. W. Wetherill

The exploration of the earth's crust con-
tinues to present us with problems of great
interest. Our general goal, of course, is a
clear recognition and understanding of the
many large-scale physical processes that,
operating during the long periods of geo-
logical time, have resulted in the present
conspicuous features of the earth, such as
the continents and the ocean depths, moun-
tain ranges and high plateaus, and the dis-
tribution of land and ocean areas. We are
also interested in other matters like the
equatorial bulge, the stability with time of
the position of the geographical poles, and
the time scale of geological uplift and ero-
sion. Perhaps we may some day have
better notions about the remote and the
immediate "causes" of such processes as
mountain building and the cyclical immer-
sion of large continental areas under shal-
low seas, but clearly the first necessity is
a comprehensive and quantitative descrip-
tion of the earth as it is today. In particu-
lar, we need data on relatively inaccessible
matters, like the horizontal and vertical
density distribution of the rocks under the
continents and under the oceans. We have
directed our efforts toward such questions

in the seismic program with explosion
waves. Our work on isotopes in rocks aims
at establishing bench marks for a time
scale in Precambrian geology.

SEISMIC STUDIES

The mathematical picture, for gravity
and seismic computations, of an earth's
crust comprised of a series of horizontal
rock layers, each several kilometers thick
and of successively increasing density, has
given way in very recent years to a much
less specific picture, namely, a crust with
horizontal and vertical inhomogeneities,
and with average characteristics that do not
change by large amounts over short hori-
zontal distances.

In Alaska we find a crust that has a
mean seismic velocity uniform with depth
to the mantle. It is easy to imagine, then,
as appears to be increasingly acceptable
among students of earth structure, that
the change in velocity is the result of a
"phase transition," a change of chemical
crystal lattice relations and resulting
change in density without change of
chemical composition. But if this is ac-
tually true, why is the transition only 11
km under the ocean surface and 30 km
under the continental surface ? The puzzle
is even more perplexing in view of the
recent heat-flow measurements through the
ocean bottom in the Atlantic and the Pa-
cific. The oceanographers find, contrary
to expectation, that ocean and continental
heat flows upward through the rocks of
the crust are almost the same. This being
so, it can hardly be an equilibrium tem-
perature and pressure distribution that de-
termines the depth of the velocity discon-
tinuity, and hence it seems improbable that
this abrupt change from crustal to "outer
mantle" rocks can be a phase transition,
unless the present heat flow is in part an
indicator of changing heat flow and chang-
ing temperature distribution.

Velocity versus Depth

We have tried to fit the crustal data ob-
tained from our measurements on the seis-

DEPARTMENT OF TERRESTRIAL MAGNETISM 95

mic waves from explosions during the past the critical reflection, calculated on the

decade to various types of possible crustal basis of SnelPs law and the measured

structures, treating the problem as one in velocities, is appreciably (10 or 15 per cent)

geometric optics in which the velocity less than the observed distance to the

varies in different specific ways with depth, abrupt appearance of the critical reflection

Our conclusion has been that the compres- of the compressional wave. This is just

sional velocity does increase with depth in what we observe in field work, but in the

some manner, probably at an increasing laboratory experiment we know that the

rate, from a mean surface value of about "crustal" velocity did not increase with

6 km/sec to a value of about 7 km/sec at depth.

a depth of, say, 30 km, and, just below this, There is at present no theory describing

abruptly to about 8 km/sec, the compres- the nature of a critical reflection of a com-

sional velocity of sound waves in the outer pressional pulse in such a medium as the

mantle. These conclusions were based earth. As a rough approximation, the criti-

upon the geometrical optics approximation cal reflection pattern could be equated to

for sound waves, which has been the only that obtained by the masking of a point

complete theory available. We have al- source by a large disk, the apparent source

ways noticed, however, that certain ex- being the mirror image of the true source

pected arrivals have not been observed, at the surface. The solution to even this

for example the refracted wave expected problem does not exist for light waves,

between 100 and 180 km. This failure has Since the source and image are not a great

been excused on the hypothesis that the many wavelengths apart in our seismic

"reverberation" level, due to interconver- work and also in the laboratory model, it

sion of wave types at topographic and sub- might be imagined that there would be a

terranean boundaries, is unfortunately kind of Fresnel interference at the edge of

enough to obscure these arrivals. This may the excluded disk (due to compressional

still be so. waves entering the mantle within the cir-

The Alaskan measurements (Year Book cle defined by SnelPs law) . Such an inter-
54, 1954-1955) are one of our more precise ference pattern, strongly dependent on
sets of data. They show well developed frequency, would predict an intensity that
total reflection from 80 to 200 km and only begins to rise at the geometrical
strong refracted waves from 160 km out. shadow edge (that is, at the critical reflec-
These values are well fitted by a simple tion distance defined by Snell's law), and
crust of a single layer of rock in which increases to a maximum at a distance well
the velocity is constant with depth down beyond this. The lateral extent of this
to 31 km, at which point it jumps abruptly interference zone can be estimated as fol-
from 6.1 to 8.1 km/sec. With this simple lows: The minimum length of the ob-
structure the critical reflection should have served critical reflection pulse in seismic
been observed at 70 to 72 km; in actual field work is 0.15 to 0.25 second, and the
fact it was observed strongly at 88 km and pulse consists of three undulations, up-
weakly at 80 km from the shot. down-up. If there is interference, it must

A model experiment carried out in the occur between at least two Fresnel half-
laboratory using impulses of microsecond wave zones generating, say, two of these
duration on a two-layered medium of brass undulations 0.12 second apart. For these,
and iron showed the same effect; namely, the distance R from source (mirror image)
no refracted wave was visible above the to surface must differ by l/2X (A = wave-
reverberation level at distances beyond the length) . That there is complete interfer-
critical reflection where it would be ex- ence at a point means a lesser degree of
pected (though perhaps weak) . interference along the surface for a distance

The distance from the impulse point to of, say, h. By simple geometry, the dis-

96 CARNEGIE INSTITUTION OF WASHINGTON

tance, perpendicular to a (slanting) ray
path, is V^RA, and along the earth's surface

h is about 1.5 V~RX.

Typical values obtained in Alaska give
£~90 km, A = 0.12X6 (km/sec) =0.7 km,
leading to a value of h of about 12 km.
Thus the spread of the interference pattern
along the earth's surface would be expected
to be in the neighborhood of 12 km. In
this interval the amplitude of the "critical
reflection" would be expected to increase
with distance from the source from a low
to a high value. The reflected pulse is
therefore expected to be weak at the critical
distance calculated from the ratio of veloci-
ties by Snell's law and could be obscured
by the reverberation background. Thus
the total reflection might not be observable
until well beyond the critical distance pre-
dicted on a simpler basis without the inter-
ference effect. Though the estimate is
crude, it is based upon well known princi-
ples and probably gives the correct order
of magnitude of the interference effect.
The difference between 70 km as calcu-
lated for the critical reflection distance by
Snell's law from the observed upper and
lower velocities and the observed distance
of 88 km when the strong reflections first
appeared is close to the added distance
estimated by this crude approximation.

Hence, the system of reflection observed
on our Alaskan expedition is indeed that
which must be expected from the sim-
plest form of a single layer of crustal rock
overlying the mantle, a crustal layer having
a constant velocity, that is, no change of
velocity with depth. The conclusions pre-
viously reached about the velocity distri-
butions in other regions will have to be
re-examined, as we have previously ac-
cepted the first appearance of the critical
reflection at distances beyond where it was
expected, on the basis of Snell's law, as
evidence of moderate increase of velocity
with depth. In general, the increases in
velocity with depth that we have previously
deduced will be diminished. They have
not been large, and this new interfer-
ence prediction may wipe them out, leav-

ing velocity constant with depth in the
crust, at least in most places. (H. E. T.,
M. A. T.)

Gravity and the Earth's Crust

The literature contains about two thou-
sand measurements of gravity whose iso-
static corrections have been calculated.
These serve as the basis of the general
belief in Airy's hypothesis that high moun-
tains are supported in hydrodynamic equi-
librium by a thicker crust floating in the
outer mantle of heavier rocks. Our seismic
measurements on the Colorado Plateau
and in Alaska (Year Book 52, 1952-1953;
Year Book 54, 1954-1955) have shown that
high land areas and thick crustal regions
do not necessarily go together. The ques-
tion then arises whether or not there is
disagreement between the two sets of
measurements, gravity and seismic.

An analysis of the gravity data shows
that the data are much too few to give de-
tailed crustal information. If all the uncer-
tainties are taken into account, and all the
data averaged together to obtain a mean
world-wide crustal depth with minimum
apparent statistical error, the depth turns
out to be 30 ±20 km. This can hardly be
considered an accurate specification of the
depth to the mantle. For any smaller por-
tion of the earth's surface, the statistical
situation is even worse. The reasons for
this inaccuracy in the specification of man-
tle depth from gravity measurements do
not seem to be widely recognized. There
are two: the data are few; gravity values
are not sensitive to the "depth of compensa-
tion" (the depth at which the excess weight
of the mountains is exactly balanced by
the defect of weight due to flotation in the
heavier rocks below).

The gravity measurements show wide-
scale isostatic and free air residuals of the
same size. Some of these regions have
dimensions of a few hundred kilometers.
The origin of the residuals is not specifi-
cally known. They indicate wide-scale
crustal inhomogeneities, or perhaps a hori-
zontal distribution of inhomogeneities in

DEPARTMENT OF TERRESTRIAL MAGNETISM 97

density in the upper mantle rocks. Further
measurements of gravity may give us some
better basis for judgment about the nature
or cause of the residuals, and possibly some
indications concerning outer earth struc-
tures. (H. E. T.)

International Geophysical Year

In co-operation with the U. S. National
Committee for the International Geophysi-
cal Year we are organizing a seismic expe-
dition to South America. Using the waves
from the large explosions regularly set off
in the copper mines there, we will en-
deavor to explore the crustal structure and
measure the crustal depths under the high
Andes and the coastal plains. The mean
heights are about 10,000 feet near the large
copper mines of southern Peru and north-
ern Chile. We therefore hope to be able
to make crustal measurements under re-
gions of great topographic heights, per-
mitting a more definitive study of the
effect of topographic height on crustal
depth than we were able to achieve in our
efforts on the Colorado Plateau and in
Alaska.

Much effort has been expended on the
complex arrangements necessary for an ex-
pedition of this nature. Outfitting six
vehicles for three months of measurements,
many of them in locations without roads,
is an arduous and time-consuming task.

Instrumental development. The seismic
instruments we have designed and made
have proved rugged and dependable. They
are in general sufficiently sensitive and
fairly portable. In some places, however,
the ground unrest is small enough so that
amplifier input provides the limiting sensi-
tivity. In others there are special oppor-
tunities to observe which we miss because
of the bulk of the equipment we must
carry. We have therefore constructed new
seismometers and prototype transistor
amplifiers for increased sensitivity and de-
creased bulk.

The new seismometer has a pendulum
mass 100 times the older seismometer (the
total weight of this new unit is 14 pounds) .

With this instrument the input noise volt-
age of the amplifiers (usually 1 to 2 micro-
volts) will be one-fifth to one-tenth the
voltage produced by ground unrest in the
quietest spots of quiet regions.

The new transistor amplifiers (devel-
oped by E. T. Ecklund, of our staff) are
small, compact, and rugged. The circuits
are stabilized by degeneration so that the
power gain (1,000,000) is largely independ-
ent of temperature and battery voltages.
The input fluctuation voltage is compara-
ble with but not quite so good as that of
a vacuum-tube system. The combination
of the new seismometer and transistor
amplifier, however, is far more advan-
tageous in signal noise and weight than
the system we have been using to date.
(H. E. T., M. A. T.)

ROCK MAGNETISM
/. W. Graham

Within the past few years the challenges
presented by the subject of rock magnetism
have been taken up vigorously by approxi-
mately a dozen research groups in many
countries of the world. It is not yet certain
what rewards will accrue from the efforts.
Briefly, the subject has become subdivided
into two schools : one maintaining that the
rock-magnetism data already at hand pro-
vide justification for believing that it is
now possible to talk with considerable as-
surance about large-scale geophysical phe-
nomena like continental drift and polar
wandering, while the other group holds
that the subject is intrinsically so complex
that no such important conclusions can be
justified in our present state of very limited
knowledge. Each school has been able to
make a strong, but not convincing, case
for its stand.

The efforts at the Department during
the past year have centered on the ques-
tion of the possible influence of stress on
the directions of magnetization of rocks.
It is at once clear that, in rocks which have
been stressed beyond the limit from which
they can recover mechanically on being
relieved, some changes of magnetic prop-

98

CARNEGIE INSTITUTION OF WASHINGTON

erties are to be expected. Such rocks, in
general, are thought to have been ade-
quately excluded from rock-magnetism
studies aimed at questions like continental
drift and polar wandering. But a more
subtle problem demands serious attention:
if a rock is in a state of nonhydrostatic
stress at the time it is magnetized (a condi-
tion that most certainly has frequently
occurred — for example, in sediments mag-
netized after deposition, and in intrusive
igneous rocks), and then the stress is re-
lieved, the magnetic moment of the sample
may change to another value, at the time
of unloading, by virtue of the property
known as magnetostriction. If the mag-
netization of rocks in general should prove
to be particularly stress-sensitive, many of
the data supposedly bearing on the prob-
lems of continental drift and polar wan-
dering would become immediately suspect,
simply on the basis that our knowledge of
the physical conditions prevailing during
the time the rocks were magnetized, and
subsequently, is woefully inadequate.

The basic question of the stress sensi-
tivity, at room temperature and atmos-
pheric pressure, of the magnetization of
rocks is being investigated experimentally
with new equipment constructed for the
purpose. The direction and intensity of
magnetization are measured with an astatic
magnetometer while the sample is held
under compressive stress of about 2500 psi
by a simple nonmagnetic mechanical "nut-
cracker." This stress is about the same as
would be acting at the base of a column of
rock 1800 feet high. Although the ultimate
stability and sensitivity of this particular
magnetometer have not been achieved
(thus precluding studies of weakly mag-
netized sediments, for example), and the
force that can be applied is well below the
limit where fracture and flow would begin,
nevertheless the observations to date are of
sufficient importance and interest to war-
rant further elaboration. They are being
made in collaboration with J. R. Balsley,
of the U. S. Geological Survey, and Pro-
fessor A. F. Buddington, of Princeton

University, utilizing the extensive suite o£
rocks on which detailed mineralogical,
chemical, and magnetic analyses have al-
ready been performed.

In this laboratory reconnaissance it has
been found that the greatest stress sensi-
tivity is displayed by rocks that contain
magnetite as the principal magnetic con-
stituent. If the original natural magnetiza-
tion of the sample is expressed as an in-
tensity-direction vector, and then the mag-
netic vector of the stressed sample is ob-
served, a vector resulting from the stress
can be inferred. Most, but not all, samples
return reversibly to their original condition
when the stress is released. For nearly pure
magnetite (96 per cent), the length of the
stress-induced vector may be as much as
40 per cent of the original vector for a
stress of only 2500 psi. In contrast, rocks
whose ferromagnetic minerals have com-
positions between the extremes of ilmenite
and hematite are quite insensitive, e.g.,
intensity change of 1 per cent or less for a
compressive stress of 2500 psi. (See figs. 3
and 4.) No general rules covering all rocks
can be formulated for the orientation or
sense of the added vector relative to the
stress direction or the original direction of
magnetization. The inference is that the
various magnetic components present in
rocks are of both positive and negative
magnetostrictive signs, and that there can
be interaction of the magnetostrictive mo-
ments with the natural moment of the
sample.

It is worth while to make an estimate of
some of the ways in which these findings
can be related to what is known of the
magnetization mechanism of rocks. The
various ferromagnetic minerals that have
been recognized in rocks are known to
cover a considerable spread in the numeri-
cal values that can be assigned to such
measurable properties as Curie point,
saturation magnetization, susceptibility,
and variation of various properties with
temperature. It is rare that in any given
rock a single magnetic species occurs;
usually two or more coexist, and they may

DEPARTMENT OF TERRESTRIAL MAGNETISM 99

be intimately intermingled or remotely
separated as discrete units. Depending on
the state of aggregation, there may be con-

25

50 75

Percent of original intensity

100

Fig. 3. Figure 3 summarizes the magnetic
changes that were produced by increasing the
axial compressive stress from 350 to 2650 psi.
The results are plotted as vectors: the original
intensity at 350 psi is taken as 100 per cent, and
its angle relative to the direction of compression
is plotted. The arrow end of the vector indicates
the intensity and direction of magnetization of
the stressed sample.

FeTiO,

Fe304

Fig. 4. Composition of the rock samples.

siderable, or little, magnetic interaction of
the different species on one another. At the
time of formation, some ferromagnetic
minerals may appear as metastable single

phases which slowly during geologic time
break down into a mosaic of two or more
mineral species. Some ferromagnetic min-
erals are prone to marked alteration by the
simple process of oxidation. It is known
that all these features are important in rock
magnetism.

To this already complicated picture we
now add the question of the effect of stress.
Reconnaissance measurements indicate that
stresses, which geologically speaking are
almost trivial, can have a marked influence
on magnetization. The response of the
magnetizations to stress is such as to indi-
cate that different species present are mak-
ing different types of contributions to the
change of the magnetization. It probably
would be possible to describe the relative
magnetostrictive contributions of the dif-
ferent ferromagnetic species in a given
sample, but such knowledge gained today
would not necessarily have any meaning
to the more interesting problem that we
have had in view, namely, attempting to
infer the past directions, not to mention
intensity, of the earth's magnetic field.
This is so for two simple reasons : that we
can give an accurate account neither of the
stress history of the rock nor of the chemi-
cal and physical evolution of the magnetic
species. It is of course clear that, the older
the rock, the greater will be the uncer-
tainty. The only obvious way out of these
difficulties is to have so many field observa-
tions from so many rocks of so many types
in so many settings that the insidious in-
fluence of magnetostriction, taken either
alone or in conjunction with time-depend-
ent parameters, is eliminated. The prospect
is hardly encouraging.

The ultimate geophysical implications of
these observations are not yet known, but
they certainly do not foster the hope that
has prevailed for decades that by way of
the techniques of rock magnetism it will
be possible to deal effectively with such
major geophysical questions as continental
drift and polar wandering. The present
observations do not assure that such hopes
are beyond reach; they do call for great

100 CARNEGIE INSTITUTION OF WASHINGTON

caution, however, in accepting rock-mag-
netism data, of the sort usually presented,
as evidence bearing satisfactorily on major
geophysical phenomena.

MINERAL AGE MEASUREMENTS

L. T. Aldrich, G. W. Wetherill, G. L. Davis,1
and G. R. Tilton 1

About ten years ago, improvements in
mass-spectrometric techniques and the de-
velopment of chemical analysis by isotope
dilution made it possible to measure the
small quantities of radiogenic daughter
products in ordinary rock-forming min-
erals. It thus appeared possible to extend
greatly the scope of mineral age measure-
ments, which had previously been limited
to rocks containing uranium and thorium.
Accordingly work was initiated in a num-
ber of laboratories with particular attention
to the measurement of radiogenic argon in
potassium minerals, and radiogenic lead
in minerals like zircon that contain only
small quantities of uranium.

In 1950 a mineral age program was
started in this laboratory. The emphasis
during the first few years was on the de-
velopment and quantitative field testing
of a mineral age method based on the
decay of rubidium into strontium as well
as the extension of the earlier work with
zircons. In 1954 potassium-argon measure-
ments began.

Until the last year workers in this field
have been primarily concerned with the
following three problems: the develop-
ment of satisfactory chemical techniques
for analysis of microgram quantities of the
parent and daughter elements; investiga-
tions of the extent to which minerals have
formed closed chemical systems with re-
spect to the parent and daughter elements;
the accurate determination of the decay
constants for the decay of the parent into
the daughter isotopes. Contributions made
by this group on these problems are sum-
marized in previous annual reports.

1 Geophysical Laboratory, Carnegie Institution
of Washington.

Advances toward their solution have
been substantial during the past several
years, and although many questions re-
main unanswered, emphasis has now
shifted to the application of the techniques
to problems of geology. The shift in em-
phasis is reflected in the present report.
Whereas in previous years most of the dis-
cussion has concerned the problems enu-
merated above, a large part of this report
is devoted to actual geological applications.
On the other hand, concern with the re-
liability of the techniques has not ended.
As was stated last year, agreement of the

0.6 0.7

0.8 09 |.'0 l.'l 1.2 1.3
Rb-Sr AGE/ K-A AGE OF MICAS

Fig. 5. Histogram showing ratio of K-A age
to Rb-Sr age of all the mica samples for which
this comparison has been made.

rubidium-strontium and potassium-argon
ages for a sample of mica is a good indica-
tion that the mineral has formed a closed
system. Therefore an age measurement is
not considered satisfactory unless such
agreement is found. The results of further
comparisons of K-A and Rb-Sr ages of
mica are shown in figure 5. Agreement
between these two ages is most common,
although not invariable. Attention is being
given to those cases in which agreement is
not found in order to see whether the lack
of agreement is related to the condition or
type of rock or to the history of a particular
geographical region.

It was reported last year that satisfactory
agreement between K-A ages of mica and
concordant U-Pb ages of uraninite could
be obtained if the specific gamma activity

DEPARTMENT OF TERRESTRIAL MAGNETISM 101

of potassium were taken to be 3.2 y/g/sec. is a long band of approximately 2600-
During the past year, an absolute counting million-year-old rocks extending from Wy-
experiment has been carried out here oming through Montana, Minnesota,
which indicates that the specific gamma Manitoba, and Ontario into Quebec. Simi-
activity is slightly higher than this, 3.39 ± lar areas of very ancient rocks have been
0.12 y/g/sec. Thus even the best micas found in Africa and Australia. Less corn-
may lose about 5 per cent of their argon, plete data indicate similar regional epi-
The experimental errors are such, however, sodes elsewhere.

that the specific activities determined by the These results will be discussed in more

counting experiment and by the geological detail in the following sections,
comparisons are not in disagreement.

The number of important geological T/ie specific Gamma Activity of Natural
problems that are amenable to study by Potassium
these techniques is overwhelming, and Until recently one of the principal diffi-
workers in other laboratories as well as culties in the determination of K-A ages
here have now turned their attention to- has been the lack of an accurate value for
ward them. During the past year our the decay constant of K40 for electron cap-
group has begun an investigation of re- ture to A40, or its equivalent, the specific
gional regularities in the ages of the Pre- gamma activity of potassium. In last year's
cambrian rock exposures. That such regu- annual report it was shown that good
larities probably exist was suggested by the agreement was obtained between K-A ages
remarkable agreement found between ages of mica and concordant U-Pb ages of
of rocks from various parts of the Grenville uraninite if a specific gamma activity of
subprovince of the Canadian Shield. Early 3.2 y/g/sec were used. It was pointed out,
work by Ellsworth and subsequent work however, that probably the true specific
by other investigators has demonstrated gamma activity was slightly higher than
that over this area of approximately 10,000 this and that a few per cent of the radio-
square miles all the rocks have an age of genie argon has been lost by the mica. To
1000 ±100 million years. Similarly in the resolve this question an absolute counting
Appalachian system almost all the igneous experiment has been carried out which
and metamorphic rocks had been found to eliminates several of the uncertainties pres-
have an age of 300 ±100 million years, ent in earlier measurements.
These results suggest the hypothesis that With the exception of the recent experi-
throughout earth history there have been ment of McNair, Glover, and Wilson
periods of a few hundred million years' {Phil. Mag., 1, 199 [1956]), all previous
duration in which extensive igneous and determinations of the specific gamma ac-
metamorphic activity took place in a given tivity have made use of counting tech-
area. The techniques of age measurement niques that did not distinguish between the
should make it possible to discover the ionization produced by the particular
time and places of these orogenic events gamma ray being studied and other gam-
and to see whether any regularities can be mas or sources of ionizing radiation. The
found in their occurrence. As an early re- difficulty is not serious for K40 because of
suit of this investigation it has been found the simplicity of its gamma spectrum. But
that,^ over a large part of Arizona, New to know the efficiency of the counter for
Mexico, Colorado, and Wyoming, rocks gamma rays of this energy (1.46 Mev) re-
were formed 1350 ± 100 million years ago, quires calibration of the counter with gam-
and that in Ontario there is a large group mas from sources of known specific ac-
of rocks 2600 million years of age. Com- tivity. Unfortunately, all the standard
bination of these results with those found sources for this energy region either have
in other laboratories indicates that there more than one gamma ray in their spec-

102

CARNEGIE INSTITUTION OF WASHINGTON

trum or have a serious uncertainty in the
fraction of the disintegrations that involve
the emission of the standard gamma ray.
These difficulties are sufficiently great to
account easily for the wide spread in the
results of earlier experiments.

In the measurement of McNair, Glover,
and Wilson, referred to above, some of the
difficulties were eliminated by means of
a Nal (Tl) scintillation spectrometer which
enabled them to use Na24 as a standard
since they could single out those counts
due to the 1.38-Mev gamma ray and reject
those due to the 2.75-Mev gamma ray.
In this way they obtained a value of 3.33
y/g/sec for the specific gamma activity.
Although this experiment is superior to the
earlier ones that did not make use of a
scintillation spectrometer, two objections
can be raised against it: the uncertainty
in guessing the difference in the efficiencies
of the counter at 1.38 and 1.46 Mev; and
the fact that their experiment was designed
to determine not the specific gamma ac-
tivity but the (3-y branching ratio. Since it
is the specific gamma activity that is im-
portant for age measurements, their value
of the branching ratio must be multiplied
by the specific beta activity, thus introduc-
ing another serious probable error.

The specific gamma activity was deter-
mined here with a scintillation spectrom-
eter; Co60 was used as a standard as well
as Na24, the efficiency of the counter being
thus determined at 1.17, 1.33, and 1.38 Mev.
The absolute activities of the standards
were known as a result of absolute beta
measurements made at the National Bu-
reau of Standards by H. M. Seliger.

A diagram of the counting apparatus is
shown in figure 6. A concentrated solution
of potassium acetate was placed in the
counting bottle above the 3 by 3 inch
Nal(Tl) crystal. A single-channel pulse-
height analyzer was used to obtain the
spectrum shown in figure 7. The peak in
the spectrum, known as the "photopeak,"
represents all the 1.46-Mev gamma rays
that lose all their energy in the crystal.
The 1.46-Mev gamma ray was counted by

setting the bias of the pulse-height analyzer
so that all pulses of higher voltage were
counted. The background was determined

6.3"

5.4'

JL

500 ml polyethylene bottle

Potossium acetate
solution

Polystyrene cylindrical
shell

3"x 3" No I (Tl) crystal

Dumont 6363
Photomultiplier

JTTTfTJr Scintillation counter

Fig. 6. Scintillation counter for the measure-
ment of the specific gamma activity of potassium.

20 22 24 26 28 30 32 34 36 38 40 42 44

Pulse height- volts

Fig. 7. Gamma-ray spectrum of K40. The
large peak on the right is the photopeak caused
by those gamma rays that have lost all their
energy in the Nal crystal.

in the same way with water instead of po-
tassium acetate in the counting bottle.
The efficiency of the counter was meas-

DEPARTMENT OF TERRESTRIAL MAGNETISM

103

ured by mixing standard Na24 and Co60 energy region from 35 to 40 volts and then
sources of relatively high specific activity using the spectrum, figure 8, to make
in with the potassium acetate. The spectra corrections for the small part of the photo-

30 32 34 36 38
Pulse height -volts

Fig. 8. Gamma-ray spectrum of Na24 showing the photopeak produced by the 1.38-Mev gamma ray.

26

28

30

38

44

32 34 36

Pulse height-volts

Fig. 9. Gamma-ray spectrum of Co60 showing the photopeaks produced by the 1.17- and 1.33-
Mev gamma rays.

obtained with these sources are shown in peak that does not occur in this interval,

figures 8 and 9. The photoefficiency of the It should be emphasized that the photoeffi-

1.38-Mev gamma ray from Na24 was de- ciency does not depend in first order on the

termined by counting all the pulses in the accuracy of the gamma-ray spectrum, but

104 CARNEGIE INSTITUTION OF WASHINGTON

on the counting rate for the 35 to 40 volt
interval. Similar measurements were made
with Co60. The two photopeaks were
counted by counting the intervals 28 to 34
and 34 to 40 volts, respectively. The cor-
rections were greater for Co60 because of
the incomplete resolution of the two photo-
peaks. Corrections were made for the
portion of the 1.33-Mev peak that lay below
34 volts by assuming that the 1.33-Mev
Co60 peak has the same shape as the 1.38-
Mev Na24 peak. Subtracting out the 1.33-
Mev peak corrected in this way from the
measured Co60 spectrum gives the 1.17-Mev
peak. A check was made by comparing
the calculated 1.17-Mev peak with the
photopeak of the 1.12-Mev gamma ray
from Zn65. The shape of the two peaks
was identical within experimental error.
Although these corrections were rather
tedious, the accuracy of the Co60 photo-
efficiencies is comparable to that of the
Na24 photoeffkiency. For both Na24 and
Co60, small additional corrections were
made for coincidence between the two
gamma rays.

The calculated efficiencies at the three
energies 1.17, 1.33, and 1.38 Mev are plotted
in figure 10. The curve through these
three points is extrapolated to 1.46 Mev.
The efficiency calculated in this way was
then combined with the potassium integral
counting data, and the specific gamma
activity was found to be 3.39 ±0.12 y/g/sec.
This value is in agreement with the result
of McNair, Glover, and Wilson (3.33 ±0.15
y/g/sec), and is about 5 per cent higher
than that given last year from measure-
ments of the radiogenic argon content of
mica samples of known age.

The Retention of Argon by Minerals

A number of analyses have been made
of potassium minerals whose age is known
from concordant U-Pb ages of cogenetic
minerals. By means of the specific gamma
activity of 3.39 y/g/sec, the ratio of A40
to K40 that would be found in the potas-
sium mineral if all the radiogenic argon
had been retained can be calculated. Com-
parison of this ratio with that found by

analysis of the potassium mineral indicates
the fraction of the radiogenic argon re-
tained by the mineral. Previously reported
measurements made by this group have
been combined with measurements made
at the University of Chicago by Wasser-
burg and Hayden to calculate the retentivi-
ties shown in table 1. Adequate data are
available only for mica and feldspar at
present.

The retentivity of mica averages about
90 per cent; that of feldspar is consider-
ably lower. The error in these retentivity

£ .7

U

| -6
g.5

|! .4

01

A.3

O

0-.2
.1

1.00

1.10

1.20

1.30

J.40

1.50

E Mev

Fig. 10. The measured efficiency of the Nal
crystal as a function of energy. The curve is
extrapolated to 1.46 Mev to determine the effi-
ciency for the K40 gamma ray.

measurements is about 10 per cent, owing
to the uncertainties in the true value of the
specific gamma activity, the comparison
U-Pb ages, and the analytic determination
of potassium and argon. Thus it is difficult
to be absolutely certain that the average
mica sample loses any argon; the data indi-
cate, however, that micas probably lose a
small fraction of their radiogenic argon
whereas feldspars lose about 25 per cent.
The mica ages in this report are calculated
using a specific gamma activity of 3.24
y/g/sec, which is approximately equivalent
to using 3.39 y/g/sec with a 5 per cent
correction for argon loss.

There are other minerals that contain
small amounts of potassium, and with
modern analytical techniques it should be
possible to measure their radiogenic argon
content and hence their retentivity. It may
be that minerals will be found that retain a
greater fraction of their argon than the
micas.

DEPARTMENT OF TERRESTRIAL MAGNETISM 105

TABLE 1. Retention of Argon by Minerals as

Indicated by Comparison with Concordant

Uraninite Ages

Specific gamma activity of 3.39 y/g/sec and
specific beta activity of 27.6 (3/g/sec

(Many of these comparisons are obtained from

the work of Wasserburg and Hayden,

Geochim. et Cosmochim. Acta, 7, 51

[1954]; 9, 153 [1956].)

Sample Location

Concordant Argon

Uranium Reten-

Age, mil- tivity,

lion years %

A. Micas

1. Portland, Conn 267 0.95

2. Glastonbury, Conn 255

3. Spruce Pine, N. C 375

4. Branchville, Conn 367

5. Parry Sound, Ont 994

6. Cardiff Twp., Ont 1020

7. Wilberforce, Ont 1030

8. Keystone, S. Dak., lepido-

lite 1600

9. Keystone, S. Dak., musco-

vite 1600

10. Viking Lake, Sask 1890

11. Bikita, S. Rhodesia 2650

B. Feldspars

1. Portland, Conn 267

2. Glastonbury, Conn 255

3. New Bedford, N. Y 355

4. Branchville, Conn 367 .75

5. Parry Sound, Ont., micro-

cline 1 994 .72

6. Parry Sound, Ont., micro-

cline 2 994 .78

7. Parry Sound, Ont., albite. . 994 .60

8. Cardiff Twp., Ont., sam-

ple 1 1010 .81

9. Cardiff Twp., Ont., sam-

ple 2 1010 .76

10. Wilberforce, Ont 1030 .77

11. Tory Hill, Ont., pegma-

tite 1030 .68

12. Tory Hill, Ont, granite... 1030 .68

13. Keystone, S. Dak 1600 .59

14. Viking Lake, Sask 1890 .79

Ages of Roc\s in the Canadian Shield

For a number of years it has been recog-
nized that the pegmatites intruding the
older rocks in the Grenville subprovince
are approximately 1000 million years of
age. More recently, measurements of ages

of igneous and metamorphic rocks in this
area have shown that the age of the
igneous intrusion and the metamorphism
is likewise 1000 million years. Rocks of
this age have also been found south of the
Grenville region in New York State, both
in the Adirondacks and in the Catskill
Mountains. Measurements of "Grenville
ages" made at the Carnegie Institution are
shown in table 2, and the location of the
Canadian samples on the map (fig. 11).
Numerous similar results have been found
in other laboratories, but the boundaries

TABLE 2. Rocks of Grenville Age

.97

,88

Sample Location

Age, million years

.99

K-A

Rb-Sr

U238-

TJ235_

.93

Pb206

Pb207

.92

.87

1.

Wilberforce, Ont. .

975

1000

1040

1050

2.

Cardiff Twp.,

.81

Ont

1010

1030

1020

1020

3.

Bancroft, Ont. . . .

890

990

.95

4.

Wavy Lake, Ont. .

1025

1075

.93

5.

Canada Hill gneiss,

.80

6.

BearMt.,N. Y..
Storm King gran-
ite, Bear Mt.,

930

1030

1020

1060

.77
.81
.77

N. Y

900

7.

Natural Bridge,

N. Y

1025

1065

of this "Grenville orogeny" have not yet
been established.

Farther to the north in Ontario, much
more ancient rocks have been found, all
approximately 2600 million years of age,
representing the oldest rocks that have
been found in North America. The data
are shown in table 3. As a result of meas-
urements made here and elsewhere, there
appears to be a belt of these very old rocks
extending from Wyoming through Mon-
tana, Minnesota, Manitoba, and Ontario
into Quebec. In eastern Ontario near the
Quebec border these rocks appear within
about 150 miles of the Grenville rocks, and
later measurements of the rocks around
Lake Timiskaming in Ontario and Quebec
may bring these rocks of such different
age into even closer contact.

A number of measurements have also

106 CARNEGIE INSTITUTION OF WASHINGTON

been made on rocks lying in the region
between these two areas. Here the K-A
and Rb-Sr ages are generally in disagree-
ment, in contrast to the other two areas
discussed, and also in contrast to our previ-
ous experience. Some of the discordance
may result from the heating and partial
recrystallization of the 2600-million-year-
old rocks during the Grenville orogeny,

but it also seems probable that rocks of
intermediate age occur in this region. The
results of these measurements are found
in tables 4, 5, and 6.

The Cutler batholith (table 4) intrudes
the Sudbury series, and thus these sedi-
mentary rocks are older than 1350 million
years. The agreement of the Rb-Sr ages
of both mica and feldspar from the peg-

Fig. 11. Map showing the location of samples from the Canadian Shield.

TABLE 3. Rocks Approximately 2600 Million
Years of Age

Age, mil-
Location and Sample lion years

Rb-Sr K-A

Hearst, Ont., pegmatite 2605 2595

Kirkland Lake, Ont., Round Lake

lamprophyre 2600 2450

Kirkland Lake, Ont., Round Lake

granite 2640 2530

Timmins, Ont., granite 2470 2520

Silver Leaf Mine, S. E. Manitoba. . 2640 2210

matites and the agreement of the K-A ages
of the micas and the Rb-Sr age of the
granite suggest that the batholith was
formed about 1750 million years ago and
reheated 1350 million years ago. The data
are insufficient to establish these events,
however.

The Copper ClifT (table 5) rhyolite is a
highly metamorphosed rock; originally it
was probably an intrusive igneous rock,
although some workers have thought it to
be a lava or a sedimentary rock. The

DEPARTMENT OF TERRESTRIAL MAGNETISM

107

Rb-Sr and K-A ages are the most dis-
cordant ones that have been found. In
particular, the K-A age of the mica is much
higher than the Rb-Sr age.

Other ages found in this region are
shown in table 6. Here are two additional
K-A ages appreciably higher than Rb-Sr
ages. The Sudbury breccia contains pieces

TABLE 4. Age Determinations on the Cutler
Batholith

Sample

Age, million years

Rb-Sr

K-A

Pegmatite 1 Muscovite 1750 1440

Feldspar 1760 1165

Pegmatite 2 Muscovite 1700 1420

Granite Biotite 1325 1380

TABLE 5. Age Determinations on the Copper
Cliff Rhyolite

Sample

Age, million years

Rb-Sr

K-A

Muscovite 1730 1390

Biotite 1220 2130

Feldspar 2360 1400

TABLE 6. Other Age Determinations in the
Sudbury Area

Location and Sample

Age, million years

Rb-Sr

K-A

Sudbury, gabbro 1325 1830

Sudbury, breccia (matrix) 1440 1870

Levack, norite 1830

Cobalt, Ont., lamprophyre 2050 2160

of the Copper Cliff rhyolite and the Mc-
Kim formation, and hence these rocks are
older than the mica in the matrix of the
breccia. Although the Rb-Sr and K-A ages
are discordant there is little doubt that the
breccia is older than 1500 million years,
and hence the rocks found in the breccia
cannot be late Precambrian. The Sudbury
gabbro intrudes the Missasagi quartzite at
Sudbury. As a result of the age measure-
ment it is found that this rock is older than

the Grenville rocks. Thus all the rocks
in this region appear to be intermediate in
age between the 1000-million-year-old rocks
in the Grenville subprovince and the 2600-
million-year-old rocks farther north, but
no conclusions can yet be drawn about
their exact ages and relationships. These
results illustrate the importance of meas-
uring both the Rb-Sr and the K-A ages of
a mica sample.

TABLE 7. Potassium-Argon and Rubidium-
Strontium Ages of Rocks from Western
United States

K-A ages are calculated from decay constants
of K40 of Ac = 0.557x 10-10 yr-1, Xj3 = 4.72x lO"10
yr-1 or a total half-life of K40 of 1.31 X 109 years.
Rb-Sr ages are calculated using a half-life for
Rb87 of 50X109 years.

Sample Location

Age, mil-
lion years

K-A Rb-Sr

1. Gneiss, Zoroaster, Grand Can-
yon, Ariz 1390 1370

2a. Lawler Peak granite, Bagdad,

Ariz 1410 1390

2b. Pegmatite in Lawler Peak

granite 1410 1500

3. Pegmatite, Wickenberg, Ariz.. 1160 1300

4. Pidlite Mine, Mora Co., N. M. . 1330 1490

5. Granite, Sandia Mts., Albu-

querque, N. Mex 1350 1340

6. Harding Mine, Dixon, N. Mex. . 1300 1300

7. Uncompahgre granite, Mesa

Co., Colo 1320 1320

8. Granite, Doyleville, Colo 1320 1310

9. Brown Derby pegmatite, Ohio

City, Colo 1330 1420

10. Granite, Sherman, Wyo 1420 1410

Another group of granitic rocks from
western United States present interesting
analyses. The ages of micas from these
rocks are given in table 7, and the locations
of the rocks are shown in figure 12. The
sample numbers in the table correspond to
the numbered locations on the map. Four
other locations are also shown on the map.
The age of the sample from each of these
locations is indicated by letter, A corre-
sponding to 2500, B to 1600, and C to 1100

108

CARNEGIE INSTITUTION OF WASHINGTON

million years. The map shows all the Pre-
cambrian rocks measured at this laboratory
from these states.

The agreement of the ages in table 7
shows that the mica in all these rocks was
formed close to 1350 million years ago, and,
therefore, that there was widespread crys-
tallization of granitic rocks at that time in

CHEYENNE

Table 8 gives the ages of several South
African micas for which both K-A and
Rb-Sr ages have been measured. Samples
from Manano and Karema were provided
by Dr. L. Cahen, of the Belgian Congo
Museum. The Muika sample was part of
the collection analyzed by L. O. Nicolaysen
several years ago here at the Carnegie In-
stitution. From geologic evidence it was
thought that the Karema pegmatite was
formed as part of a much earlier process

TABLE 8. Ages of African Rocks

Sample Location

Age, mil-
lion years

K-A Rb-Sr

Muika, Belgian Congo

Manano, Belgian Congo

Karema, Tanganyika Territory.

Bikita, S. Rhodesia

Kubuta, Swaziland

950 950

870 975

980 1100

2450 2680

2730 2920

FORT o
WORTH

SCALE, STATUTE MILES
100 200 300

(S)

Fig. 12. Map showing the location of samples
from western United States.

this area. It is suggested that this crystal-
lization accompanied a period of regional
rock formation comparable to the period
of igneous intrusion and metamorphism
in the Grenville subprovince. It is also
evident that these mica ages have been
preserved in spite of subsequent geologic
events in western United States such as the
Laramide orogeny.

than that which formed the Muika and
Manano pegmatites. These measurements
show that, if this inference is true, the
later process completely removed any trace
of the earlier history of the mica from the
Karema pegmatite. The analytical data on
the Muika mica concurred with those ob-
tained by Nicolaysen within 3 per cent.

The Bikita and Kubuta micas were also
part of the Nicolaysen collection. The
Bikita ages have been presented before,
but are included for comparison with the
Kubuta ages. The Kubuta ages are the
greatest obtained in our laboratory by
either the Rb-Sr or the K-A method.

THEORETICAL AND STATISTICAL GEOPHYSICS

S. E. Forbush

EQUATORIAL ELECTROJET

Near the magnetic equator the ampli-
tude, Ah, of the diurnal variation in the
horizontal component, H, of the geomag-
netic variations is unusually large. At
Huancayo, Peru, the abnormality exceeds
that observed elsewhere. Along the west
coast of South America, Ah was observed

in 1949 at fourteen stations by A. A.
Giesecke, Jr., Director of the Instituto
Geofisico de Huancayo. These stations ex-
tended from geographic latitude 3.4° N
to 16.2° S. The observed rapid decrease of
Ah with distance north or south of the
magnetic equator indicated a concentrated
narrow band of electric current flowing

DEPARTMENT OF TERRESTRIAL MAGNETISM 109

eastward in the upper atmosphere during tion and of the vertical field had not been

midday. This concentrated current, known measured.

as the equatorial electrojet, is superposed Continuous operation of the four vario-

on the more diffuse current distribution graphs during the IGY should provide

that accounts for the normal quiet-day data from which to determine whether

magnetic diurnal variation, SQ. electrojet effects occur in the lunar diurnal

The existence of the equatorial electrojet variation, magnetic sudden commence-

effect on the quiet-day diurnal variation ments, solar-flare effects, magnetic storm

near the magnetic equator is thus definitely changes (Dst) , etc. If it turns out that no

established. To determine whether similar electrojet effects are found for changes dur-

electrojet effects occur for other variations ing magnetic storms, then the records of

in the geomagnetic field, and if so to de- vertical intensity from the two stations

termine whether the same electrojet is in- nearest the magnetic equator promise to

volved, a chain of four temporary magnetic provide a measure for the amplitude of the

recording stations is now being established quiet-day diurnal variation and its vari-

for the International Geophysical Year on ability, which should be free of the effects

the west coast of Peru, to operate in con- of magnetic disturbance that often spoils

junction with the John A. Fleming Ob- Bartels' present measure (hw), based on

servatory of the Instituto Geofisico de the diurnal variation of H at Huancayo.
Huancayo. At each of the four temporary

TOY stations a nnrtihle Askania vario- VARIABILITY OF GEOMAGNETIC DIUR-
IGY stat ons a portable Askania vario ^^ VAJR1ATI0N AND 0F /olo-
graph will register the variations in dec- SPHERIC WINDS

lination, horizontal intensity, and vertical _ , , ., , , ,

«, . . J j ., , Last year s report described the results

intensity. The proiect is made possible - I . , r . , . , ,

. / i t 1 A or a statistical experiment showing that the

through grants approved by the Geomag- . , c F , a ° ,

. ° 6, , , r£T n T ; . 1n° magnitude or the solar Hare, or crochet,

netic Panel of the U. S. International Geo- ^ .r ^ horizontal etic com„

physical Year and through the co-operation t ^ Hu a was dennitely greater,
of the U. S. Coast and Geodetic Survey, ^ ^ ayer on^d when ^ iet_d
which has lent die four^ variographs to the ^^ variatioilj Sqj in the geomagrietic
Department of Terrestrial Magnetism, and fidd was greaten Akhough the amplitude
the help of the Instituto Geofisico de of Sq [s known t0 increase with sunspot
Huancayo, which has co-operated in the numDer, the size of crochets was found to
preliminary survey and will manage the ^e independent of sunspot number; thus
operation of the five stations. the above relation between crochet size
The preliminary survey, using two vario- and amplitude of Sq was not due to a solar
graphs, involved two or three days' con- activity effect common to both. It was con-
tinuous registration of the variations in the eluded that both effects could be ascribed
geomagnetic field at each of some sixteen to variations in the strength of the wind
stations. This survey was begun in March system that drives the dynamo responsible
and completed in May 1957. The results for Sq, since on days with stronger wind
provided a necessary guide to locate the systems both the size of crochets and the
final stations, and indicated an electrojet amplitude of Sq would be larger,
with total extension of roughly 500 miles In the last few years ionospheric wind
in the north-south direction. Moreover, velocities have been measured from the
these observations also provided data from rate of drift of meteor trails by Professor
which to calculate the height of the electro- A. C. B. Lovell and his colleagues at Jodrell
jet, previously undetermined since, near Bank. Their results showed that the semi-
the magnetic equator, the variation with diurnal variation of wind velocity, in the
latitude of the diurnal variation of declina- range of heights where it was measured,

110 CARNEGIE INSTITUTION OF WASHINGTON

had the wrong phase, according to the Zealand). The station at Cheltenham
dynamo theory, to explain the quiet-day (Maryland, U. S.) was transferred in Oc-
magnetic variation. The results obtained tober 1956 to the Fredericksburg Magnetic
also showed large variations from day to Observatory at Fredericksburg (Virginia,
day in the amplitude of the semidiurnal U. S.). Tabulations of bihourly means of
wind components (N-S and E-W) and in ionization corrected for bursts and baro-
the daily average wind velocity in the N-S metric pressure for Huancayo from 1946
and E-W directions. Professor Lovell has to 1955, and for Cheltenham from 1937 to
made available to us the amplitudes of the 1954, as well as summaries for Godhavn
semidiurnal wind component for 70 days and Christchurch have been published in
together with the daily values of the pre- volume XX of Carnegie Institution of
vailing wind components. Arrangements Washington Publication 175. These re-
have been made to obtain constants for suits, compiled in collaboration with the
recent magnetograms from Huancayo Instituto Geofisico de Huancayo, The Dan-
which will be measured with a recently ish Meteorological Office, The Department
constructed rapid scaling device to furnish of Scientific and Industrial Research, New
material from which to derive the ampli- Zealand, and the U. S. Coast and Geodetic
tude of Sg. It will then be possible to de- Survey, together with those contained in
termine statistically whether the measured earlier volumes, make available to investi-
variability of the ionospheric winds is cor- gators many of the essential data obtained
related with the variability of the ampli- since the start of the Department's cosmic-
tude of Sg. ray program.

Large ionization chamber. The large

COSMIC-RAY INVESTIGATIONS cosmic-ray ionization chamber was main-

Twenty-seven-day variation in cosmic- tained in essentially continuous operation

ray intensity and in the geomagnetic field, at Derwood during the report year. No

The phase and amplitude of the 27-day solar-flare effects have been observed since

waves in horizontal magnetic intensity, H, February 23, 1956.

at Huancayo have been determined for Co-operation in operation of cosmic-ray
about 155 solar rotations. Magnetic results meters. The successful operation of Comp-
are expected shortly from Huancayo for ton-Bennett cosmic-ray meters over a long
some of the years after 1947; these will be period at so many stations has been possi-
analyzed for the 27-day waves. With re- ble only through the wholehearted and un-
sults already at hand for the 27-day waves selfish co-operation of several organizations
in cosmic-ray intensity for 250 solar rota- and individuals. We wish to express our
tions it will thus be possible to determine appreciation to the following organizations
rather reliably whether there is any signifi- for the operation and maintenance of
cant phase difference between the maxima cosmic-ray meters : The Danish Meteoro-
of the cosmic-ray waves and the minima logical Institute and the staff of its God-
of the waves in H at Huancayo. The re- havn Magnetic Observatory at Godhavn,
suits should have a significant bearing on Greenland; the U. S. Coast and Geodetic
theories for explaining decreases in cosmic- Survey and the staff of its magnetic ob-
ray intensity associated with magnetic servatory at Cheltenham, Maryland (at
disturbance. Fredericksburg, Virginia, since October

Old cosmic-ray program. Compton-Ben- 1956); the High Altitude Observatory of

nett meters were satisfactorily operated the University of Colorado and its staff at

throughout the report year at Godhavn Climax, Colorado; the Instituto Nacional

(Greenland), Climax (Colorado, U. S.), de la Investigation Cientifica and the Uni-

Ciudad Universitaria (Mexico, D. F.), versidad de Mexico, Mexico, D. F.; the

Huancayo (Peru), and Christchurch (New Government of Peru and the staff of its

DEPARTMENT OF TERRESTRIAL MAGNETISM 111

Instituto Geofisico de Huancayo for mak- Scientific and Industrial Research and the
ing available the Compton-Bennett records staff of its Magnetic Observatory at Christ-
from Huancayo; and the Department of church, New Zealand.

LABORATORY PHYSICS

NUCLEAR PHYSICS

N. P. Heydenburg, G. M. Temmer,1 and
G. F. Pieper

During the past year, our continuing
Coulomb excitation studies of nuclear en-
ergy levels included an investigation of the
gamma radiation observed in krypton with
enriched targets, further analysis of the
gamma-ray spectrum of Fe57, and a study
of the rotational levels of the dysprosium
isotopes with enriched targets. In addition
to the Coulomb excitation work, we have
been engaged in an experimental study of
the angular distribution and angular corre-
lation of the protons and gamma radiation
emitted when fluorine is bombarded by 6-
to 6.5-Mev alpha particles. The interest
here is to determine, if possible, whether
the reaction proceeds by compound-nucleus
formation or by some direct-interaction
process.

COULOMB EXCITATION STUDIES

Krypton. In last year's annual report,
tentative isotopic assignments were given
for the four gamma rays observed from
natural krypton bombarded by 6.4-Mev
alpha particles. With six isotopes present
in natural krypton, it was desirable to have
these assignments checked with isotopically
enriched samples.

During the year two gas samples of
krypton with different isotopic enrich-
ments were made available to us from
Yale. In co-operation with C. E. Anderson,
as a Guest Investigator from Yale, these
samples have been bombarded here with
6.1- and 6.6-Mev alpha particles, using our
high-voltage equipment. With three tar-
gets having different isotopic ratios it was
possible to make the following gamma-ray

1 On leave of absence, 1956-1957, as Guggen-
heim Fellow, Saclay Laboratories, Paris, and
Institute for Theoretical Physics, Copenhagen.

energy assignments unambiguously : Kr78,
450 kev; Kr80, 620 kev; Kr82, 780 kev; and
Kr84, 880 kev. No gamma rays were ob-
served in Kr83 and Kr86. One of the tar-
gets, enriched 30 fold in Kr78 (0.35 per
cent abundant in natural krypton), made
it possible to observe the gamma ray asso-
ciated with this isotope.

The gamma rays from these even-charge,
even-mass isotopes are due to electric quad-
ruple excitation of the first spin 2+ level.
The energies of the 2+ states increase in a
systematic way with increasing mass of the
krypton isotope, whereas the gamma-ray
intensity decreases. According to nuclear
shell theory, Kr86 has a closed shell system
for neutrons (neutron number 50). Our
results for the first-excited states of the
krypton isotopes are consistent with the
now well established trends of first-excited-
state energies, which have much lower
values for nuclei having partially filled
shells and rise to quite large values at the
closed shells. Similarly, the reduced transi-
tion probabilities B(E2) decrease in ap-
proaching a closed shell. Again we observe
a strong excitation for Kr78 and have found
no evidence of a gamma ray for Kr86.

Iron. Natural iron has a high abundance
of the isotope Fe56, and the T first-excited
state of this isotope at 854 kev is readily
excited by 6-Mev alpha particles. Gamma
rays at 123 kev and 350 kev are also seen;
they are due to Fe57, which has an abun-
dance of only 2.2 per cent in natural iron.
We have investigated the energy levels of
Fe57 with a target enriched to 59 per cent.
The gamma-ray spectrum observed with
an alpha-particle energy of 4 Mev is shown
in figure 13. In the present study we have
been concerned with the properties of the
third-excited level in Fe57, since earlier in-
vestigations both here and by a group at
Oxford had established the level scheme

112 CARNEGIE INSTITUTION OF WASHINGTON

for the first and second levels. The level
scheme and associated gamma-ray transi-
tions are shown in figure 14. We have
shown that the 350-kev gamma ray is due
to the Coulomb excitation of a level at
365 kev rather than at 350 kev, by observa-
tions on the yield of this gamma ray as a
function of the alpha-particle energy.

15 20 25 30

CHANNEL NUMBER

Fig. 13. Gamma-ray spectrum from the Cou-
lomb excitation of Fe57 by 4-Mev alpha particles.
The solid line connected to the peak at 350 kev
is the shape of a single gamma ray of that
energy, obtained under similar conditions from
the Coulomb excitation of Ru104. The dashed
parabolas labeled 123 and 228 kev represent
the photopeaks of gamma rays of these energies,
after allowance has been made for higher-energy
radiations and background. The "escape peak"
from the 123-kev gamma ray is evident at about
channel 9.

These results are shown in figure 15. The
theoretical yield curves calculated from
Coulomb excitation theory for the excita-
tion of a 350-kev and a 365-kev level have
been normalized to the experimental point
at 4 Mev. The curve for AE = 365 kev is
in better agreement with the experimental
data. It should also be noticed that the
230-kev gamma-ray yield follows the curva-
ture for AE = 365 kev. That the 230-kev
gamma ray is also due to the excitation of

the 365-kev level and is a cascade transition
to the 137-kev level was shown by observ-
ing coincidences between the 230-kev and
the 123-kev gamma rays.

Since the ground-state spin of Fe57 is 1/2,
and only electric quadrupole (E2) transi-
tions are induced by Coulomb excitation,
the 365-kev level can have spin of either
3/2 or 5/2. From our results, 5/2 seems

Fe57(cc.,cc'r) Fe57

(5/2-)

I 13

J ,

(5/2-)

71

10 I

(3/2-)-
(1/2-)-

23

YI4 \i

137

228

365 KEV

350

137

14
0

Fig. 14. Proposed level scheme for Fe57.

more likely for the following reasons: the
observed branching ratio for the 350-kev
and the 230-kev gamma rays is in much
better agreement with the predicted ratio
for spin 5/2 than for 3/2; if the spin were
3/2, a transition to the ground state by
magnetic dipole (Ml) radiation would be
more probable than to the first-excited
state, whereas for 5/2 spin only E2 radia-
tion is allowed to the ground state, which
should be much weaker than Ml radiation
to the first-excited state. We observe the
transition predominantly to the first-excited
state; a gamma ray of 365 kev cannot be
present to an amount greater than 5 per
cent of the 350-kev gamma ray.

Angular distributions were also observed
for the 230-kev and 350-kev gamma rays.
The theoretical predictions are not unique
for either of the two spin choices, since

DEPARTMENT OF TERRESTRIAL MAGNETISM 113

Ea (Mev)

Fig. 15. Excitation curves for the Coulomb
excitation of Fe57 by alpha particles. The points
are the experimental gamma-ray yields. The
solid and dashed curves are the theoretical thick
target E2 Coulomb excitation functions for the
level positions indicated.

the Coulomb excitation of enriched iso-
topes of dysprosium, recently available
from Oak Ridge. In each of the two even-
even isotopes Dy162 and Dy164 we observed
only one gamma ray, with energies 82 kev
and 75 kev, respectively. These are due
to the excitation of the first of the rota-
tional levels in the sequence having spins
2+, 4+, 6+ * • * with ground-state spin 0+.
By E2 excitation we can excite only the
first T level.

For odd-A nuclei it is possible to excite
two of the rotational levels by El transi-
tions since the levels have spins 70 + l,
7o+2, 7o + 3, * * *, where h is the ground-
state spin. These levels, according to the
theory of Bohr and Mottelson, should have
energies given by

Z?=J[/(/+l)-/o(/o + l)]

where / is the spin of the level, and S is
the effective moment of inertia of the
rotating nucleus.

TABLE 9. Experimental Results from the Coulomb Excitation of Fe57

Column 1 gives the energies of the levels excited in Fe57. Column 2 gives the reduced upward
transition probability, B(E2), in units of <?2Xl0~48 cm4. Column 3 gives F, the favored factor,
the ratio of the observed B(E2) to that expected for a single particle transition. Column 4 gives
the energies of the observed gamma rays, and column 5 the branching ratios of the levels from
which they originate. Columns 6 and 7 give S and S2, where S2 is the ratio of E2 to Ml compo-
nents in the radiation.

Fe57 Level,
kev

B(E2)

Gamma-Ray

Energies,

kev

Branching,
per cent

E2

Ml

137

0.050

13

123

91

+ 0.19 0.04

137

9

PureE2

365

0.033

8.4

228

7

-0.18 0.03
+ 2.6 6.7

350

93

+ 0.28 0.08
-90 8000

another parameter, h2, the mixing ratio of
E2 to Ml radiation, is involved for the
downward transitions. Our results for
Fe57 are summarized in table 9.

Dysprosium. Continuing our investiga-
tions of the rotational bands that occur in
the region of the rare-earth nuclei, we have
observed the gamma rays resulting from

In our earlier survey of the rare-earth
nuclei we had observed two gamma rays
in natural dysprosium, one of which at 166
kev was thought to be due to the second
rotational level in one or both of the odd-A
isotopes Dy161 and Dy163. Targets en-
riched to 75 per cent were available for
these two isotopes. This enrichment proved

114 CARNEGIE INSTITUTION OF WASHINGTON

to be satisfactory for Dy163. The 166-kev
gamma ray was found to be due entirely
to this isotope. Another gamma-ray peak
at about 75 kev was found to be due in
part to Dy163 and in part to the neighbor-
ing isotopes Dy162 and Dy164. The two
gamma rays in Dy163 at 75 kev and 166
kev are believed to be due to excitation of
the first and second rotational levels (of
the ground-state rotational band). The
ratio of the energy of the second-excited
level to that of the first-excited level agrees
within our experimental errors to the pre-
dicted ratio calculated from the equation
above for 7o = 5/2, which is the known
ground-state spin of Dy163. A much weaker
gamma ray at about 93 kev was observed
which could correspond to a transition
from the second- to the first-excited rota-
tional levels. This interpretation was con-
firmed by observing coincidences between
the 93-kev and 75-kev gamma rays.

The situation in Dy161 was not so simple,
however; at first sight there did not ap-
pear to be any gamma rays associated with
this isotope, but a careful analysis did
reveal a weak gamma ray at about 103 kev.
The strong gamma ray observed at 82 kev
was accounted for by the presence of the
neighboring isotope Dy162 in the sample.
Further, the K X-ray peak which is always
excited in these rare-earth targets appeared
to be stronger than would have been ex-
pected from observations on the Dy162
target. It is known that the K X-ray peak
results primarily (for excitation by alpha
particles) from a so-called fluorescence
process. In this process, for example in
Dy162, the 82-kev transition can occur both
by gamma emission and by conversion
electron emission. In the latter case, for
K conversion, a K electron is emitted, leav-
ing a vacancy in the K shell which can be
filled from a higher shell with emission of
a K X-ray. Hence it would be expected
that the K X-ray peak should follow the
same yield as a function of alpha energy
as the 82-kev gamma. This was found to
be true for the Dy162 target but not for

Dy161. The yield of the peak, having about
the same energy as the K X-ray peak,
varied more slowly with energy than in
the Dy162 target, and indeed it followed
a theoretical yield curve for the excitation
of a 46-kev gamma ray. Hence, it seems
reasonable to believe that most of this peak
is due to a gamma ray of about 46-kev
energy. We have suggested that this ray
is due to the excitation of the first rota-
tional level and that the weak gamma ray
at 103 kev is due to the excitation of the
second rotational level. The ratio of en-
ergies is again in agreement with the pre-
dicted ratio for the ground-state spin of
5/2 for Dy161. Recently we have heard
that evidence has also been found for these
two levels in inelastic proton scattering by
Elbek at Copenhagen. The value of %
the effective moment of inertia, calculated
from the energy-level spacings for the two
isotopes Dy161 and Dy163, is quite different
— a rather striking result, since both have
the same ground-state spin.

INVESTIGATION OF THE MECHANISM OF
NUCLEAR REACTIONS

In recent years it has been reasonably
well established that nuclear reactions
other than elastic scattering proceed by a
range of mechanisms lying between the
two extremes of compound-nucleus for-
mation and direct interaction. The picture
of the compound-nucleus mechanism is
that the incoming particle is quickly amal-
gamated by the target nucleus to form a
highly excited compound system, the en-
ergy brought in by the incoming particle
being rapidly shared by all the nucleons
in the system. The compound nucleus thus
formed lasts a long enough time (perhaps
10-15 sec) to "forget" how it was formed
and then decays by particle (or quantum)
emission into an outgoing particle (or
quantum) and a residual nucleus. The
residual nucleus may be left in an excited
state, and decay to its ground state by
further particle or quantum emission. One
possible picture of the direct-interaction
mechanism is that the incoming particle

DEPARTMENT OF TERRESTRIAL MAGNETISM 115

makes a two-body collision with one of
the surface nucleons of the nucleus, caus-
ing the nucleon to be ejected, while the
incoming particle is itself then captured
by the remainder of the nucleus.

The predictions of the direct-interaction
reaction model differ from those of the
compound-nucleus theory, especially with
regard to the angular distributions of
emitted particles and the angular corre-
lations between the incoming and out-
going particles and the gamma radiation
which usually comes from the residual
nucleus if that is left in an excited state.

We have investigated the angular dis-
tributions of protons and the angular cor-
relation of protons and gamma rays in
the F19(a, p)Ne22 reaction in order to
characterize, at least to some extent, the
reaction mechanism. The bombarding en-
ergies employed (6.0 to 6.5 Mev) are in a
sense putting the direct-interaction mech-
anism to a rather severe test, since such a
mechanism is expected to apply better at
higher bombarding energies. Our results,
however, show some striking agreements
with the predictions of the direct-inter-
action model. The best agreement in pro-
ton angular distribution, shown in figure
16, occurred for the ground-state transition
at £"0 = 6.40 Mev. By itself, the close fit of
the experimental data to the predicted
yield based on the direct-interaction theory,
[jo(QR)]2* may be taken as evidence for
this reaction mechanism. We observed,
however, that a relatively small change in
bombarding energy alters the distribution
rather considerably. Since the direct mech-
anism should not show any "resonant"
behavior, this result presumably indicates
a contribution to the yield from compound-
nucleus processes. This effect is shown in
figure 17, in which five proton angular
distributions taken at closely spaced en-
ergies are presented. Of these five, only

* j0 is the zeroth-order spherical Bessel func-
tion; its argument is QR, where Q=\ Ka — Kp|,
the K's being the wave vectors of the incident
alpha particle and outgoing proton, respectively,
and R being the interaction distance.

those at 6.25- and 6.40-Mev bombarding
energies have shapes clearly interpretable
in terms of a direct interaction. The dis-
tributions at 6.00- and 6.10-Mev bom-
barding energies do show peaks at ap-
proximately the proper places for the di-
rect mechanism, although their shapes dif-
fer from the shape characteristic of the
direct process. This difference in shape
and, even more striking, the completely

Fl9(oc,p0) Ne'
Eoc '6 40 Mev

— Experiment
2

Fig. 16. Angular distribution of protons from
the ground-state transition in F19(a, p)Ne22 at
an alpha-particle bombarding energy of 6.40
Mev. The dashed curve shows the theoretical
prediction of the direct-interaction reaction mech-
anism.

different nature of the distribution at 6.55-
Mev bombarding energy are indications of
the existence of a process additional to the
direct one responsible for the result at
6.40 Mev.

The same change of distribution shape
with bombarding energy was observed for
the protons corresponding to the first-
excited-state transition in F19(a, p)Ne22,
as will be seen in figure 18. Some of the
data of the figure (e.g., the distribution at
6.10-Mev bombarding energy) can be fitted
fairly well by the appropriate direct-inter-
action prediction for this case, [J2(QR)]2,

116 CARNEGIE INSTITUTION OF WASHINGTON

but certainly not all the results can be so
treated. The reason for this rather pecul-
iar behavior on the part of these angular
distributions as a function of bombarding
energy is at present not understood in
detail.

It has been pointed out by Butler, how-
ever, that at low energies contributions to
the direct reaction from the interior of the
nucleus would tend to alter the angular

for certain direct processes has been de-
veloped by Satchler. It predicts that the
gamma-ray yield should be observed to be
azimuthally symmetric about the direction
Q = Ka— Kp, and also symmetric about a
plane through the nucleus perpendicular
to Q. It is clear that the direction Q is
specified by the momenta of the alpha

ANGULAR DISTRIBUTIONS

'(*. P.) Ne

ANGULAR DISTRIBUTIONS
Fl9t«.P0)N«22

Ew -6 55 Mev

E« «6.40 Mev

0 20 40 60 80 100 120 140 160 ISO
Gp - cm system

Fig. 17. Proton angular distributions in F19
(a, /?0)Ne22 as a function of bombarding energy
in the range of 6.00 to 6.55 Mev. The insert
shows the excitation curve at 9p = 45°; it is not
particularly representative of the total cross sec-
tion.

distributions, perhaps markedly, from the
expected form, and that under such con-
ditions a measurement of the angular cor-
relation of the gamma radiation emitted
in the decay of the residual nucleus with
respect to the direction of momentum
transfer, Q, would provide a test as to
whether the reaction proceeds directly or
via compound-nucleus formation. We have,
for this reason, begun investigating the
correlations in the reaction F19(a, /?i)Ne22a?
(y)Ne22. The theory of the correlation

0 20 40 60 80 100 120 140 160 180
6p - cm system

Fig. 18. Proton angular distributions in F19
(a, /71)Ne22 as a function of bombarding energy
in the range 6.00 to 6.55 Mev. The insert shows
the excitation curve at 0p = 45°; it is not particu-
larly representative of the total cross section.

particle and proton; thus experimentally
one measures the number of (pi, y) co-
incidences as a function of the position of
the gamma-ray detector, while selectively
detecting only pi protons at a fixed posi-
tion.

Our first correlation was done at Ea =
6.40 Mev and 0P = 4O°, <pP = 0° (0P and 4>p
are the polar and azimuthal angles speci-
fying the proton direction with respect to
the beam direction). These angles, the
bombarding energy, and the known en-

DEPARTMENT OF TERRESTRIAL MAGNETISM 117

F19 (a,/», ) Ne22* (y) Ne22 Correlations
Ea • 6 40 Mev

180 gy 90 0 90 b7 180

£y«l80° ♦y-O0

Fig. 19. F19(a, p1)Ne22a;(Y)Ne22 angular correlations at 6.40-Mev bombarding energy, for pro-
ton detection angles 0p of 0°, 40°, 90°, and 140°. The solid curve in each correlation represents one
particular (not unique) prediction of the direct-interaction theory; it is included to show that the
results at 0°, 40°, and 140° show the symmetries with respect to 0^ (called Qq in the text) predicted
by the theory, while the data at 90° do not show these symmetries. The inserts are the p± proton
angular distribution at 6.40-Mev bombarding energy and a diagram of the directions involved in
the reaction.

118 CARNEGIE INSTITUTION OF WASHINGTON

ergy release of the reaction specify Q as
having the direction 0q = 28°, <j>q = 180°.
The gamma counter was placed at several
angles 07 for <py = 0° and 4>T = 180°.

The results of this and three other cor-
relations done at 6.40 Mev for different
values of QP are shown in figure 19. The
symmetry predictions of the Satchler the-
ory are seen to be rather strikingly borne
out by the results for 0P values of 0°, 40°,
and 140°. (The solid curve shown in each
correlation has to do with the details of
the theory; it is not a unique prediction,
and somewhat better fits [e.g., at 0P = 40°]
can be obtained by using different values
of the parameters in the theory.) The
failure of the result for 6P = 90° to show
the proper symmetries is equally striking.
It is possible that we are observing one
type of direct interaction in the forward
direction and another type in the back-
ward direction (of proton emission) while
in the central region (0P = 9OO) interfer-
ence effects between the two types destroy
the correlation.

In any event, the fact that a relatively
simple theoretical development, based on
the Born approximation, fits many of the
data obtained in a quite complicated ex-
periment is very encouraging. Further
angular correlation experiments on this
reaction and on similar ones, especially
as a function of bombarding energy, should
throw additional light on the actual proc-
esses involved in nuclear reactions.

BIOPHYSICS

E. T. Bolton, R. J. Britten, D. B. Cowie,

/. /. Leahy} F. T. McClure,2

and R. B. Roberts

For ten years the aim of our studies in
biophysics has been the understanding of
the processes by which protein and nu-
cleic acid are synthesized. Although this
goal has been constant, the activities lead-
ing toward it have varied greatly. During

1 Fellow, CIW, University of California.

2 Visiting Investigator. Home address: Ap-
plied Physics Laboratory, Johns Hopkins Uni-
versity, Silver Spring, Maryland.

the first years our efforts were confined
to observations of the synthesis of rela-
tively small molecules. Later they shifted
to studies of the metabolic pools which
are the precursors of the macromolecules.
It is not sufficient, however, to consider
merely the gross chemical fractions of the
cell, the protein, nucleic acid, lipid, and
other fractions; it is equally important
to distinguish whether a specific protein,
for example, is located in the cell wall or
in the cytoplasm. For many problems it
may also be necessary to consider the or-
ganization of the cell in terms of some
framework larger than a protein molecule.
We have directed more attention this year
to the organization of the cell.

Various approaches to the problem of
cellular organization are possible. Studies
of the metabolic pools show that they are
sensitive to osmotic shock and must, there-
fore, be held in some osmotically sensitive
structure. In yeast, two types of pools can
be distinguished, of which one is an im-
mediate precursor of macromolecules and
the other is exchangeable with outside
material. The application of high pressure
causes an interconversion of the pools, pre-
sumably by alteration of cellular structure.

The cellular structure as a whole can be
altered by formation of protoplasts which
have weakened walls and lose the charac-
teristic appearance of intact cells. These
altered forms continue to synthesize pro-
tein and nucleic acid. A beginning has
been made in the study of the "particles"
of the cells and their role in the processes
of synthesis. A new approach has been
found in the study of quite large particles
that form spontaneously from disintegrated
cellular material. Through the use of
amino acid analogues, the mechanism that
selects amino acids for protein synthesis
has been investigated by observing the
"mistakes" a cell can make in protein
formation. These various items, reported
in detail below, have each contributed to
a better understanding of the functional
operation of the cell.

DEPARTMENT OF TERRESTRIAL MAGNETISM 119

AMINO ACID POOLS IN Escherichia coli

Further studies have been carried out
on the amino acid pools of E. coli in order
to understand the mechanisms of pool
formation and maintenance. The work
has been guided by the general idea that
a detailed understanding of the first step
in amino acid incorporation by the cell
will supply a foundation for investigations
of the later steps leading to macromolecule
synthesis, and may supply clues to the
mechanisms involved in the later steps.

Pool formation is an expression of the
ability of the cell to obtain nutrients pres-
ent at very low concentrations in the en-
vironment and to supply them to the syn-
thetic machinery at high concentrations.
This, perhaps, allows significant simplifi-
cation of subsequent problems of macro-
molecular synthesis.

The principal question is whether the
internally concentrated substances are free
in solution within the cell or held in a
more complex fashion. If the pool is simply
a concentrated solution that pervades the
cell, then the synthetically active structures
within the cell are bathed in this solution,
which is thus the "medium" in which syn-
thesis occurs. On the other hand, the amino
acids of the pool may be more closely as-
sociated with the substructures of the cell
responsible for protein synthesis. They
might be trapped in such substructures
(as in a brush heap) or be bound to them
by labile chemical bonds. In the latter case
it would be highly important to know the
nature of the binding sites and how in-
timately they are related to the synthetic
activities.

The experimental work described below
was aimed at distinguishing between these
alternatives. Although no directly conclu-
sive experiments were devised, the totality
of the experimental evidence obtained
showed that various simplified models
based on these alternatives were inad-
equate. The elaboration of the properties
of the pool obtained by these studies pro-
vides a list of critical requirements that

must be met in the formulation of any
satisfactory model.

Studies of the rate of formation, ex-
change, and loss from the pool. Previous
studies have shown that the pool formed
by growing cells is proportional to the
quantity of amino acid supplied, until the
pool approaches its saturation value. A
typical curve from which the pool size and
rate of pool formation are estimated is
shown in figure 20. Of more direct in-
terest for the analysis of the mechanism
of pool formation are the relationship of
the pool size to the external concentration

Time of sampling, minutes

Fig. 20. Typical curve for the uptake of C14
proline by E. coli at 25° C. Initial proline con-
centration 10-6 M; cell concentration 0.52 mg
wet cells/ml. The initial total rate of uptake of
proline may be calculated from the slope of the
dashed line. The steady-state pool size and the
external concentration at the time it is achieved
are determined from the values of am and £m
shown.

present at the time a steady-state pool is
achieved, and the initial rate of pool for-
mation. For the purposes of this discussion
the steady-state pool size is defined as the
size of the pool at the time its rate of
change is zero. A large number of experi-
ments with C14 proline over the range of
concentrations from 10~6 M to 2 X 10~4 M
have been analyzed in order to study the
variation of these two parameters with
external concentration. The steady-state
pool size varies by a much larger factor
than the initial rate of pool formation. A
mechanism sufficient to explain these re-
sults will necessarily be much more com-

120 CARNEGIE INSTITUTION OF WASHINGTON

plicated than the "permease" mechanism
proposed by Monod et al. (1956) for the
concentration of galactosides in E. coli.

Studies have also been carried out on
the rate of loss from the pool when the
external amino acid is removed. When
samples of cells at 25° C are caught on a
filter and washed with the usual growth
medium (but no glucose) for varying
periods, a rapid loss of about 20 per cent
of the pool is followed by a very slow loss.
These experiments are complicated by the
continuation of protein synthesis, and as

Exchange is observed with about the same
time constant as in A. For experiment C
another part of the culture was diluted by
a factor of 30 and correspondingly large
samples were taken. There is loss from
the pool at a slower rate than the exchange
process shown in B. Experiment D is
similar to C but in place of the dilution
the samples were washed on the filter for
varying lengths of time with the usual
growth medium (with glucose omitted).
The time constant for D is similar to that
for C.

B, add C12 proline to I0"4 molar
C, dilute by factor of 30

Time, hours

Fig. 21. Study of rate of exchange and loss of C14 proline pool in E. coli at 0° C. Cell concen-
tration 1.0 mg wet cells/ml. Suspension incubated at 25° C for 5 minutes after addition of 10~5
M C12 proline and quickly chilled to 0° C. Cells then centrifuged and resuspended in unsupple-
mented medium at 0° C. After 1 hour, C14 proline was added for experiment A. In experiment
C, a part is diluted by a factor of 30 with supplemented medium. For curve D samples were washed
on the filter for the times indicated with unsupplemented medium.

a result good quantitative measurements
of the loss at 25° have not yet been ob-
tained. The rate of loss after removal of
the external amino acid, however, is very
much slower than the initial rate of uptake
when the amino acid is added.

Similar studies have been carried out at
0° C where protein synthesis is suppressed
and more accurate rate measurements can
be made. Figure 21 shows the results of
an experiment performed at 0° C in which
the pool was labeled by exchange A. The
culture was then divided, and to one part,
B, carrier C12 proline (10~4 M) was added.

In another experiment (fig. 22) the rate
of pool formation was measured at 0° C.
It will be observed that the rate of forma-
tion of pool (time constant 4 hours) is
considerably slower than the rate of loss
(time constant <1 hour, fig. 21, D).

These results at 0° C contrast sharply
with those at 25 ° C, where pool formation
is a much faster process than loss. In
spite of this inversion of the relative rates
of formation and loss, the pool sizes at
25° and 0° C are the same within a fac-
tor of 2. It appears that the equilibrium
pool size must be determined by other

DEPARTMENT OF TERRESTRIAL MAGNETISM 121

parameters in addition to the relative rates
of formation and loss, as measured above.

The results suggest that the existence of
a pool suppresses the pool-formation step,
and that the existence of an appropriate
external concentration suppresses the loss
step. Appropriate temperature coefficients
of the four processes (pool formation and
its suppression, pool loss and suppression)
might account for the observations.

Results reported in last year's annual
report have demonstrated that exchange
between pool and external amino acid is
an important process during pool forma-
tion at 25° C with an energy source pres-

-Total C14 proline present

cur either through reactions that are an
essential part of the over-all mechanism
of pool formation or through reactions
that play no real part in that process. In
connection with the latter case, it should
be noted that a reaction of the type

A* + R-A<^A*- -R- -A A*-R + A

where R-A is some complex containing
A, would be observable in an exchange
study but would not necessarily be ob-
servable in the process of pool formation,
since the reaction causes no net change
in the amount of the complex.

TABLE 10. Approximate Rates of Formation
and Exchange

The values are expressed in millimicromoles
of proline per minute per milligram wet cells at
an external concentration of 3.5 X 10~6 M C14
proline. No glucose was present in the exchange
experiments.

Fig. 22. Proline pool formation at 0° C, 1.1
mg wet cells/ml suspension; 3.4 XlO-6 M C14
proline. Exponentially growing cells chilled to
0° C for 45 minutes before C14 proline was
added.

ent. Exchange was found to continue in
the absence of an energy source, or at 0° C,
conditions that strongly suppressed pool
formation. Current studies show that the
exchange rate has a relatively small tem-
perature coefficient whereas the rate of
pool formation has a very large one (at
least, at low temperature, table 10).

Table 10 lists approximate values for the
initial rates of pool formation and ex-
change at 25° and 0° C.

Any satisfactory model of the amino
acid pool must certainly allow for the
occurrence of exchange in the absence of
an energy source and for the strikingly
different temperature dependence of the
exchange process and the process of pool
formation. The exchange process may oc-

Tem-

Rate of

perature,
°C

Pool Formation

Exchange

0
25

0.0074
2.0

0.18
0.63

Miscellaneous properties of the amino
acid pool. As mentioned in the last annual
report, cells suspended in high-osmotic-
strength media have higher saturation
values of pool size. For example, when
proline (0.07 mg/ml) is supplied to cells
suspended in 1.3 osmolal medium (four
times the strength of the usual growth
medium), a pool of about 1000 uM/g dry
cells is observed. When 20 mg/ml of ca-
sein hydrolysate is supplied in the same
medium (1.4 osmolal), however, the total
pool for all the amino acids is only about
1000 uM/g dry cells. Thus, whereas small
amino acid pools are unaffected by the
presence of other amino acids (except for
certain interactions of similar amino acids,
such as valine, leucine, and isoleucine) the
large amino acid pools interact strongly,
and are clearly not held by mechanisms
specific for each amino acid.

122 CARNEGIE INSTITUTION OF WASHINGTON

In order to gain insight into the nature
of pool binding, experiments were carried
out to determine the temperature change
in a thick suspension of cells when a large
amino acid pool was released by osmotic
shock. The results indicated that very
little heat was released or absorbed during
the process. The experimental limit was
about 2 kcal/mole of amino acid pool.

Some exploratory studies have been
made on the reaction of hydroxylamine
with the pool amino acids of E. colt. Hy-
droxylamine reacts rapidly with acyl phos-
phates, anhydrides, and halides to form
hydroxamic acids; the corresponding car-
boxylates, amides, and peptides react very
slowly. If the pool amino acids were
present as activated forms such as acyl
phosphates, a fairly efficient conversion to
the corresponding amino hydroxamic acids
would be expected in the presence of high
concentrations of hydroxylamine. Experi-
ments to test this possibility show very
small yields of amino hydroxamic acids.
In the two cases which were tested with
great sensitivity, identifiable quantities of
leucine and tyrosine hydroxamic acids have
been observed. These quantities, however,
correspond to a very small fraction of the
total pool of these amino acids.

In "Mg-deficient" media the ability to
form amino acid pools in E. coli was
markedly reduced, as was the rate of incor-
poration into the TCA-precipitable frac-
tion. Reductions of as much as a factor of
5 in pool size have been observed, but the
results are quite variable, presumably as
a result of uncontrolled traces of Mg
present in the "Mg-deficient" media.

When E. coli cells are exposed for a few
minutes to a hydrostatic pressure of 20,000
psi, a large part of the amino acid pool
is released to the medium. Even though
growth of the cells is inhibited for an hour
or so after the pressure is removed, most
of the released pool is quickly reincorpo-
rated. The sensitivity of the pool to hydro-
static pressure presumably results from the
distortion of the structures holding the
amino acid pool, and is probably related

to the sensitivity of the pool to osmotic
shock.

Osmotic effects. Further studies have
been carried out on the effect of osmotic
strength on the amino acid pool in the
hope of obtaining insight into the mech-
anisms that hold the pool.

The pool size at low amino acid con-
centration is independent of the osmotic
strength of the medium. When a sudden
drop in osmotic strength occurs, however,
the pool is partly removed and after a
short time recovers to its previous value
if the amino acid concentration remains
constant. These results were reported in
last year's annual report without interpre-
tation. In combination with later evidence,
they lead to the following picture of the
process of osmotic shock.

The cell is initially in osmotic equilib-
rium with its normal medium. When the
osmotic strength is suddenly reduced there
is first a flow of water into the osmotically
sensitive structures along with a slow loss
of solute from the cell. The consequent
stretching of the structures due to the in-
ternal pressure increases the permeability
to the solute, allowing a faster rate of loss
of solute molecules. These two processes
together finally lead to a new osmotic
equilibrium. The dynamics of this process
are such that the cell passes through a
transient state in which the structures are
distended. The loss of pool is associated
with this loosening up. The structures
responsible for the holding of the pool may
be either the cell wall or membrane or
internal constituents normally constrained
by the cell wall.

Since competing rates of flow of water
and solute are involved, this picture sug-
gests that a slow change in the osmotic
strength might be less effective than a
rapid shock. Observation bears out the sug-
gestion. In an experiment with a large pro-
line pool a reduction in osmotic strength
of a factor of 3 was made in four steps of
equal concentration ratio. As a result, 70
per cent of the pool was removed. If the

DEPARTMENT OF TERRESTRIAL MAGNETISM 123

same final osmotic strength was achieved
in a single step, 90 per cent was removed.
Further, when the same final strength was
achieved through a concentration change,
graded continuously over several minutes,
only 50 per cent of the pool was removed.

Another implication of this description
of the process of osmotic shock is that
molecules which do not ordinarily enter
the osmotically sensitive structures might
be able to diffuse in during the transient
period when the permeability is increased.
In our laboratory jargon this process has
been called "trick or treat," since it was
suggested just after Halloween and brings
to mind the children's trick of throwing
in orange peels while the door is open.

This transient permeability was tested
experimentally by giving a thick suspen-
sion of cells at 0° C a sudden osmotic
shock in the presence of radioactive S04"
or PC>4=. The suspension was then diluted
without osmotic shock to remove diffusible
label, and filtered. A small amount of
SO/ or P04= was taken up, corresponding
to about 5 per cent of the cell volume at
the external concentration of SO/ or PO/.
Various controls showed that this uptake
was indeed due to the sudden downward
change in osmotic strength. Upward os-
motic shocks neither remove the pool nor
cause the trick-or-treat phenomenon.

This phenomenon may shed some light
on earlier observations of the paradoxical
conditions required for efficient produc-
tion of mutations with Mn++, as described
in the annual report of 1951-1952 (Year
Book 51). Effective production of muta-
tions occurred when cells suspended in
saline were transferred to low-osmotic-
strength medium containing Mn++.

Studies have been made of the osmotic
effectiveness of various solutes for protec-
tion of a preformed pool or its removal by
osmotic shock. The experiments were per-
formed by collecting on a filter cells con-
taining a C14 proline pool, and briefly
washing (5 seconds) with the following
types of solutions: (A) the solute at the

same osmotic strength as the growth me-
dium (0.37 osmolal) dissolved in growth
medium; (B) the solute at this osmotic
strength dissolved in water; and (C) solu-
tion as in A, followed by a wash with the
usual growth medium. Table 11 summa-

TABLE 11. Osmotic Effects of Various Solutes
on E. coli Proline Pool

Solutes

Percentage of Pool Removed

A,
upshock

Butanol 77

Diethylene glycol . . 30

Ethyl acetoacetate . . 50

Methanol 20

Ethanol 20

Propanol 20

Acetone 10

Propionamide 10

Succinimide 20

Acetamide 20

Dioxan 5

Glycerol 5

Urea 5

NaCl 0

NaAC 3

Diethylamine HC1 . . 6

Tris HC1 8

Glycine 1

Alanine 3

Valine 5

Proline 5

Glucosamine HC1 ... 0

Xylose 0

Glucose 0

Galactose 0

Sucrose 0

c,

B,

upshock

protec-

and

tion

down-

shock

100

77

70

95

90

50

100

20

100

20

100

20

100

15

100

20

100

20

100

50

95

40

99

70

97

63

38

55

60

50

35

60

35

55

16

47

12

50

12

60

30

60

5

50

10

52

10

52

10

45

5

70

A. Washed on filter with solute at 0.37 os-
molal in growth medium.

B. Washed on filter with solute at 0.37 os-
molal in H20.

C. As in A followed by wash with growth
medium.

rizes the results. Since, in general, upward
osmotic shock has no effect on the pool, A
simply measures possible chemical or de-
structive effects on the cells. Compounds
such as butanol remove most of the pool.
Washes of type B measure the effectiveness

124 CARNEGIE INSTITUTION OF WASHINGTON

of the solute as an osmotic protector for
short periods of time. It will be seen that,
in general, high-molecular-weight com-
pounds are most effective, although the
zwitterion glycine (M.W. 75) is an effec-
tive protector but glycerol (M.W. 92) is
not. It is presumed that solutes which
do not act as protectors are able to enter
the osmotically sensitive structures rapidly.
The resulting excess water activity outside
causes water to flow in until the distension
allows internal osmotic constituents to leak
out, re-establishing equilibrium.

Washes of type C measure the effect of
a sudden doubling of the osmotic strength
followed rapidly by a return to the usual
medium. Compounds like acetone which
enter the cell rapidly without causing
damage (as shown by the figures in col-
umns A and B) do not cause significant
removal by downward osmotic shock. On
the other hand, compounds like urea and
glycerol, which are not effective protectors,
are capable of removing the pool by os-
motic shock.

A sudden increase of the osmotic
strength of the medium causes a quite
different effect from a sudden decrease.
When a growing cell suspension is sud-
denly mixed with an equal volume of
medium containing 1 M sucrose, synthesis
of nucleic acid and protein stops. After
a period of roughly 8 minutes, the syn-
thetic activities resume; during this period
the metabolic pool materials continue to be
incorporated. Similar effects are observed
with glucose and sodium chloride, but
proline causes no major effect.

If the cells are washed with Tris me-
dium just after sucrose is added, there is
no loss of P32-labeled pool materials. At
the end of 8 minutes a Tris wash removes
half the pool. It appears that the sucrose
penetrates slowly into some structure of
the cell associated with holding the P32-
labeled pool. Before the sucrose has pen-
etrated, the structure is perhaps dehy-
drated, causing the cessation of synthesis.
After the sucrose has penetrated, the struc-

ture can be osmotically shocked by wash-
ing the cells with Tris. It is striking that
the period for this action is quite different
from the period required for sucrose to
penetrate the cell wall of lysozyme-treated
cells in the procedure for protoplast for-
mation. It may also be significant that
glucose has the same action, whereas it is
less effective in the protoplast procedure
(see below).

An observation of considerable interest
from the point of view both of the osmotic
properties of the cell and of the mechanism
of pool formation is that different types
of pools appear to have different sensitivity
to osmotic shock. Both amino acid pools
and P32-labeled pools are completely re-
moved by a quick water wash at room
temperature; at 0° C, however, while the
amino acid pools are still completely re-
moved, only half of the P32-labeled pool is
removed. Osmotic shocks (at room tem-
perature) that remove half of the amino
acid pool will remove considerably less of
the P32-labeled pool. It also appears that
when very large amino acid pools, formed
at high osmotic strength with casein hy-
drolysate, are partially removed by osmotic
shock, the amino acid distribution is con-
siderably altered.

These observations indicate that various
pool materials are organized in different
ways within the cell, and perhaps are asso-
ciated with different substructures.

Some studies have been made, by means
of freezing-point measurements, of the re-
lease of the total osmotically active material
of the cell by osmotic shock. The results
are broadly similar to those of studies of
amino acid pools. Water washes remove
the total osmotically active constituents al-
most completely. Boiling of the cells after
water washing releases only traces of addi-
tional material effective in depressing the
freezing point of water. When the osmotic
shock is performed in small steps, the re-
lease of the total osmotically active material
is similar to the release of amino acid pools,
though perhaps a somewhat greater per-

DEPARTMENT OF TERRESTRIAL MAGNETISM

125

centage is released for the same shock.
The total released from the cell indicates
that, if this material is osmotically active
when present in the cell, the osmotic pres-
sure within the cell is slightly greater than
that of the medium and is dependent on
the osmotic strength of the medium. If
this material were concentrated in regions
smaller than the whole cell, the osmotic
pressure would be proportionately higher.

Unusual behavior is exhibited by a thick
water suspension of water-washed cells. If
such a suspension is cooled somewhat be-
low 0° C (in a vigorously stirred freezing-
point cell), and an ice crystal is added to
seed the crystallization, the temperature
rises initially as in a typical freezing-point
experiment. The temperature rises higher
than 0° C, however, and continues to rise
slowly and steadily as heat is extracted.
When the temperature reaches about
+ 0.7° C, stirring of the mixture becomes
difficult. If the system is now removed
from the cooling bath and allowed to stand
(with stirring) in the air, the temperature
falls. The fall can be made more precipi-
tous by immersing in hot water. If the
system is returned to the cold bath (—2°
C), the temperature rises. This anomalous
behavior, an apparent negative heat ca-
pacity, is reversible provided that the tem-
perature is kept in the range of 0 to +0.7°
C. The cause of this apparent negative
heat capacity has not yet been determined.

Summary of the properties of the met-
abolic pool in E. coli. The properties of
the pool that must be incorporated in any
model may be summarized as follows.

1. Energy is required for formation of
the pool but not for maintenance of a
preformed pool.

2. Pools are formed very slowly at 0° C,
and preformed pools are maintained.

3. Exchange occurs independently of any
energy source. The temperature coefficient
of the exchange process is small; that for
formation is very large.

4. Small pools are not influenced by the
presence of other amino acids, but large
pools are suppressed.

5. When the external amino acid is re-
moved, the pool is lost at a much slower
rate than that at which it was initially
formed.

6. The initial rate of pool formation and
the pool size do not have the same de-
pendence on amino acid concentration.

7. Pools are removed by osmotic shock,
but not all pool compounds are removed
equally.

8. The saturation value of the pool is
roughly proportional to the osmotic
strength of the medium.

As yet no model incorporating all the
features necessary to provide for the eight
properties listed above has been worked
out in sufficiently critical detail to be tested
against the available data.

AMINO ACID POOLS IN YEAST

Kinetic studies have shown that in
Candida utilis pool formation is a pre-
liminary and necessary step in macro-
molecule formation. For nucleic acid syn-
thesis two chemically distinct and func-
tionally different purine pools are known.
The first, a concentrating pool, accumulates
nucleic acid bases within the cell at levels
exceeding their external concentrations.
This pool is evident only when the syn-
thetic medium is supplemented with bases.
The pool size is variable and dependent
upon external concentration. Once concen-
trated, these bases may provide material
for the second, a nucleotide pool, which
is always present and remains constant in
size during exponential growth. Here con-
version of one nucleotide to another oc-
curs, furnishing the appropriate molecules
for nucleic acid synthesis.

Evidence has accumulated indicating the
existence of two functionally distinct
amino acid pools (in yeast) analogous to
these pools of purine compounds.

Kinetic studies of pool formation from
exogenous amino acids. It has been shown
previously that during exponential growth
in media containing fructose as the sole
carbon source 13 per cent of the cellular
carbon is contained in an amino acid pool.

126 CARNEGIE INSTITUTION OF WASHINGTON

It is now known that the total amount of
amino acid pooled in the cell can be in-
creased by the addition of amino acids to
the growth medium.

Exogenous amino acids are quickly in-
corporated by exponentially growing cells.
Figure 23 shows the time course of incor-
poration of trace quantities of C14 glutamic
acid. The tracer "pulse" appears first in
the cold trichloroacetic acid (TCA) -soluble
fraction, and, as the small quantity of
exogenous amino acid becomes exhausted,
the transfer of pool radiocarbon (cold-
TCA-soluble fraction) to the protein frac-
tion is observed. These results resemble

Total

3-

•

s

Cold -TCA- preopitable fraction

£ 2'

1

1

1/ /

>< i-

,Cold-TCA-soluble traction

i

§

Q:

0-

Fig. 23. Time course of incorporation of tracer
quantities of C14 glutamic acid. At time = 0, the
exogenous concentration was 0.002 mg glutamic
acid/ml medium.

those obtained in the investigation of the
kinetics of incorporation of exogenous
purines, and demonstrate that the incor-
porated amino acid is contained in a met-
abolic pool which may supply amino acid
carbon for protein synthesis. Similar re-
sults have been obtained with a variety of
C14-labeled amino acids supplied to C.
utilis growing exponentially in C medium.
When concentrations of exogenous
amino acids are higher, the quantity of
amino acid contained in the cold-TCA-
soluble fraction increases and exceeds the
external concentration. Table 12 shows the
accumulation of C14 threonine in the cold-
TCA-soluble fraction with increasing con-
centrations of exogenous threonine. For
comparison, table 13 shows the steady-state
distribution of pool amino acid and protein
amino acid found in cells grown in C

medium when fructose was the sole carbon
source. These data demonstrate that the
size of the pool concentrated from exoge-

TABLE 12. Distribution of C14 Threonine

Carbon among Pool and Protein

Amino Acids*

C14

Threonine,

Concentration
of Exogenous

[iM/g dry weight cells f

C14 Threonine,

Cold-TCA-

Cold-TCA-

(aM/ml medium

Soluble

Precipitable

Fraction

Fraction

50.0

.... 442

520

8.4

.... 354

446

6.7

.... 268

328

5.0

.... 190

272

3.4

.... 104

149

1.7

.... 32

82

0.8

9

41

* Cells grown from light inoculum to about
2.8 mg wet weight of cells per ml medium.

t Calculated on the basis that all the radio-
carbon incorporated remained C14 threonine.

TABLE 13. Steady-State Distribution of Radio-
carbon among Pool and Protein Amino Acids*

Pool Protein

Component Quantity rf Quantity of

r Compound, Compound,

uM/g dry uM/g dry

Isoleucine-leucine 117 785

Lysine 24 625

Glutamic acid 290 640

Aspartic acid 9 762

Valine 65 512

Alanine 240 695

Threonine 8 455

Serine 8 600

Proline 7 287

Arginine 63 210

Glycine 108 488

Per cent accounted for. . 87 85

* Data obtained from cells growing exponen-
tially in C medium containing C14 fructose.

nous amino acids depends on the external
concentration and may greatly exceed the
pool formed endogenously from fructose.
Chromatographic examination of hy-
drolysates of the cold-TCA-soluble and

DEPARTMENT OF TERRESTRIAL MAGNETISM 127

precipitable fractions of cells grown at the
highest concentration of exogenous thre-
onine (50 uM/ml medium) showed that
more than 90 per cent of radiocarbon in
the TCA-soluble fraction was contained
in threonine, with the remainder in iso-
leucine. In the protein fraction, however,
the ratio of radioactivity in these two
amino acids was about one-to-one.

Amino acid pool characteristics. The
size of pool formed from fructose carbon
is constant. Furthermore, the addition of
high concentrations of exogenous amino
acid produces no immediate dilution or ex-
change with the existing pool. The trans-
fer to protein of these pool amino acids is
uninterrupted by the addition of the amino
acid supplements. On the other hand, the
exogenous amino acids are rapidly incor-
porated by the cell (fig. 23) and accumu-
lated to levels exceeding their external con-
centration. Thus, there are two pools of
amino acids, one formed from fructose, the
other from exogenous amino acids; for
convenience, they will be called the in-
ternal and expandable pools of amino
acids, respectively.

The amino acids of the internal pool are
not lost to the medium during exponential
growth, nor are they exchanged with
exogenous amino acids. In the expandable
pool, however, the situation is quite differ-
ent, as can be seen from the results of ex-
periments on cells grown in C medium
containing fructose and supplemented by
a high concentration of C14-arginine (1.0
mg/ml medium). The kinetics of incor-
poration of radiocarbon into pool and
protein, followed during this labeling
growth period, showed that approximately
two-thirds of the protein arginine was be-
ing derived from fructose carbon and one-
third from the amino acid supplement.
After 3l/2 hours of growth in this medium
the cells were harvested and washed twice
with C medium, and an aliquot was trans-
ferred to C medium containing nonradio-
active arginine (1.0 mg/ml, culture A).
Since the quantity of amino acid incor-
porated during growth in the labeled me-

dium was quite small in comparison with
the total quantity present, the cells were
exposed to an essentially constant environ-
ment of exogenous arginine. Another
aliquot was added to C medium containing
no arginine (culture B). At the time of
transfer, 57 per cent of the incorporated
radiocarbon was contained in the cold-
TCA-soluble fraction. Optical-density
measurements indicated that growth pro-
ceeded after the transfer with no delay in
the unsupplemented culture (B) and with

80 r

60 80 100

Time , minutes

Fig. 24. Upper curves = loss of radioarginine
from cells after transfer to nonradioactive C
medium containing 1.0 mg C12 arginine/ml
{A) or unsupplemented C medium (B). Lower
curves = incorporation of radioarginine carbon
into protein fraction of C. utilis after transfer
of labeled cells to nonradioactive C medium
(A and B as above).

only a slight delay (10 min) in the other
(A).

Both groups of cells lose radiocarbon to
the medium, as is shown in figure 24, but
the loss in B (no supplement) is less rapid
than in A. In B, reincorporation of the
undiluted radioactivity becomes evident in
about an hour. These results may be inter-
preted as follows. Where no supplement
was added, the expandable pool decreases
for two reasons : there is a continuous flow

128 CARNEGIE INSTITUTION OF WASHINGTON

into the internal pool and thence into pro-
tein, and there is an approach to a new
equilibrium with the medium. When the
expandable pool has reached a small
enough value through these two processes,
and the external concentration has reached
a large enough value, incorporation of
radiocarbon into protein again increases.
In experiment A, the loss of expandable-
pool radiocarbon is due to transfer through
the internal pool into protein, and to ex-
change with the nonradioactive arginine in
the medium. This exchange process is
faster than the loss of pool material to the
medium in B. These results are similar to
those obtained with E. coli discussed ear-
lier in this report. The continuous and
large dilution in experiment A results in a
steadily decreasing specific radioactivity of
both pools. Consequently, there is a stead-
ily decreasing rate of radiocarbon incor-
poration into the protein. The protein-
incorporation curve can be quantitatively
predicted on the assumption that the cell
processes as measured in the S^-hour pre-
labeling continued (except for the slight
lag), and the only change in circumstances
was the introduction of the exchange of
radioactivity between the pool and me-
dium.

The rate of incorporation of radiocarbon
into protein in the case of the unsupple-
mented medium (B) is not immediately
changed by the transfer, but it falls as
radiocarbon from the expandable pool is
lost to the medium. Apparently the pro-
portion of the flow of arginine carbon
through the internal pool, which is derived
from the expandable pool (as contrasted to
that derived from fructose), is a function
of the size of the expandable pool. As the
flow from the expandable pool decreases,
the specific radioactivity of the internal
pool also decreases, and there is a conse-
quent reduction in the rate of incorpora-
tion of radiocarbon into protein.

In figure 25 the logarithms of the radio-
activity in the cold-TCA-soluble fraction
in the two experiments described above are
plotted against the time after the transfer

to the nonradioactive medium. In both
experiments the data can be approximated
by a pair of straight lines. In A the early,
fast component presumably represents the
combined effects of exchange and transfer
to protein via the internal pool, whereas
the slow component reflects the decrease in
specific radioactivity of the internal pool.
Since the internal pool is constant in size
and undergoes no exchange, the rate of
dilution of its specific radioactivity is
limited by the rate of protein synthesis.
In B, the slower component again reflects

50 1-

Cold-TCA- soluble fraction

40 60 80 100

Time, minutes

140

Fig. 25. Loss of pool radioarginine after
transfer of labeled cells to nonradioactive C
medium {A and B as in fig. 24) .

the transfer of internal pool carbon to pro-
tein while the faster component includes
the approach to the new equilibrium be-
tween the expandable pool and the
medium.

Effect of hydrostatic pressure. Studies
have been made of the effect of pressure
treatment on the stability of the amino
acid pools in yeast. Exponentially growing
cells were labeled by a 2-minute immer-
sion in C medium containing carrier-free
C14 fructose. The cells were washed by
centrifugation in nonradioactive C me-
dium, and then resuspended in C medium
without fructose. These cells contained
87 per cent of the incorporated radiocarbon
in the TCA-soluble fraction.

DEPARTMENT OF TERRESTRIAL MAGNETISM

129

An aliquot of these cells was compressed
in a pressure cell at 30,000 psi. This pres-
sure was rapidly applied and released 10
times over a 5-minute interval. C12 fructose
was added to the treated culture, the cells
were aerated at 30° C, and the growth and
fate of the labeled pool carbon were meas-
ured. Figure 26 shows that 64 per cent
of the incorporated radiocarbon was lost
from the cells to the medium during the
first 90 minutes of aeration. The remain-

TIME IN MINUTES

Fig. 26. Loss of radiocarbon from cells (up-
per curve) and metabolic pools (lower curve)
after cells were subjected to a pressure of 30,000
psi.

ing pool radiocarbon was transferred to
protein or nucleic acid. No increase in
optical density was observed until after
100 minutes, when growth resumed, and
the cells were soon growing at the optimal
rate.

A control culture of these labeled cells
not subjected to pressure continued to grow
exponentially upon addition of fructose
and lost only a few per cent of incorporated
radiocarbon to the medium during the 90-
minute interval.

Simple rupture of cellular membranes
cannot be responsible for the release of

internal pool amino acids to the medium.
No loss of the pool is observed after the
pressure treatment unless fructose is added
to the culture. Thus, both energy and time
are required for the release of pool ma-
terials. Furthermore, microscopic examina-
tion during the course of the experiment
showed no cell fragments or visible cellular
alterations.

One explanation of the observed results
is that the pressure treatment disrupts the
pool organization and reorganization must
occur before growth resumes. Evidence
supporting this conclusion was obtained

20 40 60 80 100 120 140 160

TIME IN MINUTES

Fig. 27. Kinetics of incorporation of C14 fruc-
tose in C. utilis after cells were subjected to a
pressure of 30,000 psi.

in another experiment by following the
kinetics of C14 fructose incorporation im-
mediately after the pressure treatment.
Figure 27 shows that rapid incorporation
of fructose carbon occurs. Initially this in-
corporation is due to pool reconstruction,
since most of the incorporated radiocarbon
appears in the TCA-soluble fraction. The
pool appears to reach saturation in 90 min-
utes, at which time the quantity of radio-
carbon contained in the TCA-soluble frac-
tion corresponded to two-thirds of the nor-
mal steady-state value of the internal pool.
This incorporation is approximately equal
to the quantity of pool material lost from
the cells (fig. 26) and therefore represents
replenishment of the pool. No increase in
the optical density of the cells was ob-
served until 120 minutes after the addition
of the fructose.

130 CARNEGIE INSTITUTION OF WASHINGTON

The effect of osmotic shoc\. The in-
ternal pool of amino acids is relatively
insensitive to osmotic shock. C14-fructose
labeled cells lose from 3 to 8 per cent of
their internal pool material when sus-
pended for half an hour in distilled water.
About the same loss is observed when
these labeled cells are transferred to non-
radioactive C medium. The amino acids
in the expandable pool, on the other hand,
behave quite differently. Cells were grown
for several hours in medium containing
6 uM C14 threonine/ml, then harvested
and washed twice in unsupplemented me-
dium. An aliquot was chemically frac-
tionated, and the quantity of C14-threonine
carbon found in the TCA-soluble fraction

ably due to a distortion of the structures
binding the pool.

Conclusions. Two functionally distinct
amino acid pools exist in C. utilis. The
major characteristics of these two amino
acid pools are compared in table 14.

The fact that these two pools display
such different characteristics and do not
rapidly equilibrate with each other in the
cell indicates that they are physicochemi-
cally distinct. It seems unlikely that phos-
phorylated forms could account for this
difference, since there is insufficient pool
phosphorus (300 uM/g dry weight cells)
available even for the 1000 uM of amino
acids in the internal pool. The expandable
pool can be even larger than the internal

TABLE 14.

Major Characteristics of Amino Acid Pools in Candida utilis

Characteristic

Expandable Pool Internal Pool

Function Concentrates exogenously supplied ami-

no acids.

Size Variable and dependent upon exoge-

nous amino acids concentration.

Stability Sensitive to osmotic shock. Exchanges

with exogenous amino acids.

Interconverts and selects amino acids
for protein incorporation.
Fixed.

Insensitive to osmotic shock. Not ex-
changeable with exogenous amino acids.

corresponded to 56 uM/g dry weight of
cells. This value is seven times the con-
centration found for threonine in the in-
ternal pool (see table 13). When another
aliquot of these labeled cells was suspended
in water the cells immediately lost 50 per
cent of the total pool radiocarbon to the
water.

The internal pool of amino acids in cells
subjected to a pressure of 30,000 psi be-
comes very sensitive to osmotic shock.
Cells labeled with a 5-minute immersion
in medium containing C14 fructose were
harvested and found to have 80 per cent
of the incorporated carbon in the TCA-
soluble fraction. An aliquot of these cells
after the pressure treatment lost 31 per cent
of the internal pool carbon when sus-
pended in distilled water. A control sus-
pension (no pressure) lost 6 per cent to
the water wash. The sensitivity of the in-
ternal pool to hydrostatic pressure is prob-

pool, but no additional incorporation of
P32 occurs when large quantities of exoge-
nous amino acids are rapidly accumulated
in the expandable pool. Thus, neither the
internal nor the expandable pool can be
explained on the basis of phosphorylation
of the amino acids. Some correlation be-
tween phosphorus turnover and amino
acid accumulation would be expected if
phosphorus were associated with these
amino acids.

It has been suggested that the internal
pool is complexed with macromolecular
components of the cell. The only class of
substances present in sufficient quantity to
accommodate the internal pool amino acids
are the cellular proteins. A one-to-one
complex between protein amino acids and
pool amino acids would require that 25
per cent of the protein be so involved.

From the known molar distribution of
internal pool and nucleic acid carbon there

DEPARTMENT OF TERRESTRIAL MAGNETISM

131

would be 3.3 times too many amino acids
for a one-to-one correlation between pool
amino acid and nucleotide residues. A
mixture of protein and nucleic acid could,
of course, serve as a complexing system.

It seems likely that macromolecules, most
probably the proteins, are the sites of asso-
ciation for the internal pool amino acids.
These amino acids are on the main line
of protein synthesis, and are more tightly
bound than those in the expandable pool.
The internal pool is more closely linked to
the mechanisms of final protein incorpora-
tion, for it is here that amino acid inter-
conversions and selection for protein incor-
poration occur.

In the expandable pool the amino acids
accumulated do not interconvert, as can be
seen from the data of table 12. The thre-
onine concentrated by the cells grown on
the highest exogenous threonine concentra-
tion remained, for the most part, as pool
threonine, but a small portion was con-
verted to pool isoleucine. This isoleucine
and a small quantity of the total pool thre-
onine are probably components of the in-
ternal amino acid pool. The demonstration
in this experiment that protein threonine
and isoleucine both become about equally
radioactive even though there is a large
disparity in their pool concentrations re-
inforces this conclusion. Figure 28 is a
schematic diagram of the flow of carbon
from exogenous amino acids and from
fructose, via the metabolic pools, to pro-
teins as interpreted from these studies.

These kinetic investigations describing
the flow of exogenous carbon through
metabolic pools and hence into protein
provide a more definite picture of some of
the preliminary steps in protein formation.
In addition, they reveal that the amino
acid composition of the cell may vary,
reflecting the kind and quantity of exoge-
nous molecules. This altered composition,
in turn, affects the endogenous flow of
fructose carbon which alone, in the ab-
sence of other exogenous organic mole-
cules, is a satisfactory carbon source. Such
studies demonstrate the great capacity of

some living cells to utilize ever-changing
environments economically.

PROTOPLASTS

The effects of lysozyme on the structure
and function of E. coli. In order to study
the macromolecular composition of the
bacterial cell by physicochemical methods,
such as ion exchange and ultracentrifuga-
tion, it is necessary to disrupt the cell and
release its contents. The use of the enzyme
lysozyme, together with osmotic shock,
has proved an efficient way of breaking

Exogenous -
threonine «-

Fructose -

Cold -TCA- soluble
fraction

Fig. 28. Flow of threonine carbon through
metabolic pools in the synthesis of protein.

E. coli. The conditions for lysis and physi-
cal studies of the lysates are described be-
low. Lysozyme and osmotic shock may
also be used to inflict controlled damage
upon E. coli and thus provide altered cells,
spherical forms which are usually called
"protoplasts." Study of these forms should
prove helpful in evaluating what struc-
tures of the cell must remain intact in
order for macromolecule synthesis to occur
at normal rates. Metabolic studies on the
spherical forms are described in the follow-
ing paragraphs.

Lysis due to lysozyme treatment and
osmotic shoc\. Cultures of E. coli growing
or resting in various media (C medium,
C medium diluted with 9 volumes of
water, or C medium containing 18 per cent
sucrose) show no response to added lyso-

132 CARNEGIE INSTITUTION OF WASHINGTON

zyme (100 ug/ml). They neither lyse, complete and rapid lysis. On the other

show an impaired growth rate (15°, 23°, hand, when the osmotic pressure of a lyso-

or 37° C), nor exhibit morphological zyme-sucrose suspension of cells was slowly

change visible in the phase-contrast or lowered by the dropwise addition of dis-

dark-field microscopes. When such cul- tilled water, lysis failed to occur. Such

tures are harvested, washed free of excess "decompressed" cells were morphologically

lysozyme by centrifugation, suspended in and metabolically indistinguishable from

0.5 M sucrose, and suddenly diluted with untreated bacteria.

10 volumes of distilled water, the bacteria From these results it was evident that
continue to maintain their normal rodlike the presence of lysozyme and a consider-
structure and upon subculture grow as well able osmotic shoc\ were necessary to cause
as the corresponding control (lysozyme- lysis. Since osmotic shock results from a
omitted) cultures. It appears that under relatively high intracellular concentration
these conditions lysozyme does not seri- 0f solute, the opportunity was afforded to
ously damage E. coli. If, however, lyso- test the rate at which various solutes en-
zyme is injected into a suspension of cells tered and left the cell. When E. coli were
in 0.5 M sucrose a few seconds before sud- suspended in 0.5 M sucrose-lysozyme solu-
den dilution with 10 volumes of water, tion for varying lengths of time and sud-
nearly complete lysis occurs and a highly denly diluted, complete lysis occurred only
viscous solution containing numerous free after 30 to 60 seconds' equilibration of the
cell walls results. An electron micrograph cells with the suspending medium. Al-
of several of these cells walls is shown in ternatively, when bacteria that had been
figure 29, plate 3. Ultracentrifugal analysis equilibrated with a 0.5 M sucrose solution
of the viscous solution shows a pattern of were first suddenly diluted and then lyso-
sedimenting components typical of those zyme was added after increasing intervals,
observed in cell juices of E. coli. A sedi- lysis occurred only in those cells that had
mentation diagram is shown in figure 30, less than 1 minute's exposure to the dilute
plate 4. medium. Thus, enough sucrose to cause
The influence of molecular size of the osmotic lysis can enter or leave E. coli
solute particle on the ability of E. coli to cells in about a minute,
lyse was tested by suspending cells in Amino acid utilization by E. coli proto-
various solutions isotonic with 0.5 M plasts. E. coli protoplasts may be efficiently
sucrose, adding lysozyme, and diluting the prepared from exponentially growing cul-
suspensions suddenly with 10 volumes of tures by suspending the cells in 0.5 M
water. The degree of lysis was found to sucrose in C or Tris medium, adding lyso-
increase with increasing molecular weight Zyme (100 ug/ml), and suddenly diluting
of the solute particle. Thus, inorganic salts the suspension with an equal volume of
(NaCl, Na2S04, C medium salts) and water. In about 20 minutes at 25° C (10
low-molecular-weight organic compounds min at 37° C) the suspension consists of
(urea, glycerol, Tris, sodium succinate, 85 to 95 per cent protoplasts. Under these
sodium glutamate, casamino acids) caused conditions protoplast formation requires
no change in the suspension of cells. With considerable time and is a temperature-
glycine, however, lysis occurred slowly, sensitive process. The walls of these forms
Xylose, sorbitol, mannitol, inositol, and enclose volumes from 2 to 50 times those
glucose did not allow lysis, but the marked of the rodlike forms. The protoplasmic
streaming birefringence of the suspensions body within the wall may occupy the
was much reduced, and observation of the entire volume of the protoplast, may fre-
bacteria in the dark-field microscope re- quently be observed in a doubled condition
vealed that 50 to 100 per cent of the cells within a single outer wall, or may adhere
had assumed a spherical shape. Lactose, to one side of the wall, yielding a saucer-
sucrose, maltose, and raffinose allowed shaped inner structure. Since such proto-

DEPARTMENT OF TERRESTRIAL MAGNETISM

133

plasts are morphologically altered E. colt,
it was pertinent to inquire whether the
structural changes would influence the
maintenance and formation of amino acid
pools and the synthesis of protein.

When growing E. coli were allowed to
metabolize C14 proline in the presence or
absence of lysozyme, no differences in the
utilization of the amino acid could be de-
tected. Equal C14-proline pools were
formed, and protein synthesis continued,
as shown in figure 31. Lysozyme-treated

• r Lysozyme treated

600-

0= No Lysozyme

Total -^__^ y.

400-

•/

-

Or

•TCA Soluble

•

200-

4*

^

^O

^TCA

ppt.

c

1 1 1

5 10 15 ;

1
>0

1

25

1

30

Minutes after adding C14- proline

Fig. 31. C14-proline utilization by cells grow-
ing in the presence of 100 ug/ml lysozyme.

cultures grow indefinitely at the rate char-
acteristic for E. coli. Thus, lysozyme treat-
ment in itself had no apparent effect upon
amino acid utilization or growth of E. coli.

Cells that had previously formed a C14-
proline pool and synthesized protein con-
tinued to maintain a large fraction of the
original pool, increased the size of the pool,
and continued to synthesize protein at a
high rate even though they were undergo-
ing the morphological change, rod-to-
sphere, which followed lysozyme-sucrose
osmotic-shock treatment. These findings
are illustrated in figure 32.

The results of a more stringent test of
the capacity of E. coli protoplasts to form
an amino acid pool and to synthesize pro-
tein are shown in figure 33. For this ex-
periment a protoplast suspension contain-
ing 96 per cent spherical forms was
prepared and allowed to utilize C14 proline.
It is evident that an amino acid pool is
formed and that protein is synthesized.
It was also found that washing the proto-

plasts on membrane filters with isotonic
media removed the labeled amino acid pool
without causing lysis. Thus the amino
acid pool is held so loosely that even me-
chanical mishandling of the protoplasts
causes its loss,

• = Lysozyme treated
0= Control

10 20 30 40

Minutes after osmotic shock

Fig. 32. C14-proline utilization by E. coli after
osmotic shock in the presence of lysozyme (solid
circles). Percentages indicate the proportion of
cells that were spherical as determined by direct
observation in the phase-contrast microscope.
Lysozyme was omitted in the control culture
(open circles) . The larger pool size for cells that
received no shock or lysozyme treatment is shown
by the dotted line.

Minutes after addmcj C - proline

Fig. 33. C14-proline utilization by a suspen-
sion of cells containing 96 per cent spherical
forms (solid lines). Washing with isotonic
medium removes the pool but does not lyse the
cells (open circles).

FRACTIONATION OF CELL JUICES ON ION-
EXCHANGE COLUMNS

Proteins and nucleic acids of E. coli have
been separated and characterized by means
of cellulose ion exchangers.1 Small amounts

1 For preparation of the exchangers see E. A.
Peterson and H. A. Sober, /. Am. Chem. Soc,

134 CARNEGIE INSTITUTION OF WASHINGTON

of TCA-soluble bacterial amino acids, glutamic acid, leucine, and isoleucine for
which occur in apparent association with the purpose of (a) tracing the behavior of
high-molecular-weight components of cell any free amino acid of bacterial origin, and
juice, have also been isolated by means of (b) distinguishing bacterial amino acids
the exchanger. Various proteins having that were not free to equilibrate with the
different enzyme activities have been sepa- C14-labeled substances under these condi-
rated. (3-Galactosidase activity has been tions. This radioactive mixture was im-
found distributed among different proteins mediately passed into a column of the
separated by the exchanger. Bacterial anion exchanger DEAE cellulose (di-
nucleic acid has been separated from bac- ethylaminoethyl cellulose),
terial protein, and the nucleic acid has been Washing the column with several col-
further fractionated. umn volumes of 0.25 M sucrose in Tris

Association of bacterial amino acids with medium completely removes the C14-la-

high-molecular-w eight components of E. beled amino acids, the bulk of the bacterial

coli. It has become evident from studies amino acid, and a small fraction of the

on protoplasts that the pool holding ca- protein and nucleic acid. Elution with 1 M

pacity of the E. coli cell lies beneath the NaCl dissolved in the sucrose-Tris medium

cell wall. Normal pools are still formed releases nearly all the remaining protein

when the cell wall has been fundamentally and nucleic acid from the column. When

altered by lysozyme and osmotic shock, the latter substances and salts are removed

Unfortunately, the inner architecture of from the eluate by TCA treatment and ion

the cell is only crudely known. There is exchange, amino acids can be recovered

not yet completely satisfying evidence of and identified by paper chromatography

the physical reality of a plasma membrane (fig. 34) . These amino acids are found

or of a nucleus. On the other hand, it is to be nonradioactive. Thus, they are clearly

clear that much of the substance of the cell of bacterial origin but differ in their be-

is organized into macromolecules or com- havior on the ion exchanger from the bulk

plexes of macromolecules such as the 20 S of the TCA-soluble bacterial amino acids

and 40 S ribonucleoprotein particles; see and also from the admixed C14-tracer

figure 30. An attempt was made, there- amino acids. In quantity they correspond

fore, to learn whether pool amino acids to a few per cent of the pool amino acids

could be found in association with macro- originally present in the cells. A definite

molecular components. tie or association between these amino

For this purpose the experimental pro- acids and macromolecules seems the most

cedure was as follows. Exponentially probable explanation for this behavior,

growing E. coli were suspended in 0.5 M Depending on the criterion employed,

sucrose dissolved in Tris medium, and these associated amino acids resemble or

were disintegrated with a modified French differ from the metabolic pool amino acids

pressure cell. The resulting suspension was 0f E. coli. For example, like the metabolic

centrifuged, the pellet was discarded, and pool amino acids, they cannot be recovered

the supernatant fluid was filtered through from cells that have been osmotically

collodion membranes to yield a clear am- shocked by washing with distilled water,

ber cell juice. This juice contains all the Alternatively, when cell juice is prepared

material of the original suspension except in 0.5 M sucrose solution lacking salts, and

the cell walls and a small amount of un- then passed over the ion exchanger, the

broken cells. To the juice was added a free neutral amino acids pass through the

solution of carrier-free C14-labeled alanine, column. The anionic amino acids, glu-

^n wi /meA \ 7- • \ ~ tamic and aspartic acids, and also a small

78.751 (1956); for application to plasma protein r L t . -i ri • i

separations, see H. A Sober et al., /. Am.Chem. amount of neutral amino acids of bacterial

Soc, 78, 756 (1956). origin, however, remain adsorbed. These

DEPARTMENT OF TERRESTRIAL MAGNETISM 135

neutral amino acids can be released by
washing the column with distilled water
and be identified chromatographically
(fig. 35) . Proteins and nucleic acids are not
released by this treatment. Thus, the ion-
exchanger-macromolecule complex holds
the associated amino acids in a labile fash-
ion reminiscent of the way in which the
metabolic pool amino acids are held in the
cell. On the other hand, the macromole-
cule-associated amino acids differed in that
they could not be labeled by exchange with

Cell juice +
Cl4-qlu,-ola,-L +11

cellulose ion-exchange columns. Several
examples of the separation of E. coli pro-
teins and nucleic acids are shown below.
Figure 36 illustrates the separation of S35-
labeled components of pressure cell juice
prepared from cells grown to a high den-
sity from a small inoculum in the presence
of S3504=. The effluent volume is given on
the abscissa, and the amount of radioac-
tivity (upper line) or amount of protein
(lower line, chemically determined with
the Folin phenol reagent), on the ordinate.
It is evident that the radioactivity and the

DEAE— Cellulose
onion exchanger

/4 M SUCROSE
IN
TRIS

1

1

1/4M

+ IM

S-T

NoCI

Ao'

OG

J O

L+1L

Fig. 34. Ion-exchange separation of bacterial
amino acids from admixed C14 amino acids.
Solid circles indicate the location of labeled amino
acids on two-dimensional paper chromatograms ;
open circles indicate unlabeled amino acids.
A = alanine, G=glutamic acid, V = valine; L-f-IL
= leucine, isoleucine.

C14 amino acids, by using either cell juices,,
as described above, or undamaged cells
at 0° C.

It is not yet clear what role the macro-
molecule-associated amino acids play in
the metabolism of the cell. Although they
constitute only a few per cent of all the
TCA-soluble amino acids in the cell, they
may represent a transitional phase lying
between the metabolic pool amino acids
and the protein end product.

Protein and nucleic acid fractionation on

W'.'.'.W

(*-SUC ROSE -4*DISTILLED*|«- SUC ROSE-*]
WATER + TRIS

• o
o o

• o

GLU
ALA
VAL
L+IL

= Radiooctive

O - Non radioactive

Fig. 35. Separation on DEAE cellulose of bac-
terial amino acids from C14-labeled admixed
amino acids by means of distilled-water wash-
ing. The amount of C14 eluted is plotted against
the volume of eluent. Fractions shown as shaded
areas at the centers of each region were chromato-
graphed on paper in one dimension. The result
is depicted immediately below the corresponding
shaded area.

protein of the cell can be resolved into
several components. The largest S35 peak
is most probably due to the peptide gluta-
thione; the remaining peaks are due to
jproteins. Comparison of the specific radio-
activities for several of the components of
the diagram demonstrates that the com-
ponents differ in the amount of sulfur that
they contain relative to each chemical unit
(principally the amino acid tyrosine)
which reacts with the phenol reagent.

136 CARNEGIE INSTITUTION OF WASHINGTON

Thus, different proteins are separated. Ion-
exchange separations of the proteins o£
cells exposed to S3504= for only a few
minutes reveal several regions of even
higher and several regions of lower specific
radioactivity. Such a finding bears the
implication that the sulfur of some kinds
of proteins passes into other proteins. The
ion exchanger, thus, provides a tool for the
study of possible generic relationships
among the bacterial proteins.

a separation is shown in figure 39. There
is little correlation between the nucleic
acid (ultraviolet absorption at 260 mu)
distribution and the protein distribution
(S35 content). Since ultracentrifuge and
electron-microscope studies have shown
that much of the ribonucleic acid is found
in spherical particles of about 1 million
molecular weight it appears that these
cellulose ion exchangers, as we have em-
ployed them, destroy the original nucleo-

50

Effluent volume (ml)

Fig. 36. Elution diagram of S35-labeled components and proteins of E. coli. One-milliliter efflu-
ent samples were analyzed for each region. The elution schedule was as follows:

a = 05 M sucrose

b=

a (

+ 0.05 M Tris-

succinate

pU 7.7

c =

(<

+ 0.1

M

«

(< <<

d=

((

+

<

«

pH 7.0

e =

((

+

<

<<

« <«

+ 0.05 M NaCl

f=

« <

+

<

<<

<< n

+ 0.1 M "

ff=

(< 1

+

<

(<

a a

+ 0.2 M "

h =

u

+

<

M

cc ((

+ 0.4 M "

Figure 37 illustrates the separation of
two enzyme activities, acetokinase and
(3-galactosidase.

Higher resolution among the protein
components can be achieved by eluting the
column with a smoothly increasing con-
centration of salt as shown in figure 38.
In this run, activity corresponding to the
adaptive enzyme, 3-galactosidase, was
found in a well defined region of 120 to
150 ml effluent volume.

Nucleic acids may also be separated
from one another and from protein by
means of cellulose ion exchangers. Such

protein associations. It should be possible,
therefore, to use the ion exchangers to de-
termine the complexity of subcellular par-
ticulates such as the microsomal particles.

INCORPORATION OF AMINO ACID ANA-
LOGUES INTO PROTEINS

Through collaboration with Dr. Georges
N. Cohen, of the Institut Pasteur, it has
been shown that selenomethionine can
completely replace methionine for expo-
nential growth of a methionine-requiring
mutant of E. coli. Active (3-galactosidase
is found under this condition. Experiments

DEPARTMENT OF TERRESTRIAL MAGNETISM

137

are in progress to find whether this
(3-galactosidase has the same specific mo-
lecular activity as the normal enzyme.

The demonstration that exponential
growth occurs when selenomethionine re-
places protein methionine reflects the syn-
thesis of all the essential enzymes. Work-
ers at the Institut Pasteur have found that,

20
Effluent volume (ml)

Fig. 37. Ion-exchange separation of acetoki-
nase (A) and (3-galactosidase (B). The total
protein eluted is shown as a continuous line.
The elution schedule followed was:

with other structural analogues of amino
acids (p-fluorophenylalanine, (3-2-thienyl-
alanine, norleucine), linear growth, instead
of exponential growth, always occurred,
accompanied by the suppression of the syn-
thesis of one or more active essential
enzymes. In each case the analogues were
incorporated into the proteins, replacing
structurally related natural amino acids.
These results and the selenomethionine
data demonstrate that the amino acid com-
position of protein can be altered by the
presence of amino acid analogues in the
medium. Since the incorporation of amino
acid analogues into protein results either
in the synthesis of fully active enzymes or
in the suppression of synthesis of some
enzymes, the analogues become powerful
tools in the study of the protein-forming
mechanisms.

Incorporation of selenium into proteins.
While selenomethionine can completely
replace methionine, selenium cannot en-
tirely replace sulfur for growth of E. coli;
but, when the cell contains glutathione,
which can serve as an internal sulfur reser-
voir, or when traces of exogenous sulfate
added, selenium partially replaces

are

a = 0.02 MPO,

b=

c =

d=

e =

f=

+ 0.005 M cysteine + 0.0 MNaCl
0.1 M "
0.2 M "
0.3 M "
0.4 M "
0.6 M "

sulfur. Hydrolysates of proteins obtained
from cells grown in media containing
radioactive selenite and trace quantities of
sulfate are found to contain a radioactive
material with chromatographic properties
similar to those of cysteine. The incor-

/3-Goloctosidase
H 8 octivity

80 100 120

Effluent volume,(ml)

140

160

180

200

Fig. 38. Elution diagram of E. coli protein on DEAE-cellulose column. A linear salt gradient
was used as shown on the broken line. The protein pattern (continuous line) is very complex,
and (3-galactosidase (insert) has been resolved into three apparent components.

138

CARNEGIE INSTITUTION OF WASHINGTON

poration of selenium is proportional to the
increase in bacterial mass. Selenium is also
incorporated into the alcohol-soluble pro-
teins, but, unlike the sulfur of this protein,
cannot be transferred from it to the resid-
ual proteins during sulfur starvation.

A possible reason for the inability of
selenium to totally replace sulfur for the
synthesis of proteins might be the inability
of selenocysteine to form Se-S or Se-Se
bridges, or the incompatibility between
proteins containing such linkages and a

diameter) is the most prominent. A sec-
ond group, which is sometimes resolved
into two groups, has roughly one-half this
value.

As it appeared possible that one group
of particles might be the product of the
other, an attempt was made to observe
kinetic relationships among the groups.
Cells were labeled with S35 and P32 during
a long exposure to the tracer to give a
steady-state condition, or during a short
exposure to place the tracer materials in

200

- .6

- .4 c

- .2

100
Effluent volume (ml)

200

Fig. 39. Elution diagram of nucleic acid (O.D. at 260 mu) and protein (S35 radioactivity) on an
Ecteola-cellulose column. A linear salt concentration gradient was used for elution. The occurrence
of (3-galactosidase is shown by the shaded area.

normal catalytic activity. It is also possible
that the responsible enzymes of E. coli are
incapable of transforming selenocysteine
to selenomethionine. This possibility is
strengthened by the finding that seleno-
glutathione is not formed from selenite.
Perhaps some of the biosynthetic enzymes
might thus show a more strict specificity
than the protein-synthesizing mechanism.

PARTICLES

Various lines of evidence point to the
importance of small particles in the syn-
thetic activities of the cell. Sedimentation
analysis shows several groups of particles
in the juice of E. coli. In most of the pub-
lished work a group having a sedimenta-
tion constant of 40 S (corresponding to a
molecular weight of about 1 million, 200-A

the early product of synthesis. The cells
were then broken, and the unbroken cells
and cell walls were centrifuged out. The
supernatant fluid containing the small par-
ticles was then centrifuged at 105,000g
for various periods of time. The cellular
material was separated into cell walls,
particles, and nonsedimentable material,
but no separation of the groups of particles
was achieved. Different runs gave quite
different rates of sedimentation; the parti-
cles seem very sensitive to the composition
of the suspending fluid.

It did appear that, when the tracer was
supplied for a short time only, the specific
radioactivity of the fraction containing the
particles was higher than that of the other
fractions. Also, chromatography of the
particle fraction showed that a small but

DEPARTMENT OF TERRESTRIAL MAGNETISM 139

definite portion of the amino acid pool is
carried down with the particles. Further
work is planned for next year to establish
the kinetic relations between the particles
and the other parts of the cell.

Model

During the year we have attempted to
arrange the information about the activities
of the cell in terms of definite models.
Even if they are completely wrong, models
are of some value, as they suggest numer-
ous experiments. The models attempt to
interpret two known features of the cell's
operation : the role of particles in synthesis,
and the role of organization in the cell.

A large fraction of the cell mass is found
in the form of ribonucleoprotein particles.
Most of the RNA of the cell and a con-
siderable fraction of the protein can be
centrifuged down; indeed, it seems quite
possible that most of the protein and RNA
of the cell is associated with particles,
some of the protein being tightly bound
and some being released during the proc-
ess of disrupting the cell. Accordingly,
the synthesis of a new cell involves as a
major activity the synthesis of new parti-
cles. In one model we have interpreted the
different classes of particles observed in the
ultracentrifuge (20, 26, and 40 S) as stages
in a process wherein small particles grow
to large particles which then disintegrate
into several small ones. Such a model
emphasizes the necessity for experiments
to determine the composition and kinetic
relations of the different classes of particles.

According to this model, each particle
would synthesize and contain only one
type of protein. Therefore, the organiza-
tion of enzymes must result from a spatial
organization of the particle with respect to
other particles. The cell wall does not pro-
vide enough area to hold the particles.
Consequently, it seems reasonable to as-

sume that they might be held in a definite
framework by the deoxyribose nucleic acid.
Calculation shows that there is sufficient
DNA to provide a framework with the
DNA arranged in coils (300-A diameter,
200-A pitch) with their axes parallel. Such
a structure could account for the relation-
ship between the order of genes and the
order of biochemical events which was re-
ported last year by the Department of
Genetics of the Institution. It also appears
possible for such a structure to duplicate
all its individual components and then to
divide by extrusion into two identical
structures.

This model summarizes our present
knowledge of the cell. It shows that we are
beginning to relate structure and function,
but it also emphasizes how little is known
of either.

Hoagland Compound

During the past year a compound has
been reported by Dr. Hoagland in which
amino acids are linked to nucleic acid. It
is reasonably stable in acid but unstable
in basic solution. Such a compound would
appear in the "TCA-precipitable" fraction
and might be mistaken for protein. Tests
were therefore made to determine how
much of this compound is present in grow-
ing E. coli. A radioactive amino acid was
added to a growing culture of cells, and
the usual samples were taken to measure
the total and TCA-precipitable fractions.
In addition, samples were injected into
NaOH to release any amino acids bound
to nucleic acid. The samples were then
neutralized and mixed with TCA to pre-
cipitate protein. In E. coli this procedure
showed that only a minute fraction of the
amino acids could be bound in alkali labile
compounds. With carrier-free amino acids
and washed cells from which all pre-exist-
ing pools were removed there was an indi-
cation that such compounds were present,
but only in quantities of about 1 uM/g
cells. The usual pool would be about 100
uM/g, and the amino acids in the proteins,
of course, would amount to several thou-

140

CARNEGIE INSTITUTION OF WASHINGTON

sand micromoles per gram of cells. In
yeast (C. utilis) a considerably greater
fraction showed the characteristics expected
of amino acids bound to nucleic acid. The
quantity in this case was large enough
(15 uM/g dry cells) to offer an oppor-
tunity for its isolation and for further
study of its kinetic characteristics.

Protomorphs

For some time it has been apparent that
organization is important in the bacterial
cell. The cell is by no means a "bag of
enzymes"; disruption of the cell may have
little effect on individual enzymic activi-
ties, or even may increase them, but many
of the more elaborate functions that re-
quire the co-operative action of several
enzymes cease. In particular, crushing cells
greatly reduces their ability to incorporate
amino acids into protein.

It is also highly probable that the or-
ganization of the cell stems from a spatial
arrangement of the constituent molecules.
The localization of many of the Krebs
cycle enzymes into the mitochondria of
mammalian cells has been observed. A
large part of the bacterial protein and
ribonucleic acid is organized into micro-
somal particles, and similar particles have
been found to play a major role in protein
synthesis in mammalian cells. In mam-
malian cells the microsomal particles are
arranged on the surfaces of membranes;
in the bacterial cell it is likely that they
are also held in a definite spatial arrange-
ment, perhaps in a framework of deoxyri-
bose nucleic acid.

Since, presumably, the forces responsible
for these organized systems are properties
of the macromolecular constituents of the
cells, it seems possible that thoroughly dis-
organized cellular material might, under
the proper conditions, reconstitute some of
the spatial arrangements of the cell analo-
gously to the reconstitution of tobacco
mosaic virus from its protein and RNA
components.

Having these thoughts in mind, we were
very much interested to observe the de-

layed appearance of particles in a clear
preparation of cellular juices. These parti-
cles have several properties that distinguish
them from coacervates formed from the
gelatin and nucleic acid; it even seems
possible that they offer a new form for
study which is intermediate between the
highly organized cell and disorganized
cell juice. Because they are formed from
protoplasm and have a distinctive shape, it
seems suitable to refer to them as "proto-
morphs," to avoid confusion with the
many other particles of cellular origin.

Preparation. These particles develop in a
preparation made as follows : Wash 10 g of
E. coli growing in a glucose mineral salts
medium (C) with a solution (TSS) 0.25
M in sucrose buffered with trishydroxy-
methylaminomethane (Tris) 0.01 M and
brought to pYL 8 with succinic acid. Centri-
fuge out the cells, and resuspend them
with 10 ml TSS. Break the cells by forcing
the suspension through a small hole at
10,000 psi. Bring the volume up to 50 ml
with TSS, and centrifuge 10 minutes at
40,000^. This treatment removes any un-
broken cells and the larger fragments of
cell walls. Decant the supernatant fluid,
and centrifuge at 105,000^ for 15 minutes
to remove the smaller fragments of cell
walls. Again decant the supernatant fluid,
which should be quite clear, and dilute
to 100 ml with TSS. Add MgCl2 and
MnCl2 to make the solution 0.005 M in
each. Leave overnight, and the originally
clear solution becomes quite cloudy owing
to the formation of protomorphs. The
yield is variable but may be as high as
0.5 g. The pH of the solution must be 7
to 8; below pH 7 the protomorphs do not
form, and below pH 6 a precipitate de-
velops when the manganese is added. All
operations following the breaking of the
cells are performed at 0° to 5° C.

In this procedure some variation is prob-
ably permissible; only a few of the vari-
ables have been studied. The addition of
manganese is essential; magnesium is not
essential but improves the yield. The pur-
pose of the low temperature during the

DEPARTMENT OF TERRESTRIAL MAGNETISM 141

formation period is to reduce bacterial con-
tamination and denaturation of protein.
When the cell juice is centrifuged for
2l/2 hours at 105,000^ to remove the micro-
somes (absorption at 260 mu reduced by
67 per cent), it gives a much lower yield of
protomorphs. The microsome pellet when
resuspended in TSS, however, gives a
precipitate but no protomorphs on addition
of manganese. Other variables, such as de-
pendence on the presence of Tris, suc-
cinate, or sucrose, have not been investi-
gated.

Appearance, composition, stability. The
protomorphs formed by this procedure ap-
pear as shown in figure 40, plate 5. (Our
thanks go to Dr. Wm. Duryee for making
this photomicrograph.) They are very
nearly spherical, and range in diameter
from about 1 to 5 microns. Estimated by
the absorption at 260 mu (1 mg nucleic
acid/ml gives optical density = 25), the
nucleic acid content is roughly 20 mg/g
wet weight. The Folin reaction indicates
that there is approximately 50 per cent
more protein per unit of 260-mu absorption
than in the cell juice from which the
protomorphs formed; thus, the protein
content is roughly 120 mg/g wet weight.
Paper chromatograms run in butanol/
formic acid/water show the presence of
lipid material in about the same proportion
as appears in preparations of microsomes
from E. coli. Thus, the protomorphs con-
tain several of the constituents of bacterial
protoplasm.

In contrast to simple coacervate particles,
which can exist only in a narrow range of
/7H, the protomorphs are quite stable.
They can be centrifuged down and resus-
pended without disintegrating; they main-
tain their form in H20, 1 M NH4OH, or
5 per cent trichloroacetic acid (TCA), or
after shaking with ether. Sodium phos-
phate buffer (0.07 M, pH 7) dissolves out
much of the nucleic acid of the proto-
morphs, leaving a less optically dense
residue; casein hydrolysate (5 per cent)
has much the same effect. Ethylenedi-
aminetetraacetic acid (EDT) (0.01 M) dis-

solves the protomorphs, leaving a clear
solution.

Functions of the protomorphs. Suspen-
sions of the protomorphs in TSS supple-
mented with 5 X 10~3 M Mg were incu-
bated with P3204 (4xl0~5 M) at 22° C.
Samples were taken periodically and fil-
tered. Other samples were mixed with
equal volumes of 10 per cent TCA and
subsequently filtered. The concentration
of P3204 in the protomorphs increased
steadily with time; all this phosphate was
extracted with TCA, however, and there
was no sign of incorporation of P32 into
nucleic acid.

Similar experiments were carried out us-
ing a mixture of C14-labeled amino acids
from hydrolyzed Chlorella protein. A
steady increase in the C14 content of the
protomorphs was observed. In this case,
however, the major portion was not ex-
tracted with TCA.

As this incorporation rate is very small
compared with that of bacteria, it is neces-
sary to show that bacterial contamination
could not cause the incorporation. Plate
counts and microscopic examination of the
culture indicated that the bacteria present
could at most account for only 1 per cent
of the observed incorporation.

Chemical tests were made to determine
whether the radioactivity of the TCA-
insoluble material was bound in protein.
The radioactivity was precipitable with
TCA after the insoluble material was
dissolved in 1 N NaOH. On paper chro-
matograms the radioactivity had the same
movement as protein; after hydrolysis it
was found with the free amino acids.

Clearly, from this preliminary report,
much work remains to be done to establish
the nature of these protomorphs. Both the
yield and the synthetic ability are highly
variable from one experiment to another.
Not only do the conditions of formation
require further exploration, but also the
functional properties must be investigated
in more detail. It is not certain whether
the incorporation of amino acids represents
net synthesis or some form of exchange.

142 CARNEGIE INSTITUTION OF WASHINGTON

The protomorphs may have a functional
performance inherently different from the
disorganized cell juices from which they
are formed, or they may merely represent
the separation of functional material from
some inhibitory substance in that medium.
Studies are in progress that should clarify
these points.

PROTEIN SYNTHESIS IN MOUSE TISSUES

During the year our collaboration with
Drs. L. B. Flexner and J. B. Flexner, of the
University of Pennsylvania, has continued.
Observations on amino acid and protein
synthesis in the liver and cerebral cortex
of the mouse have been extended to in-
clude adult as well as newborn animals.
Kinetic data obtained from 4 or 5 litter-
mates sacrificed at various times up to 1
or 2 hours after injection of glucose or
amino acids labeled with C14 showed the
concentrations of labeled glucose in blood
and tissues and the concentration of la-
beled amino acids in blood, tissue pool,
and tissue protein.

One of the unanticipated and striking
findings is the high rate of utilization of
carbon of glucose for synthesis of nonessen-
tial amino acids in the cerebral cortex. A
complete, quantitative analysis of the con-
tribution made by all sources to any one
of these nonessential amino acids is lack-
ing. The importance of glucose, however,
is made evident by the finding that it
supplies ten times more carbon to glutamic
acid and glutamine of cortex than does
blood glutamic acid. This is true for both
the newborn animal and the adult. In
contrast, the quantities of glutamic acid
derived from glucose by the liver of new-
born and adult is about the same as that
derived from blood glutamic acid. In both
tissues, the rate of incorporation of glucose
carbon into carbon of amino acids is far
higher in the adult than in the newborn,
and correspondingly the rate of degrada-
tion of the amino acids is higher in the
adult than in the newborn.

At present, a major concern of this work
is to gain a measure of the relative rates

of protein synthesis and degradation at
different stages of growth. The nerve cells
of the cerebral cortex of the newborn
mouse must synthesize protein to provide
both for the increase in cell size and for
the increase in number of nerve processes
that accompanies development. The nerve
cells of the adult animal, however, have
reached their final size; they do not divide;
and consequently they must synthesize
protein only to compensate for that lost
by degradation. Much the same basic dif-
ference exists between the liver of the
newborn and that of the adult, although
in this organ the situation is complicated
by the demands made upon it for the syn-
thesis of plasma proteins of the blood.

The experimental observations made so
far permit a first step in analyzing the
effect of growth and degree of maturation
on rate of protein synthesis. Calculations
have been made of the rates at which indi-
vidual amino acid pools furnish their
amino acids for protein synthesis. There
is a surprisingly small difference between
the newborn and adult animal. In the
cerebral cortex, the pools of the newborn
are drawn upon at only twice the rate of
the adult. In the liver even less difference
between the two age groups is found. Fur-
ther analysis of these findings in terms of
protein synthesis depends upon determina-
tion of the concentrations of individual
amino acids in their respective pools.

VISITORS

During the year we benefited by a num-
ber of visitors who worked with us in the
laboratory for short periods of time. They
included Dr. Julius Marmur, Harvard Uni-
versity; Mr. Charles Stroebel, University
of Minnesota; Dr. J. R. Vallentyne, Geo-
physical Laboratory; Mr. Ellis Kempner,
Yale University; Mr. S. K. Roberts, Prince-
ton University. Co-operative work was
carried out with Dr. S. Bernhart, Naval
Medical Research Institute; Dr. Louis
Flexner, University of Pennsylvania; and
Dr. Bill H. Hoyer and Dr. Edgar Ribi, of
Rocky Mountain Laboratory, U. S. Public

DEPARTMENT OF TERRESTRIAL MAGNETISM 143

Health Service. Dr. F. T. McClure, Ap- Program, contributed greatly to the bio-
plied Physics Laboratory, Dr. G. Cohen, physics section. We also wish to thank
Pasteur Institute, and Dr. John Leahy, several members of the National Institutes
who were here as part of the Fellowship of Health for the use of their equipment.

OPERATIONS AND STAFF

CO-OPERATIVE WORK OF THE
DEPARTMENT

The Institution's policy of co-operation
has been continued with institutions in
this country and abroad, including the
Applied Physics Laboratory, Associated
Universities, Inc., Bernard Price Institute
of Geophysical Research (Africa), Brook-
haven National Laboratory, Chalmers Uni-
versity of Technology (Sweden), U. S.
Coast and Geodetic Survey, Department
of Defense, Geological Survey of Canada,
U. S. Geological Survey, Geophysical In-
stitute of Huancayo (Peru), Johns Hop-
kins University, Lamont Geological Ob-
servatory, Mount Wilson and Palomar Ob-
servatories, National Bureau of Standards,
National Institutes of Health, National Re-
search Council, National Science Founda-
tion, Ontario Department of Mines (Can-
ada), Pasteur Institute (France), Rocky
Mountain Laboratory of U. S. Public
Health Service, Universities of Missouri,
Pennsylvania, and Virginia, University of
Melbourne (Australia), and Yale Univer-
sity. We continued our collaboration with
the American Geophysical Union, the In-
ternational Union of Geodesy and Geo-
physics, and the International Scientific
Radio Union. Foreign and domestic vis-
itors have used our facilities, some on
fellowships of the Institution from Eng-
land, France, and various parts of the
United States.

Research work concerned with mineral
ages using isotope measurements was con-
tinued with the Geophysical Laboratory.

We have had continued assistance in our
cosmic-ray investigations from the obser-
vatories at Cheltenham, Maryland (to Sep-
tember 30, 1956) ; Christchurch, New Zea-
land; Climax, Colorado; Fredericksburg,
Virginia (from October 1, 1956); God-

havn, Greenland; Huancayo, Peru; and
Mexico, D. F.

Contracts with the government (no
overhead charges) have been continued
for the investigation of the earth's crust
and of cosmic rays. Licenses are in force
with the Atomic Energy Commission for
the procurement of isotopes for nuclear
studies, measurement of mineral ages, and
biophysical investigations.

Three grants from the National Science
Foundation for the activities of the NSF
Advisory Panel on Radio Astronomy are
being administered, without charge.

The National Science Foundation has
also provided funds for the partial cost
of projects, to be carried out in South
America as our participation in the Inter-
national Geophysical Year, concerned with
seismic crustal investigations and with the
height in the ionosphere of equatorial
electrojets.

We have supplied ionium collectors to
the National Meteorological Service of
Argentina and to the Instituto Geofisico
de Universidade do Porto of Portugal.

Several staff members have continued
their service on panels of the U. S. Na-
tional Committee of the International Geo-
physical Year, and one is a member of the
Executive Committee of that Committee;
until December 31, 1956, one member con-
tinued full time with the National Acad-
emy of Sciences on studies concerning a
world data center in connection with the
International Geophysical Year. One mem-
ber continues as Chairman of the Advisory
Panel on Radio Astronomy of the National
Science Foundation. Two members are
on the Committee on Growth of the Na-
tional Research Council, and one is on the
Council of the Biophysical Society and con-
tinues to serve on the Advisory Committee

144 CARNEGIE INSTITUTION OF WASHINGTON

to the Federal Civil Defense Administra-
tion. One member continues as Chairman
of the U. S. A. National Committee, In-
ternational Scientific Radio Union. An-
other staff member (resigned December
31, 1956) continued his full-time research
work for the government.

One staff member visited many institu-
tions abroad on a world trip sponsored
by the Carnegie Corporation of New York.
Another participated in an electrojet sur-
vey in South America as a part of one of
the Department's projects for the Inter-
national Geophysical Year. Another mem-
ber spent several months at the Pasteur
Institute in Paris. One staff member is
abroad on a Guggenheim Memorial Fel-
lowship.

Department staff members attended the
Twentieth International Geological Con-
ference in Mexico City, the symposium on
Polar Atmosphere at Oslo, Norway, the
International Conference on Rock Mag-
netism in London, the Geophysics Collo-
quium at Cambridge, England, and the
symposium on Electromagnetic Phenom-
ena in Cosmical Physics at Stockholm,
Sweden.

ADMINISTRATION AND OPERATION

The Department continues to publish
the Journal of Geophysical Research, partly
subsidized by the Institution.

We have continued to rent part of one
farm for the activities of the radio as-
tronomy group.

BIBLIOGRAPHY

Aldrich, L. T. The measurement and applica-
tions of mineral ages. Program Am. Geo-
phys. Union, Washington Meeting, p. 12,
1957.

Aldrich, L. T., G. L. Davis, G. R. Tilton, and
G. W. Wetherill. Radioactive ages of min-
erals from pegmatites. (Abstract.) XX
Intern. Geol. Congr., Mexico Meeting, pp.
207-208, 1956.

Aldrich, L. T., G. L. Davis, G. R. Tilton, G. W.
Wetherill, and P. M. Jeffery. Evaluation of
mineral age measurements. I. Natl. Acad.
Sci.-Natl. Research Council, Nuclear Sci.
Series Rept. 19, Publ. 400, p. 147, 1956.

Aldrich, L. T., G. W. Wetherill, and G. L.
Davis. Determinations of radiogenic Sr87
and Rb87 of an interlaboratory series of
lepidolites. Geochim. et Cosmochim. Acta,
10, 238-240 (1956).

Aldrich, L. T., G. W. Wetherill, and G. L. Davis.
Occurrence of 1350 million-year-old granitic
rocks in western United States. Bull. Geol.
Soc. Am., 68, 655-656 (1957).

Aldrich, L. T., G. W. Wetherill, G. R. Tilton,
and G. L. Davis. Half-life of Rb87. Phys.
Rev., 103, 1045-1047 (1956).

Aldrich, L. T. See also Davis, G. L.; Tilton,
G. R.; Wetherill, G. W.

Anderson, C. E. See Pieper, G. F.

Britten, R. }. Effect of the osmotic strength of
the growth medium on the amino acid pool
of Escherichia coli. (Abstract.) Science, 124,
935 (1956).

Burke, B. F. Building an amateur's radio tele-
scope. Junior Astronomer, 6 (no. 4), 2-5
(1956).

Burke, B. F. Systematic distortion of the outer
regions of the galaxy. (Abstract.) Astron.
J., 62, 90 (1957).

Burke, B. F., and M. A. Tuve. Interferometers
in radio astronomy. Smithsonian Contribs.
to Astrophys., 1 (no. 1), "New Horizons in
Astronomy," 31-36 (1956).

Burke, B. F. See also Forbush, S. E.

Cohen, G. N., and D. B. Cowie. Remplacement
total de la methionine par la selenomethio-
nine dans les proteines ^Escherichia coli.
Compt. rend. acad. sci., Paris, 244, 680-683
(1957).

Cowie, D. B., and B. P. Walton. Kinetics of
formation and utilization of metabolic pools
in the biosynthesis of protein and nucleic
acid. Biochim. et Biophys. Acta, 21, 211-
226 (1956).

Cowie, D. B. See also Cohen, G. N.

Davis, G. L., G. R. Tilton, G. W. Wetherill, and
L. T. Aldrich. Radioactive ages of min-
erals from granites (Abstract.) XX Intern.
Geol. Congr., Mexico Meeting, p. 212, 1956.

Davis, G. L. See also Aldrich, L. T.; Tilton,
G. R.; Wetherill, G. W.

Erickson, W. C. Generation of polarized light
in the Crab Nebula. Nature, 179, 773 (1957).

Firor, J. W. Inferences from radio signals from
the sun and planets. Proc. Natl. Acad. Sci.
U.S., 43, 2-8 (1957).

Forbush, S. E. Sobre la ausencia de efectos
meteorologicos en la variacion, debida a las
manchas solares, de la ionization por rayos
cosmicos. Rev. mex. fis., 4, 125, Mexico,
1955.

DEPARTMENT OF TERRESTRIAL MAGNETISM 145

Forbush, S. E. Solar influences on cosmic rays.
Proc. Natl. Acad. Sci. U.S., 43, 28-41 (1957).

Forbush, S. E. The U. S. program for cosmic
ray investigation during the IGY. Presented
at Special Meeting U. S. National Commit-
tee, Intern. Geophys. Year, June 27-29,
1957.

Forbush, S. E., and B. F. Burke. Absorption of
cosmic radio noise at 22.2 Mc/sec following
solar flare of February 23, 1956. (Letters to
Editor.) /. Geophys. Research, 61, 573-
575 (1956).

Franklin, K. L. Outer galactic structure from
O, B, carbon stars, and neutral hydrogen.
Astron. ]., 62, 15 (1957).

Graham, J. W. Paleomagnetism and magneto-
striction. /. Geophys. Research, 61, 735-
739 (1956).

Hayden, R. J. See Wetherill, G. W.

Heydenburg, N. P., and G. F. Pieper. Coulomb
excitation of Fe57. Bull. Am. Phys. Soc,
ser. II, 2, 61 (1957).

Heydenburg, N. P., and G. M. Temmer. Alpha-
alpha scattering at low energies. Phys. Rev.,
104, 123-134 (1956).

Heydenburg, N. P., and G. M. Temmer. Cou-
lomb excitation and cascade decay of rota-
tional states in odd-mass nuclei. Phys. Rev.,
104, 981-989 (1956).

Heydenburg, N. P., and G. M. Temmer. Ex-
citation of nuclei by charged particles. Ann.
Rev. Nuclear Sci., 6, 77-116 (1956).

Heydenburg, N. P. See also Pieper, G. F.;
Temmer, G. M.

Jeffery, P. M. The radioactive age of four
W. Australian pegmatites by the potassium
and rubidium methods. Geochim. et Cos-
mochim. Acta, 10, 191-195 (1956).

Jeffery, P. M. See also Aldrich, L. T.

Nicolaysen, L. O. See Tilton, G. R.

Pieper, G. F., and N. P. Heydenburg. Angular
distributions of protons in the F19 (a, /?)Ne22
reaction. (Abstract.) Bull. Am. Phys. Soc,
ser. II, 2, 182 (1957).

Pieper, G. F., N. P. Heydenburg, and C. E.
Anderson. Coulomb excitation of enriched
Kr samples. (Abstract.) Bull. Am. Phys.
Soc, ser. II, 2, 69 (1957).

Pieper, G. F. See also Heydenburg, N. P.

Scott, W. E. John Adam Fleming, 1877-1956.
/. Geophys. Research, 61, 589-592 (1956).

Scott, W. E. List of recent publications. /. Geo-
phys. Research, 61, 581-587, 759-767 (1956) ;
62, 175-182, 335-342 (1957).

Tatel, H. E. Structure of the earth's crust from
gravity measurements. (Abstract.) Science,
124,941 (1956).

Tatel, H. E., and M. A. Tuve. Seismic investi-
gation of crustal structure. Rome Meeting,
Intern. Union Geod. Geophys., Aug. 14-

Sept. 25, 1954. Compt. rend., Strasbourg,
no. 11, p. 140 (1955).

Temmer, G. M., and N. P. Heydenburg. Con-
tribution of Coulomb excitation to inelastic
scattering between nuclear resonances. Phys.
Rev., 104, 989-992 (1956).

Temmer, G. M., and N. P. Heydenburg. Cou-
lomb excitation of medium-weight nuclei.
Phys. Rev., 104, 967-980 (1956).

Temmer, G. M., and N. P. Heydenburg. Ex-
citacion Coulombiana Cu, Zn, As, Se, Y. Ru.
Rev. mex. fis., 4, 133, Mexico, 1955.

Temmer, G. M. See also Heydenburg, N. P.

Tilton, G. R. Interpretation of lead ages by acid
washing experiments. Natl. Acad. Sci.-Natl.
Research Council, Nuclear Sci. Series Rept.
19, Publ. 400, p. 79, 1956.

Tilton, G. R. Uranium and thorium dating.
(Abstract.) Program New Yor\ Meeting,
A A AS, Sec. E, Geol. and Geog., p. 21, 1956.

Tilton, G. R., G. L. Davis, G. W. Wetherill, and
L. T. Aldrich. Isotopic ages of zircon from
granites and pegmatites. (Abstract.) Pro-
gram Am. Geophys. Union, Washington
Meeting, p. 39, 1957.

Tilton, G. R., G. L. Davis, G. W. Wetherill, and
L. T. Aldrich. Isotopic ages of zircon from
granites and pegmatites Trans. Am. Geo-
phys. Union, 38, 360-371 (1957).

Tilton, G. R., and L. O. Nicolaysen. The use
of monazites for age determination. Geo-
chim. et Cosmochim. Acta, 11, 28-40 (1957).

Tilton, G. R. See also Aldrich, L. T.; Davis,
G. L.; Wetherill, G. W.

Tuve, M. A. See Burke, B. F.; Tatel, H. E.

Vestine, E. H. The Aurora Australis and re-
lated phenomena. Antarctica in the IGY.
Geophys. Monograph no. 1, pp. 91-106,
1956.

Vestine, E. H. Exploring the atmosphere with
a satellite-borne magnetometer. Scientific
Uses of Earth Satellites, J. A. Van Allen
(Ed.). Univ. of Mich. Press, Ann Arbor,
pp. 198-214, 1956.

Vestine, E. H. Introductory remarks, geophysics
symposium, Natl. Acad. Sci. Proc. Natl.
Acad. Sci. U. S., 43, 1-2 (1957).

Vestine, E. H. John Adam Fleming. Trans.
Am. Geophys. Union, 37, 531-533 (1956).

Vestine, E. H. Observational and theoretical
aspects of magnetic and ionospheric storms.
Proc. Natl. Acad. Sci. U. S., 43, 81-92 (1957).

Vestine, E. H. Relations between fluctuations in
the earth's rotation, the variation of latitude
and geomagnetism. Intern. Assoc Geomag.
and Aeronomy, Bull. 15a (1956).

Vestine, E. H. Theoretical geophysics. Science,
124, 234-236 (1956).

Walton, B. P. See Cowie, D. B.

Wasserburg, G. J. See Wetherill, G. W.

146 CARNEGIE INSTITUTION OF WASHINGTON

Wells, H. W. Flux measurements of discrete
radio sources at frequencies below 30 mega-
cycles. /. Geophys. Research, 61, 541-545
(1956).

Wells, H. W. Upper atmospheric winds, absorp-
tion and other special projects in the U. S.
program in ionospheric physics. Presented
at Special Meeting, U. S. National Com-
mittee, Intern. Geophys. Year, June 27-29,
1957.

Wetherill, G. W. Potassium-argon age determi-
nation method. (Abstract.) Program New
Yor\ Meeting, AAAS, Sec. E, Geol. and
Geog., p. 22, 1956.

Wetherill, G. W., G. L. Davis, and L. T. Aldrich.
Age measurements on rocks north of Lake
Huron. (Abstract.) Program Am. Geophys.
Union, Washington Meeting, p. 39, 1957.

Wetherill, G. W., G. R. Tilton, L. T. Aldrich,
and G. L. Davis. Interpretation of dis-
cordant uranium-lead and thorium-lead ages.
(Abstract.) XX Intern. Geol. Congr., Mexico
Meeting, p. 230, 1956.

Wetherill, G. W., G. R. Tilton, G. L. Davis, and
L. T. Aldrich. Evaluation of mineral age
measurements. II. Natl. Acad. Sci.-Natl.
Research Council, Nuclear Sci. Series Rept.
19, Publ. 400, p. 151, 1956.

Wetherill, G. W., G. J. Wasserburg, L. T.
Aldrich, G. R. Tilton, and R. J. Hayden.
Decay constants of K40 as determined by the
radiogenic argon content of potassium min-
erals. Phys. Rev., 103, 987-989 (1956).

Wetherill, G. W. See also Aldrich, L. T.; Davis,
G. L.; Tilton, G. R.

PERSONNEL

Director: M. A. Tuve.

Chairmen of Sections: Biophysics, R. B. Rob-
erts; Physics of the Atmosphere, H. W.
Wells; Physics of the Earth's Crust, H. E.
Tatel; Laboratory Physics, N. P. Heyden-
burg; Analytical Geophysics, E. H. Vestine
(on leave of absence to December 31, 1956,
to engage in work for the National Acad-
emy of Sciences; resigned January 31,
1957), S. E. Forbush (from June 1, 1957).

Staff Members: L. T. Aldrich, E. T. Bolton,
R. J. Britten, D. B. Cowie, S. E. Forbush,
J. W. Graham, N. P. Heydenburg, E. A.
Johnson (on leave of absence to engage in
research work for the government; resigned
December 31, 1956), R. B. Roberts, H. E.
Tatel, G. M. Temmer (on Guggenheim
Memorial Fellowship abroad from August
1, 1956), E. H. Vestine (on leave of ab-
sence to December 31, 1956, to engage in

work for the National Academy of Sci-
ences; resigned January 31, 1957), H. W.
Wells. Temporary: B. F. Burke, J. W.
Firor, G. W. Wetherill.

Fellows and Visiting Investigators: G. N.
Cohen, Institut Pasteur, Paris, France (from
April 10, 1957); E. H. Greaser, University
of Cambridge, England (to August 3, 1956,
and September 2-October 1, 1956); W. C.
Erickson, University of Minnesota (entire
report year); K. L. Franklin (to August
31, 1956, and May 27-31, 1957); H. L.
Heifer (part time; full time from Septem-
ber 1, 1956); J. J. Leahy, University of Cali-
fornia (entire report year); F. T. McClure,
Applied Physics Laboratory (November 1,
1956-May 31, 1957); George Pieper, Yale
University (from September 1, 1956); Har-
old Weaver, University of California (from
October 1, 1956).

Guests and Associates: Carl E. Anderson,
Yale University (October 1-5, 1956); H.
Faul, U. S. Geological Survey (part time);
L. B. Flexner and Mrs. J. B. Flexner, Uni-
versity of Pennsylvania (part time); W. K.
Ford, Jr., University of Virginia (part
time); S. Gorodetzky, Institute of Nuclear
Research, Strasbourg, France (April 10-12,
1957); D. N. Kundu, National Research
Council (October 22-November 20, 1956);
H. Lenhoflf, Walter Reed Hospital (part
time, September 17-November 8, 1956);
J. Marmur, Harvard University (Septem-
ber 17-October 15, 1956); Carl L. McGin-
nis, National Research Council (October
22-November 20, 1956); Mrs. I. Z. Roberts
(part time); S. Roberts, Princeton Univer-
sity (November 2, 1956); Manik Talwani,
Columbia University (October 12-15,
1956); G. J. Wasserburg, California Insti-
tute of Technology (June 17-21, 1957).

Research Assistants: J. B. Doak, E. T. Eck-
lund, C. A. Little, Jr., W. E. Scott.

Computer: Miss I. Lange (retired March 1,
1957).

Laboratory Assistants: Miss R. E. Bresnahan
(from April 15, 1957), S. J. Buynitzky,
Miss E. F. French, P. A. Johnson, Miss
B. D. North (from May 6, 1957), R. W.
Reuschlein, Mrs. B. P. Walton (to June 14,
1957).

Chief of Section: Administrative and Operat-
ing, M. B. Smith. Chief of Subsection,
Main Shop and Maintenance, W. F. Steiner.

DEPARTMENT OF TERRESTRIAL MAGNETISM 147

Librarian: Mrs. A. P. Moffett (to August

31, 1956), Mrs. L. J. Prothro (July 16-

August 3, 1956, and from September 1,

1956).
Accountant: Miss H. E. Russell.
Draftsman and Photographer: C. S. Leonard,

Jr. (to December 15, 1956).
Director's Secretary and Office Manager:

W. N. Dove; Stenographers: Mrs. C. C.

Ator, Mrs. C. W. Windmuller.
Senior Instrument Makers: B. J. Haase, L. A.

Horton, J. G. Lorz.
Machinist: F. J. Caherty.

Foreman of Buildings and Grounds: C. Bal-

sam.

Caretaker: E. Quade; Assistant Caretakers:
C. R. Domton, C. R. Forshier (from Octo-
ber 23, 1956), S. Gawrys (from June 17,
1957), P. M. B. de Macedo (to September
15, 1956), S. Swantkowski.

Part-Time and Temporary Employees: Mrs.
L. Beach, R. W. Bluehdorn, T. J. Delaney,
W. K. Ford, Jr., B. Z. Kile, Jr., R. C. Kile,
J. E. McLeod, R. L. Peppell, Jr., Mrs. M. T.
Sheahan, C. F. Stroebel, W. W. Wyatt.

Plate 2

Department of Terrestrial Magnetism

Fig. 1. A portion of the 330-mc helix array used for locating sources of radio radiation on the
solar disk. The two helical receiving elements are mounted on a common ground screen and
are connected to an open-wire transmission line. The present array contains 30 units like the one
shown above and occupies a line 2000 feet long at the River Road field station.

Plate 3

Department of Terrestrial Magnetism

Fig. 29. Electron micrograph of cell walls from E. coli treated with lysozyme and osmotically
shocked. The walls appear as flattened disks which show a high incidence of double structure as
though the original condition were of the bag-within-a-bag type. The dark rods are unbroken bac-
teria about 1.5 microns long.

Plate 4

Department of Terrestrial Magnetism

W

8(DNA)

:$.,: ;

Fig. 30. Sedimentation diagram of E. coli extract prepared by lysozyme
treatment and osmotic shock. The picture was taken about 30 minutes after
the rotor reached a speed of 59,000 rpm. The most obvious components are
the constituents labeled 40 S, 26 S, 20 S, and 8 S. The sharp spike is due to
deoxynucleic acid. Such a simple pattern contrasts sharply with the exceed-
ingly complex diagrams of the protein composition of E. coli as revealed by
ion exchange (cf. fig. 36).

Plate 5

Department of Terrestrial Magnetism

Fig. 40. Photomicrograph of aggregated protomorphs (magnification, X 330),

GEOPHYSICAL LABORATORY

Washington, District of Columbia PHILIP H. ABELSON, Director

CONTENTS

page

Introduction 151

Order and Disorder 151

Evidence for order and disorder in

minerals 151

Short- and long-range order 153

The linear run model of short-range

ordering 153

Dimensions in the run model 154

The numbers and locations of nearest

neighbors 155

Some preliminary experimental re-
sults 155

Basalt Magmas 156

Stability of Annite 161

The Age of Roc\s and Minerals 164

The decay constants of Rb87 and K40 . . 165

Problems in the Precambrian 165

Study of a point in the fossil time

scale 168

Zircon age work 169

Simple Absolute Measurement Technique
for Beta Radioactivity; Applica-
tion to Naturally Radioactive Ru-
bidium 171

The method 172

Results 178

Application to naturally radioactive ru-
bidium 178

General applications 178

Paleobiochemistry 179

Effects of ultraviolet light on the

"primitive environment" 179

Thermal degradation of amino acids. . 185

Ore Minerals 186

The Fe-S system 187

Stability relations of pyrite 187

The FeS-S join 191

The Cu-S system 195

The upper stability curve of covel-

lite 195

The Ni-S system 197

The NiS-NiS2 join 197

page
Phase relations in the Fe-S-O system . 198

The Fe-S-Se system 200

The CoAs2-NiAs2-FeAs2-As system. . 201
Relations between composition of ore

minerals and ore solutions 204

Feldspars 206

Ternary feldspars 206

The system CaAl2Si208-Si02-H20. . 214
Optical properties of heated plagio-

clases 216

The Crystallization of Roc\-Forming
Minerals from Magmas and the
Nature of the Residual Liquid . . . 217

Granitic Pegmatites 222

Pyroxenes 223

The join MgSi03-CaMgSi206 223

Chloritoid 225

Al\ali Amphiboles 228

Reconnaissance in the System FeO-

Fe203-Si02-H20 230

Isograd Problems in Metamorphosed Iron-
Rich Sediments 232

Crystallography 237

Synthetic nephelines 237

A sodium nepheline in nature 240

Solid solution 242

Phosphates 242

Disorder in crystals 243

Polymorphism versus isomerism 244

Digenite 245

Crystallochemical analysis 245

Application of the "Morey-Schreine-
ma\ers' Theorem of Coinci-
dence" 246

Miscellaneous Administration 247

Penologists' Club 247

Seminars 247

Symposium on high pressures 247

Lectures 248

Summary of Published Wor\ 249

Bibliography 251

Personnel 252

Carnegie Institution of Washington Year Boo\ 56, 1956-1957

INTRODUCTION

The Geophysical Laboratory continues history. Studies in paleobiochemistry po-
as one of the active centers of research in tentially may lead to new and detailed un-
earth science. Through application of the derstanding of the origins and develop-
tools and viewpoints of physical science we ment of life as well as to a better grasp of
are achieving new understanding of the such problems as the origins of petroleum,
nature and history of the processes by When equipment currently under develop-
which the earth evolved to its present state, ment is operable, this Laboratory will be

The principal lines of activity at the able to participate very actively in advanc-

Laboratory are studies of phase-equilib- ing one of man's great frontiers — high-

rium relations of the major mineral groups, pressure studies. It is abundantly clear that

research on ore minerals, radioactive age a new type of chemistry awaits discovery

measurements, paleobiochemistry, chemical and exploitation.

reactions at extremely high pressures, crys- Crystallography is becoming one of the

tallography, and investigations of order- most vital fields in science today. By means

disorder phenomena. of neutron diffraction, nuclear and para-

Phase-equilibrium relations among the magnetic resonance, and X-ray diffraction,

major mineral groups are providing a the structural chemist is learning how the

whole series of geological "thermometers" precise determination of molecular ar-

which may be applied to igneous and meta- rangement leads to real understanding of

morphic rocks, furnishing tools for ac- chemical reactivity. The nature of the

quiring a great deal of information about chemical bond as manifested in covalent,

the conditions under which such processes metallic, ionic, or intermediate types is

as mountain building occurred. Similar closely related to structure,

types of thermometers are being developed During the past year Chayes has achieved

for ore minerals. The ultimate accumula- what may be a breakthrough in under-

tion of a number of these will permit cross standing various types of order-disorder in

checks and reliable determinations of the crystals. Using an optical analogue he has

temperatures present during ore formation, produced diffraction patterns similar in

which, in turn, will lead to better under- type to those obtained from X-ray studies

standing of the processes involved and in- of crystals. It is too early to evaluate the

evitably to knowledge having economic practical consequences, but fundamental

importance. knowledge in such an area as the physics

The co-operative radioactive dating pro- of the solid state will certainly lead to use-
gram continues to open new vistas. Re- ful applications. Detailed description of
liable dates may now be assigned to events Chayes' studies and of other work of the
occurring in any age throughout the earth's Laboratory follows.

ORDER AND DISORDER

F. Chayes

EVIDENCE FOR ORDER AND DISORDER nealing temperatures, and it was largely

the interpretation of these latter differences

Striking differences in physical proper- that led to the crystallographic superlattice

ties are often noted in metals or metallic and the closely related notion of long-range

compounds, depending on whether anneal- ordering. The crystallographer's superlat-

ing has taken place above or below some tice in turn led directly to the "interpene-

critical temperature. X-ray diffraction pat- trating-lattice" models of ordering used in

terns may also differ markedly with an- statistical thermodynamics, in which long-

151

152 CARNEGIE INSTITUTION OF WASHINGTON

range ordering is usually treated as an end
product of short-range ordering.

Recent progress in experimental petrol-
ogy has focused attention on solid-state
transitions, and in the course of the last
decade mineralogists and penologists have
gradually adopted much of the viewpoint
and vocabulary developed in connection
with order-disorder studies of metallic
compounds. This trend is perhaps par-
ticularly true of feldspar studies. The
suggestion that the various known modi-
fications of alkali feldspar are attributable
to ordering actually dates from well before
the war. Immediately after the war, as
experimental data became available, the
notion was extended to the plagioclase
feldspars. Synthetic feldspars are similar
to those found frequently in volcanic rocks;
the forms of alkali feldspar and Ca-poor
plagioclases found in plutonic and meta-
morphic rocks have not yet been success-
fully synthesized. For a variety of reasons
it is now rather generally supposed that
volcanic rocks form at temperatures higher
than those characteristic of plutonic or
metamorphic processes. On the usual as-
sumption that the mineral associations
found in common rocks are equilibrium
or near-equilibrium assemblages, the forms
characteristic of volcanic environments are
regarded as stable at "high temperature,"
whereas those found in the plutonic and
metamorphic rocks are considered stable at
"low temperature." 1

In few nonmetallic minerals are the
physical differences between the high- and
low-temperature forms pronounced, and in
many they can be detected only with great

1 Although on the whole quite reasonable and
conformable with much experimental evidence,
this position can sometimes be maintained only
by rather dexterous argumentation; in some
recent high-pressure base-exchange experiments
reported by Wyart, for instance, what many
workers now consider the high-temperature
forms of feldspar are retained at rather low
temperatures but the allegedly low-temperature
forms do not survive except at rather high
temperatures.

difficulty. Frequently, as in the plagioclase
data described below, it is necessary to rely
on minor differences in X-ray diffraction
patterns. It is now almost a convention
to regard the form apparently stable at the
higher temperature as disordered and to
develop an ex post facto argument which
makes it seem reasonable that the differ-
ences between the X-ray diffraction pat-
terns of the high- and low-temperature
forms are such as might occur if the low-
temperature form were in some way or-
dered. There is usually no sound optical
theory indicating which (if either) of the
patterns is actually characteristic of order
(or disorder) . Students of the layered min-
erals, for instance, are fairly well agreed
that the appearance of streaks or of weak,
poorly resolved additional reflections is
indicative of disorder; but in feldspar pat-
terns additional poorly resolved reflections
are called subsidiaries, and some workers
regard them as indicative of order.

In this latter connection it may be
pointed out that, although the hc subsidi-
aries of intermediate plagioclase are now
often considered evidence of ordering, their
displacement from the principal layer lines
can be predicted quite nicely from com-
position on the assumption that the dis-
tribution of Al and Si among tetrahedral
sites is such as would be expected at com-
plete short-range disorder.

If a is the proportion of tetrahedral sites
occupied by Al (as calculated from the
chemical analysis), the expected number
of "runs" of Al in any randomly chosen
consecutive series of N such sites is Na
(1 — a), where a run is defined as a se-
quence of i Al ions bounded at both ends
by an Si ion. Since the entire series con-
tains Na Al's, the average length, E(Iai),
of runs of Al ions is thus Na/Na(l — a) =
1/(1 — a) ; similarly, E(/si) = 1/a. Figure 1,
based entirely on the data of P. Gay, shows
the remarkable correlation of the pseudo-
periodicity, indicated by the displacement
of hc from the principal layer line, with
the average run length of Si at short-range

GEOPHYSICAL LABORATORY 153

disorder, calculated from the composition.
This is of course not to be construed
as proof that the plagioclases studied by
Gay are disordered. It is fair to point out,
however, that although close dependence
of he spacing on composition has been
known for some time, no argument based
on ordering has provided a reasonable ex-
planation of the spacing actually found.
To the limits of error of the data, the slope
of the solid line shown in figure 1 is unity
and its intercept is zero; the average run
length of Si at complete disorder and the
pseudoperiodicity indicated by the hG dis-

Fig. 1. Sc separations in intermediate plagio-
clase as a function of Si run length. X=E(iSi) —
or1, Y = 36Q S^1. (A, Y = X; B, line of best fit

calculated from data of P. Gay, Y= 1.052 X
-0.222.)

placement are evidently estimates of the
same parent parameter.

SHORT- AND LONG-RANGE ORDER

According to now standard definitions,
short-range ordering concerns the immedi-
ate environment of an atom as judged by
the numbers of "right" and "wrong" pairs
it forms with its "nearest neighbors,"
whereas long-range order concerns the
emergence and strength of a new over-all
periodicity in the crystal. In the ingenious
and often rather terrifying "interpenetrat-

ing-lattice" models proposed by statistical
thermodynamicists, the elimination of
wrong pairs automatically generates what
is called long-range order by thermody-
namicists and a superlattice by crystallog-
raphers. It is not the elimination of wrong
pairs as such that generates the long-range
order in these models, however, but the
mechanism by which they are eliminated.
The alleged relation between long- and
short-range ordering is not to be regarded
as something that emerges from arguments
based on interpenetrating-lattice models;
it is built into them. Models in which
long- and short-range ordering are in gen-
eral quite independent can be constructed,
and in view of current interest in mineral-
ogical order-disorder problems considera-
tion of such models seems eminently
worth while.

THE LINEAR RUN MODEL OF SHORT-RANGE
ORDERING

In the sequence AABAAABBABBAB
there are seven A's and six B's. Defining a
"run" as a sequence of elements of one
kind bounded at each end by an element
of the other kind (or by the beginning or
end of the series), the sequence above con-
tains eight runs, two of length 1, one of
length 2, and one of length 3 in A; two of
length 1 and two of length 2 in B.

Run sequences of this type have been
studied in considerable detail, perhaps
chiefly in connection with statistical quality
control, and several of their characteristics
useful in a discussion of short-range order-
ing are now well known. Chief among
them is the expected number of runs, from
which may be calculated both the average
length of run in either element and the
numbers of right and wrong pairs at com-
plete disorder or complete short-range
order.

Given a sequence of sufficient length, in
which the probability that any particular
site will be occupied by an A is simply the
ratio of the number of A's to the number
of (y4's + J5's), the expected number of runs

154 CARNEGIE INSTITUTION OF WASHINGTON

is E(d)=2Na$, where N is the number of
elements (or sites), a = T,(A)/X,(A + B),
andP = l-a.1

Now the probability of occupancy speci-
fied above is precisely that which obtains,
in theory, at complete disorder. From the
definition of a run it is obvious that the
last element in each run and the first ele-
ment of the succeeding run form a right
pair, and that there can be no other right
pairs. Since the last element of the last run
remains unpaired, the expected number of
right pairs at complete disorder is thus 1
less than the number of runs, or

E(PAB)=E(d)=2Na$-l

Coupling each site with its next adjacent
site, it is evident that a sequence of N sites
yields a total of (N—l) pairs, and since
(2Na$ — 1) of these pairs are right, the
remaining N(l — 2a(3) =Na2+Nfi2 must
be wrong, i.e. must juxtapose two like ele-
ments as nearest neighbors. It is easily
shown that ATa2 of these are in A, IV$2
in B.

At perfect short-range order there are by
definition no wrong pairs in the minor ele-
ment of the sequence, which we shall take
as A. Since right pairs must involve only
the first or last element of a run, it follows
that at perfect short-range order all runs
in A are of length 1. This means that there
will be as many runs in A as there are A's.
If boundary corrections are ignored, there
will also be as many runs in B as there are
in A, and so the expected number of runs
in the entire sequence will be IN a. This
is again the number of right pairs, for each
A forms a right pair with each of its two
bounding B's. All the remaining pairs are
wrong pairs in B, and their number is
N(l — 2a). These results are shown in
table 1. The frequencies of right and

1 From the definition we may also write

Z(A)L(B)
UW=L Z(A+B)

Although this and all succeeding estimates are
large-sample approximations, the approximations
are very good even for N much smaller than is
likely to be required in crystal chemistry.

wrong pairs in the two limiting cases are
in agreement with those calculated from
the interpenetrating-lattice models. Noth-
ing has been said, and except as the quan-
tity (3 — a)— >0 nothing is implied, about
long-range order.

TABLE 1. Numbers of Right and Wrong

Pairs at Complete Disorder and Complete

Short-Range Order

Pair Type

Disorder

Short-Range
Order

Wrong: (A A) Na2 0

(BB) N(32 N(l-2o)

Right: (AB) 2Na(3 2Na

DIMENSIONS IN THE RUN MODEL

At complete disorder the run model is
essentially independent of direction. It is
only necessary that each site occur once
and only once, and that in some specified
operational sense the nth. site always follow
the (n — l)th and be followed by the
0 + l)th.

In principle, at least, complete short-
range disorder ought to be spherically iso-
tropic in a three-dimensional array. As
ordering develops, however, its level or
intensity varies vectorially, and the lattice
of the run sequence must be assigned some
direction in the crystal. We may then deal
with sets of intersecting run sequences,
each site belonging to a member of each
set. The members of each set are parallel
to each other and intersect members of
each other set at a common angle. Since
every site is included in one member of
each set, and an element of either type may
fall on any site, there is still no a priori re-
lation between long- and short-range order.
A knowledge of the level of ordering along
one set offers no indication of the level of
ordering along other sets; in particular,
there may be perfect short-range order
along one set and complete disorder along
all others. (In the run model this could be
true of long-range ordering as well.)

GEOPHYSICAL LABORATORY 155

THE NUMBERS AND LOCATIONS OF
NEAREST NEIGHBORS

The interpenetrating-lattice models used
in the standard treatments are so con-
structed that no site may have as nearest
neighbor another site on the same sublat-
tice. In the run model, on the contrary,
the nearest neighbors of the nth. site are
always the (n — l)th and (w + l)th sites
of the same run sequence. Thus if a site
is a member of only one run sequence it
has only 2 nearest neighbors. In a plane
net a site may be a member of two or three
intersecting run sequences and may accord-
ingly have 4 or 6 nearest neighbors. In
three dimensions any site may be a mem-
ber of three or four intersecting run se-
quences, and thus may have 6 or 8 near-
est neighbors. The interpenetrating-lattice
models are not so limited or inflexible with
regard to the number of nearest neighbors,
a parameter of considerable importance in
thermodynamic calculations.

SOME PRELIMINARY EXPERIMENTAL RESULTS

Instrumentation. Enough has been said
to indicate that knowledge of the influence
of disorder on diffraction effects would be
of considerable value in studies of solid-
state transitions. Direct mathematical anal-
ysis is rather forbidding, and at present
there seems no other approach to the prob-
lem in three dimensions. To produce two-
dimensional models (or masks), however,
is comparatively simple, and the diffraction
patterns of such masks are readily observed
in an optical diffractometer of the kind
first described by Taylor, Hughes, and
Lipson. In this apparatus the optical train
consists of a point source, filters, a collimat-
ing lens, a collecting lens, and a microscope
or camera. The mask is placed in the
collimated beam (between collimator and
collector), and its (Fraunhofer) diffraction
pattern is observed or photographed at the
focal plane of the collector. During the
last two months of the report year a small
instrument of this type was constructed,
with a 2-watt concentrated arc light as
point source. The lenses are of 1-inch

working diameter, and their focal lengths
are such that the image is brought to final
focus less than 2 feet from the point source.
Its dimensions make the instrument both
inexpensive to build and convenient to use.
The relatively short focal lengths of the
collimator and collector generate a rather
small final diffraction pattern, a limitation
not critical for present purposes. The nar-
rowness of the collimated beam is a more
serious handicap, for it requires the use of
rather small masks.

For the production of suitable masks a
procedure combining punching and photo-
graphic reduction has been developed. A
more complete description of the instru-
ment and an account of the procedure for
making masks is being prepared for pub-
lication elsewhere. To date only masks
containing a single "atom" — i.e. in which
all openings are of the same diameter —
have been prepared, and emphasis has been
placed on the effect of "mistakes" in the
stacking of layered structures. Layering is
particularly useful as a starting point, not
only because masks portraying varying
levels of ordering are easily prepared but
also because it has been examined theo-
retically.

Equal numbers of layers in each of two
positions. This is the case on which A. J.
C. Wilson builds his theory of the optical
effect of mistakes in layered structures. The
a and b axes lie in the plane of layering,
the c axis normal to this plane, all the lay-
ers are supposed identical, and the mistake
consists of a shift of the layer by half the b
axis. The diffraction mask modeling this
situation would be a be section, with lines
containing equal numbers of equally
spaced holes in the b direction and a ran-
domly distributed offset of b/2 from one
layer to the next along c. Figure 2a shows
such a mask, and figure 2b its diffraction
pattern.1 The diffuse streaks cutting the
principal layer lines are in accord with the
reciprocal lattice calculated by Wilson; the

1 Figures 2 and 3 are on plate 1. Plates are
collected, facing page 192.

156 CARNEGIE INSTITUTION OF WASHINGTON

weak but clearly resolved reflection mid- layers in one position followed by one in
way between layer lines occurs in a region the other, e.g. BB ABB ABB A. . . . This
in which his theory evidently indicates no arrangement is characterized by perfect
diffraction, either diffuse or resolved. In short- and long-range order; e.g., all As
a mask in which short-range ordering was occur in runs of length 1 along c, and the
perfect — that is, one in which there was "repeat distance" along c is 3 layers. There
an offset between each pair of layers — this are four principal reflections at the corners
central spot would be of the same intensity of a square, an exceedingly weak one be-
as the principal reflections and there would tween each pair of principals on the layer
be no diffuse scatter crossing the layer lines, and two bright ones symmetrically
lines. spaced at % and % of the distance between

The contrast between the observed and layer lines,

expected effect of disorder on diffraction Figure 4# 2 is a 2 : 1 (actually 35 per cent)

is so striking as to require further study, mask in which there is perfect short-range

and this is now in progress. Since the but no long-range order, and figure Ab

transform of the perfectly ordered array is its diffraction pattern. The principals

would have a strong central spot, it is and weak subsidiaries are as in figure 3a,

tempting to suppose that the presence of but instead of two bright reflections at l/3

a weak spot in this position in the trans- and % of the layer line interval there is a

form of the disordered array merely sig- single, evidently composite, large one at l/2

nalizes the presence of an undue concen- that interval.

tration of ABAB . . . sequences in the Figure 5a is a 35 per cent mask in which

mask. There is excellent reason to sup- there is complete disorder— i.e. the prob-

pose, however, that this is not the correct ability that any layer is an A is simply the

explanation in this particular case. proportion of A's in the parent population

Preponderance of layers in one position, oi layers. Figure 5b is the diffraction pat-
In the standard derivation, based on Wil- tern of this mask. The four principals are
son's analysis, the number of layers in each as in figures 3 and 4, but there is much
position is, at least by implication, the diffuse scatter, and there are also two well
same. The remaining work to be reported resolved subsidiaries at 1/6 and 5/6 the
on is concerned with masks in which there layer line interval.

is a preponderance of layers in one of the The patterns yielded by perfect long-
two positions. Crystallographers have evi- (and snort') ranSe order> Per£ect short>
dently not considered this possibility in ranSe order> and complete disorder are
connection with order-disorder diffraction thus clearly differentiable. Investigation of

effects, but even the preliminary results the nature of the passage from one pattern

to either or the others, as well ~ ■ !

of similar diffraction effects
A : B ratios, is now in process.

ciiclls, uul even liic uiciiminai y icbuiu> . , r , * ^ .. . * ,

•i ii . ,i . •. • i . i to either or the others, as well as the study

now available suggest that it might be r . ., ,._- . *• r i-n-

or similar difir action effects tor different

worth close study.

Figure 3 is the diffraction pattern of a

mask containing a rigid succession of two 2 Figures 4 and 5 are on plate 2.

BASALT MAGMAS

H. S. Yoder, Jr., and C. E. Tilley

Each of the two major basalt magma others view the various magmas as unre-

types, the tholeiitic basalt type and the lated, each giving rise to its own rock types

alkali basalt type, has been nominated as independently.

the parental or primitive magma. Some New experimental data bear on the prob-

believe that a third magma as yet unob- lem of the alleged primary magma. Speci-

served is in fact the primary magma, and mens of natural basalt representative of

GEOPHYSICAL LABORATORY 157

three magma types, (a) tholeiite, (b) alkali
basalt, and (c) high-alumina basalt, were
selected for thermal study at atmospheric
pressure. Types (a) and (£), represented
by specimens from Hawaii, conform to
the two chief magma types recognized by
the authors of the Mull Memoir, and (<r),
represented by a nonporphyritic Warner
basalt from California, is chemically com-
parable to the Porphyritic Central type of
Mull. The chilled marginal facies of the
Skaergaard layered gabbro of East Green-
land conforms closely to this type (Tilley,
1950) . Analyses of four rocks of the three
types are given in table 2, and the norms in
table 3.

The first two analyses are of tholeiitic
type: (1) the 1921 Kilauea lava, an olivine-
enriched example discussed in last year's
report; and (2) a prehistoric lava from
Kilauea, a typical silica-saturated basalt
corresponding more closely to the primitive
Kilauean liquid (Powers, 1955).

The pyroxenes of the four rocks have
been isolated and analyzed; the results of
the analyses are expressed in terms of the
principal pyroxene components in figure 6.
Three fractions of pyroxene (A, B, C) ob-
tained from the 1921 lava and two (D, E)
from the prehistoric lava provide evidence
that some fractionation took place during
the crystallization.

The pyroxenes have compositions typical
of those from the three magma types. In
general, the pyroxenes of the alkali basalts
are along the Di-Hd join, while the
tholeiitic pyroxenes are more pigeonitic.

Each of these basalts, as well as others,
was held at various temperatures through
the courtesy of J. F. Schairer. Each basalt
is a single bulk composition in a multicom-
ponent system, and its crystallization his-
tory can be determined in the usual way
by the quenching method. Small samples
were dried at 110° C, sealed in platinum
tubes, and held for 1 to 24 hours at constant
temperature and atmospheric pressure. Be-
cause of loss of iron to the platinum con-
tainer and some oxidation of the iron, the
temperatures stated should be considered

approximate. The results are given in
figure 7.

For the 1921 flow of Kilauea, the liq-
uidus is at 1235° C, and olivine is the first
major phase to appear. At about 1190° C
pyroxene comes in, and at 1170° C these
are joined by plagioclase. At 1090° C the
charge is all crystalline. The temperatures

TABLE 2. — Chemical Analyses of Basalts

1

SiO, 49.16 51.18 48.27 49.28

Al263 13.33 14.07 18.28 15.98

Fe2Os 1.31 1.35 1.04 4.11

FeO 9.71 9.78 8.31 7.94

MnO 0.16 0.17 0.17 0.19

MgO 10.41 7.78 8.96 4.44

CaO 10.93 10.83 11.32 9.55

Na,0 2.15 2.39 2.80 3.47

K26 0.51 0.44 0.14 1.26

H00+ 0.04 0.10 0.15 0.39

H20~ 0.05 0.01 0.07 0.53

PX>5 0.16 0.15 0.07 0.23

TiO, 2.29 2.10 0.89 3.06

Cr203 0.09 0.05 .... 0.02

Total 100.30 100.40 100.47 100.45

1. Tholeiite basalt: 1921 lava, Kilauea caldera,
Hawaii, No. 57364 (Cambridge University Col-
lection). Analyst: J. H. Scoon.

2. Tholeiite basalt: prehistoric flow, National
Park Quarry on highway 0.75 mile NE of Vol-
cano Observatory, Kilauea No. 57358 (Cambridge
University Collection). Analyst: J. H. Scoon.

3. High-alumina basalt: Warner flow, 4 miles
SE of East Sand Butte, Medicine Lake High-
lands, Calif., No. 127-ML-295. Collected by
C. A. Anderson. Analyst: J. H. Scoon.

4. Alkali basalt: Papalele Gulch, near highway
NE of Mauna Kea, Hawaii, No. 60464 (Cam-
bridge University Collection). Analyst: J. H.
Scoon.

observed by Jagger at the Halemaumau
crater in 1921 were 1190° and 1200° C.
The above results are in accord with his
observations and suggest that the water
pressure of the magma was very low. The
most significant observation, however, is
that all three major phases appear within
65° C. Dip samples of the lava in 1911
showed all three phases present even
though the crystal content was only a few

158 CARNEGIE INSTITUTION OF WASHINGTON

per cent. The index of refraction of the
glass so obtained was given as 1.605 by
Merwin. The glass prepared in the labora-
tory by completely melting the 1921 flow
basalt had an index of refraction of 1.608.
Note also that in the laboratory the basalt
had a small interval of crystallization of
about 150° C. A prehistoric lava from
Kilauea (fig. 7), which has less olivine,
shows approximately the same results but
has a liquidus at 1195° C. All three major
phases appear within a small range of tem-
perature, 40° C. The total range of crys-
tallization is only 130° C. The liquidus
for the 1887 basalt from Mauna Loa is
about the same temperature as the Kilau-
ean basalts, each producing olivine as the
first major phase. The alkali basalt from
Mauna Kea has a liquidus of 1185° C,
and plagioclase is the first major phase to

; a-

Mouna Kea

oMedlcine Lake

Oc<bC

Fig. 6. Plot of pyroxenes from rocks of table 2
in the system diopside (Di)-hedenbergite (Hd)-
enstatite (En)-ferrosilite (Fs). The fractions A,
B, and C are from the 1921 flow, and D and
E from a prehistoric flow, of Kilauea.

appear. Olivine and pyroxene come in to-
gether at about 1160° C, and at 1040° C
the charge is all crystalline. Again, a nar-
row range of temperature in which all
three major phases appear, 25° C, is ob-
served. The total range of crystallization
is only 150° C. A representative of the
third major basalt type is the high-alumina
basalt from Medicine Lake Highlands,
California. This basalt yields approxi-
mately the same results.

The significant conclusions from this
cursory study of representatives of the
three major basalt types are: (1) all three
major phases appear in a short interval of
temperature under the experimental con-
ditions; (2) all three major phases appear

1 1

1 1

n of glass

Mauna Kea

XTL

a

PX+OL PL

® — A

1.597

Prehistoric Kilauea

XTL

PL PX OL

1.597

1887 Mauna Loa

?<E

OL

o

1.597

Medicine Lake

XTL

PX

PL+OL

1.592

1921 Kilauea

XTL

a —

PL PX

A— -X

1

OL

— o

1.608

IIOO 1200

TEMPERATURE °C

Fig. 7. Results of thermal treatment of se-
lected basalts representative of major basalt
magma types. The abbreviations are for olivine
(OL), plagioclase (PL), pyroxene (PX), and all
crystalline (XTL).

TABLE 3. Norms of Basalts Given in Table 2

Tholeiite
1921 Kilauea

Tholeiite

Prehistoric

Kilauea

High-Alumina

Basalt
Medicine Lake

Alkali Basalt
Mauna Kea

Qz

0.30

• • • •

Or

.. 2.781

2.221

0.561

7.501

Ab

.. 17.82 f 45.90 20.96 h 49.31 23.58 h 61.11 29.34 \- 61.16

An

. . 25.30J

26.13J

36.97J

24.32J

Di

. . 22.93

22.04

15.23

17.97

Hy

. . 15.35

22.44

6.43

Ol

9.14

20.55

1.83

Mt

. . . 2.09

1.86

1.39

6.03

11

4.41

3.95

1.67

5.78

Ap

. . . 0.34

0.34

0.17

0.34

H20*

. . . 0.09

0.11

0.22

0.92

Total

100.25

100.35

100.34

100.46

GEOPHYSICAL LABORATORY 159

together at about the same temperature
(1160° to 1170° C), irrespective of the bulk
composition of the basalt; (3) the total
range of crystallization is small, of the
order of 150° C; (4) olivine or plagioclase
appears on the liquidus to the exclusion of
pyroxene for the basalts studied. Powers
(1955) records the fact that no augite
phenocrysts are found in most of the lavas
of the shield-building stage at Hawaii.

These data can be interpreted in terms
of the synthetic system of simplified basalts
as determined in the laboratory. To repre-

Fig. 8. The "simplified basalt" system diop-
side (Di)-albite (Ab)-anorthite (An)-forsterite
(Fo) based on the experimental studies of Os-
born and Tait (1952), Bowen (1915), and
Schairer (unpublished data). Certain liberties
were taken near the corner Ab, which are ampli-
fied in part by figure 9.

sent the compositions of a simple basalt
(pyroxene + olivine + plagioclase), the tet-
rahedron diopside (Di)-forsterite (Fo)-
albite (Ab)-anorthite (An) was con-
structed from the studies of Osborn and
Tait (Di-Fo-An), Bowen (Di-Ab-An),
and Schairer (unpublished data: Fo-Ab,
Di-Ab). The system given schematically
in figure 8 is, of course, only pseudoqua-
ternary, since the compositions of all the
phases cannot be represented by the com-
ponents chosen. If the spinel field is neg-
lected, there are three volumes, each rep-
resenting a major phase — diopside, forster-

ite, and plagioclase. Two phases are in
equilibrium with liquid along the shared
surfaces, and all three phases are in equi-
librium with liquid along the one common
curve, the four-phase curve. In the heating
experiments described, all three phases ap-
peared over a small range of temperature.
For this reason the composition of the
basalts must lie very close to a similar four-
phase curve. Whether olivine or plagio-
clase (or even pyroxene) precipitates out
first is of small importance. What is im-
portant is the temperature at which all
three phases begin crystallizing together.
Each composition should on cooling reach
the four-phase curve at a characteristic
temperature. Liquids at higher tempera-
tures on the four-phase curve can give rise,
through fractionation, to liquids at a lower
temperature on the curve. Liquids at the
lower temperatures cannot yield a magma
having a higher temperature on the four-
phase curve. From the data on the natural
rocks given in figure 7, it is seen that
neither the alkali basalt nor the tholeiite
basalt can be specified as the parent, since
they reach the "four-phase curve" at about
the same temperature. On the basis of the
data presented, it may be tentatively con-
cluded that all these basalts are themselves
a product of the same melting process, the
diversity arising as a result of different
initial bulk compositions.

If the simple basalts represented in figure
8 fractionated, the liquid would descend to
the point marked ~1090° C. Certain liber-
ties were taken to simplify the relations
near the Ab corner, and now these must be
examined more closely. Schairer has re-
studied the join Ab-Di in the larger aspect
of the nepheline-silica-diopside join, which
is not ternary (fig. 9). Here enters an-
other difficulty in choosing the alleged
parental basalt. Consider a bulk composi-
tion in the diopside field on the Ab-Di
join. The liquid stays in the plane until it
hits the plagioclase field boundary curve.
Then, if the liquid were slightly saturated
in silica, its fractionated liquids would
trend toward silica. If the liquid were

160 CARNEGIE INSTITUTION OF WASHINGTON

slightly undersaturated, it would go to the
nepheline side. In examining the original
liquid, very careful analytical work would
be required to determine in advance which
way it would go. A similar situation arises
when compositions near the Ab-An join
in the nepheline^-silica-anorthite system
(Schairer, unpublished) are considered. In
all the systems involving potash, such as
An-leucite-silica (Schairer and Bowen),

of the two major magma types, and its be-
havior bears on their relationships. The
tholeiite magmas characteristically show
the reaction relation with olivine; the alkali
basalt magmas do not. These observations
can be interpreted in terms of Bowen's
diagram for Di-Fo-silica (fig. 10). The
tholeiites must start their crystallization
in the olivine field in the area formed by a
line joining Mg2Si04 to the point where

DIOPSIDE
1391 6° @ CoMgSigOg

NEPHELINE
CARNEGIEITE
NaAlS.04

JADEITE" ,,,„._„

COMPN. Il18-3

WEIGHT PERCENT

Fig. 9. The system CaMgSi206-NaAlSi04-Si02 (Schairer, unpublished data given in Yoder, 1950,
with permission). The join albite-diopside, which is not binary, is not shown.

Di-leucite-silica (Schairer and Bowen),
Fo-leucite-silica (Schairer), and fayalite-
leucite-silica (Bowen and Schairer), the
liquids trend toward silica and not toward
the feldspathoid. It is clear from these
considerations that no one magma can go
to both sides if only the simple major
phases are considered. On these grounds
it must be concluded that a single magma
cannot produce both a tholeiite trend and
an alkali trend by fractionation. The dif-
ferences between the magmas capable of
producing such trends, however, may be so
small as to be not readily detectable.
Olivine plays an interesting role in each

the pyroxene boundary curve crosses the
MgSi03-CaMgSi206 join and the pyroxene
boundary curve. The alkali basalts appar-
ently begin their crystallization in the re-
maining area of the olivine field, or in the
pyroxene field near the diopside corner to
the left of the MgSi03-CaMgSi206 join.
To support this view, Powers notes that the
lavas showing augite phenocrysts are usu-
ally assigned to the declining stage of ac-
tivity in the Hawaiian volcanoes, the
alkali-rich stage.

Water must be considered among the
agents effective in altering the course of
crystallization, and the oxidation or reduc-

GEOPHYSICAL LABORATORY 161

tion of iron may account for alternative barrier between the saturated and unsat-

courses. The problem of the alleged paren- urated magma types. The effects of water

tal basalt seems to hinge on finding suit- as a possible mechanism are now under

able mechanisms to bridge the apparent study.

CaMgSi206

Mg2Si04

MgSi03

SiO

Fig. 10. The system CaMgSi206-Mg2Si04-Si02 (Bowen, 1914). The join CaMgSi206-MgSi03,

which is not binary, is dashed.

STABILITY OF ANNITE
H. P. Eugster

Preliminary data on the stability of the
ferrous biotite annite were reported last
year. Since then experimental work on this
mineral has been completed up to 2000
bars total pressure. The revised equilib-
rium diagrams are presented in figures 11,
12, 13, and 14. Of the three independent
variables, temperature T, total pressure
Ptot ( = Ph2o), and partial pressure of
oxygen Po2, only two, namely T and Ptot,
can be varied continuously. Po2 is defined

by the five buffers that surround the semi-
permeable sealed platinum tube. The
buffers are mixtures of iron oxides, fayal-
ite, and quartz. Hydrogen formed by the
dissociation of water and passing through
the platinum tube acts as a transfer agent
in equalizing Po2 between buffer and sam-
ple. Ph2 and Po2 are not independent
variables, since the dissociation of water is
constant for a given T and Ptot. The five
buffers used are (a) fayalite + iron + quartz,

162 CARNEGIE INSTITUTION OF WASHINGTON

(b) iron + wiistite, (c) wiistite + magnetite, presented in figures 11 and 12. In terms

(d) magnetite + quartz + fayalite, and (e) of the three variables T, Ptot, and Po2,

hematite + magnetite. annite occupies a volume within which it

The Po2-T curves for these pairs are represents the stable phase. This volume

500 600 700

TEMPERATURE. °C

800

900

Fig. 11. Isobar ic section of the stability field of annite for Ptot = Ph2o = 2000 bars. The curves

are Po2-T curves for the following buffers: (a) iron + fayalite + silica, (b) iron + wiistite, (c) wiist-

ite + magnetite, (d) fayalite + magnetite-}- silica, and (e) magnetite + hematite. For further expla-
nation see text.

600
TEMPERATURE,

800

Fig. 12. Isobaric section of the stability field
of annite + quartz for Ptot=PH2o = 2000 bars.
For the significance of curves {a), (b), (c), (d),
and (e) see legend to figure 11 and text.

is separated by curved surfaces from four
other volumes representing the phase
assemblages hematite + sanidine + vapor,
magnetite + sanidine + vapor, fayalite + leu-
cite + kalsilite + vapor, and iron + sanidine
+ vapor. Figure 11 shows a section
through these volumes at Ptot =Ph2o = 2000
bars. The effect of Po2 and T on the loca-
tion of the reversible equilibria is clearly
discernible. For points A, B, C, D, and E,
reversibility of the reactions has been
demonstrated and the temperatures are
known within ±5° C. Figure 12 shows a
section also at Ptot =Ph2o = 2000 bars, but
for the bulk composition of annite +
quartz. Fayalite + sanidine + quartz can co-
exist over a rather wide Po-T field, thereby
restricting the stability of annite in the
presence of quartz considerably.

Figure 13 shows Ptot-T curves for the
reactions studied to demonstrate the influ-

GEOPHYSICAL LABORATORY 163

ence of total pressure. Along the univari-
ant curves, Po2 is not constant but changes
according to the Po-T curves of the buffer
assemblages used for a specific reaction.
The Ptotr-T curves show slopes common to
curves of most hydration-dehydration reac-
tions.

Figure 14 is a three-dimensional drawing
showing the stability volumes of the five
phase assemblages for the annite bulk

which annite represents the stable phase.
It is convenient to present the stability re-
lations of hydrous iron silicates in terms
of total pressure, partial pressure of oxy-
gen, and temperature as independent vari-
ables. But it should be remembered that a
representation in terms of Ptot, Pn2, and T
is equally justified. This is particularly
significant, since hydrogen and not oxygen
equalizes gradients between buffer and

or

=>
o->
oo

LU

<r
a.

2500

2000 -

1500 -

1000 -

500 -

400

500 600 700
TEMPERATURE. °C

800

900

Fig. 13. Temperature-total pressure (^>tot = jf>H20) diagram for systems annite and annite + quartz.
For each of the selected univariant curves the partial pressure of oxygen is equal to that of the buffer
used (brackets) and changes with temperature accordingly.

composition. Two additional volumes are
not represented, because their existence
could not be verified experimentally. From
phase-rule considerations we know that the
magnetite + sanidine + vapor volume must
be separated from the f ayalite + leucite -h
kalsilite + vapor volume by a narrow vol-
ume for magnetite + f ayalite + leucite +
vapor, whose width is narrower than the
experimental error (±5° C). A second
narrow volume for iron + f ayalite + leucite
+ vapor must separate the volumes for
fayalite + leucite + kalsilite + vapor and for
iron + sanidine + vapor. These modifica-
tions do not affect the volume within

sample. In the geologically important re-
gion the partial pressures of hydrogen
range from tenths to several hundred bars.
Annite is the first hydrous phase for
which the relationships in a portion of the
Ptot-Po2-T space have been worked out.
The reactions involved are hydration-de-
hydrations combined with reduction-oxi-
dations. The two main conclusions de-
rived from the work on annite can be
stated as follows: Iron silicates possess a
definite range of partial pressures of oxy-
gen over which they are stable. This range
changes with change in temperature. In
the case of annite the upper limit for Po2

164 CARNEGIE INSTITUTION OF WASHINGTON

is that of the magnetite-hematite boundary, oxygen between some of the participating
The equilibrium temperatures of all reac- phases will depend on the magnitude of
tions that show a transfer of hydrogen or the partial pressure of oxygen.

o

Q-

Q- 2000

CO

<

CD

IjJ

DC
3
CO
CO
UJ
CC
Q_

o

1000

500

600 700 800
-> TEMPERATURE

Fig. 14. Ptot-Po2-T model of the stability volume of annite, presented in isobaric sections. The
curved surfaces separating the five individual volumes are defined in this drawing by Po2-T and
Ptot-T curves.

THE AGE OF ROCKS AND MINERALS

(A co-operative program of the Geophysical Laboratory and the Department
of Terrestrial Magnetism of the Carnegie Institution of Washington)

G. L. Davis, G. R. Tilton, L. T. Aldrich} G. W. Wetherdl,1 and H. Paul 2

We have to remember that while nature is
complex with time-less subtlety, human
thought issues from the simple-mindedness
of beings whose active life is less than half
a century. — A. N. Whitehead, 1919

Now that reliable methods of age de-
termination have been developed a num-
ber of additional problems have become
accessible to study. Fossils have been used
to establish relative ages in post-Precam-
brian times, but they can give only a rough
indication of the actual periods of time

1 Department of Terrestrial Magnetism.

2 Visiting Investigator, U. S. Geological Survey.

involved. The fossil time scale currently
employed is based on only four points, all
of questionable value, either because of
lack of concordancy in the isotopic ages
used in calibration or because the strati-
graphic position of the samples cannot be
accurately established. The recently de-
veloped ability to measure ages on granite
greatly increases the number of samples
suitable for gaining information for this
time scale.

In Precambrian rocks, fossils are absent,
and relative ages are difficult to establish
over any great distances in the field.

GEOPHYSICAL LABORATORY 165

Younger orogenic belts, such as the Appa-
lachian Province, contain considerable
bodies of igneous and metamorphic rocks
that were formed at about the same time.
Perhaps, when a sufficient number of age
measurements are completed in the Pre-
cambrian, vestiges of a number of older
orogenic belts will be found. For example,
the wide prevalence of 1000-million-year-
old igneous rocks in the Grenville is gen-
erally believed to indicate such a belt. A
major long-range goal of work on age
determination is to establish the space-time
relationship of orogenic chains of the past
as a prelude to understanding the role
played by such belts in the development of
the continents.

Reliable age determinations depend on
finding concordant ages — that is, agree-
ment between two uranium-lead ages and
between uranium-lead, rubidium-stron-
tium, and potassium-argon ages. Such
agreement is not always found. Often the
rubidium-strontium and potassium-argon
ages agree with each other but the two
uranium-lead ages are discordant. Some-
times even the rubidium-strontium and
potassium-argon ages are discordant. Po-
tentially, useful information is buried in
these results. It may be possible in the
future to specify within reasonably narrow
limits the conditions that brought about
the discordancies and thereby gain infor-
mation on the post-crystallization history
of these rocks. Workers in the field of age
determination have not yet begun a serious
attack on this problem.

In the past year some progress has been
made in all the phases of activity men-
tioned. Precambrian problems have been
studied in the Rocky Mountains of western
United States and at Sudbury, Ontario.
Information of potential use to the fossil
time scale has been obtained from the
Hercynian Chain in western Europe and
from the Wichita Mountains in Oklahoma.
A study was begun to test the influence, if
any, of orogenies in producing discordant
ages in the mineral zircon. These results
will be discussed in more detail below.

THE DECAY CONSTANTS OF Rb87 AND K40

A year ago the report of this group
stated that the uncertainties in the decay
constants of Rb87 and K40 appeared to be
resolved by comparing K40-A40 and Rb87-
Sr87 ratios in micas with the concordant

TJ238_pb206 an(J U235„pb207 giyen by

uraninites from the same mineral assem-
blages. Since the decay constants of U238
and U235 are known to within 2 to 3 per
cent, it was possible to calculate the decay
constants for Rb87 and K40 from the
uranium-lead age. The calculated con-
stants were in agreement for six mica-
uraninite assemblages ranging in age from
370 to 2700 million years. This agreement
led to the belief that alteration of the Rb-Sr
and K-A ratios in the micas by processes
other than radioactive decay was not a
serious problem. Crystal counting experi-
ments on K40 by Wetherill (see Report of
the Director of the Department of Ter-
restrial Magnetism) during the past year
have confirmed the geologic value of Xe for
the decay of K40 to A40 within 5 per cent,
assuming that none of the K40 decays di-
rectly to the ground state of A40. Wether-
ill's value for Xe is enough higher than the
geologic value to indicate that even the
best micas may have lost on the average
4 or 5 per cent of the A40. The geologically
determined constants are used for pur-
poses of age determination, but these con-
stants must agree closely with the labora-
tory values if any great confidence is to be
placed in age measurements using them.
The agreement now appears to be close
for most micas.

PROBLEMS IN THE PRECAMBRIAN

The occurrence of 1300- to 1400 -million-
year-old granitic roc\s in western United
States. Rubidium-strontium and potas-
sium-argon ages have been measured on
eleven micas, including biotite, muscovite,
and lepidolite, from Precambrian granites
and pegmatites occurring in the Cordil-
leran System in Arizona, New Mexico,
Colorado, and Wyoming. The ages are
given in table 4. All these micas have ages

166 CARNEGIE INSTITUTION OF WASHINGTON

between 1300 and 1400 million years, indi-
cating that there was a widespread crystal-
lization of granitic rocks at this time and
that the micas have preserved their ages in
spite of more recent events, including the
Laramide orogeny. These mica ages are
believed to date a period of regional rock
formation of a type comparable to the
later periods of igneous intrusion and

TABLE 4. Age Determinations on Micas from
Western United States

K-A ages are calculated from decay constants
of K40 of Xe = 0557x lO"10 yr-1, X/5 = 4.72 X 10"10
yr-1, or a total half-life for K40 of 1.31 X 109 yr.
Rb-Sr ages are calculated using a half-life for
Rb87 of 50X109 yr.

Location

Age,
million years

K-A Rb-Sr

1. Gneiss, Zoroaster Creek,

Grand Canyon, Ariz 1380 1350

2a. Lawler Peak granite, Bagdad,

Ariz 1410 1390

2b. Pegmatite in Lawler Peak

granite 1420 1500

3. Pegmatite, Wickenberg, Ariz. . . 1160 1300

4. Pidlite Mine, Mora Co., N.M.. 1330 1490

5. Granite, Sandia Mts., Albuquer-

que, N. M 1350 1340

6. Harding Mine, Dixon, N.M... 1300 1300

7. Uncompahgre granite, Mesa

Co., Colo 1320 1320

8. Granite, Doyleville, Colo 1320 1310

9. Brown Derby pegmatite, Ohio

City, Colo 1330 1420

10. Granite, Sherman, Wyo 1420 1410

11. Silver Plume granite, Colo 1280

metamorphism in the Appalachian and
Grenville Provinces.

Zircon ages, reported previously from
the Lawler Peak and Quartz Creek gran-
ites, have been found to be discordant in
such a way that the Laramide orogeny
could have been partly responsible.

Precambrian ages in Ontario. The Lau-
rentian Shield in central and eastern
Canada and northern Michigan, Wiscon-
sin, and Minnesota has long been recog-
nized as a favorable area for the study of
problems in Precambrian geochronology.

Approximately 2 million square miles of
glaciated volcanic, plutonic, and sedimen-
tary rocks are exposed. To the northwest,
rocks of the Timiskiming-Keewatin Prov-
ince are cut by pegmatites having ages of
about 2600 million years. To the southeast
in the Grenville Province, numerous in-
trusive igneous rocks have ages of about
1000 million years. Between these two
provinces is another, narrower belt, the
Huronian series, which is known to be
younger than the Timiskiming-Keewatin
rocks, but is of unknown relation to the
Grenville rocks since the Huronian and
Grenville are bounded by a fault zone.

There are, then, several reasons for mak-
ing detailed age studies on rocks in the
Laurentian Shield. Precambrian rocks are
exposed over a great area, and they repre-
sent a long span of time, at least 1000
million to 2500 million years ago. There
is a need to determine the relations of the
Huronian to the Timiskiming-Keewatin
series on an absolute time scale and, if
possible, to fit the Grenville series into this
sequence. Studies were commenced on
these problems in the past year, with par-
ticular reference to the Huronian series.
Valuable information has been obtained,
although no conclusive solutions have re-
sulted as yet.

Mica ages have been determined for a
number of igneous and metamorphic rocks
of known stratigraphic relation to the sedi-
mentary rocks around Sudbury: the
Huronian, Sudbury, and Keewatin series.
Their absolute time sequence and their
geographic extent have been among the
major unsolved problems in Precambrian
geology. The present studies have at-
tempted to place limiting values on the
ages of the various series by studying the
ages of micas from rocks known to be
older or younger than a particular series.
This approach is necessary since there is no
proven way at present to date the sedi-
ments directly.

The results obtained so far appear in
tables 5, 6, and 7. Immediately evident are
the numerous discordances between the

GEOPHYSICAL LABORATORY 167

rubidium-strontium and potassium-argon
ages for many of the micas, in sharp con-
trast to our past experience. Moreover, the
few discordances found previously were of
the type for which the potassium-argon
age was less than the rubidium-strontium
age; they were considered an indication of
argon leakage. Several of the Sudbury
micas give potassium-argon ages that are
much greater than the rubidium-strontium

ages. As yet there is no explanation for
these rather surprising inconsistencies.

Although the data contain a large num-
ber of discordant ages, some conclusions
may be drawn. The Wavy Lake granite
represents an outlier of igneous rock of
the same age as the igneous rocks that in-
trude the Grenville sediments in the Gren-
ville subprovince. The sediments called
Huronian in this area appear to be more

TABLE 5. Age Determinations on Micas from the Sudbury District

Location and Sample

Stratigraphic Position

. Age>
million years

Rb-Sr K-A

Wavy Lake granite Intrudes Huronian 1075 1025

Sudbury gabbro Intrudes lower Huronian 1325 1830

Sudbury breccia (matrix) Younger than Sudbury series and Copper Cliff

rhyolite 1440 1870

Levack norite 1830

Haleyburian lamprophyre Pre-Cobalt (middle Huronian) 2050 2160

Hearst pegmatite Intrudes Keewatin 2595 2605

Round Lake lamprophyre Pre-Huronian, Post-Keewatin 2600 2450

Round Lake batholith, granite Pre-Huronian, Post-Keewatin 2640 2530

Timmins, granite Pre-Huronian, Post-Keewatin 2470 2520

TABLE 6. Age Determinations on the Cutler
Batholith

(Intrudes the Sudbury series)

Sample

Age,
million years

Rb-Sr K-A

Pegmatite 1 Muscovite 1750 1440

Feldspar 1760 1165

Pegmatite 2 Muscovite 1700 1420

Granite Biotite 1325 1380

TABLE 7. Age Determinations on the Copper
Cliff Rhyolite

(Stratigraphic position uncertain)

Sample

Age,
million years

Rb-Sr K-A

Muscovite 1730 1390

Biotite 1220 2130

Feldspar 2360 1400

than 1300 million and less than 2600 mil-
lion years old. They are actually less than
2150 million years old if the nearly con-
cordant age obtained for the Haleyburian
lamprophyre is significant and if the age
of the mica from the lamprophyre does not
represent a period of metamorphism subse-
quent to its formation. Since so many dis-
cordant ages have been found in this area,
the 2150-million-year limit must be viewed
with caution until further samples of the
same age are found. The Sudbury series
is older than 1400 million years because it
is intruded by the Cutler batholith. It is
possibly older than 1750 million years since
most of the rubidium-strontium ages ob-
tained from the Cutler give this value.
The Keewatin series is older than 2600
million years. It should be emphasized
that these conclusions apply only to these
series as they are identified around Sud-
bury. The results cannot be extrapolated
with certainty to form conclusions regard-

168 CARNEGIE INSTITUTION OF WASHINGTON

ing rocks called "Huronian" or "Keewa- The large (30-lb or more) rock samples

tin" elsewhere. were crushed, ground, and separated into

The close proximity of the 2600-million- their principal mineral constituents with

year-old rocks at Timmins and the Round particular attention to zircon and mica.

Lake batholith at Kirkland Lake to the The separations were made at the institutes

lOOO-million-y ear-old rocks at Wavy Lake of mineralogy and physics at the Uni-

is interesting. Between these areas, only versity of Strasbourg.

150 miles apart, rocks of intermediate age Analyses of selected zircon samples for

occur, so that repeated igneous activity uranium, thorium, and lead and of some

must have occurred in this rather restricted of the mica samples for potassium, rubid-

area. The large number of discordant ages ium, strontium, and argon have been made

found in the micas may bear some relation by the isotope dilution methods described

to this fact. in previous reports, as part of the co-opera-

The 2600-million-year-old rocks listed in tive program with the Department of

table 5 have about the same age as a num- Terrestrial Magnetism,

ber of intrusions found elsewhere. Rocks The zircon data obtained thus far are

of this age have now been found in north- summarized in table 8. They show that

ern Wyoming, southern Montana, north- all the zircons contain original lead, with

ern Minnesota, and southeastern Manitoba, the possible exception of the Oslo sample

as well as at Hearst and north of Sudbury (which contains too little lead to determine

in Ontario. Rocks with ages of 2600 mil- whether or not original lead is present in

lion to 2700 million years on the continents significant proportion) . In agreement with

of North America, Africa, and Australia our previous experience, reliable age in-

are still the oldest reliably dated rocks. It formation cannot be obtained from zircons

seems probable that reliable dates for older with appreciable common lead content,

rocks will be determined in the future, and these rocks cannot be dated by analyz-

because these old rocks are intruded into ing zircon alone.

sediments that must have been derived in As part of the zircon program, a suite of

turn from substantially older igneous galenas from the Vosges Mountains was

rocks. analyzed for lead isotopes (see table 9).

As expected, the composition of these leads

STUDY OF A POINT ^N THE FOSSIL TIME is very similar t0 that Q£ the common leads

from Germany, analyzed by Geiss, so that
While serving as Fulbright lecturer to we cannot justifiably calculate the zircon
the University of Strasbourg during the jata into concordance by accepting some
academic year 1954-1955, Faul collected a unusUal isotopic composition for the origi-
suite of granitic rocks from the general nai ieacj tney contain,
area of the Hercynian Chain and some of Rubidium-strontium and potassium-
its suspected outliers. Samples were taken argon analySes of micas from some of these
in the Oslo area, the Harz Mountains, the same rocks show a much more consistent
Schwarzwald Mountains, the Vosges picture (see table 10) . Within the limits of
Mountains, the Alpine complex, and the error tne age is the same for all the rocks,
Massif Central. Field work in each area about 345 million years,
was carried out in close co-operation with, The stratigraphic age of these rocks is
and where possible in the company of, a known to be pre-Westphalian (pre-middle
local geologist who was particularly inter- Carboniferous), and they are usually as-
ested in the igneous petrology of the area, signed to the Dinantian (lower Carbon-
The field assistance of Drs. Barth, Denkel, iferous). According to the U. S. Geologi-
Gjelsvik, Hiigi, Roques, Siat, Wedepohl, cal Survey version of the Holmes time
Weil, Wenk, and Wimmenauer is grate- scale, the Carboniferous began 265 million
fully acknowledged. years ago. By this time scale the present

GEOPHYSICAL LABORATORY 169

measurements would place the Hercynian
rocks in the middle Silurian. It follows
that either the time scale is not correct
here or the accepted stratigraphic assign-
ment must be in error.

ured as well. In general, the new deter-
minations have not altered the conclusions
stated a year ago — that discordant isotopic
ages are accompanied by the presence of
common lead in the samples, and zircons

TABLE 8. Age Determinations on Zircons from the Hercynian Chain

Location

Pb

(total),

ppm

Pb
(origi- U, Th,

nal), ppm ppm
per cent

Age, million years

TJ238/ JJ235/ Th232/ Pb207/

pi^OG p[}207 Pb208 Pb206

Natzwiller, Vosges 73 32

Col de la Grosse Pierre, Vosges. .166 20

Wembach, Schwarzwald 46 11

Martinskapelle, Schwarzwald . . 163 59

Martinskapelle, Schwarzwald,

leached with hot HC1 86 34

Halbmeil, Schwarzwald 78 40

f 18
Oslo nordmarkite S . 7

T989
1991
2850
740
1730

1251

784
J363
\365

534

865
382
343

345

177

J388

\396

292

293 312
337 326
247

292

374

267 282

263

291 440 ±60

283 260 ±60

272

285 .......

225 395 ±60

Braces show duplicate determinations.

Errors (standard deviations) are less than 2 per cent except where shown otherwise.
When ages involving Pb207 are not given, too much primary Pb207 was present to permit accu-
rate determination of radiogenic Pb207.

TABLE 9. Isotopic Composition of Galenas from the Vosges Area

Location

206/204

Ste. Marie 18.80

±0.14
Musloch, Ste. Croix 18.37

±0.16
La Croix 18.88

±0.27
"Donner" mine 18.37

±0.05
"Aurora" mine 18.65

±0.08
Steinbach 18.74

±0.065
Wegscheid 18.24

±0.23
Auxelles-Haut 18.72

±0.05

Isotopic Ratios

206/207

206/208

207/204 208/204

1.206

0.485

15.58

38.73

±0.005

±0.008

1.186

0.482

15.49

38.08

±0.002

±0.004

1.211

0.481

15.60

38.90

±0.007

±0.001

1.184

0.483

15.52

38.03

±0.007

±0.003

1.189

0.481

15.68

38.73

±0.004

±0.0015

1.200

0.487

15.62

38.45

±0.0015

±0.0015

1.157

0.467

15.76

39.00

±0.009

±0.005

1.190

0.483

15.72

38.77

±0.003

±0.007

The errors indicated are the observed mean deviations of usually about ten sets of ratios.

ZIRCON AGE WORK with no detectable common lead give con-

Isotopic age determinations have been cordant, or nearly concordant, isotopic

completed for six zircons, excluding those ages. Among the zircons reported in table

from the Hercynian Chain. In three, the 11, those from Conway, Canada Hill, and

ages of associated micas have been meas- Finland contained no detectable common

170 CARNEGIE INSTITUTION OF WASHINGTON

TABLE 10. Mica Age Results

Location

Age,
million years

Rb-Sr K-A

336

344

Vosges

Lac Blanc * i|40

\346

Natzwiller 350

Col de la Grosse Pierre 333

Schwarzwald

Wembach 345

Martinskapelle 341

Halbmeil 334

Sasbach-Walden 331

Massif Central

Royat 351

328

at present for the cause of the relation be-
tween the common lead content of a zircon
sample and the age results.

Our previous work showed that three
zircons from the Grenville subprovince in
Ontario gave concordant isotopic ages in
spite of the fact that the samples varied
greatly in crystal size and amount of radia-
tion damage. The two zircons from the
Fenno-Scandian Shield have likewise given
quite satisfactory age results. The result
for the Rapakivi granite is somewhat dis-
cordant, but not so discordant as the re-
sults found for zircons containing common
lead, in which the U238-Pb206 and Pb207-

"DU206 „^.^n U/-.-.r^ A1CC*.~£±A U», C^~*-^~r. ^C

* Duplicate determination.

neat

ly

2.

TABLE 11. New

Age Results for Zircon

i and Associated Biotite

Mineral

Age, millior

t years

Location

u238/

U235/

Pb207/

Th232/

Rb87/

K40/

p^206

Pb207

p^206

p^208

Sr87

A40

Conway, N. H.,

Zircon

187

184

140 ±60

190

granite

Biotite

185

182

Wichita Mts., Okla.,

Zircon A

520

527

550

506

pegmatite (zircon)*

Zircon B

514

522

550

493

granite (biotite)

Biotite

500

480

Bodom granite,

Zircon

1590

1625

1675

1540

Finland f

Rapakivi granite,

Zircon

1165

1350

1650

1050

Finland f

Canada Hill gneiss,

Zircon

1020

1060

1150

Bear Mt., N. Y.

Biotite

1030

930

Hybla, Ontario,

Cyrtolite

1350

1190

900

435

McDonald Mine

* Zircons A and B are separate zones separated from a single large crystal. A has uranium and
thorium contents about six times those of B and correspondingly more radiation damage,
t The Bodom and Rapakivi determinations were made by O. Kouvo at these laboratories.

lead and all give reasonably concordant
ages. The Wichita Mountains samples are
a notable exception to the generalization,
for while they give nearly concordant ages
the lead in sample A contains 10 per cent
common lead and that in sample B con-
tains 5 per cent common lead. These are
the exceptions that have been found in the
16 zircons analyzed to date. The Hercyn-
ian zircons must be left out of the present
discussions until studies can be completed
in that area. No explanation can be given

The report of a year ago mentioned the
observation that discordant isotopic ages
were found for zircons from Precambrian
granites from the Cordilleran System in
western United States, where orogenies
have occurred in more recent (Mesozoic)
times. In contrast to the zircons, micas
separated from the same granites appear to
have preserved their ages through whatever
events were responsible for the discordant
zircon ages. One observation was made in
the Appalachian Province to test further

GEOPHYSICAL LABORATORY 171

the possible influence of orogenies on zir-
con age results. The Canada Hill gneiss
from the Hudson highlands has been stud-
ied. The rock is of Precambrian age and
has been folded during the Taconic orog-
eny. From past experience it was ex-
pected that the mica might preserve the
original age of the rock but the zircon
might give discordant age results which
would have some connection with the
orogeny. Actually, the biotite and zircon
ages appear to have been relatively un-
affected by the orogeny, although the
potassium-argon age of the biotite is prob-
ably somewhat low.

Geologic implications of the ages. The
Wichita Mountains biotite was separated
from the Lugert granite, and the zircon
came from a pegmatite in the same granite.
The results from table 11 give an age of
about 500 million years for the granite.
This granite has been called "Precambrian"
in published works by geologists. Recent
discussions with Professor Clifford Merritt,
of the University of Oklahoma, and Dr.
William Ham, of the Oklahoma Geolog-
ical Survey, indicate that some uncertainty
attaches to this stratigraphic assignment.
If the granite is in fact Precambrian, it
would be the youngest Precambrian rock
known and would thus be of considerable
importance to the fossil time scale.

The northern and central Appalachian
Mountains have provided many mica age
measurements grouping at 300 to 350 mil-
lion years, but the zircon and mica from

the Canada Hill gneiss in the heart of the
chain are 1000 million years old, the same
age found for a zircon in the Adirondack
Mountains just to the west and for the
igneous rocks that intrude the Grenville
sediments in Ontario. It would appear that
some parts of the Appalachian Mountain
rock were not affected by the Taconic
orogeny about 350 million years ago.

The Bodom and Rapakivi granites ap-
pear to be 1600 million to 1650 million
years old. They are considered to be of
post-Karelidic age.

ACKNOWLEDGMENTS

J. E. Thomson, of the Ontario Depart-
ment of Mines, and T. C. Phemister, of
Aberdeen University, Scotland, provided
intimate knowledge of field relations in
the Sudbury District and supervised the
collection of samples there. P. W. Gast, of
the Lamont Geological Observatory, as-
sisted us in the collection of the Canada
Hill gneiss and the Storm King granite.
The Lugert granite from the Wichita
Mountains was obtained through the ad-
vice and assistance of W. B. Hamilton and
E. E. Glick, of the U. S. Geological Survey.

Several of the general statements in this
report are based on the data of our col-
leagues at the University of Minnesota, the
Lamont Geological Observatory, the Mas-
sachusetts Institute of Technology, the
University of Toronto, and the California
Institute of Technology, as well as on our
own data.

SIMPLE ABSOLUTE MEASUREMENT TECHNIQUE FOR

BETA RADIOACTIVITY; APPLICATION TO

NATURALLY RADIOACTIVE RUBIDIUM

W. F. Libby

The earlier method of Suttle and Libby
for routine simple absolute assay of solids
did not specifically take account of the fact
that the back-scattered radiation is some-
what softer than the original 3 radiation
and depends in both intensity and softness
on the atomic number of the back scatterer.
This effect has been recognized during the
course of the present research, and agree-

ment with the known standard 3 samples
as furnished by the National Bureau of
Standards and the Oak Ridge National
Laboratory has been improved.

An important new finding is that the
rough surface of a crystalline powder re-
quires a larger correction for geometry
than a smooth surface. This fact has been
shown by direct calculation and proved

172 CARNEGIE INSTITUTION OF WASHINGTON

experimentally. The geometrical effect of
the surface roughness of a powdered solid
is most marked for soft (3 rays, for which
the surface looks much rougher than for
hard (3 rays. Empirically, for powders as
ordinarily prepared, a half-thickness of
about 7 mg/cm2 seems to be a good divid-
ing line; (3 radiations of smaller half-thick-
ness require a geometry factor some 40
per cent larger than those of larger half-
thickness. The geometry factor for hard
(3's is the same one calculated for a smooth
surface.

With these changes, the technique of
measuring the absolute radioactivity of
solids and liquids by placing them in a
cylindrical position around an ordinary
Geiger counter gives results agreeing with
the true absolute assays within 5 per cent.

The technique was applied to the meas-
urement of the half-life of naturally radio-
active rubidium. The value found was
50.7 ±2 billion years, in good agreement
both with that determined by Aldrich,
Wetherill, Tilton, and Davis on old rocks
by the uranium-lead method and with the
latest value determined by Huster, Rausch,
and Geese-Bahnisch by 4tt counting of
very thin deposits of rubidium salts.

This technique should have wide appli-
cation in the development of new uses of
isotopes, particularly in introducing iso-
topes into the ordinary chemistry class-
room.

THE METHOD

Suttle and Libby showed that, under
conditions of cylindrical geometry in
which the sample lies on the surface of a
cylinder whose axis is identical with that
of the Geiger counter used to measure the
radiation, (3 radiations resulting from a
single transition between two nuclear en-
ergy states are absorbed exponentially, even
though the transition may be highly for-
bidden as in K40. The cylindrical position
of the sample is essential for the control of
the very large effects of (B-ray scattering;
it is for this reason that the popular end-
window type of counter with its flat sample

does not give exponential absorption with-
out special orifice windows to control the
scattering. These ordinary counters can be
used for absolute counting only with rigor-
ous controls and exacting disposition of
sample which present considerable diffi-
culty in the usual laboratory; as a result,
applications of isotopes that involve the use
of absolute counting have not been gener-
ally made.

The fact that nearly exponential absorp-
tion curves can be obtained under certain
conditions has long been known. As has
been shown earlier, if the absorption of
the radiation is exponential, the total self-
absorption in an ordinary solid or liquid
sample, which has finite thickness and
therefore can be readily made and handled,
can be easily calculated and the relation
between the absolute disintegration rate
and the observed count rate obtained. It
was assumed that the effects of the self-
scattered radiation would be encompassed
in the geometrical constant G used in the
formula. This point is examined in the
present research.

Let o = absolute specific radioactivity,
disintegrations / min / mg of
sample.
1/Xs — absorption coefficient of the
radiation in the material of
the sample, cm2/mg.
l/Xw — absorption coefficient in coun-
ter-wall material, cm2/mg.

/=wall thickness of counter, in-
cluding the air between the
surface of the sample and the
counter wall, mg/cm2.

G = geometry factor, the ratio of
4tt to the average solid angle
subtended by the inner sur-
face of the cylindrical counter
wall at the sample surface.

x = sample thickness (less than

__ saturation), mg/cm2.

Z = atomic number of the sample
on a weight average basis.

v\ = back-scattering coefficient for
close geometry.

A — area of sample, cm2.

GEOPHYSICAL LABORATORY 173

Then, for a layer of sample at depth
y (mg/cm2) below the top, and of thick-
ness dy, the count rate will be:

dR=(Aa/G)(l + Y])e^/x»-l/x^ dy (1)

or, integrating over the sample thickness,

R=(Ao/G)\s(l + v\)(l-e-x/K)e-l/K (1)'

Seliger has shown that the back-scattered
radiation is of lower energy and softer in
penetrating power than the original radia-
tion, the softening depending on the angle
of scattering as well as on the atomic num-
ber of the back-scattering material. Muller
has very carefully studied the variation of
back scattering with the atomic number
of the material causing the back scattering
under a particular set of geometrical condi-
tions, with applications to analytical chem-
istry in mind. It is generally agreed that
the back-scattered radiation is softer than
the original, that for materials of Z below
15 the factor by which the absorption coeffi-
cient of the back-scattered radiation meas-
ured under 2n conditions is increased is
about 2, and that for larger values of Z
the factor decreases essentially linearly to
about 1.2 at atomic number 90. Therefore,

for ordinary materials in which Z is less
than 15 we can write a new equation for
the relation between the count rate and
the absolute specific activity:

R (cPm) = ^ (1 + \ e-l/x*>)e-VK •

(l-e-*/K) (1)"

On the right-hand side of this equation the
first term in parentheses takes account of
the fact that the counter wall and the air
between the sample and the counter wall
will absorb the back-scattered radiation
more than they do the original. It also of
course takes account of the magnitude of
the back-scattered radiation.

The corresponding formula for larger
values of Z is easily obtained by replacing
the coefficient 2 in this term by the appro-
priately smaller softening factor of the
back-scattered radiation and including in
the exponential term in the parentheses

the value of this new coefficient less 1.
For samples that are thick with respect to
Xs, the last parenthetical factor (the satura-
tion term) disappears, and for material of
Z less than 15, the softening of back scat-
tering can be combined with the geometry
factor into a new factor, G\ to obtain the
formula of Suttle and Libby:

R=(Ao\/G')e~l/K (2)

Bothe and Danziger in theoretical stud-
ies of 3 absorption deduced the main fea-
tures of the whole phenomenology, and

their curve for y\ versus Z agrees well with
the experimental data of others. Since it
is clear from the theory that the scattering

effect for materials of low Z should be
particularly simple, it is not surprising
that for these materials the very simple
equation 2 is nearly as accurate as the more
detailed equations 1' and 1".

An empirical equation for v\ which has
been used in this research and which fits
quite well both the experimental and the
theoretical relations between y\ and Z for
close geometry conditions approaching 2n
in the solid angle subtended by the counter

is

y]=0.65(l-^/35) (3)

The absorption coefficient, 1/X, has been
found to depend on the maximum energy
of the (3 spectrum, E, in the following way :

X (mg/cm2) =55E% (4)

or

/i/2 (mg/cm2) =38E% (4)'

where h/2 is the half-thickness in alumi-
num, and X similarly refers to the recipro-
cal of the absorption coefficient in alumi-
num.

Lerch has shown that X, the reciprocal
of the absorption coefficient, depends on
the average atomic weight, M, of the ab-
sorbing medium. The relation is:

Am = Ao/[1 + (M/100)] (5)

Equation 5 has been used in this research,
and figure 15 shows the adequate degree to
which it fits the experimental data for X
given in table 12.

174 CARNEGIE INSTITUTION OF WASHINGTON

2.00

1.00

tr

50

• CI14 155 kev

hS35 167 kev

oCI36 716 kev

®TI204770kev

xP32 1708 mev

Solid curve:

*l/2

(/l/2)Al

1

®

.^c>

/ At^wtA

V ioo J

1

(-if)

O

100

Atomic weight of absorber

150

Fig. 15. Half-thickness versus atomic weight of absorber.

200

TABLE

12. Absorption

Data

Isotope

Maximum

Energy

of (3

Spectrum,

Mev

Absorbing
Material

Half-Thickness,

mg/cm2

Reciprocal of

Absorption

Coefficient,

mg/cm2

T

Zr93

0.0189

0.060

He
Al
Al
Al
Mylar plastic*
Al*
Mylar*
Al
Al*
Al
Al
Al*
Cu*
Sn*
Pb*
Al
Al
Cu*
Sn*
Al*

0.050
0.35
0.63
1.9

2.2*
2.3*
2.7*
4.85
4.9*
6.09

22

32*

26*

21*

18*

67

84

60*

50*
130*

0.0720
0.506

Sm151

C14

0.0755

0.155

0.91
2.74

g35#

0.167

3.16*
3.3*

Rb87

0.270

3.9*
7.0

Ca45*

0.255

7.1*

Tc"

0.296

8.8

^204
Q]36*

0.762

0.716

32
46*

K40

p32

1.36

1.708

37*
30*
26*
96
122

y90#

2.275

86*

72*

189*

* New data.

GEOPHYSICAL LABORATORY 175

Of course, the fact that the back-scattered
radiation is softened relative to the original
radiation means strictly that the absorption
curve cannot be entirely exponential for
the thinnest layers of absorber, since the
absorption relation must be as given in
equation 1". The verification of this state-
ment is to be seen in figure 16, which
shows actual absorption curves of various
soft (3 emitters taken in the screen wall
counter, which allows the very softest ra-
diation to be measured. Here we see that

*f.U

>

2.0

» \

• \

1.0
0.8

© ^

0.6

-\

\ \Co45 (CoS04)
o\ ^v.

0.4

-

\ \c14 (C0CO3)

0.2

-

\S35 (No2S04)

01

1

0 5

Mylar absorber thickness, mg/cm2

10

Fig. 16. Absorption curves for bare soft (3's.
Sources in screen-wall counter.

there is a soft component that is quickly
absorbed out, and then the long normal
exponential absorption curve is left, which
is the only curve observed with Geiger
counters of usual wall thickness. There-
fore, the absorption is not strictly exponen-
tial for soft |3 emitters. Exponential curves
will be found for hard (3 emitters, for in
this case the back-scattered radiation is lost
in the large percentage of hard radiation
that is present from thick solids. Clearly,
however, for hard radiations and thin
sources the absorption term should not be
strictly exponential.
The absorption curves for the naturally

radioactive element rubidium observed in
the earlier work of Suttle and Libby
showed a nonexponential character in a
way that was difficult to understand at that
time. This, we now see, is the result of
the large Z for rubidium compounds in
increasing the back-scattering coefficient /),
and of the fact that the radiation from ru-
bidium is soft (cf . table 12) . For counters
with walls as thin as 2 mg/cm2, the absorp-
tion curves normally are exponential as

Emax = 255 kev
Substrate Al half- thickness
/cm2
9

n on thick Pb

ly thin on thick Cu

thin on thick Al

0 10 20

Thickness of absorber, mg/cm2

Fig. 17. Ca45 absorption curves in aluminum.

shown in figure 17, in agreement with fig-
ure 18 for the corresponding values of
absorber plus counter wall thickness.

The value of G, the geometry factor in
the full equation 1", is precisely the ratio
of 4n to the average solid angle subtended
at the sample surface by the inner wall of
the counter. If the counter is long relative
to the sample, and if the radius of the inner
wall is p and the radius of the sample sur-
face relative to the counter wire is r, then
it can be shown that

G^n/tn/2-cos-1 (p/r)] (6)

is the equation for the geometrical value
of G.

176 CARNEGIE INSTITUTION OF WASHINGTON

A very important question had to be kindness of Mr. S. A. Reynolds). By

answered in the course of the attempt to means of equations 1', 1", and 2, G was

apply this simple absolute assay technique calculated from the observed count rate R

to soft (3 emitters. It was found that the and the known specific radioactivity o.

results obtained for the specific radioac- The results are given in table 13.

tivity were normally and uniformly about The counter used consisted of a thin

30 per cent low for all isotopes with X metallized plastic cylindrical wall inflated

values below about 10. The effect was the by the counting gas gently flowing through

purely geometrical one of the roughness at pressure slightly in excess of atmos-

of the surface of a crystalline powder. For pheric. The counter had a wall thickness

soft (3 rays for which the range in the solid of 1.82 mg/cm2, and the sample was placed

was less than the thickness of the crystals, around the counter on the inner surface of

the only radiation escaping was from the a plastic cylinder on which was placed a

surfaces of the crystals. With harder (3's, sheet of rubber 1.5 mm thick with a square

however, the entire crystal emitted and the or circular hole of accurately known area

surface-effect roughness disappeared. punched in it. The distance between the

The effect of roughness of the sample is counter wall and the surface of the sample

important. The expression (6) is applica- was 0.27 mg/cm2 of air. By the use of the

ble only to a smooth sample constituting rubber sheet, the sample area was accu-

the wall of a cylinder of inner radius r, rately known. The sample powder was

or a portion of the wall of such a cylinder, placed in the recess of the rubber sheet

Consider a normal crystalline powder con- and smoothed with a spatula. Under these

sisting of cubes 50 \x on edge. If the density conditions the counter wall radius, p, was

were 2 gm/cm3, then the cube edge would 1.5 cm and the inner sample surface radius

correspond to an x value in equations 1' was 1.8 cm.

and 1" of 10 mg/cm2. Therefore, we see The larger G factor for soft radiation as
immediately that, for hard 3 emitters with shown in table 13 was further established
X values larger than 10 mg/cm2, the sur- experimentally by making a finely divided
face of the solid powder of randomly ori- Na2COs which had C14 in it. This was
ented cubes would appear to be smooth and done by powdering Na2C#03*H20 and
the powder would have a G value close to then dehydrating it at low oven tempera-
that for a smooth surface as given by tures so that sintering did not occur. Under
equation 6. For soft 3 emitters, on the these conditions the value of G was the
other hand, only the surfaces of the crystals smaller one for hard 3's rather than the
can emit, and the surface therefore must larger one found for the same salt before
appear rough. The fact that roughness dehydration, as shown in table 13. In all
causes a reduction in the total outward other cases for soft 3 radiations the G was
flux of radiation relative to that from a true higher, and the average for all soft 3's
smooth cylindrical surface of the same ma- with X values less than 10 was 3.9 versus
terial at the same specific radioactivity may 2.72 for the hard 3's. Table 13 gives the
not be obvious, but detailed calculation for final G values as determined according to
various likely powders, such as randomly the three equations 1', 1", or 2.
oriented cubes or hexagonally packed It is clear, of course, that the necessity
spheres, shows that this is a general result for deciding which value of G applies to
and that the magnitude of the effect agrees a particular solid sample (and 3 radio-
with the results on the various 3-radiation activity) being measured is a point of con-
standards obtained from the National cern. How can one tell? For 3's with X
Bureau of Standards (through the kind- values well above 10 mg/cm2, the rough-
ness of Dr. W. B. Mann) and the Oak ness necessary for the larger value of G will
Ridge National Laboratory (through the be easily visible and easily destroyed by

GEOPHYSICAL LABORATORY 177

grinding with a mortar and pestle. There-
fore, for this class of radioactivity suc-
cessive measurements after grinding will
bring the count rate to a constant high
value independent of the degree of fine-
ness of the solid which is characteristic of
the smooth surface and the lower G value.
For the softer (3's, however, it is neces-
sary to do the opposite — to grow the crys-

The procedure for converting the (3
standards, which were solutions of very
high specific activity, to solid form for
measurement was to add a solution of an
appropriate salt to a known volume of the
standard, mix, evaporate or precipitate
chemically, and grind the resultant solid.
Sometimes it was difficult to obtain solids
that were chemically identical with the

TABLE 13. Experimental Geometry Factors, G

Isotope

Substrate

mg/cm2

Fully Corrected

for Back

Scattering and

Softening,

equation 1"

Partly
Corrected,
equation V

No Back-
Scattering
Correction,
equation 2

P32 Na2S04 129 2.65±0.02

(NH4)H2P04 141 2.87 ±0.03

K40 K2S04 98 2.73±0.06

CI36 NaCl 45 2.85±0.03

AgCl 34 2.74 ±0.03

BaCl2 34.4 2.61 ±0.03

Tl204 T1C02H 27.1 3.09 ±0.04

Average
for A>10 2.72

Ca45 CaC03 7.5 4.12 ± 0.04

CaS04-2H20 7.9 4.18 ±0.04

CaO 7.0 4.28 ±0.04

CaS04 7.4 3.73 ±0.03

S35 Na,S04 3.5 3.46±0.3

BaS04 3.0 4.4 ±0.2

C14 CaC03 2.9 3.4 ±0.1

Na2C03-H20 3.1 4.15 ±0.16
Very fine

Na2C03 3.0 2.55 ±0.08

Average
for X<10
(omitting fine Na2COs) 3.9

2.73 ±0.02

2.85 ±0.03
2.83 ±0.06
2.97 ±0.03

2.86 ±0.03
2.71 ±0.03
3.11 ±0.05

2.82

4.41 ±0.05
4.47 ±0.04
4.60 ±0.04
4.01 ±0.03
3.78 ±0.3
5.0 ±0.3
3.76 ±0.1
4.56 + 0.18

2.78 ±0.09

4.2

2.33 ±0.02
2.02 ±0.03
2.28 ±0.05

2.43 ±0.03
1.99 ±0.03
1.87 ±0.03
2.01 ±0.05

2.20

3.72 ±0.04
3.76 ±0.04
3.71 ±0.04
3.28 ±0.03
3.16 ±0.3
3.5 ±0.3
3.16±0.1
3.95 ±0.1

2.44 ±0.09

3.5

tals larger and larger by sintering or other
device, and thus to reach a constant count
rate independent of crystalline size. Any
doubt can be settled by a cursory examina-
tion with a microscope, the relative mag-
nitude of the crystal size and X being
borne in mind. It appears that the soft 3
geometry factors for various powders are
essentially the same, as can be seen in table
13, though there is some evidence of scat-
ter, which could be due to the size or
shape of the particular crystals.

radioactive molecules, and attempts were
made to use substitutes with which the
radioactive species was likely to form
mixed crystals. For the harder (3 radia-
tions, the requirement that mixed crystals
be formed seemed to be less necessary. For
example, radioactive phosphate containing
P32 (Aai= 122 mg/cm2) was measured on
Na2S04 powder. It seems unlikely that
any substrate not chemically identical can
be used in the case of soft 3's.

178 CARNEGIE INSTITUTION OF WASHINGTON

RESULTS

In table 14 are shown the results of the
application of the method to a series of
(3-radiation standards furnished by the Na-
tional Bureau of Standards and the Oak
Ridge National Laboratory. From these
data it appears that the method is good to
about 5 per cent. Equation 2, which is the
simplest, does nearly as well, except for
the large atomic numbers, as the more
complete equations 1' and 1" '.

was applied. The results are given in table
15.

In calculating the final error for the
half-life as determined on the three rubid-
ium samples, the average deviation of
3.3 per cent for equation 1" as applied to
soft (3's with a G value of 3.9 was used.
The average deviation of 9.5 per cent for
a single determination as given in table
14 was divided by the square root of the
number of determinations to determine

TABLE 14. Results for Standard Sources
(Per cent deviation) calculation

Isotope and
Substrate

Fully Corrected,
equation 1"

Partially
Corrected,
equation V

Uncorrected for

Back Scattering,

equation 2

P32 Na2S04 +1.1 +0.7

(NH4)H2P04 + 5.6 +1.2

(NBS)

CI36 NaCl +4.7 +5.4

AgCl +0.7 ave.3.9 +1.5 ave.4.4

BaCl2 -4.0 -3.9

(ORNL)

T1204 T1C02H + 13.6 + 10.4

(NBS)

Ca45 CaC03 +5.6 +5.1

CaS04-2H20 +7.1 +6.5

CaO +9.6 +9.5

CaS04 -4.4 -4.5

(ORNL)

S35 Na2S04 -11 ave.9.5 -10 ave.9.8

BaS04 +13 +20

(NBS)

C14 CaC03 -13 -10

Na2C03-H20 +6.4 +7.4

(NBS)

+ 5.5
-8.0

+ 10.5

-9.4 ave.8.3
-15

-8.7

+ 6.5
+ 7.6
-6.1
-6.3

-8.6 ave.7.8
+ 0.3

-9.7
+ 13

APPLICATION TO NATURALLY RADIOACTIVE
RUBIDIUM

Aldrich, Wetherill, Tilton, and Davis
compared the ratio of radiogenic Sr87 to
Rb87 found in several minerals differing
in rubidium content in rocks of known
age as determined by the uranium-lead
method. They calculated the half-life of
natural radioactive Rb87 to be 50 ±2 billion
years. Strassman and Walling found 63
billion by a similar method. Because of the
importance of this determination to geo-
chronology, the technique described above

the average deviation of the mean from
the true value. The agreement among the
three different rubidium samples in table
15 indicates that the error of the deter-
mination is the error in the determination
of G, which as explained should be 3.3
per cent.

GENERAL APPLICATIONS

Equation 2 certainly is simple enough
for use in high-school laboratory experi-
ments. In this way isotopes of real chemi-
cal interest, convenient lifetime, and low

GEOPHYSICAL LABORATORY 179

enough specific activity to be completely
safe can be introduced into the ordinary
high-school chemistry course. Among these
isotopes are C14, CI36, S35, and Ca45. Thus
the radioactive forms of acetic acid, hydro-
chloric acid, sulfuric acid, and the calcium
salts can be placed on the reagent shelves
and the label can carry the specific activity,
so that the students by using a known
volume can introduce a known amount of
radioactivity, and by subsequent absolute
counting of the various solids produced in
the experiment can calculate an isotopic
balance to compare with the ordinary ma-
terial balance. This technique makes pos-
sible the ready application of isotopic dilu-
tion techniques.

In industrial applications the possibility
of keeping an analytical check on a known
amount of C14 added in appropriate chemi-
cal form at the beginning of, or during, an
industrial organic chemical process cer-
tainly affords many opportunities for con-
trolling the process. The labeling of a
particular constituent of the crude feed
for an oil refinery for a fixed period would
make possible a detailed examination of
the flow rates and patterns throughout the
plant, e.g. the contribution of this constit-
uent to coke in the catalytic crackers, the
completeness of the burn-off of the coke
from this constituent in the burn-off cycle,
or the general holdup in various stages in
the plant.

TABLE 15. Determination of the Half-Life of Rb87

Sample K n Specific Activity, Half-Life,

mg/cm2 dpm/mg billion years

RbCl 3.85 0.384 53.1 ±1 49.6 ±1

Purified by Dr. Suttle

RbCl 3.85 0.384 51.7±0.8 51.0±0.8

Johnson, Mattie & Co., London;

spectrographically pure

Rb2COs 3.83 0.361 54.0±0.5 51.1 ±0.5

Average 50.7±2

PALEOBIOCHEMISTRY

EFFECTS OF ULTRAVIOLET LIGHT ON THE
"PRIMITIVE ENVIRONMENT"

P. H. Abelson

The origin of life is a topic of tran-
scending interest which has drawn the at-
tention of many investigators. Many of
them have sought to isolate a part of the
problem both crucial and capable of solu-
tion— the synthesis of biological building
blocks, especially amino acids, from simple
chemicals of the environment. To accom-
plish these reactions a number of energy
sources have been employed, including a,
3, and y radiation and electrical discharges.
A variety of test chemicals has been em-
ployed, many of which make convenient
experimental objects but could hardly have
existed in substantial quantities on the
primitive earth.

It is difficult to be certain concerning
processes that might have occurred three
billion years ago. It is feasible, however,
to set some limits on the areas of permis-
sible speculations. An analysis of the na-
ture of the primitive environment could
be expected to point up the importance of
a few key compounds and reactions. Study
of the effects of energy on these substances
might then disclose reactions of major im-
portance on the primitive earth.

The relative scarcity of the gases neon,
argon, krypton, and xenon in our present
atmosphere is a most significant phenom-
enon. The abundance curves of isotopes
of the various elements are fairly well
known, and neon has been observed as
an important constituent of stellar atmos-
pheres while being scarce on earth. Brown

180

CARNEGIE INSTITUTION OF WASHINGTON

and Suess have estimated that neon is
present on our planet to an extent only
10~10 of probable cosmic abundance, and
that similarly argon, krypton, and xenon
are relatively absent. We do not know the
nature of the processes by which the earth
was formed, whether these inert gases
were lost during the accumulation period
or subsequently; but at any rate they were
lost, and it seems reasonable that such
other volatile constituents as hydrogen, ni-
trogen, methane, and carbon monoxide
would also have been lost at the same time.
Rubey has calculated the amounts of
volatiles that have appeared in the atmos-
phere, hydrosphere, or biosphere since
weathering first began. He advances many

TABLE 16. Inventory of Organic Matter

oxygen or oxidation to match the reduced
substances in the sediments. It is possible
that this unaccounted-for oxygen was con-
sumed in the oxidation of carbon monox-
ide and hydrogen issuing from volcanoes.
The reducing nature of these gases is a
consequence of the physical-chemical equi-
libria of water and carbon dioxide with
the reduced iron compounds and possibly
other substances.

A calculation shows that 3 atm of hydro-
gen would be in equilibrium with 1000 atm
of water at 1200° K, and that at the same
temperature 1 atm of hydrogen would be
in equilibrium with 100 atm of water

TABLE 17. Inventory of Oxygen

O,

Grams

Element

Grams

Carbon 68xl02°

Oxygen 25 X 102°

Hydrogen 9.6 X 1020
Oxygen needed to

burn to C02 + H20 235 X 1020

powerful arguments that these volatiles
were not originally present on the surface
of the earth but came from the interior of
the earth through a gradual degassing
process. Of interest to the present argu-
ment are data that he has provided on
the inventory of organic matter in sedi-
ments, shown in table 16. These permit
construction of an oxidation-reduction bal-
ance. To measure the reducing side of the
ledger, the amount of oxygen required to
burn this organic matter to water and
carbon dioxide can be calculated. The
amount of oxygen in the atmosphere at
present and that consumed in the oxida-
tions of ferrous to ferric iron and sulfur
to oxidized sulfur may be noted in table 17.
Considerable uncertainty attaches to the
amount of oxygen consumed in oxidation
of sulfur, since the relative proportions of
the original forms of this substance are
not known; the value quoted represents
an upper limit. There is not sufficient

In atmosphere
Consumed in oxidation

FeO-^Fe2Os
Consumed in oxidation

S->SOs

Total accounted for
Unaccounted for

Total

12X1020

14X1020

41X1020

67 X1020
168 X1020

235 XIO20

along the wiistite, fayalite, and magnet-
ite join. The equations governing equi-
libria involving carbon dioxide, carbon
monoxide, water, and hydrogen in the
presence of fayalite, wiistite, and magnet-
ite may be combined to eliminate the
common oxygen component. As a result
there can be obtained an equation involv-
ing the equilibrium of water, carbon di-
oxide, hydrogen, and carbon monoxide.
If the volatile constituents that have ap-
peared at the surface were in equilibrium
with one another, and with a basaltic
crust, the original chemical form of some
of the constituents can be estimated, tak-
ing into account the oxidation-reduction
balance. Results can be noted in table 18,
where it is apparent that, since water is
the major volatile that has been released,
hydrogen was the major reduced substance
accompanying such a mixture. Later it will

GEOPHYSICAL LABORATORY 181

be seen that a highly important feature of
the composition of these volatiles is the
comparative amounts of carbon dioxide
and hydrogen.

Relatively soon after the earth was
formed it cooled to about its present tem-
perature. Hence, only a very small pro-
portion of water remained in the atmos-
phere. The oldest rocks that have been
dated contain the same minerals as those
found today. Weathering of these rocks
led to sediments similar in composition
to those formed recently. There are argu-
ments and evidence making it reasonable
to assume that the pH of the primitive
oceans was not much different from that
today. As carbon dioxide was liberated
most of it dissolved in those waters, leav-

TABLE 18. Composition of Volatiles on the
Basis of Equilibrium Conditions

H20. A wide variety of compositions
yielded mixtures of amino acids.

Analysis, however, shows that corpuscu-
lar radiation and electrical discharge are
relatively limited as sources in comparison
with the energy coming in sunlight. In
table 19 is shown the relative energy avail-
able per year. It can be noted that energy
from sunlight far exceeds that from other
kinds of sources, lightning is relatively in-
significant, and cosmic rays are of very
little consequence in terms of their energy
content. Levels of radiation due to natural

TABLE 19. Energy Input to Earth

Source of Energy

cal/cm2/yr

Sunlight
Lightning
Cosmic rays

260,000

0.9

1.5 XlO"3

Gas

Moles

H20
H2

co2

CO

N„

920 XlO20
10.3 XlO20

21 XlO20
0.24 XlO20

1.5 XlO20

TABLE 20. Rocket Measurements by the Naval

Research Laboratory of Energy Incident

on Earth

A

Energy,
watts/cm2/u

ing hydrogen, nitrogen, and carbon mon-
oxide as the principal constituents of the
atmosphere. This mixture was altered fur-
ther through the action of various forms of
energy.

In attempts to set up models for produc-
tion of compounds in the primitive en-
vironment, several different kinds of en-
ergy sources have been tried. Earlier at
the University of California the cyclotron
a-particle beam was employed. Recently, |3
and y radiation have been used as sources.
The experiments of Stanley Miller have
demonstrated the production of amino
acids in an environment of methane, am-
monia, and water as a result of the action
of an electric discharge. Abelson has con-
firmed these findings and extended them
by testing the effects of electrical discharges
on twenty other mixtures of gases, includ-
ing CO, N2, H2, H2O; C02, NH3, H2,

4500
3000
2800
2600
2400
2200

0.22

0.061

0.024

0.013

0.0058

0.003

radioactivity are variable, but in many
areas are about equivalent to those due to
cosmic rays.

Only part of the sun's energy is in
the ultraviolet region, but this portion is
particularly effective in causing chemical
transformations. Rocket experiments car-
ried on by the Naval Research Laboratory
have measured the energy incident upon
the top of the atmosphere with results
shown in table 20. From these data it can
be estimated that 468 cal/cm2 of wave-
lengths shorter than 2540 A fall on the
top of the atmosphere annually. Such
radiation is absorbed by carbon dioxide.
Simultaneously 19 cal/cm2 of wavelengths

182 CARNEGIE INSTITUTION OF WASHINGTON

shorter than 1800 A strike the atmosphere.
This radiation can be absorbed by water.
Carbon monoxide exhibits some slight
band absorption at 2056 A, but major in-
teraction begins at 1546 A. Only 5.5 cal/
cm2/yr are found in shorter radiation.

Another way of assessing the possible
role of ultraviolet radiation is to match an-
nual production of chemicals and radia-
tion. Such a comparison may be noted in
table 21, where it is evident that the energy
available far exceeds the annual produc-
tion of chemicals.

Ultraviolet light decomposes carbon di-
oxide to carbon monoxide plus atomic
oxygen. In the presence of hydrogen,
atomic oxygen reacts to form OH + H.
OH in turn reacts with hydrogen to form

TABLE 21. Annual Production of Chemicals
and Energy per Square Centimeter

Quanta

Moles

X<2900

1.4 X1022

co2

8 XlO16

A < 2540

2.3 xlO21

CO

9 XlO14

X= 1100-1345

3X1018

N2

6 XlO15

H2

4 XlO16

water plus more atomic hydrogen. One
may speculate that, as carbon dioxide of
the primitive atmosphere was used up,
being split into carbon monoxide plus
oxygen, the partial pressure of carbon di-
oxide was maintained by the carbonate
buffering system of the oceans. As long
as there were large amounts of molecular
hydrogen in the atmosphere, the oxygen
liberated by the breakup of carbon dioxide
would be converted into water. However,
after most of the hydrogen had been used
up, small amounts of free oxygen would
appear, and they would recombine with
carbon monoxide or any other reducing
substance. Carbon dioxide must have acted
as a sort of buffer in the oxidation-reduc-
tion system, guaranteeing that the atmos-
phere could never have been very reducing.
It seems impossible to visualize any great
concentration of substances such as meth-
ane present together with carbon dioxide.

The values given by Rubey indicate that
in the last 3 billion years considerably more
carbon dioxide has been liberated than is
needed to supply oxygen to react with all
the hydrogen or other reducing substances.

In an atmosphere consisting predomi-
nantly of carbon monoxide, nitrogen, and
hydrogen, the action of short ultraviolet
light could lead to a number of additional
substances, including ammonia and hydro-
gen cyanide. Under any reasonable as-
sumptions of the nature of the primitive
oceans, all ammonia and hydrogen cyanide
would be found in the aqueous phase with
almost none remaining in the atmosphere.
Carbon monoxide is slightly soluble in
water, being slowly converted into for-
mate. The interactions of ultraviolet light
with the atmosphere and subsequent ab-
sorption of products in the oceans thus
would modify the composition of both
atmosphere and oceans. Later in this re-
port some effects of ultraviolet light on the
substances dissolved in the oceans will be
considered.

In enumerating the kinds of organic
compounds, formation of which would be
crucial to the creation of building blocks
essential for life, one is impressed by the
possible important role of the aldehydes.
This class of compounds could serve as
important materials in the building-up of
long carbon chains through reactions that
can occur in the aqueous phase. No other
set of organic substances can perform in
quite such a manner. Strecker in 1850
found that, if formaldehyde, hydrogen
cyanide, and ammonia are all simulta-
neously present in solution, glycinonitrile
is formed. This substance subsequently is
hydrolyzed by water to form the amino
acid glycine. The Strecker synthesis is not
limited to formaldehyde but may be car-
ried out with other aldehydes. It is inter-
esting that this type of synthesis gives rise
to a-amino acids, which are the principal
type of amino acids used in biological
processes.

It seemed important to determine
whether ultravolet light acting on for-

GEOPHYSICAL LABORATORY 183

mate could lead to the extremely impor-
tant substance formaldehyde. In alkaline
solutions formate has little absorption at
wavelengths longer than 2540 A; above
2700 A it is virtually transparent. It was
recalled that iron is a ubiquitous constitu-
ent of the crust, and that, if there was
little or no oxygen in the atmosphere,
appreciable concentrations of ferrous iron
might be present in solution. AIM solu-
tion of formate containing 0.004 M ferrous
iron at a pH of 8.3 was irradiated, in an
evacuated vessel, by ultraviolet light of
wavelength 2536 A. After 1 hour, a non-
condensable gas was noted, and formalde-
hyde could be detected in the solution. At
the end of 3 hours, approximately three
times as much noncondensable gas was
present but the same amount of formalde-
hyde was determined. In order to obtain
a better measure of the rate of formation
of formaldehyde, experiments were per-
formed using a trapping mechanism de-
signed to take formaldehyde out of the
scene of destruction. For this purpose the
Strecker synthesis was employed. Solu-
tions 0.4 M in formate, 0.4 M in ammo-
nium hydroxide, 0.2 M in sodium cyanide,
and 0.0016 M in ferrous sulfate were ir-
radiated for periods of an hour to several
days, and the resulting glycinonitrile was
hydrolyzed and determined as glycine.
Relatively large quantities of this amino
acid were found.

In another series of experiments the
hydantoin reaction was employed as the
trapping mechanism. For this purpose a
solution 0.1 M in ammonium carbonate,
0.1 M in formate, 0.1 M in sodium cyanide,
and 0.001 M in ferrous sulfate was irradi-
ated for 2 days with 2536 A radiation. The
resulting hydantoin was hydrolyzed with
NaOH, and glycine was isolated. A 10
per cent yield based on formate was ob-
tained.

An important series of reactions is the
condensation of aldehyde and cyanide fol-
lowed by hydrolysis. The kinetics of the
first of these two reactions has been studied
using initial concentrations of 10~3 M cya-

nide and 10~4 M formaldehyde. Some
results are shown in table 22. The reaction
proceeds rapidly until the concentration of
formaldehyde drops to 10~7 M or 3 parts
in 109. The effect of pH on the reaction
was also investigated. The optimum pW
for the reaction was found to lie in the
range 8 to 9.5. The reaction thus proceeds
best at a pYL near that of sea water. Hy-
drolysis of glycolonitrile to glycolic acid
is an irreversible reaction which progresses
at a moderate rate at 20° C. With the
appearance of glycolic acid the stage is set
for further synthesis. Irradiation by ultra-
violet light again could yield an aldehyde;

TABLE 22. Reaction Kinetics at 20° C

H2
HCN + CH20<=±HOC - C = N

10~3 M NaCN, 1.05 X 10-4 M CH20, pH 8.1

Time,

CH20,

Time,

CH20,

min

10-4mole

min

10~4 mole

0

1.05

6.0

0.100

1.1

0.64

7.0

0.078

2.1

0.433

9.0

0.035

3.1

0.320

10.0

0.025

4.0

0.207

12.0

0.010

5.0

0.147

20.0

0.001

evidence for this step is discussed later.
Condensation with cyanide would result in
a dihydroxynitrile, glyceronitrile. A series
of condensations followed by hydrolysis
could lead to 6-carbon compounds very
similar to carbohydrates.

Several experiments have tested the ef-
fect of ultraviolet light on glycolic acid.
The irradiation was carried out in a so-
lution containing hydrogen cyanide, am-
monia, and a small amount of ferrous iron
in addition to the glycolic acid. No at-
tempt was made to attain a maximum
yield of amino acids. Nevertheless hydrol-
ysis of the nitriles formed revealed a yield
of amino acids of 15 per cent, based on
the original amount of glycolic acid. Ex-
amination of the products by paper chro-
matography produced a rather interesting
result. Serine, the expected product, was

184 CARNEGIE INSTITUTION OF WASHINGTON

present, but alanine and glycine were also ing which creatures were forced to develop
identified, as well as other substances. Fur- photosynthesis suddenly to survive. Rather
thermore, substantial quantities of more a sharp competition for a limited food sup-
complex colored substances were formed, ply favored those creatures that could de-
These were opaque to ultraviolet light and velop alternative energy sources such as
relatively easily adsorbed on talc. In na- photosynthesis.

ture, complex organic substances would be One of the products of this research was
adsorbed and ultimately buried in sedi- an observation of the effect of ultraviolet
ments. Here is a possible mechanism for light on ferrous iron. Many who have
production of carbonaceous, organic sedi- considered Precambrian geology have cited
ments that does not invoke the action of the widespread occurrence of ferric iron
living creatures. as evidence that in those times an oxygen
Others who have speculated on the ori- atmosphere was present. Recent experi-
gin of life have postulated that a thick ments have brought this assumption into
organic soup was formed and that when question. Dilute ferrous iron at about pH
living creatures were available they quickly 8 was placed in a silica flask, which was
depleted this broth. Once the broth was subsequently evacuated and subjected to
gone the creatures had no alternative but ultraviolet light of wavelength 2536 A.
to develop photosynthesis in order to sur- Hydrogen was formed, and a precipitate
vive. Consideration of the model that has of ferric iron noted. Search of the litera-
been presented here does not lead to quite ture then revealed that similar effects of
these same conclusions. If all the available ultraviolet light on ferrous iron had been
hydrogen were used up in processes lead- demonstrated by Chastaing in 1877 and
ing to formate, the maximum concentra- that a threshold of 2900 A had been estab-
tion of this substance would be 0.6 M; lished for the reaction,
even allowing for the reducing capacity of Others have pointed out that ozone
sulfur and ferrous iron, the figure would which today absorbs ultraviolet light of
rise to no more than 0.8 M as a maximum, wavelengths shorter than 3000 A prob-
The concentration of organic substances ably did not appear in quantity until living
in the aqueous phase after a time would creatures invented photosynthesis. Com-
actually be considerably less as the longer, paratively large amounts of energy in the
more complicated organic compounds were band 2540 to 2900 A were hence available
precipitated or adsorbed out of the system at the surface to act on the new ferrous
while relatively little of the diluting water iron released by weathering each year. No
was lost. It can be conceived that in the claim is made that these experiments prove
early stages of the earth as the oceans that ferric iron in the Precambrian was
gradually grew in magnitude there was a due to the action of ultraviolet. On the
relatively constant production of various other hand, it seems clear that the presence
kinds of organic molecules, some of which of ferric iron is not necessarily proof that
were lost by precipitation. Owing to this either oxygen or photosynthesizing organ-
loss and to the continuing addition of isms were present at the time the ferric
water to the system, the concentration of iron was laid down,
organic molecules would actually diminish The result of this experiment with iron
somewhat in spite of the constant rate of suggested still another, employing a solu-
production. Under this picture the first tion of sodium sulfite at pH 8 together
living creatures which could extract or- with a small amount of ferrous iron. Again
ganic compounds from the medium had the solution was placed in a silica vessel,
available an annual supply of organic which was evacuated and irradiated with
molecules, and hence there was not neces- 2536 A ultraviolet light. Sulfate appeared
sarily a catastrophic starvation period dur- in the solution. It would seem that the

GEOPHYSICAL LABORATORY 185

occurrence of sulfate in the Precambrian is
not necessarily diagnostic of oxygen in the
atmosphere.

THERMAL DEGRADATION OF AMINO ACIDS
/. R. Vallentyne

Approximately 1017 g of organic matter
is synthesized annually on the earth's sur-
face by plants. Although most of this is
recycled in the biosphere, a small fraction
becomes buried in sediments and soils,
thus leaving the biochemically active part
of the biosphere. Once sedimented, this
organic matter is subjected to further de-
composition, dependent on subsequent bio-
logical and chemical attack. If the fossil
compounds occur in materials that are
protected from biological attack (such as
a calcite shell), then to a first approxima-
tion the system could be treated in terms
of chemical kinetics. It is now well known
(Year Book 53, p. 99) that amino acids
occur in fossils as old as the Devonian.
It is also known that certain amino acids
are geologically less stable than others, for
example serine, threonine, and phenyl-
alanine as compared with the more stable
glutamic acid, glycine, and alanine.

From a knowledge of the relative de-
composition rates of amino acids under de-
fined conditions in the laboratory, coupled
with analyses of fossil materials, geological
temperatures of preservation might be in-
ferred if age is known. Two difficulties
of interpretation should be clearly stated,
however. In the first place, the medium
of preservation (for example, calcite, bone,
or shale) must be expected to influence
the decomposition rates of amino acids.
For the exact interpretation of data from
fossil materials kinetic experiments must
be conducted in media that simulate their
geological counterparts. For example, the
decomposition rates of amino acids in
modern shells could be determined by an-
alyzing for amino acids before and after
a given treatment. Second, since kinetic
experiments are limited in duration to a
few years at the most, a high degree of

extrapolation is required to obtain esti-
mates of decomposition rates at low tem-
peratures. Extrapolation is always danger-
ous. Extremely sensitive methods for the
analysis of decomposition products aid in
extending experimental values to low tem-
peratures, but do not overcome the diffi-
culty completely. We will probably have
to rely on the type of argument used by
geochronologists in measuring age : if two
independent measurements lead to a pre-
diction of a single age, that value is more
reliable than the same value determined by
either method alone. Thus, if a tempera-
ture determination based on amino acids
agrees with another based on a different
set of compounds, a qualitative feeling of
confidence results.

A preliminary attack on some of these
problems has been made by studying the
decomposition rates of amino acids in di-
lute aqueous solution. These data will
serve as a base line with which future data
(with the medium as the variable) can be
compared. Decomposition rates for a va-
riety of reactions can be expressed in terms
of the Arrhenius equation:

h = se-AHa/RT

The specific reaction rate constant, \, is the
reciprocal of t±/e, where ti/e is the time
(in seconds) required to decompose 63
per cent of the initial amount of the com-
pound, at a given temperature. The fre-
quency factor, s, depends on the type of
reaction. It has a value of 1013 sec-1 for
unimolecular reactions and much lower
values for bimolecular reactions. In log-
arithmic form the Arrhenius equation is:

log h/e = AHa/2303RT-\og s

An increase in AHa, the activation en-
ergy, will thus markedly increase ti/e, but
an increase in s leads to a decrease in ti/e.
For a-alanine (Year Book 53, p. 101),

^ = 3X1013 e-^000/RT

The comparable equation for phenyl-
alanine (0.002 M solution) is:

^ = 1.7xl08^-3O'8OoABr

186 CARNEGIE INSTITUTION OF WASHINGTON

with both sets of data shown graphically
in figure 18. At temperatures below 280°
C, phenylalanine decomposes more quickly
than alanine under the same conditions.
It can be predicted that after 5 to 10 mil-
lion years' storage at 30° C most of the
phenylalanine will have been degraded.

portance of extending the approach to
other amino acids.

That water is involved in the decomposi-
tion of phenylalanine under the conditions
employed is evidenced by the fact that the
decomposition follows first-order kinetics
over the concentration range 2 X 10~4 M to

(lO,0yr)-

10'

(I08 yrh

J4
(lO%r)
o

I !0'2
I" (lO4 yr)"

10"

•5 x IOy yr

O o a " Alanine

• • Phenylalanine

I x 10° yr

1 io»

fe (lO2 yr)-

"- I08
"tf 0 yO-l

10*

10'

10 :

30

50

75 100

Temperature, °C

150

200

250

300

Fig. 18. Time required for 63 per cent decomposition of a -alanine and phenylalanine (in dilute
aqueous solution) at various temperatures. The scale of the horizontal axis is in units of the recip-
rocal of the absolute temperature.

Abelson (Year Book 53, p. 101) has noted
that recent shells of the clam Mercenaria
mercenaries contain both alanine and phen-
ylanine, whereas in Miocene shells of the
same species alanine is present but phen-
ylalanine is lacking. This qualitative agree-
ment between analytical and (extrapo-
lated) experimental data suggests the im-

4xl0-3 M, with the value of the specific
reaction rate constant, \, independent of
concentration. Only the Arrhenius equa-
tion reveals that the reaction is bimolecu-
lar. Since water is involved in the decom-
position of phenylalanine, it may be im-
portant to record the water contents of
fossil materials before analysis.

ORE MINERALS

Results important to our understanding from systematic laboratory studies of the
of mineral associations found in nature subsolidus relations among some of the
have been obtained during this past year more common ore minerals. The investi-

GEOPHYSICAL LABORATORY 187

gations of the stability relations of the most
common sulfide, pyrite, and of the impor-
tant ore mineral covellite have been com-
pleted. Studies of the composition of pyr-
rhotites formed in equilibrium with pyrite
at various temperatures and pressures are
near completion. In addition, the phases
and solid solutions occurring in the CoAs2-
NiAs2-FeAs2-As system, as well as those
in the Fe-S-O system, are nearly finished.
Studies are also progressing on the Cu-
Fe-S, Fe-Ni-S, Fe-Zn-S, and Fe-S-Se
systems.

Some of the geological thermometers
based on sulfide assemblages appear to be
well established as useful tools. The FeS-
ZnS system has been employed in measure-
ments on more than a hundred ore de-
posits. As studies of more systems are
completed in the laboratory, it should be
possible to obtain cross checks between
temperatures determined by observations
of different assemblages occurring in the
same deposit.

Detailed studies of ore deposits, aided
by such laboratory tools, may in the future
materially assist the field investigator in
his interpretation of mineral associations
and textures and will enable him to deter-
mine temperature, and possibly pressure,
of formation of a host of mineral assem-
blages. These data are essential if attempts
are to be made to estimate the composi-
tion of the solutions that transported the
ore minerals to their site of deposition and
to determine the direction of movement
of these solutions.

THE Fe-S SYSTEM

Stability relations of pyrite (Kullerud,
Yoder) . The upper stability limits of py-
rite, FeS2, have been carefully determined,
and the role of this binary compound in
the Fe-S system can now be specified.
Interpretation of the significance of the
existence of an invariant point in which
the phases pyrrhotite, pyrite, liquid, and
vapor are stable unites many seemingly
unrelated experiments. For example, pre-
vious work at pressures less than 1 atm
has been related to new studies at high

pressures. A critical theoretical evaluation
of the P-T-X diagram of the system Fe-S
has brought forth the advantages and limi-
tations of the various types of experimental
techniques employed. These theoretical
deductions and experimental techniques,
some applied for the first time to sulfides,
permit a new understanding of many of
the ore-forming mineral assemblages.

The revised upper stability curve of py-
rite, FeS2<^Fei-;rS + L, is given in figure

75.000 -

_ 30,000 -

15.000 -

5,000 -

400

500 600 700 800 900
Temperature °C

1000

1100

Fig. 19. Revised upper stability curve of py-
rite, FeSo^Fe^S + L.

19. It is to be noted that the reaction does
not involve a gas phase. The experimen-
tally determined points are 815° C, 75,000
psi; 770° C, 30,000 psi; 755° C, 15,000 psi;
748° C, 5000 psi. The curve terminates
at the invariant point c, which lies at 743°
± 3° C and about 180 psi. All these points,
with the exception of c, were determined
using collapsible gold tubes (see Year
Book 55, p. 181). Point c, where the four
phases pyrrhotite, pyrite, liquid, and vapor

188

CARNEGIE INSTITUTION OF WASHINGTON

coexist, was determined in evacuated, rigid,
silica-glass tubes. Since some of the experi-
mental methods restrict the possible prod-
ucts, it was necessary to consider the rela-
tion of these products to the Fe-S system.
In figure 20 is given the schematic
pressure-temperature diagram for the Fe-S
system as deduced from published and
new data using principles based on Gibbs'
phase rule. Because of the large pressure
range, only a schematic diagram is pos-
sible, although an attempt was made to
maintain the temperature scale where pos-

"the minimum temperature of liquefac-
tion," of the primitive system Fe-S. The
four univariant curves originating from a
can be deduced by means of the principles
outlined by Morey and Williamson (1918).
One of these, the curve labeled L+V + S,
terminates at the invariant point b, which
has been found by experiment to lie at
115° C and 0.018 mm Hg (Tuller, 1954).
Point b is the triple point of monoclinic
sulfur. Portions of the curves L+S and
V+S have been investigated previously,
and the curve L+V, which terminates at

800 1000

Temperature, °C

1800

Fig. 20. Schematic pressure-temperature diagram for the Fe-S system. The pressure Px is ap-
proximately 1 atm; P2, several thousand atmospheres.

sible. The pressure Pi is approximately 1
atm, and P2 may be considered equal to
several thousand atmospheres. Tempera-
ture-composition sections at these two pres-
sures are given below. The details of
construction of the P-T diagram are pre-
sented in order of increasing temperature.
For clarity of presentation the polymorphic
phase changes have been neglected. The
phases Fe2S3 and Fe3S4 (smythite) have
also been neglected because their stability
fields, if any, are not known with certainty.
The invariant point a is the temperature
and pressure at which the four phases 1
FcSo + L+V + S coexist; it is the eutectic,

1 The phases are arranged in the order of their
composition, from the most iron-rich first to the
most sulfur-rich last.

a critical point, has been estimated. The
critical point of sulfur is believed to be
about 1040° C and approximately 116 atm,
and the boiling point (1 atm) is 444.6° C,
according to West (1950). The curve FeS2
+L+V, for which few experimental data
are known to the writers, terminates at the
point c, 743° C and about 180 psi. The
pressure of point c was ascertained by ex-
trapolating the data on the curve Fei-^S +
FeS2 + F given by Allen and Lombard
(1917) to 743° C, the temperature deter-
mined by the present writers for the break-
down of pyrite in the presence of vapor.
Some data for the remaining three curves
originating at c have been obtained. The
curve Fei-tfS + FeS2+L is that given in
figure 19. Allen and Lombard (1917) as

GEOPHYSICAL LABORATORY 189

well as others give data on the curve Fei-#S
+ FeS2 + F in the region of about 1 to 680
mm Hg. They indicate that this curve
probably reaches 1 atm at about 689° C.
No data are available for the solubility
curve Fei-tfS+L+1/ except in the vicinity
of e. The continuation of this curve, L+
Fei^S + F, now a decomposition curve,
terminates at d. The phases remain the
same; the liquid becomes more iron rich
than the crystalline phase, however.

The point e was carefully investigated
by Jensen (1942) ; it marks the congruent
melting of the binary compound Fei-^S.
The temperature is given as 1190° C;
moreover, the pressure, which was not
determined, is that of the vapor of the
system. Jensen also determined the curve
L + Fei-arS + F and the invariant point d
where Fe, L, Fei-*S, and V are in equi-
librium. Point d is given as 988° C, and
the pressure, not determined, is that of the
system. No data are known for the curves
Fei-tfS + L, Fe + Fei-*S+L, or Fe + Fei^S
+ V. The curve Fe + L+F has been
studied by Friedrich (1908, 1910) and
others. The point / is the triple point of
iron. The melting point of iron under 1
atm of helium is 1539° C, according to
Roeser and Wensel (1942). The vapor-
pressure curve of pure iron, Fe+F, has
been calculated by Jones, Langmuir, and
MacKay (1927), and the vapor pressure
at the melting temperature is estimated to
be about 0.03 mm Hg. On the basis of
these data the point / lies near 1539° C
and 0.03 mm Hg. The same authors give
the boiling point of iron as about 3202° C,
and the critical point, therefore, must lie
at a temperature greater than 3202° C and
at a pressure in excess of 1 atm. The criti-
cal point of iron, not indicated in figure
20, is joined by the critical curve (dashed)
to the critical point of sulfur.

Two isobaric sections through the P-T
diagram are presented in figures 21 and
22 for pressures Pi and P2, respectively.
The T-X diagram of figure 21 is believed
to represent the equilibrium relations that
would be observed at a pressure of about

1 atm. The numbered points on the P-T
diagram (fig. 20) along the line Pi may
be located on the T-X diagram (fig. 21).
Similarly the schematic T-X diagram for
Pi is believed to represent the relations at
a pressure of several thousand atmospheres.
These sections differ from the customary
projection of the so-called "condensed"

i l r

L + V

Fe * FeS

Fe,.vS« V

Fe,_xS<

FeS2

Fe S2 ♦ V

FeS2 « L

FeS2

0

Fe

30/ 40/ \50 \60 70

FeS Fe3S4 Fe2S3^FeS2
Weight per cent

Fig. 21. The T-X diagram for the Fe-S sys-
tem at the pressure P13 approximately 1 atm (see
fig. 20).

diagram (see, for example, Ricci, 1951,
p. 63) given in figure 23, wherein the sys-
tem is under its own pressure and gives
those phases in equilibrium with vapor.
The term "condensed" is, therefore, a mis-
nomer, especially in systems containing
volatile components, since vapor is present
even though it is neglected. A truly con-
densed diagram would be one from which
vapor (or gas) is absent for all assem-
blages. Such a T-X diagram at constant
P is not possible for the Fe-S system. The
diagram of figure 23 is produced by pro-

190 CARNEGIE INSTITUTION OF WASHINGTON

jecting onto the T-X plane those curves
in figure 20 that contain a vapor phase.
The pressure is not constant, and is fixed
only when two phases in addition to vapor
are present as given by the three-phase
curves. When only one or two phases are
present the pressure is indeterminate un-
less the volume of the system is specified.
With the aid of these diagrams the na-

is limited by the strength of the heated
silica-glass tube. The tube may be sup-
ported by an external pressure in some
applications in order to increase its usable
pressure range. It should be emphasized
that the rigid silica-glass-tube experiments,
regardless of an external supporting pres-
sure, cannot yield data for those reactions
in which vapor is absent, as the curve in

Fe + FeS

0 o-

o

Fe

FeS,+ S

-L-o-o— L-6

1600

1400

1200

1000

400

200

Fe-FeS'V

Fe,xS.

FeS2«

V

30 / 40/ \50 V 60

FeS Fe^ Fe2S3 FeS2
WEIGHT PER CENT

100

S

10 20 30 / 40/

FeS

FeSo'L'V

i-v-s^

Fe,S

FeS ^V'S

60 70 80 90

100
S

Fe,S,

FeS,

Fig. 22. The T-X diagram for the Fe-S sys-
tem at the pressure P2, several thousand atmos-
pheres (see fig. 20).

ture of the various types of experiment
employed can now be elucidated. The
evacuated rigid silica-glass tube is the con-
tainer most commonly used for systems
involving sulfur. Since the tube is rigid
(^constant volume), and free space is
available, a vapor phase is always present.
For this reason the silica-glass tube can
yield only data which pertain, for example,
to a diagram of the type given in figure
23, the so-called condensed diagram. The
pressure is usually unknown and, of course,

3^4 " " "2

Weight per cent

Fig. 23. The so-called "condensed" diagram
of the system Fe-S. Vapor is present in all
assemblages, and the pressure of the system is
not constant.

figure 19 and the other essentially vertical
curves in figure 20.

A second type of experiment employs
collapsible gold tubes. Here the walls of
the container are deformed by application
of an external pressure: the internal pres-
sure is taken to equal the external pres-
sure. If the applied pressure is greater
than the vapor pressure of the chemical
system, then the vapor is condensed and
no vapor is permissible. On the other hand,

GEOPHYSICAL LABORATORY 191

if the applied pressure is less than the composition of pyrrhotite when coexisting
vapor pressure of the chemical system, then in equilibrium with pyrite is a function
a vapor may exist. By this technique the of temperature and pressure. Two experi-
entire P-T-X space may be investigated, mental methods are used for studying the
yielding data both on the regions in which equilibrium relations between the coexist-
vapor is permitted and on those in which ing phases. The first employs silica-glass
it is prohibited. tubes as sample containers; as a result, a
A third type of experiment employs two vapor exists at each temperature above the
silica-glass bulbs connected by a tube. Sul- solid and/or liquid phases by virtue of the
fur is maintained at a given temperature presence of a vapor space in each tube,
in one bulb, and its vapor pressure is The second method involves the use of
thereby fixed. In the other bulb the sulfur- collapsible gold tubes as sample containers,
containing system to be investigated is These two methods and the results ob-
maintained at a series of temperatures. In tained are discussed separately,
this way the system in the latter bulb is The equilibrium relations between pyr-
held at a constant vapor pressure. Such rhotite and pyrite were studied by the
experiments would yield data for those silica-tube method from 325° to 743° C
regions in figure 21, for example, where at vapor pressures from a few millimeters
vapor is permitted. The regions in which of mercury to about 10 bars, respectively;
vapor is prohibited (namely, Fe + FeS, Fe pyrite and liquid were studied up to 743°
+L, Fei-#S, Fei-tfS4-FeS2) could not be C, where the vapor pressure was about 10
studied by this technique. bars; and pyrrhotite and liquid were stud-
In a fourth technique the pressure of the ied from 743° to 785° C at vapor pres-
system is measured by the deflection of the sures from 10 to 25 bars, respectively. The
spiral of a spiral silica-glass-tube pressure magnitude of the vapor pressure above
gauge. Here again free space exists, and pyrrhotite and pyrite at a specific tempera-
only those regions in which vapor is per- ture is estimated from the data of Allen
mitted can be studied. and Lombard, De Rudder, D'Or, Roedder,

Since the compositions of the coexisting Juza and Biltz, and Rosenqvist.
sulfides in nature indicate that ore deposits Equilibrium at each temperature was
may form in either the absence or the approached from two directions. All runs
presence of a sulfur vapor (or gas), it is withdrawn from a furnace were immedi-
of paramount importance that the complete ately quenched in cold water. Tempera-
P-T-X space be investigated for the sulfide ture measurements were within 4° C un-
sy stems. These results in conjunction with less otherwise indicated. Pyrrhotite corn-
hydrous systems will have important bear- positions were determined to a precision
ing on the problem of how metals are of ±0.08 atomic per cent iron by the
transported to the site of accumulation. lattice-spacing technique described in last
The FeS-S join (Arnold). The phase year's report. Pyrite, as pointed out in that
relations involving pyrrhotite, pyrite, liq- publication, shows no measurable variation
uid, and vapor are being studied between in its cell dimensions at a considerable
the compositions stoichiometric FeS and range of temperatures, indicating little or
pure sulfur from 325° to 785° C. The no variation in its metal-to-sulfur ratio,
relationship between pyrrhotite and pyrite, Figure 24 is a diagram for a condensed
which is a portion of this system, promises system at 30 bars pressure (refer to the
to provide a method for estimating the discussion by Greig, Year Book 54, p. 131),
temperature of formation of naturally oc- which summarizes the equilibrium rela-
curring pyrrhotites and pyrite assemblages, tions between pyrrhotite, pyrite, and liquid
a pressure of formation being assumed, as calculated on the basis of this study and
The method is based on the fact that the by the work of Jensen. Although the ex-

192 CARNEGIE INSTITUTION OF WASHINGTON

-

T

1

1

i 1 —

"

■

2

-

C

o
o

£

>*
a

T3
<D

-i

o
o.

CD

i

1

M

o>

S

CD

h

</>

-

>

c
<d

3
.O

c

1

g

>

.. D

r

cd

cd

cd

•n

c
|

|

-

0>

CD

o

£

<d

c

c

0)

u.

i

'

•D

J^

■D

>

a.

-

o

o
o

1

o
a

a.

_J

CL

o

Plate 1 Geophysical Laboratory

» • •

• I . i .

♦ i

• • %

(b)

(a)

Fig. 2. "Layers" of equal scattering power with displacement of b/2 ran-
domly distributed; equal numbers of "layers" in each of the two positions.
a, the mask; b, diffraction pattern of 2a.

* ♦

• • •

• ♦

Fig. 3. Diffraction pattern of a perfectly ordered sequence of two lines in
one position followed by one line in the other (BB ABB ABB A . . .).

Plate 2 Geophysical Laboratory

• mm
f •

• • * ♦
• ■ f

• • •

4W

4 (a)

• • ! • f • •* •

•!•!

t * * w -

•

5 (*) 5 (a)

Fig. 4. Short-range ordering in a "2:1" mask: «, a mask in which NB =
2NA, each A is neighbored by Fs but the run lengths of B's are random;
b, diffraction pattern of 4a.

Fig. 5. Disordered "2:1" mask: a, a mask in which NB = 2NA, but run
lengths and sequences of both A and B are random; b, diffraction pattern
of 5a.

Plate 3

Geophysical Laboratory

Fig. 25. Photograph showing exsolution lamellae in pure pyrrhotite.
Slight granularity due to the second dark phase may be seen in the host.
The veinlet in the center of the photograph is pyrite. X 1200.

GEOPHYSICAL LABORATORY 193

periments were made at various pressures
lower than 30 bars, it is practicable to con-
struct a section at this pressure, because
the effect of pressure on the melting points,
field boundaries, and inversion tempera-
tures is comparatively small. For instance,
a calculation based on the data presented
by Kullerud and Yoder in last year's re-
port shows that the melting point of py-
rite at 30 bars is only about l/2° C higher
than at about 10 bars. Also by calculation
based on the data presented in figure 27
the position of the field boundary aPo(Py)
at 30 bars is shifted only 0.006 atomic per
cent iron towards the iron side of the
diagram.

The supercell and monoclinic forms of
pyrrhotite stable below 138° C discussed
by Haraldsen have been omitted from the
diagram because the relationships in this
area are not understood. Also marcasite
(FeS2) and smythite (FesS^ have not
been included, as their fields of stability
are not known. The melting relations of
pyrrhotite were determined by Jensen. The
datum points plotted on the diagram rep-
resent compositions determined from the
silica-tube runs projected on this isobaric
section.

The notation suggested by Greig for des-
ignating the fields, field boundaries, and
isothermal lines representing the condi-
tions under which one or more phases
exist in equilibrium has been found to be
very useful, as the relations between the
various phases become immediately ob-
vious. For example, aPo(Py) designates
the line indicating the conditions under
which a-pyrrhotite may coexist in equilib-
rium with pyrite; Py(aPo) designates
the conjugate line indicating the condi-
tions under which pyrite can coexist in
equilibrium with a-pyrrhotite.

As is indicated in figure 24, pyrite melts
incongruently at 743° C to pyrrhotite hav-
ing a composition of 45.95 atomic per cent
iron and a liquid close to pure sulfur in
composition. The compositions of the
liquid that can coexist with pyrrhotite and
of the liquids that can coexist with pyrite

have not been determined. The arrange-
ment for the liquidus curves shown dia-
grammatically, however, is correct for the
case that pyrite melts incongruently.

An inversion in the most sulfur-rich
pyrrhotites was observed between 666° and
675° C in four runs; it is tentatively placed
at 670° ± 5° C. The inversion tempera-
ture is believed to decrease with increasing
iron content in pyrrhotite. The low-tem-
perature form designated a-pyrrhotite is
quenchable and has the normal NiAs-type
structure at room temperature. The high-
temperature form designated y-pyrrhotite
is apparently nonquenchable, and inverts
to the a form on quenching.

Preparations of pyrrhotite coexisting
with pyrite, quenched from above the
inversion temperature, frequently showed
exsolution lamellae when examined under
the microscope (fig. 25) .* The lamellae
have the following general characteristics:
they generally possess straight parallel
sides, but occasionally occur in irregular
blebs, both up to several microns in width;
in some grains they are arranged in three
or four intersecting sets; they have a re-
flectivity higher than the host, but much
lower than pyrite; they are softer than the
host, and are anisotropic. Both the pyr-
rhotite host and the lamellae may contain
a second separate phase having a lower
reflectivity and a greater hardness than
either the host or the lamellae. This phase
appeared as dispersed black dots in both
the lamellae and the host, being preferen-
tially concentrated in the host and impart-
ing a granular texture to it. Occasionally
this phase in the host also appeared as very
fine subparallel lines producing a crepe-
paper-like textural effect. Pyrrhotite struc-
tures with similar physical characteristics
have been described from natural occur-
rences by Ramdohr, by Schneiderhohm
and Van Der Veen, and by Scholtz.

Powder camera photographs of these
exsolved pyrrhotites at room temperature
show no additional reflections that can be

1 Figure 25 is on plate 3.

194 CARNEGIE INSTITUTION OF WASHINGTON

attributed to a second phase with a differ-
ent symmetry or composition. Insufficient
data are available to explain the presence
of the various phases or to decide whether
the lamellae were exsolved as the a or the
Y form. Very likely this pyrrhotite ex-
solution is related to the inversion that
takes place at 670° ± 5° C.

The presence of the inversion and re-
lated exsolution raises the question of the
applicability of the lattice-spacing-compo-
sition curve for evaluating pyrrhotite com-
positions. Because the pyrrhotites on which
the spacing curve was based were homo-
geneous, this curve is pertinent only to
homogeneous pyrrhotite. Only composi-
tions of homogeneous pyrrhotites were
used as a basis for constructing the equi-
librium diagram.

The pyrrhotite-pyrite equilibrium was
also studied between 325° and 600° C at
1000 and 2000 bars pressure. The appara-
tus was the standard cold-seal bomb fed
by flexible capillary tubing. Quenching
was effected by plunging the hot bomb
directly from the furnace into water while
simultaneously releasing the pressure. All
runs were made in collapsible gold tubes,
which transmitted the water pressure, es-
sentially hydrostatically, to the contained
sulfides. Equilibrium was approached from
two directions at each pressure and tem-
perature. The results are shown in figure
26, together with the equilibrium results
determined at <1 atm included for com-
parison. The results show that pressures
up to 2000 bars do not measurably alter
the equilibrium relations between pyrrho-
tite and pyrite as determined under their
own vapor pressure at 325° C and lower.
However, at 600° C and 1000 and 2000
bars the equilibrium composition of pyr-
rhotite coexisting with pyrite was 46.18 and
46.46 atomic per cent iron, respectively,
while at <1 atm pressure at 600° C the
pyrrhotite composition was 46.03 atomic
per cent iron.

The fact that the composition of pyr-
rhotite when coexisting in equilibrium
with pyrite is a function of temperature

and pressure raises the possibility of the
application of these relations to problems
of geological thermometry. In applying
this "thermometer" in its present state of
development, however, at least one diffi-
culty should be kept in mind.

Because the experimental system con-
tained only iron and sulfur, the effect of
impurities on the equilibrium value of the
metal-to-sulfur ratio of pyrrhotite should

650 —

600

500

400

300

Limit of pyrrhotite solid solution
at 2000 bars, 1000 bars. < I bar.

1000 bars-,
2000 bars

<l bar

49.00 48.50

47.50 4700 46.50

— Atomic percent Fe

46.00

Fig. 26. Curves showing the effect of pres-
sure of < 1, 1000, and 2000 bars on the com-
position of pyrrhotite that can coexist in equi-
librium with pyrite.

be investigated. Quantitative analyses of
natural pyrrhotites indicate that cobalt and
nickel are the most significant impurities,
although rarely occurring in combined
concentrations greater than 2 weight per
cent.

An attempt was made to study the effect
of 0.65 to 6.50 per cent cobalt on the
equilibrium between pyrrhotite and pyrite
at 600° C. It was hoped that the compo-
sition of the resulting cobalt-rich pyrrho-
tite coexisting in equilibrium with cobalt-
rich pyrite could be determined by means

GEOPHYSICAL LABORATORY 195

of an X-ray spacing technique. The curve
determined at 730° C and intended for
this purpose is given in figure 27. It shows
the change of the pyrrhotite ^(102) as
various amounts of cobalt are substituted
for iron in the stoichiometric pyrrhotite
structure. In nonstoichiometric pyrrhotite,
measurements indicated that a phenom-
enon additional to the replacement of iron
by cobalt had taken place. Apparently a
specific amount of cobalt as a function of
pressure and temperature filled vacant iron
positions in the nonstoichiometric pyrrho-
tite structure, producing a large increase
in the d(\02); additional cobalt replaced
iron and caused a decrease in d(l02) pro-
portional to the amount available.

2 0920 t-

g 2.0900 -

2.0880

Wt. percent Co in (Fe,Co)S »

Fig. 27. Curve showing the effect of substi-
tuting cobalt for iron on the ^(102) of stoichio-
metric FeS. Amounts of cobalt in excess of 1.95
weight per cent cause a decrease in the d(\02).

Because of this additional replacement
effect, the proposed X-ray spacing tech-
nique could not be employed and the
influence of cobalt on the equilibrium be-
tween pyrrhotite and pyrite could not be
evaluated. Because nickel will no doubt
produce similar problems, at this stage this
thermometer should be applied only to
deposits containing no more than small
amounts of cobalt and nickel.

THE Cu-S SYSTEM

The upper stability curve of covellite
(Kullerud) . Covellite is one of the most
important sulfides of copper. It occurs in
association with other copper minerals such
as chalcopyrite, chalcocite, or digenite. It
is found in zones of secondary alteration,
or as a primary mineral associated with
chalcocite or digenite (Butte), and has
even been reported as a sublimation prod-
uct (Vesuvius). Knowledge of the sta-

bility field of covellite, therefore, would
contribute to an understanding of the con-
ditions existing during the formation of
many copper-bearing sulfide ore deposits.

Covellite may be readily synthesized in
the dry way in silica tubes by mixing
copper with appropriate amounts of sul-
fur. Copper and sulfur in the atomic
ratio of 1 : 1 will, in evacuated, sealed silica
tubes, react even at 20° C to form some
covellite in a few hours. Even after 5
months at this temperature, however, the
tubes still contained small amounts of
unreacted copper and sulfur. At 100° C
all copper had reacted to form covellite
in about 5 weeks, and at 200° C in about
2 weeks.

The lengths of the silica tubes were
adjusted so that the copper-sulfur mixtures
occupied a third to a half of the tube
volume. It was noticed that at 235° C
some digenite (CuoS5) was already formed
with the covellite in the tubes. Thus covel-
lite under such conditions starts to break
down to digenite and sulfur vapor below
235° C. Sulfur vapor produced by this
process builds up pressure to stabilize the
remaining covellite. The dissociation pres-
sures of covellite at temperatures ranging
from 400° to 490° C were determined by
Wasjuchnowa, by Preunner and Brock-
moller, and by Allen and Lombard.

By adding sulfur beyond the 1 : 1 copper-
to-sulfur ratio in the silica tubes so that
liquid as well as vapor is always present,
covellite was found to be stable up to 507°
±3° C. Above this temperature digenite
+ liquid + vapor are stable. The pressure
at the invariant point of 507° ±3° C where
CuS + CuoSs + L+I7 coexist may be deter-
mined by extrapolation of the P-T curve
given for dissociation of covellite by the
above-mentioned workers. Such extrapo-
lation gives a pressure of 880 mm Hg for
the invariant point. At this point the fol-
lowing four curves intersect: (1) CuoS5 +
CuS + F; (2) CuS + L+F; (3) Cu9S54-
CuS + L; and (4) Cu9S5 + L4-F. The
situation at the invariant point is shown
schematically in figure 28. Curve 3, C119S5

196 CARNEGIE INSTITUTION OF WASHINGTON

+ CuS + L, is the upper stability curve of
covellite. It was determined by using sealed
collapsible gold tubes as described in last
year's report for the determination of the
upper stability curve of pyrite.

The upper stability curve of covellite is
shown in figure 29. Points on the curve
are: 507° C at about 880 mm Hg (invari-
ant point c), 510° C at 7500 psi, 515° C
at 15,000 psi, and 525° C at 30,000 psi.

Covellite below the invariant point c
breaks down to digenite + vapor, not to
chalcocite + vapor as reported by earlier
workers. Above the invariant point c,

urable variations in the cell dimensions.
Digenite formed in these experiments by
breakdown of covellite commonly grows
as well developed crystals. They always
occur as octahedra, which occasionally are
extremely malformed and frequently mod-
ified by cube faces. Sometimes two oppo-
site octahedral faces predominate over all

400

500
Temperature °C

600

Fig. 28. Curves showing schematically in-
variant point c where the four phases Cu9S5 +
CuS + L+V are stable.

covellite, in collapsible tubes, breaks down
to digenite + liquid. In the diagram the
stability field of covellite is on the left of
the curve and the field of digenite + liquid
is on the right.

Covellite grown in silica tubes is dark
blue, is commonly massive, and rarely
shows crystal faces. Single crystals were
in a number of experiments grown at high
pressure in collapsible gold tubes. These
crystals, less than 0.1 mm long, appear to
be hexagonal plates. Powder X-ray dif-
fraction studies of covellite crystals grown
under varied pressure and temperature
conditions (from 100° C at less than 1 mm
Hg to 520° C at 2000 bars) show no meas-

30,000

-

1 1 1

1 1
■ 00

-

-

Covellite (CuS)

4-

Digenite +■ liquid

-

-

3
0

■1-
m

in
0

-

15.000

0

;

7.500

«

-

1 1 l.<

1 1

200

300

400 500 600

Temperature °C

700

800

Fig. 29. The upper stability curve of covel-
lite CuS<r±Cu9S5+L.

other faces to such an extent that the
crystal looks like a flat hexagonal plate.
Other times, observed at 15,000 and 30,000
psi, cube faces dominate. The digenite
crystals vary in size from about 0.1 to 4
mm, depending on time, temperature, and
pressure. The larger crystals were grown
in silica tubes, and the smaller in collapsed
gold tubes. The cell dimensions of dige-
nite formed at various temperatures and
pressures were determined by X-ray dif-
fraction methods. The cell size, however,
remained constant within the limits of the

GEOPHYSICAL LABORATORY 197

experimental accuracy, an indication that
digenite is essentially a stoichiometric com-
pound.

The liquid formed by decomposition o£
covellite when quenched appears white to
yellow and consists of almost pure sul-
fur. X rays at room temperature of such
quenched liquids, produced at various
pressures and temperatures, always gave
the pattern of orthorhombic sulfur.

crystals and NiS2 were described. A dia-
gram was also presented showing variation
in d spacings for the (102) X-ray reflec-
tions as a function of mix-crystal (Nii-^S)
composition. This curve was used to de-
termine the composition of the Nii-^S
when formed in equilibrium with NiS2.
Experiments with mixtures of Ni and S
in the atomic ratio of 3:4 have established
the subsolidus relations between the two

34 NlS2 32

At. per cent Ni

Fig. 30. Subsolidus phase relationships in the NiS-NiS2 binary join.

THE Ni-S SYSTEM

The NiS-NiS2 join {Arnold, Kulle-
rud). This study of the subsolidus rela-
tions in the millerite (NiS)-vaesite (NiS2)
binary join is part of an investigation of
the entire Ni-S system. A systematic in-
vestigation of this system is necessary be-
fore a detailed exploration of the extremely
important Fe-Ni-S system can be under-
taken.

In last year's report methods of prepara-
tion of NiS as well as of Nii-^S mix-

phases in the 300° to 700° C temperature
range. When stoichiometric NiS is heated
a rapid inversion from the millerite to the
hexagonal Ni-As structure occurs at 374°
C. This inversion temperature is lowered
markedly by omission of Ni from the NiS
lattice. Thus when Nii-^S is formed in
equilibrium with NiS2 the inversion takes
place at about 325° C.

The Ni3S4 phase, reported in the litera-
ture to be stable up to 325° C, did not
appear at 300° C even after 6 months.

198 CARNEGIE INSTITUTION OF WASHINGTON

No sign of melting was detectable in
Nii-a>S and NiS2 mixtures at temperatures
as high as 790° C, in spite of the fact that
stoichiometric NiS melts at 800.5 ±1° C.

Figure 30 shows the phase relations in
the NiS-NiS2 join. The amount of nickel
omission solid solution in Nii-^S when
this phase forms in equilibrium with NiS2
is seen to vary from 1.7 atomic per cent

V in silica tubes, as well as with NiS + S
and NiS2 in collapsible gold tubes at 2000
bars, have shown that NiS2 under such
conditions is stable beyond the melting
point of NiS (800.5° ±1° C). The con-
clusion is that the four phases Nii-a?S +
NiS2 + L+F cannot coexist at any pres-
sure or temperature.
Powder X-ray diffraction studies of NiS2

2(gas)

Fig. 31. Phase relations in the Fe-S-O system: (a) below about 675° C.

Ni at 325° C to about 2.4 atomic per cent
Ni at 700° C, giving rise to a very steep
solvus curve.

Biltz (1936) determined the curve for
dissociation of NiS2: at 650° C, 42.5 mm
Hg; at 700° C, 154 mm Hg; at 720° C,
260 mm Hg; at 730° C, 324 mm Hg; and
at 760° C, 649 mm Hg. NiS2 in this range
breaks down to Nii-«?S + V. Extrapolation
of these data gives breakdown of NiS2
under 1 atm sulfur pressure at 765° C.
Preliminary experiments at 850° C with
NiS2 with excess sulfur to give NiS2 + L+

grown at temperatures from 350° to 854°
C and at pressures ranging from a few
millimeters of mercury to 2000 bars show
an apparent variation in the cell lengths
of the NiS2 crystals from a = 5.685 ±0.002
A at 350° C and pressure of less than 1
mm Hg to *= 5.690 ±0.002 A at 854° C
and 2000 bars.

PHASE RELATIONS IN THE Fe-S-O SYSTEM
G. Kullerud

The Fe-S-O system includes pyrite
(FeS2), the most common sulfide mineral,

GEOPHYSICAL LABORATORY 199

and pyrrhotite (Fei-^S), as well as the
very important oxides hematite (Fe2Os)
and magnetite (Fe304). A series of pre-
liminary experiments with mixtures of
iron oxides and sulfur in evacuated, sealed
silica tubes have reproduced some of the
interesting mineral assemblages common

hematite, and pyrite form a stable assem-
blage, and hematite may be transformed
into pyrite and S02 gas by introduction
of sulfur. Further, hematite and S02 may
be produced from pyrite by introduction
of oxygen, as perhaps takes place on the
Isle of Elba. It is seen that hematite below

T= 675°- -700 °C

2 (gas)

Fig. 31. Phase relations in the Fe-S-O system:
(b) between about 675° and 700° C, (c) be-
tween 700° and 743° C, (d) above 743° C.

in nature. In the following discussion of
the phase relationships shown in figure 31,
solid solutions in FeS, Fe203, and FesCX
have been neglected.

Figure 3la shows that at temperatures
below about 675° C pyrite and pyrrhotite
can coexist with magnetite. This relation-
ship has been observed in localities such as
the Coeur d'Alene district. Magnetite,

T = — 700-743 °C

S°2 (gas)

this temperature is not stable in the pres-
ence of pyrrhotite. The above-mentioned
phase relations are expected to remain un-
changed at the lower temperatures at
which many ore deposits are believed to
have formed.

At about 675° C the following reaction
takes place : pyrite + magnetite^pyrrho-
tite + hematite. Figure 31£ shows that py-
rite, pyrrhotite, and hematite now are
stable together, and that hematite, mag-
netite, and pyrrhotite may coexist, but mag-
netite and pyrite do not form a stable
mineral assemblage above this temperature.

Further changes in the phase relations
occur at about 700° C with the reaction:
pyrite + hematite^pyrrhotite + S02. Fig-
ure 3\c shows that neither hematite nor
magnetite now is stable with the pyrite +
pyrrhotite assemblage, whereas pyrrhotite,
hematite, and magnetite still can coexist.

At 743° C pyrite is no longer stable in
a rigid silica tube regardless of the amount
of sulfur present. (See discussion of the
stability limits of pyrite elsewhere in this
report.) Figure 3ld shows the phase rela-

200 CARNEGIE INSTITUTION OF WASHINGTON

tions above 743° C, where pyrite no longer
exists. Hematite, magnetite, and pyrrho-
tite remain a stable mineral assemblage.
The geological significance of these phase
relations is far-reaching, since these min-
erals are of almost ubiquitous occurrence.

THE Fe-S-Se SYSTEM
G. Kullerud and H. L. Barnes

Until recently the selenide minerals had
received relatively little attention from
geologists and laboratory workers, but the
rapidly developing industrial and elec-
tronic applications of selenium have fo-
cused increasing attention on the modes
of occurrence of this element. Thus, fer-
roselite (FeSe2), which was first recog-
nized as a mineral about 2 years ago, has
already been reported from 15 to 20 lo-
calities on the Colorado Plateau alone.
By chemical analyses of natural minerals,
FeSe2 and FeS2 (pyrite) have been found
to enter into solid solution with each other.
These Fe(S,Se)2 mix-crystals, which often
occur with uranium minerals, may ulti-
mately serve as geological temperature in-
dicators when the FeS2-FeSe2 system has
been studied in the laboratory. The present
investigation is also designed to explore
the distribution of selenium at various
temperatures and pressures between pyr-
rhotite and pyrite.

Stoichiometric FeSe was synthesized at
various temperatures from 325° to 925° C.
X-ray diffraction patterns made at room
temperature showed that eskebornite in
this temperature range is of the NiAs
structure type (provided that nonquench-
able polymorphs do not exist). FeSe and
FeS, therefore, are isostructural. FeSe, like
FeS, can omit iron from its structure, and
the chemical formula of natural eskeborn-
ite, therefore, should be written Fei-^Se.
Although FeS, according to the literature,
cannot take excess iron in solid solution,
eskebornite was found to take about 2
weight per cent iron in solid solution at
800° C. Mixtures of FeS and FeSe were

made up at 10 weight per cent intervals
and heated at 800° C for 60 days. X-ray
diffraction and polished-section studies
showed that the solubility of FeSe in FeS
at this temperature is between 30 and 40
weight per cent. The solubility of FeS in
FeSe is between 40 and 50 weight per cent.
X-ray diffraction studies of Fe(S,Se)
mix-crystals of varying composition show
appreciable variation in the (102) d spac-

40 50 60

70

80

90

100
FeSe

Mol per cent Fe Se ►

Fig. 32. The curve showing variation in d
spacings for the (102) X-ray reflections as func-
tion of mix-crystal Fe(S,Se) composition.

ings. The (102) d spacings versus com-
position have been plotted in figure 32.
Sodium chloride (#=5.62869 kX) was used
as internal standard for all measurements.
The dotted part of the spacing curve shows
where the immiscibility gap occurs at
800° C.

Experiments with FeS and FeSe mix-
tures at temperatures between 850° and
400° C established the limits of the binary
solid solution within a few per cent. The
top of the solvus curve is situated at about
850° C and at a composition of approxi-
mately 55 weight per cent FeS and 45

GEOPHYSICAL LABORATORY 201

weight per cent FeSe. The solubility of
FeSe in FeS decreases rapidly with de-
creasing temperature, and at 600° C is
about 10 weight per cent. The solubility
of FeS in FeSe at 600° C is about 30
weight per cent. The two-phase field in
this system is not binary. Beyond the limit
of FeS solubility in FeSe and of FeSe sol-
ubility in FeS the two phases Fei-a;(S,Se)
and Fei+#(S,Se) coexist. The amount of
iron deficiency in the first and excess of

The crystal chemistry of these arsenides
is of particular interest, as most chemical
analyses have suggested that compositions
with marked departures from stoichio-
metric proportions might be common. The
different arsenides are commonly zoned
or intergrown with one another, however,
and can be distinguished only by means
of the polarizing reflecting microscope or
X-ray techniques. The few analyses that
have been made on material checked for

o<8 26 —

Composition, mole percent

Fig. 33. Curves showing unit cell dimensions of synthetic skutterudites as a function of cobalt-
to-nickel ratio, cobalt-to-iron ratio, and cobalt-to-nickel-and-iron ratio (where the nickel-to-iron ratio
is 1:1).

iron in the second of these compounds
at various S : Se ratios must be carefully
studied before the tie lines can be deter-
mined in the ternary system.

THE CoAs2-NiAs2-FeAs^As SYSTEM
E. H. Roseboom, Jr.

Ore deposits mined for arsenic, cobalt,
nickel, and silver, although not abundant,
are of world-wide distribution and consti-
tute a distinctive type. They are remark-
ably low in sulfur, with cobalt, nickel, and
iron present as arsenides or sulfarsenides
and with silver and bismuth present in
the native state.

homogeneity have indicated nearly stoi-
chiometric compositions for diarsenides,
but metal-to-arsenic ratios as low as 1 : 2.65
for skutterudites, (Co,Ni,Fe)As3-*. Such
ratios might vary sufficiently with tem-
perature to provide a means of determin-
ing the temperature of formation of ore
deposits.

Very little was known about the system
Co-Ni-Fe-As, which contains eight natu-
rally occurring minerals. The present study
was undertaken to explore the portion of
the system pertaining to natural deposits.
It was hoped that phase changes suitable
for establishing limits of temperature and

202 CARNEGIE INSTITUTION OF WASHINGTON

pressure during formation might be dis-
covered for this group of minerals. As
most natural skutterudites contain cobalt,
iron, and nickel, and as so little was known
of the phase relations, it was thought that
a reconnaissance of the arsenic-rich por-
tion of the quaternary system would be of
more value than a detailed study of a
binary or ternary system in the quaternary.
In preliminary attempts to make skutter-
udites of mixed cobalt, iron, and nickel

amount of arsenic that combined with the
metal was determined by weighing the
inner tube after the run. The contents of
the tube were reground and reheated in
the same way until there was no further
increase in weight. The maximum arsenic
contents given below for cobalt skutterud-
ite (CoAs3-a?), loellingite (FeAs2), and
rammelsbergite (NiAs2) are each the av-
erage and standard deviation of three
samples. The minimum arsenic contents

Phases stable with excess arsenic at 800 C
(opprox. 23atm of vapor pressure).

Light lines and numbers in the friarsenide field
indicate lines of equal cell dimension in angstroms

Dashed lines indicate field boundaries

Heavy lines connect compositions of coexisting
diarsenides and triarsenides

CoAs2

True shape of projected surface

is shown by cross halching

Rammelsbe

Fig. 34. The cobalt-nickel-iron content of solid phases in the system Co-Ni-Fe-As which are
stable with excess arsenic and a vapor phase at 800° C.

composition, it was decided to work at
800° C after runs of several days' duration
at 600°, 700°, and 750° C failed to ap-
proach equilibrium.

The present work is an isothermal pro-
jection of the solid phases stable with a
vapor at 800° C and with total pressure
equal to the equilibrium vapor pressure
of the solid phases present. The runs were
made in evacuated sealed silica-glass tubes.

The maximum arsenic content was de-
termined for the phases stable with arsenic
by heating at 800° C a known amount
of metal in an open tube inside a larger
closed tube containing excess arsenic. The

were determined by making runs of suc-
cessively lower arsenic content until the
presence of another phase was detected by
microscopic examination of polished sec-
tions of the samples. The arsenic contents
determined in this manner may be too
large by as much as 0.01, as a much larger
proportion of arsenic than cobalt is in-
volved in the vapor phase.

At 800° C cobalt skutterudite has a min-
imum arsenic content between CoAs2.94
and CoAs2.95 and a maximum of
CoAs2. 960 ±o,ooi. Loellingite has a lower
limit between FeAsi.97 and FeAsi.os and
an upper limit of FeAsi.998 ±0.002. Rammels-

GEOPHYSICAL LABORATORY 203

"S3*

o *> <

<, << <u

oj 00

<

moo 2

Ll GC T3

<N„ "> i-t

g S3 jj

w, ON < ^5

S a,o52 >

<u GO ^

4— > f— 4

wU^3

^

.5*5

o

</>

<\l

•—

<

(A

2

+

<

jQ

7*

01

C\J

fl>

<

0

E

E

0

q:

° <*> CS

v\ >a <3 e^

£^ Svo

rt coPh ^

-3 "2 CO

"O Z l/^ <n

OvQ N^

g ^£ S

■'-, ps « *<^

"55 as <u O

c > fc ^
o 2 <u

<u "

y < o

in

,r-| o re

.5 b] fe

204 CARNEGIE INSTITUTION OF WASHINGTON

bergite has a lower limit above NLA.S1.99 FeAs2 80 per cent CoAs2, and reaches a

and an upper limit of NiAs1.999iO.001. In maximum at CoAs2. NiAs2 and CoAs2,

no case was there a measurable difference and FeAs2 and CoAs2, apparently form a

in X-ray d spacings between the phases complete solid solution series. The series

made with excess arsenic and the same NiAs2-FeAs2 is discontinuous at 800° C

phases made with deficient arsenic. because of an asymmetrical solvus with

The name skutterudite is used for all a crest estimated at about 900° C and about

cubic "triarsenides" of cobalt, iron, and 15 per cent FeAs2 85 per cent NiAs2.

nickel. Thus the formula may be ex- The condensed diagram of the systems

pressed approximately as (Co,Fe,Ni)As3-;r. CoAs2-NiAs2-As, NiAs2-FeAs2-As, and

Skutterudites of varied cobalt, iron, and FeAs2-CoAs2-As at 800° C is shown in

nickel content were made in the presence figure 35 as the unfolded sides of a tetrahe-

of excess arsenic. Figure 33 shows the dron with arsenic at the apex. The dia-

effect of composition on cell size for vari- gram is based on relatively few runs. After

ous series of skutterudites. Figure 34 shows the diarsenide and skutterudite solid solu-

the cobalt-iron-nickel content of phases tion series were established, it was possible

stable with excess arsenic at 800° C. The to locate the corners of the three-phase

lines of equal cell dimension for the skut- fields from the d spacings of the phases,

terudites are also shown. In the iron and As arsenic melts at about 818° C under

nickel corners of the triangle there are its own vapor pressure, a eutectic between

fields in which loellingites and rammels- arsenic and the arsenides of highest arsenic

bergites of varied compositions are stable content might extend below 800° C. Hence

with excess arsenic. Between these fields [t [s possible that at 800° C there are some

and the skutterudite field, a skutterudite \[qu[^ fields in the arsenic-rich portion of

is stable with either a loellingite or a the diagram, although no evidence of them

rammelsbergite, plus arsenic. was 0bservecL

Runs made at 800° C in the diarsenide

+ skutterudite + arsenic fields and annealed relations BETWEEN COMPOSITION OF ORE

at 600° and 700° C for 3 months indicated MINERALS AND ORE SOLUTIONS

that the limits of the binary series of skut- h. L. Barnes and G. Kullerud

terudites decrease by 1 to 3 per cent except „ . , . 111

£ i_. • 1 1 c ft • 1 1 • Economic geologists and laboratory

for the iron-rich end or the nickel-iron . 11 • • c .i_

, ,. 1 • 1 • ee scientists agree that the majority or the

skutterudites, which increases to 55 per ,,, .5, ... ' 1 u

„-. • t 1 rAAo ^r worlds sulfide ore bodies must have been

cent iron 45 per cent nickel at 6UU C . . , , , .. a • 1

^, ,. . r £ . 11 1 .11 deposited from aqueous solutions or fluids.

The limits or iron, cobalt, and nickel _,r fl .1 L 1

c 1 t-r • 1 1 J a-^ These aqueous rluids must have contained

content of these artificial skutterudites con- 1 1 1 r 1 • 1 • 1 r

. on r 1 • • 1 metals and sulrur combined in the rorm

tain 87 per cent or the remaining analyses , . .,

for natural skutterudites compiled from °£ «>mPlex ^ns Such complexes could
the literature by Holmes (1947) after he b[eak ^ a;> the 1'esultJof changej. m
had eliminated analyses made on material Physical and chemical conditions, leading
that was probably not skutterudite. Only t0 sulfide deposition.
6 per cent of these analyses were more than In some ore bodies limits can be placed
5 per cent beyond the limits. on tne conditions existing at the time of
Although both loellingite and rammels- °^ formation on the basis of subsolidus
bergite are orthorhombic, cobalt diarsenide relations and mineral solubilities, but even
appears to be monoclinic. The evidence where the phase relations give relatively
for this was the splitting of the 110 and accurate physical data, the chemistry in-
Ill d spacings into two spacings. The volved during transport and precipitation
split begins at about 25 per cent NiAs2 is not well understood. A complete under-
75 per cent CoAs2 and at about 20 per cent standing of the chemistry of such trans-

GEOPHYSICAL LABORATORY 205

porting agents could undoubtedly be de-
veloped into a powerful tool for prediction
of ore occurrence.

Although the considerable economic and
theoretical importance of an understand-
ing of the chemistry of ore solutions has
been clear to numerous workers in the
field of geology, the ore solutions involve
extremely complicated chemistry which of
necessity has been commonly oversimpli-
fied in its treatment.

Thus, many recent papers using mineral
compositions for data on the composition
of hydrothermal ore solutions have been
based on simple ionic chemistry without
a clear statement of the implicit assump-
tions involved. By summarizing the physi-
cal and chemical conditions which must
necessarily be evaluated, these assumptions
become self-evident:

1. Equilibrium conditions must occur
during deposition of the minerals. In other
words, metastable phenomena such as su-
persaturation or coprecipitation must not
modify conditions of precipitation, or cal-
culations of equilibria would be mean-
ingless.

2. If the ratio of concentrations of any
given element in two minerals is to be
used for defining the ore solution, the two
minerals must have been precipitated un-
der physically and chemically identical con-
ditions. Practically, this limitation requires
either that the two minerals were precipi-
tated simultaneously or that there has been
no change in the composition, pressure, or
temperature of the ore solution between
the time of precipitation of the first and
second mineral.

3. There must be no post-depositional
changes in the composition of mix-crystals
used for data on depositional environment,
though these mineral compositions may be
far out of equilibrium with subsequent
environments. Neither exsolution nor so-
lution is permissible, although both proc-
esses are favored by recrystallization or
changes in chemical environment, pressure,
or temperature, especially during slow

cooling of deep-seated deposits or heating
in later metamorphism.

4. The relative distribution and types
of ions present in the ore solution must be
known. Specifically, information on the
extent to which dissociation, hydrolysis,
and complexing take place is important in
order to estimate the total ionic strength
and to ascertain the metal transporting
ions.

5. Activity coefficients of both the aque-
ous ions and the mix-crystals need to be
evaluated as functions of the four inde-
pendent variables temperature, pressure,
ionic species, and ionic strength.

Lack of information necessitated the as-
sumption of specific conditions for theo-
retical treatment even without direct evi-
dence. Geologic data on the history of a
particular mineral sample are very difficult
to collect, and, at present, there exist vir-
tually no chemical data obtained under the
difficult experimental conditions approxi-
mating those at which ore deposits were
formed.

Knox (1908) and Dickson and Tunell
(1955) have studied the chemical behavior
of HgS under varied chemical conditions
simulating ore transport. Experimental
work on the solubility of ZnS in H20 un-
der elevated pressures and temperatures
in a H2S atmosphere is now in progress in
this Laboratory.

If sufficient data were available, four
simultaneous equations could be derived
to relate these four independent variables
to mix-crystal concentration ratios. For a
unique solution at least four accurate con-
centration ratios are necessary. These ra-
tios may be of two types : (a) the fractional
content of one trace element in one min-
eral may be expressed and used as a ratio
provided that there is an excess of the trace
element present at the time of deposition
to assure that the mineral is saturated with
the trace element at that temperature and
pressure; and (b) the distribution ratio
of one trace element between two min-
erals may be used where neither is satu-
rated with the trace element. The four

206

CARNEGIE INSTITUTION OF WASHINGTON

ratios required could be measured theoreti-
cally for any combination of the two types
where the distribution ratio could be meas-
ured for one trace element in several min-
erals, several trace elements in the same
two minerals, or any intermediate com-
bination.

From a practical point of view, the trace-
element concentration in a mineral can be
used only if it is sufficiently high for ac-
curate analysis and if the mineral can be

effectively separated from potentially con-
taminating foreign minerals. In order to
use these ratios, the minerals involved
must have been deposited contemporane-
ously. It is doubtful whether minerals
fulfilling these conditions and providing
a sufficient number of ratios for the solu-
tion of the simultaneous equations can be
found in actual ore deposits; it remains a
crucial question, however, and must be
investigated.

FELDSPARS

TERNARY FELDSPARS

H. S. Yoder, D. B. Stewart, and J. R. Smith

The ternary feldspar study briefly out-
lined last year has advanced to a very
fruitful stage. The relations of the various
feldspars, of greatest import to the petrolo-
gist, may now be defined more closely
and the results applied to natural rocks.
The principal feldspars lie in the system
NaAlSi308 (Ab) -KAlSi308 (Or) -CaAL-
Si2Og (An) ; this system has been investi-
gated with water under a pressure of 5000
bars, mainly using glasses prepared by
Franco, Schairer, and Bowen.

Projections of the bounding ternary sys-
tems Ab-An-H20, Or-An-H20, and Ab-
Or-H20 are given in figures 36, 37, and
38, respectively. The uncertainty of the
nature of the melting of compositions near
anorthite mentioned last year has been
resolved. The "3-alumina" of unknown
composition that appeared in some runs
has dissolved with longer runs, and it may
be stated with confidence that anorthite
melts congruently at 5000 bars water pres-
sure. Much effort was put into fixing the
limits of the solvus in the Or-Ab-H20
system (fig. 38). The solvus as pictured
is based on 1-month runs using only the
results from glass starting materials. The
number of phases was identified by means
of powder X-ray diffraction patterns, since
optical recognition was not always possible.
Equilibrium was demonstrated by holding
previously crystallized glasses known to
consist of a single phase at the same tem-

peratures and pressure. In the unmixing
region as outlined by runs using glass as
a starting material most of the composi-
tions began to unmix. Unfortunately, sin-
gle feldspars formed previously from a
glass did not unmix to the same composi-
tions of feldspars as those obtained from
glass crystallized in the unmixing region
even in a month's time on the basis of the

201 spacings of the X-ray diffraction pat-
tern. The 201 spacings do not give a suit-
able measure of the composition in un-
mixed feldspars, according to Coombs and

others, and so the 201 spacing could be
used as only an approximate measure of
the attainment of equilibrium. However,
the incompleteness of unmixing of those
single feldspars previously formed from
glass was well outside the errors of esti-
mating the composition by the 201 spacing.
From runs made to test the possibility of a
solvus in the Ab-An-H20 system, only
single-phase crystalline products were ob-
tained. The peristerites may be the result
of complex phase relationships arising in
the Ab-Or-An system.

Because the lines in the diagrams in
figures 36, 37, and 38 are projections of
boundary surfaces between assemblages in
equilibrium with gas, they cannot be read
in the same manner as anhydrous dia-
grams of similar appearance. In order to
illustrate this fact, two sections determined
experimentally at constant temperature and
pressure are given for Ab-Or-H20, one
of the bounding systems. In figure 39 is

GEOPHYSICAL LABORATORY 207

1300

Pu o = 5000 bars

700 —

600

J I

NaAISUO

3W8

20

40 60

WEIGHT PER CENT

80

CaAI2Si20g

Fig. 36. Projection of the ternary system NaAlSi308-CaAl2Si208-H20 at 5000 bars H20 pressure.

i r

1300

1200

N00

o5 1000
W

h-900A- O O
0_

Ph20 " 5000 bars

o o o o

LU

500

700

600

+

KAISi308

bO

CaAI2Si208

40 60

WEIGHT PER CENT
Fig. 37. Projection of the ternary system KAlSi308-CaAl2Si208-H20 at 5000 bars H20 pressure.

208

CARNEGIE INSTITUTION OF WASHINGTON

the 720° C, 5000 bar section, and all these
data project onto the 720° C line in figure
38. Attention is called to the large amount
of H20 in the liquids (~11 per cent) and
the resulting large area in which a gas
phase is prohibited. The second section, fig-
ure 40, at 710° C, 5000 bars, shows some-
what the same relations, and in addition in-
dicates that the solvus has been transected
in the region where a gas phase is prohib-
ited. The maximum on the solvus does not
appear on the projection (fig. 38), since

that they contain about 1 to 2 per cent of
the feldspar components (see figs. 39 and
40).

With the knowledge gained from the
three bounding ternary systems, it was then
possible to study the quaternary Ab-Or-
An-H20 at 5000 bars. If the problem is
described loosely in terms of the anhydrous
systems, the study consisted of tracing
the way in which the three fundamental
bounding systems, one of the continuous-
series type (fig. 36), another of the eutec-

1000

900

PH 0=5000 bars

Liquid + gas
__ Feldspar + liquid + gas

500

400

0

Ab

10

20

70

80

90

100
Or

30 40 50 60

WEIGHT PER CENT
Fig. 38. Projection of the ternary system NaAlSi308-KAlSi308-H20 at 5000 bars H20 pressure.

only those assemblages in equilibrium with
gas are indicated. On the basis of the data
in these two sections, the maximum on the
solvus at 5000 bars total pressure is placed
at 715° ±5° C and Ab55Or45±3 weight
per cent. Bowen and Tuttle determined
the maximum of the solvus at 1000 kg/cm2

to be 660° C and Ab550r45, using the 201
method for two points and determining
the number of phases for a third point at
the maximum. With the latter datum, the
pressure raises the maximum on the solvus
about 14°/1000 bars. Although the compo-
sition of the gases has not been studied in
great detail, preliminary results suggest

tic type (fig. 37), and the third of the
minimum type (fig. 38), merge in T-X
space. The projection of the determined
liquidus diagram when gas is present is
given in figure 41. One of the most sig-
nificant features is the generally low tem-
peratures, temperatures readily available
in the crust, as compared with the an-
hydrous system determined by Franco and
Schairer. Also of importance is the OrAb-
rich nature of the liquids along the four-
phase boundary curve. This observation
has bearing on many geological problems,
particularly those involving partial melting
or metasomatism. The four-phase bound-

GEOPHYSICAL LABORATORY 209

H,0

PT , , = 5000 bars
Total

Or

Weight per cent

Fig. 39. The 720° C section at 5000 bars of the system NaAlSi308 (Ab)
KAlSi308 (Or) - H20.

PTotal = 500° bQrs
T = 710 °C

Weight per cent

Fig. 40. The 710° C section at 5000 bars of the system NaAlSi3Os (Ab)
KAlSi308 (Or) - H20.

210 CARNEGIE INSTITUTION OF WASHINGTON

ary curve becomes a three-phase trough (in- join. With fractionation, the range of

cheated by short dash) at a fixed tempera- simultaneous crystallization would be ex-

ture between 695° and 698° C. The change tended.

takes place at the temperature at which Some notion as to the way in which
two feldspars react with liquid and gas to the tie lines connecting coexisting feld-
produce one feldspar, liquid, and gas. The spars in equilibrium with liquid and gas
one feldspar produced has the composition sweep across the 5000-bar projection may
of the point of contact of the solidus and be gained from figure 44. The family of
the solvus. such tie lines, which include the four deter-
In order to understand more fully the mined experimentally, generates a surface,
relations in the complex quaternary sys- the boundary curves (dashed) marking
tern, three isothermal, isobaric projections the maximum solid solution of coexisting
have been studied. The 770° C, 5000 bar feldspars in equilibrium with liquid and
projection, given in figure 42, presents only gas. A third isothermal, isobaric projec-
those assemblages that are in equilibrium tion is now being investigated at 660° C,
with gas. The chief observations are the 5000 bars, to determine the relations when
composition of feldspars in equilibrium only crystals and gas are present. It is
with liquid and gas and the orientation realized that many of the phases obtained
of the tie line (light line) which separates in these studies undergo transitions of the
the two feldspars + gas field from the first order and of higher orders. Although
two feldspars + liquid + gas field. The powder X-ray diffraction patterns have
tie lines that connect coexisting feldspars been taken of many of the synthetic crys-
or coexisting feldspar and liquid in the tals, none of the crystals was suitable for
three-phase regions do not appear on the single-crystal X-ray study. Minor changes
diagram, since no data are known for fix- may, therefore, be required in the equi-
ing the composition of ternary feldspars librium diagrams when sufficient knowl-
except along univariant curves or at in- edge of the exact structural form of each
variant points. At invariant points or along phase is obtained. In addition, hydrous
univariant^ curves the points were fixed phases will appear at temperatures lower
by observing the number and kind of than those investigated,
phases about the point. In advance of these data the subsolidus
In figure 43 the relations of the 720° C, regions of one and two feldspars, those re-
5000 bar projection are given for only those gions of most importance to the geologist,
phases in equilibrium with gas. The ori- may be illustrated schematically. Figure
entation of the tie line fixing the composi- 45 outlines the field boundary surfaces at
tion of the two feldspars in equilibrium 5000 bars which separate the two-feldspar
with liquid and gas should be noted and region (inside the truncated "dome") from
compared with the similar tie line in fig- the one-feldspar regions. The top surface
ure 42. In general, the change of composi- Gf the "dome," the intersection of the sol-
tion of the AbAn-rich feldspar with tern- vus ancl the solidus, represents the ex-
perature is greater than that change for perimentally determined relations already
the OrAb-rich feldspar. The small tri- given in figure 44. For comparison, a
angular area of the projected four-phase similar picture (fig. 46) is presented for a
region in both isothermal, isobaric projec- water pressure of about 2000 bars. This
tions (figs. 42 and 43) indicates the limited schematic figure was deduced mainly from
range of temperature through which two the work of Bowen and Tuttle on the
feldspars may crystallize simultaneously Ab-Or-H20 system at 2000 bars. In fig-
from a given magma under equilibrium ure 46 it is seen that the surface mark-
conditions. The range is zero at the Or- ing the solvus-solidus intersection does not
An-H20 join and near the Or-Ab-H20 meet the Ab-Or join. The resulting ori-

PH 0=5000 bars

876°

695- Qr

Weight per cent

Fig. 41. Projection of the quaternary system NaAlSi308 (Ab) - KAlSi3Os
(Or) - CaAl2Si208 (An) - H20 at 5000 bars H20 pressure. The circles indicate
compositions for which a bracket was obtained. The dots indicate composi-
tions for which a bracket was not obtained; significant temperatures were
investigated, however.

5000 bars

Weight per cent

Fig. 42. Projection of the NaAlSi308 (Ab)-KAlSi308 (Or) -CaAl2Si208
(An)-H20 system at 770° C and 5000 bars H20 pressure. Only those assem-
blages in equilibrium with gas are given.

211

= 5000 bors

Liquid + Gas

Or

Weight per cent

Fig. 43. Projection of the NaAlSi308 (Ab)-KAlSi308 (Or)-CaAl2Si208
(An)-H20 system at 720° C and 5000 bars H20 pressure. Only those assem-
blages in equilibrium with gas are given.

5000 bars

Weight per cent

Fig. 44. Plot of experimentally determined tie lines connecting coexisting
feldspars in equilibrium with liquid and gas. These appear from the highest
to the lowest temperature in figures 37, 42, 43, 38, respectively.

212

GEOPHYSICAL LABORATORY 213

entation of the tie lines connecting coexist-
ing feldspars is therefore different from
those at 5000 bars water pressure. Further-
more, the extent of solid solution increases
on this surface with lower pressure for a
given bulk composition. In general, the
tie lines make a smaller angle to the Or-
An side at high temperatures and low
pressures than at low temperatures and
high pressures. If the pressure or tem-
perature were known, then the tempera-
ture or pressure of formation, respectively,

In figure 47 are assembled some of the
analyzed pairs of coexisting feldspars re-
corded in the literature. Most of the sam-

Fig. 45. Schematic presentation of the field
boundary surfaces separating the two-feldspar
region (inside the truncated "dome") from the
one-feldspar and the two-feldspar + liquid + gas
regions of the Ab-Or-An-H20 system at 5000
bars H20 pressure.

could be estimated from knowledge of the
compositions of the coexisting feldspars.
Since the field geologist can often make
an estimate of the depth at which a rock
may have formed, assuming that Ph2o =
Ptotai, the compositions of the coexisting
feldspars can yield a measure of the tem-
perature of formation, provided that equi-
librium was attained. If the pressure of
formation cannot be specified, but is be-
lieved to be constant, the relative change
of the compositions of coexisting feldspars
in a series of rocks indicates the relative
temperatures of formation.

Fig. 46. Schematic presentation of the field
boundary surfaces separating the two-feldspar
region (inside the "dome") from the one-feld-
spar and the two-feldspar + liquid + gas regions
of the Ab-Or-An-H20 system at 2000 bars H20
pressure. Based on the work of Bowen and
Tuttle (1950) on the Or-Ab-H20 system.

WEIGHT PER CENT

Fig. 47. Plot of chemically analyzed pairs of
coexisting feldspars recorded in the literature.
Homogeneity of phase has not been tested by
X rays in most cases.

pies were not available to test for homo-
geneity by X rays. The orientations of
the tie lines, with the exception of two
having the most An-rich plagioclases, are

214 CARNEGIE INSTITUTION OF WASHINGTON

in accord with those obtained experimen-
tally. The two exceptions are indicative
of a lower pressure and a higher tempera-
ture; in fact, the feldspars from these speci-
mens are phenocrysts from surface flows.
However, another specimen from a surface
flow, having a plagioclase of Ann, cannot
be readily identified as having formed at
low pressure, and it is likely that the
properties of these feldspars are inherited
from an environment at depth.

The present study, giving quantitative
information on the number of feldspars,
their relative proportions, and the extent
of their solid solutions, emphasizes the
need for determining the composition of
each feldspar phase in a rock. The com-
positions of the feldspars, or any pair of
minerals sharing components, may yield
specific clues to the temperature and pres-
sure of formation. Inasmuch as more than
half of the earth's crust is composed of
feldspar, the feldspars are potentially most
valuable geothermometers and geopiezom-
eters.

THE SYSTEM CaAl2Si208-Si02-H20
D. B. Stewart

Despite the dominance of the feldspars
as constituents of the earth's crust, rocks
composed entirely of feldspar are not com-
mon. In addition to femic minerals, most
feldspar-bearing rocks contain significant
amounts of either free silica or more rarely
a feldspathoid. The phase relations of feld-
spars and silica have been studied inten-
sively at this Laboratory by Schairer and
Bowen, and Tuttle and Bowen have inves-
tigated the systems NaAlSi308-Si02-H20,
KAlSi308-Si02-H20, and NaAlSi308-
KAlSisOs-SiOs-HsO ("synthetic granite")
to pressures of 4000 bars.

The present study complements these
investigations, as the great bulk of all
granitic rocks contain the CaAl2Si20s com-
ponent, and the work on ternary feldspars
reported in this Year Book indicates the
profound effect of small amounts of this
component on the phase relations of the
feldspars. A feldspar, silica polymorph,

gas, and liquid saturated with H20 coexist
at a given pressure at only one tempera-
ture. One of the purposes of this investi-
gation of the system CaAl2Si208-Si02-
H20 is to determine the locus of four-
phase points as the pressure is changed.
The system CaAl2Si208-Si02-H20, to-
gether with the results of Tuttle and
Bowen for the same pressure, fixes the
end points of the H20-saturated liquidus
where feldspar and silica coexist. At this
fixed pressure the assemblage including
anorthite as one of the four phases coexists
at a higher temperature than any similar
assemblage of feldspar, silica, saturated
liquid, and gas. To the extent that the
three limiting ternary systems are known
at various pressures, the direction and mag-
nitude of the changes of the liquidus
caused by variations of H2O pressure dur-
ing crystallization can be described.

All the present results were obtained
with Yoder's apparatus, and his unpub-
lished results for the system CaAl2Si208-
H2O have been utilized and confirmed.
The Si02-H20 liquidus passes through
the points 1130° ±5° C at 2000 bars H20
pressure and 1065° ±5° C at 5000 bars.
These data modify and extend the diagram
given by Tuttle and England (1955), and
such high temperatures and pressures in-
dicate that quartz veins or the quartz cores
of pegmatites could not have crystallized
in situ from a hydrous Si02 melt. The rate
of lowering of the saturated liquidus in
the interval 2000 to 5000 bars is about 22°
C per 1000 bars H2O pressure, a rate two-
thirds of the rate of lowering of the satu-
rated liquidus of any feldspar in the same
interval.

A projection of the results at 2000 bars
H20 pressure is given in figure 48. The
only silica polymorph found was high
quartz; the saturated liquids contained
approximately 6 per cent H20 by weight.
The four-phase point is at 922° ±3° C, and
the ratio of the solid phases at this point
is CaAl2Si208 37 : Si02 63.

Schairer and Bowen (1947) found the

GEOPHYSICAL LABORATORY 215

eutectic of the anhydrous system to be at
1368° ±2° C and CaAl2Si208 50.5 : Si02
49.5 weight per cent. The shift of this
point toward increasing Si02 at low H20
pressures is a consequence of the greater
lowering of the saturated liquidus of the
silica polymorphs cristobalite and tridymite
relative to that of CaAl2Si208. A trend
toward a lower Si02 content at the four-
phase point was anticipated at pressures
higher than 2000 bars because at such pres-
sures the rate of lowering of the saturated

conditions the proportion of quartz crys-
tallizing from the liquid relative to feld-
spar crystals will decrease on that large
portion of the liquidus where CaAl2Si2Os-
rich plagioclase is the first feldspar phase
to appear with quartz.

When a silica mineral and feldspar are
crystallizing simultaneously, the ratio of
silica mineral to feldspar crystallizing will
also vary with the H20 pressure, increas-
ing as the H20 pressure increases to about
500 bars and decreasing as it increases

1500

1400

1000

Liquid + gas

PH2Ol2000barS

1500

1400

Anorthite + liquid + gas

Anorthite + /3quartz + gas

j 1 L

CoAI2Si202

20 30 40 50 60

Weight per cent

70

80

90

- 800
Si02

Fig. 48. Projection of the ternary system CaAl2Si208-Si02-H20 at 2000 bars H20 pressure.

liquidus of anorthite with increasing H20
pressure is greater than the rate of lower-
ing of the saturated liquidus of quartz.
Preliminary results on the four-phase point
at 5000 bars indicate a shift toward a lower
ratio of Si02 and suggest that the ratio
is close to An 42 : Si02 58. A similar shift
of the four-phase point toward Si02 at low
H20 pressures and reversal at higher pres-
sures was shown by Tuttle and Bowen in
the systems NaAlSi308-Si02-H20 and
KAlSi308-Si02-H20. The ratio of Si02
to CaAl2Si208 is greater than that for
either NaAlSisOs or KAlSi308 at corre-
sponding pressures. Accordingly, during
isobaric crystallization under equilibrium

above this pressure. These relations sug-
gest a possible geological barometer for
H20 pressure. The difference between the
Si02 content at the field boundary in the
anhydrous system and the maximum ratio
of Si02 in the corresponding hydrous sys-
tems is of the order of 10 per cent, and
should be easily detectable from the rela-
tive abundances of silica mineral and feld-
spar phenocrysts in lavas. In a sequence
of flows of the same composition it may
prove possible to use the variation of the
ratio of these phenocrysts to each other as
a measure of the H20 pressure on the
magma chamber where the phenocrysts
formed, and to demonstrate a relation be-

216 CARNEGIE INSTITUTION OF WASHINGTON

tween the intensity of volcanic activity and clases. In view of the occurrence of dif-

H20 pressure. ferent structural modifications of plagio-

Another application is indicated by J. }. clases in rocks, it is important to know
Norton, of the U. S. Geological Survey, what differences, if any, exist between the
through his detailed mineralogical studies optical properties of the different modifi-
of the Hugo pegmatite. This pegmatite cations. In last year's report (Year Book
contains quartz and feldspar in all its 55), it was shown that the optic axial
zones, and Norton's data indicate that angle (217) of natural plagioclases through-
starting from the wall zone the composi- out most of the composition range is
tions of successive zones of the pegmatite, changed significantly by heat treatment,
neglecting mica, first increase in Si02 and 2V is therefore a convenient measure of
then decrease. As the CaAl2Si208 content the structural state of a plagioclase. Meas-
is low, these observations suggest that the urements of the changes produced by heat
H2O pressure was changing, presumably treatment in refractive indices and in total
increasing, during crystallization of the birefringence have now been made. The
successive pegmatite zones. methods of measurement, described in de-

If the effect of other components is tail elsewhere (in press), are such as to

neglected, the fact that granitic rocks in give an estimated accuracy of ±0.0004 for

general plot near the silica-feldspar field refractive indices and ±0.0002 for total

boundary of the anhydrous "synthetic birefringence.

granite" system could be taken to mean Ten samples of chemically analyzed nat-
that the liquids from which these granitic Ural plagioclases whose optical properties
rocks crystallized formed at low H2O are accurately known were heated either
pressures. This need not be so, however, in tne dry way or hydrothermally until
as a similar position for this boundary also X-ray powder diffraction patterns showed
occurs at H2O pressures between 2000 and them to be inverted to the "maximum"
5000 bars, a value well within geologically high-temperature forms. The optical prop-
probable limits. The Si02 ratio apparently erties were then measured again by the
continues to decrease at higher pressures, same methods. The data are shown in
so that crystallization of granitic rocks at fi 49? wMch is j 1 self-explanatory;

JeT grlbdedPreSSUreS the m°St imPortant Point t0 note is that
TTTf. 1 * ^aio-^o-^ the difference in Nx between the high-
Work in the system CaAl2Si208-Si02- 11 r • r u*.
TT^-ii-ii •• 1 1 and low-temperature rorms is very slight
H20 will include quantitative data on the 11 1 • J .?
TT r\ £ Jl v »j £ j j or zero throughout the entire composition
H2U content or the liquids formed and . ° r tTt mi
mi • j- 1 ,1 . • ^1 range, so that measurements or Nx will
will indicate how this varies as the tern- , "' .. . . . r .

1 1 n-i 1. • give a reliable estimate or the composition

perature and pressure change. Ihe liqui- r 1 • 1 n r • 1

dus will be determined at 5000 and 10,000 o£ a P^gioclase regardless of its structural

bars. These data will be useful for quan- sAtate' For plagioclases more calcic than

titative aspects of the theory of the effects An20' ^ o£ ^ three refractive indices

of H20 in silicate systems. maY be usecL The composition being

known, the structural state can be deter-
OPTICAL PROPERTIES OF HEATED mined by the optical or X-ray methods
plagioclases described in previous Year Books (54, 55) .
/. R. Smith Studies of natural plagioclases by these
Optical measurements continue to be the methods will undoubtedly aid in the in-
most widely used method of determining terpretation of the crystallization and cool-
compositions of the ubiquitous plagio- ing history of many rocks.

GEOPHYSICAL LABORATORY 217

0.0140
0.0130
0 0120
0 0110
0.0100
0.0090
0.0080
0.0070

1.560

i i i i i I 1 1 1 1 1 1 1 1 1 1 1 1 1 i i 1 1 1 1 i i i i 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 i | i 1 1 1 1 i i 1 1 1 1 1 1 1 1

I 580

.550

1.540

1.530

Refractive indices

I 320 ' ' ' ] ' ' ' ' ' ' ' ' ' ' '

I ' I i i i i : i I i i I i i I i i i I I i I I i I I i I i i I i I I I I I i I I I I I i I I i i I I I I I I I I I I I

I 560

0

10

30

40

50

60

70

80

90

100

Mole per cent anorthite

Fig. 49. Light lines are total birefringence and refractive indices of natural plagioclases from
large plutonic intrusions. The circles and the heavy lines represent total birefringence and refrac-
tive indices of some of the same plagioclases after they had been inverted to the "maximum" high-
temperature modifications by heating.

THE CRYSTALLIZATION OF ROCK-FORMING MINERALS FROM
MAGMAS AND THE NATURE OF THE RESIDUAL LIQUID

/. F. Schairer

All the important rock-forming minerals
of the igneous rocks, with the exception of
quartz, have a variable composition. They
are solid solutions which undergo progres-
sive changes in chemical composition dur-
ing progressive cooling and crystallization
of the magma. Even when there are no
interruptions in the cooling cycle, there
are continuous or discontinuous changes
in the compositions of individual minerals
and in the kinds of minerals crystallizing
from the liquid phase. A mineral or as-
semblage of minerals stable at an early
stage in the crystallization process may
become unstable at a later stage and un-
dergo transformation to new mineral as-
semblages with changes in both chemical
composition and crystal system. Because
of the very complexity of the mutual melt-
ing and stability relations, much informa-

tion can be gleaned from the nature of the
minerals and mineral assemblages.

Two years ago (Year Book 54, pp. 141-
142) we discussed the relations between
early- and late-crystallizing minerals from
melts and the nature of residual liquids
from crystallization. During the past year
we have acquired specific information on
mineral assemblages, the direction of
change of composition of the liquid phase
during crystallization, and the nature of
the residual liquid in a large part of the
quaternary system Na20-MgO-Al203-
Si02.

The three volumes albite-corundum-
spinel-silica, albite-forsterite-cordierite-spi-
nel, and albite-cordierite-spinel-silica con-
stitute a large portion of the regular (equi-
lateral) tetrahedron employed to describe
the phase-equilibrium relations in the

218 CARNEGIE INSTITUTION OF WASHINGTON

quaternary system Na20-MgO-Al203- piercing points of curved lines within the

Si02. These volumes include that portion tetrahedron in the plane (join) which the

of the quaternary system of most interest particular diagram represents. Along these

to the geologist or petrologist who is con- curved lines, three curved surfaces (faces

cerned with the origin of igneous rocks or of adjacent primary phase volumes) inter-

with the mineral assemblages of meta- sect. Such a curved line is called a quater-

morphic rocks. All the crystalline phases nary univariant line or quintuple line,

found in these volumes are common rock- Along such univariant lines three solid

forming minerals or accessory minerals of phases are in equilibrium with the liquids

either igneous or metamorphic rocks. whose compositions lie on such lines. In

That portion of the quaternary system figures 50, 51, and 52, light lines show
lying between albite, corundum, forsterite, contours of temperature (isotherms) .
spinel, and silica is being studied by a series Figure 50 shows the phase-equilibrium
of triangular joins. Some years ago at this relations in the join albite-cordierite-silica.
Laboratory Greig made a reconnaissance This join cuts the phase volumes of mul-
of the system albite-forsterite-silica, a lite, cordierite, spinel, albite, tridymite, and
ternary system within this quaternary sys- cristobalite. Three univariant lines pierce
tern (unpublished data of J. W. Greig). this join at the points C, G, and I, re-
He located two ternary invariant points, spectively. The point / is a ternary eutectic,
albite + forsterite + magnesium metasilicate and is, therefore, a temperature maximum
+ liquid and albite + tridymite + magne- on the univariant line albite + cordierite +
sium metasilicate + liquid, which lie quite tridymite + liquid. The point F is the
close in composition to the binary side line binary eutectic albite + cordierite + liquid,
albite-silica. During the past year we pre- Mixtures whose compositions lie near the
pared an extensive series of melts in three ternary eutectic are viscous and quite diffi-
triangular joins; we present the data on cult to crystallize completely. They require
them now. runs of several weeks' duration to attain

In figures 50, 51, and 52, the phase- equilibrium between crystals and liquid,

equilibrium data obtained by the method Experiments are still in progress to locate

of quenching are presented graphically, the position and temperature of point /

In these figures, open double circles repre- more accurately.

sent the compositions of chemical com- Figure 51 shows the phase-equilibrium

pounds, and black dots represent the com- relations in the join albite-forsterite-cordi-

positions of mixtures studied. The dia- erite. This join cuts the phase volumes of

grams are divided into areas by heavy forsterite, albite, spinel, mullite, and cordi-

curves. The areas labeled mullite, cordi- erite. This last phase volume is cut twice

erite, spinel, albite, etc., are plane sections by this join. Three univariant lines pierce

of phase volumes (with curved faces) this join at the points N, 0, and P, re-

within the tetrahedron in the join which spectively. The point P is a ternary eutectic

the particular diagram represents. Each and is, therefore, a temperature maximum

phase volume within the tetrahedron gives on the univariant line albite + forsterite +

the compositions of all quaternary liquids cordierite + liquid. The point Q is the

(one liquid phase) in equilibrium with a binary eutectic albite + forsterite + liquid,
single crystalline phase. The heavy curves Figure 52 shows the phase-equilibrium

on the joins are traces of the curved bound- relations in the join albite-magnesium

ary surfaces between two adjacent primary metasilicate-cordierite. This join cuts the

phase volumes within the tetrahedron in phase volume forsterite, protoenstatite,

the plane (join) which the particular dia- albite, cordierite, spinel, and mullite. Three

gram represents. Where three heavy curves univariant lines pierce this join at the

meet at a point in the join, these points are points U, V , and W, respectively.

GEOPHYSICAL LABORATORY 219

H /062*3"
3°

CORDIERITE

2Mg0.2AI?03.5SiO;

40 SO 60D 70

/374*3°
WE/ GHT PERCEN T

80 E 90 F ALBITE

tests' ,098+3° Na20 AI203.6SiO2

Fig. 50. Equilibrium diagram of the join albite-cordierite-silica, showing compositions studied,
primary phase volumes cut by this join, piercing points of quaternary univariant lines, and isotherms.

CORDIERITE
. @2Mg0.2AI203.5Si02

/468t?jt\m \

/O,
/44S*2J\

20k

I \V

\ ° \ A

WW

I890±20\
FORSTERITE
2MgO.Si02

40 SO

WE/GHT PERCENT

1098*3"' ALB IT E

Na20.AI20,.6Si0,

Fig. 51. Equilibrium diagram of the join albite-forsterite-cordierite.

220 CARNEGIE INSTITUTION OF WASHINGTON

The data just presented on three triangu-
lar joins indicate within approximate limits
the temperatures and compositions of seven
quaternary invariant points and describe
the crystallization behavior of compositions
in the volumes albite-corundum-spinel-
silica, albite-forsterite-cordierite-spinel, and
albite-cordierite-spinel-silica. The relations
between univariant lines and ternary and
quaternary invariant points are shown
diagrammatically in figure 53. All uni-

A, B, C, and D from the other quater-
nary invariant points and that D is the crys-
tallization goal of the liquid in the vol-
ume albite-corundum-spinel-silica; simi-
larly, that the temperature maxima in CE
and EF separate the quaternary invariant
point E from the other quaternary invari-
ant points and that E is the crystallization
goal of the liquid in the volume albite-
forsterite-cordierite-spinel; and, similarly,
that the temperature maxima in EF and

CORDIERITE

^2Mg0.2Ai203.5Si02

Mg METASILICATE /a

MgO.Si02

WEIGHT PERCENT

-1097 ±3° ALBITE

Na20.AI203.6Si02

Fig. 52. Equilibrium diagram of the join albite-magnesium metasilicate-cordierite.

variant lines within the tetrahedron are
curved lines. For simplicity, they are
shown as straight lines in the figure, which
is not spatial but merely depicts the rela-
tions of quaternary invariant points to
one another and to certain ternary invari-
ant points lying in a face of the tetra-
hedron. The lengths of the univariant lines
in figure 53 are arbitrary and without sig-
nificance. Arrows on the univariant lines
indicate the direction of falling tempera-
ture.

An examination of figure 53 shows that
the temperature maxima in CE and DG
separate the quaternary invariant points

DG separate the quaternary invariant
points F and G from the other quaternary
invariant points and that G is the crystal-
lization goal of the liquid in the volume
albite-cordierite-spinel-silica.

An examination of figure 50 shows that
the ternary eutectic albite + cordierite +
tridymite + liquid (/ of fig. 50) lies very
close in temperature and composition to the
binary eutectic albite + tridymite + liquid
(H of fig. 50). An examination of figure
53 and a study of the tetrahedral model of
the quaternary system shows that both D
and G (fig. 53) must also lie close in tem-
perature and composition to each other

GEOPHYSICAL LABORATORY 221

(MgO-AI203-3i02)

146015°

(MgO-AI203-Si02)

i -^

SAP
CORD

(Na20-Al203-Si02)

(AB-COR-SP)

AB
COR

145315°

1104+3°

AB

COR

MU

/

AB
COR
MU
SP

AB

MU

(MgO-AI203-Si02) )4I575±5

COR

MU
SP

(AB-FO-SP)

FO
SP

1360 + 5°

FO
PR

(MgO-AI203-Si02)

MU

CORO

SAP

MU

CORO

SP

MU

CORO

SAP

MU (Mg°-Al203-Si02)

<SAP £

148213°

(No20-AI203-Si02)
► 1050110°

AB
MU
TR

kSP

/

MU
CORO
SP
AB

MU
CORO

AB

MU
TR
CORO

MU
.. TR

(MgO-AI203-S'02)
_f 144315°

AB

CORO
*SP

/

AB

CORO
SP
FO

FO
CORD

137015°

(MgO-Ai203-Si02)

AB
■ CORO
, FO

AB

CORO
^/FO
p PR

AB
CORO

A8

TR
CORD

AB
TR

CORD

PR AB
PR

TR

(A6-F0-SiG2)

CORD

PR

TR

g© 134515°
(MgO-Al203-Si02)

(AB-FO-Si02)

Fig. 53. Diagram showing univariant lines and their relation to ternary invariant points (small
black dots and letters a through /') in the limiting systems to quaternary invariant points (large black
dots and capital letters). These lines and points do not lie in a plane. Only their relations to one
another are shown in this diagram, which is not intended to depict their angular spatial relations.
The length of the lines and the position of a temperature maximum on a line are arbitrary and
without significance. Arrows indicate the direction of falling temperature. Abbreviations for crys-
talline solid phases along the lines and at the points: AB, albite; COR, corundum; MU, mullite;
SP, spinel; CORD, cordierite; SAP, sapphirine; TR, tridymite; FO, forsterite; PR, protoenstatite.

222 CARNEGIE INSTITUTION OF WASHINGTON

and to the ternary eutectic albite + cordier-
ite + liquid. Thus we see that during
crystallization all compositions in two large
volumes albite-corundum-spinel-silica and
albite-cordierite-spinel-silica proceed to-
ward a similar goal for the composition of
the residual liquid. This goal is a soda
granite in composition in these potash-
and lime-free mixtures.

Some of the liquids in the two volumes
just discussed are not too far removed from
possible magmas. The "simplified mag-
mas" in Na20-MgO-Al203-Si02 are, o£
course, anhydrous instead of having at
least a small water content. There is only
soda present with potash lacking, only
magnesia present with ferrous oxide lack-
ing, and only alumina present with ferric
oxide lacking; there is no lime. In spite

of these deficiencies, it can be seen that a
large range of rock-forming compositions
would give a common end product of
crystallization, particularly if differentia-
tion proceeded by fractional crystallization.
Note the significant fact that this same end
product, a soda granite, would be reached
even if these simplified magmas were con-
taminated with numerous small fragmen-
tal xenoliths of basic rocks such as peri-
dotites or dunites, or if they were con-
taminated by assimilation, in whole or in
part, with xenoliths of highly aluminous
sediments.

Thus these studies of soda-bearing melts
add further quantitative evidence to sup-
port Bowen's arguments for the impor-
tance in petrogenesis of KAlSiCX-NaAl-
Si04-Si02 as "petrogeny's residua system."

GRANITIC PEGMATITES

P. M. Orville

The origin of granitic pegmatites and
the unusual textures and structures within
them have been the subject of geological
investigation and speculation for many
years. Recent laboratory investigations in
the Ab-Or-Si02-H20 ("synthetic gran-
ite") system enable the course of crystal-
lization in melts which closely approxi-
mate the granitic pegmatites in composi-
tion to be considered on a quantitative
basis.

Pegmatites enriched in potassic feldspar
in their upper portions are common in the
Black Hills, South Dakota; Pala, Cali-
fornia; Colorado; New England; and the
southern Appalachian states. In 1954 at
Yale University a study of a complex of
layered pegmatites in the southern Black
Hills, South Dakota, showing such a dis-
tribution of potassic feldspar, was begun
in the hope that the field and laboratory
data together might lead to an under-
standing of the process by which this
segregation took place.

The pegmatites studied are thin, steeply
tabular bodies separated parallel to their
median plane into an upper unit contain-

ing large perthitic microcline crystals set
in a fine-grained plagioclase-quartz matrix
and a lower unit consisting of plagioclase-
quartz-muscovite aplite. Such sharply lay-
ered pegmatite dikes are gradational into
dikes having a uniform composition and
aplitic texture from wall to wall and a
mineralogic and bulk chemical composi-
tion identical with that of the layered
pegmatites.

The bulk compositions of the pegmatites
(neglecting muscovite) fall on, or very
near, the field boundary between the
quartz and alkali feldspar fields approxi-
mately halfway between the "granitic mini-
mum" composition and the Ab-Si02 side
line of the liquidus in the system Ab-Or-
Si02-H20 at 2000 bars. The composition
of the upper unit corresponds closely to
the composition of the minimum in the
synthetic granite system at this pressure,
and the composition of the lower unit
falls near the Ab-Si02 side line on the
quartz-alkali feldspar field boundary.

The presence of a few per cent An in the
melt raises the crest of the ternary feldspar
solvus by a large amount, and the solvus

GEOPHYSICAL LABORATORY 223

intersects the liquidus within a short dis-
tance of the Or-Ab side line (Yoder, Stew-
art, and Smith, this report). The plagio-
clase from these pegmatites is a calcic
albite (Ab93An7), and this amount of An
may be sufficient so that crystallization
of two feldspars takes place from the
magma.

It is possible that the separation of these
pegmatites into two compositional units

is the result of fractional crystallization of
a hydrous silicate melt. The aplitic unit
might represent the first stage of crystal-
lization as quartz and albite crystallize to-
gether along the quartz-plagioclase field
boundary. The coarse-grained unit con-
taining potassic feldspar might represent
the crystallization of the rest liquid that
approaches the composition of the synthetic
granite minimum.

PYROXENES

For a petrologist concerned with the
genesis of igneous and high-grade meta-
morphic rocks there is no more important
system than the pyroxene quadrilateral
MgSiOs-CaMgSisOe-CaFeSisOe-FeSiOs.
Pyroxenes whose compositions (neglecting
minor components) lie in this quadri-
lateral are major constituents of basalts,
gabbros, andesites, and important groups
of metamorphic rocks such as the iron
formations. Rhyolites often contain iron-
rich pyroxenes in minor, though geneti-
cally significant, amounts. The pyroxenes
also form monomineralic and bimineralic
aggregates in the pyroxenites and perido-
tites. Laboratory study of the pyroxenes
will not only provide quantitative data
about the physical conditions under which
these rocks have formed but will also give
us a better understanding of the courses
of chemical fractionation in magmas.

The importance of pyroxenes has long
been recognized by experimental penolo-
gists. The melting relations along the join
MgSi03-CaMgSi206 were first studied at
this Laboratory in 1914. The joins
CaFeSisOe-FeSiOs and MgSi03-FeSi03
have been studied more recently. Investi-
gations along the side lines of the quadri-
lateral MgSi03-CaMgSi206-CaFeSi206-
FeSiOs have yielded data of vital interest,
but inasmuch as most natural pyroxene
compositions fall well within the quadri-
lateral the major gains have yet to be made.

It is our intention to carry through a
study of both the melting and subsolidus
equilibria in the entire quadrilateral. The

/. F. Schairer and F. R. Boyd, Jr.

data discussed below, on the join MgSiOs-
CaMgSi206, represent our initial effort in
what will be a continuing program in com-
ing years.

THE JOIN MgSiO;r-CaMgSi206
F. R. Boyd, Jr., and J. F. Schairer

The crystal-liquid equilibria along the
join MgSi03-CaMgSi206 were determined
many years ago by Bowen (1914) in his
study of the system forsterite-diopside-
silica. Bowen's work has been checked in
detail and found to be substantially ac-
curate. We have determined liquidus tem-
peratures on compositions between MgSi03
and CaMgSi2Oe prepared at 5 weight per
cent intervals. All our data lie precisely
on the curves of Bowen. For those com-
positions with a forsterite liquidus we have
determined the temperature of appearance
of pyroxene. These data are in agreement
with those interpolated from Bowen's data.
The beginning of melting was determined
for all these compositions, and the results
check those of Bowen closely.

Atlas (1952) has investigated the sub-
solidus region in the system MgSiOs-
CaMgSi2Oe using a lithium fluoride flux.
Our work confirms that of Atlas on the
existence and approximate position of the
solvus. Our locations of the solvus curves
differ from those in the Atlas diagram
largely in that we have found that the
solvus intersects the solidus; i.e., there is
not a complete solid solution between
MgSiOs and CaMgSi2Oe.

224 CARNEGIE INSTITUTION OF WASHINGTON

Our work on the join MgSiOs-
CaMgSi206 is not yet finished. We have
presented in figure 54 only those portions
of the subsolidus diagram that we think
are unlikely to undergo any further modi-
fication. Our principal uncertainty re-
mains the relations involving the various
polymorphs of MgSiOs.

We have used a variety of techniques
in determining the subsolidus phase rela-

liquid or crystals + liquid

orfhoenstatite ♦ diopside

Q Point on so'vus boundary I

Q Point on solvus
^ ot 500 bars n
Q* Solvus boundory li

a Single-phase njn . dry

0 Single -phase njn , 500 bars H-0

to about 1200° C on the diopside side of the
solvus and down to about 1100° C on the
enstatite side.

The determinations of compositions of
phases along the solvus curves have been
made by X-ray methods. We have devel-
oped four curves relating X-ray spacing to
composition in this system. Two of these,
which have proved to be superior in prac-
tice and have been used almost exclusively,
are presented in figures 55 and 56.

Figure 55 shows the variation in spacing

of the 311 peak in diopside as a function
of composition. This curve is linear be-

MgSiOj K> 20 30 40 50 60

Weight per cent

70

90 CoMqSi206

Fig. 54. Subsolidus diagram of the system
MgSi03-CaMgSi206.

tions along this join. Runs on the solvus
curves (A, B, C, and D in fig. 54) have
been made dry in the temperature range
between 1100° and 1400° C. Runs have
been made in hydrothermal quenching
apparatus from a temperature of 1150° C
down to about 850° C. We have also used
a heating stage on an X-ray diffractometer
up to a temperature of about 1350° C.

Our starting material for both dry and
hydrothermal runs has generally been
glass. We have, however, been able to
unmix homogeneous, crystalline phases
(prepared at temperatures just below the
solidus temperature) at temperatures down

Fig. 55. Variation of the 311 spacing of di-
opside with composition in the system MgSiOs-
CaMgSi2Oe.

tween Dig5En5 and DieoEn4o. Between
DigsEns and pure diopside there is an
abrupt change in slope, which does not ap-
pear to be an inversion since the X-ray
parameters of both pure diopside and
DigsEn5 prepared hydrothermally at about
900° C show no significant differences
from the parameters of samples prepared
dry above 1300° C. The solvus curve
ceases to be of interest in the composition
region between DigsEns and pure diopside;
accordingly, we have not further pursued
the matter of this break in slope.

Figure 56 shows the variation in spacing
with composition of a clinoenstatite peak
at about 57° 20 CuKa. This spacing has
been used to determine the position of the
solvus curve C.

Points on the solvus boundaries deter-

GEOPHYSICAL LABORATORY 225

mined with the aid o£ these X-ray curves
are plotted as circles in figure 54: open
circles represent data from dry runs; circles
with a horizontal bar, data from hydro-
thermal runs at 500 bars (H20). Hydro-
thermal data and dry data overlap at 1150°
C on curve A. The agreement is within
experimental error. Along curves A and B
each point represents an average of up to
five runs using, generally, two or more

Weight per cent CoMgSi0206

Fig. 56. Variation of a clinoenstatite spacing
with composition in the system MgSi03-
CaMgSi2Os.

bulk compositions. The spread of data at
a given temperature is seldom more than
2 per cent in composition. Along curve C
we have had trouble with X-ray interfer-
ence from the coexisting diopside phase,
and only bulk compositions in the two-
phase region closest to the solvus curve
were used to fix its composition.

We have checked our locations of the
solvus curves by making runs in the single-
phase regions close to the curves over a
wide range in temperature. These runs
are indicated in figure 54 by squares or
rectangles. The agreement in X-ray spac-

ing of these runs with those made at
1365° C, just below the solidus, has been
within X-ray error (about ±0.02° 20).

There is a pronounced change in slope
of the solvus boundary on the diopside side
at about 1075° C. It is probably produced
by the inversion of orthoenstatite to a
higher-temperature form.

We have not yet established a determina-
tive curve for orthoenstatite solid solutions.
In the temperature range below 1050° C,
however, diopside appears in our hydro-
thermal runs whose composition is richer
in CaMgSi2Oe than EngoDiio. Curve D
must, therefore, lie approximately as
shown.

We are not yet sure of the stable crystal-
line form of the pyroxene in the single-
phase region bordering curve C. In dry
runs bordering curve C we have generally
obtained clinoenstatite; in hydrothermal
runs (all below 1100° C) we obtain proto-
enstatite. We have located curve C by
means of the clinoenstatite spacing in fig-
ure 56. Whatever form of enstatite turns
out to be stable along curve C, it is improb-
able that the position of the solvus curve
itself will have to be modified. The inver-
sions involving clino- and protoenstatite are
very rapid, whereas the unmixing reaction
is relatively sluggish. Hence, it is most un-
likely that any unmixing takes place dur-
ing inversion in the quench.

In our hydrothermal studies on pure
MgSi03 we have prepared well crystallized
orthoenstatite and protoenstatite. Unlike
protoenstatite prepared dry, the hydro-
thermal material shows little tendency to
invert to clinoenstatite. We shall withhold
our detailed results on this aspect of the
problem, however, until we are sure of the
stable fields of these various polymorphs.

CHLORITOID

L. B. Halferdahl

Chloritoid is a geologically significant mineralogists. Many petrologists believe
mineral which has been described from that it can form only under restricted con-
many occurrences, and yet its properties ditions, notably those in which stress plays
and composition are still disputed among a part, despite the fact that its growth has

226 CARNEGIE INSTITUTION OF WASHINGTON

been recorded in other geological environ-
ments. Some preliminary results of an in-
vestigation of the chemical composition,
the optical and X-ray properties, the sta-
bility, and the nature of the occurrences of
chloritoid were reported last year. Addi-
tional results are given here.

Five new chemical analyses of chloritoid,
including one so-called "ottrelite" from
Salm Chateau, Belgium, and all the other
available analyses that are considered re-
liable indicate that the composition of
chloritoid can be represented by the gen-
eral formula FeO • Al2Os • SiOa • HaO with
the following ranges in the amounts of
substitutions now on record :

Substitution

Atomic Per Cent

Mg ->Fe"

0to40

Mn ->Fe"

0tol7

Fe'" -»A1

OtolO

F ->OH

Less than 0.25

Fe'" -»Fe"H

Small

2Fe'"-»3Fe"

Very small

orientation of the optical indicatrix, how-
ever, one or even two principal optic direc-
tions may chance to lie in (001) with
neither coinciding with the b axis. There-
fore, an observation that one principal
optic direction lies in (001) may not be
sufficient to establish monoclinic symmetry
for the chloritoid grain in question, par-
ticularly if no principal optic direction lies
in (001) in other grains.

Optical measurements made on samples
of chloritoid from 11 localities including
both triclinic and monoclinic polymorphs
show that Z makes angles varying from
2° to 30° with the normal to (001). In

TABLE 23. Unit Cell Parameters of
Chloritoid Polymorphs

Locality

Cstiiaciksaring,
Turkey

Chibougamau,
Quebec

Polymorph Monoclinic

Triclinic

It is important to realize that these may
not represent the maximum possible
amounts of substitutions. Nevertheless, the
purity of the chloritoid used in all previ-
ously published analyses showing more
than this amount of manganese is subject
to question.

X-ray investigations have revealed two
chloritoid polymorphs with the unit cell
parameters given in table 23. On a geo-
metrical basis it is possible to consider the
unit cell of the structure of monoclinic
chloritoid determined by Harrison and
Brindley (1957) as consisting of two tri-
clinic unit cells in a zigzag arrangement.
Such cells have parameters very similar to
those of the triclinic chloritoid from Chi-
bougamau, Quebec. This structural rela-
tionship is similar to that of the clino- and
orthopyroxenes.

Measurements made on the universal
stage show that the orientation of the op-
tical indicatrix varies considerably in differ-
ent grains of the same chloritoid sample.
In triclinic chloritoid no principal optic
direction is required by symmetry to lie
in (001). Because of the variations in the

a 9.48

b 5.48

c 18.19

a 90° 0'

0 101° 46'

Y 90° 0'

G(meas.) .. 3.79

G(calc) ... 3.80

9.50
5.48
9.16

96° 53'
101° 49'

90° 2'
3.79
3.79

the monoclinic polymorphs measured,
X=b. In the triclinic polymorphs, X and
Y make angles from 0° to 30° with (001),
but in many Y is closer to (001) than X.
The orientation of X and Y with respect
to the a and b axes in triclinic chloritoids
was not determined, because no crystallo-
graphic directions other than the (001)
cleavage were identified with certainty in
the thin section and grain mounts of the
specimens studied. In chloritoids from 11
localities the optic angle about Z ranges
from 36° to 125°. Most, however, are
optically positive with 2Vz in the range
45° to 60°. The refractive indices of chlori-
toid measured in this study and those
previously published show that a varies
from 1.713 to 1.730, (3 from 1.719 to 1.734,
and y from 1.723 to 1.740. All indices de-

GEOPHYSICAL LABORATORY 227

crease with increasing magnesium content
of the chloritoids.

The results of hydrothermal experiments
on natural chloritoid and on some oxide
mixtures of the chloritoid composition are
presented in figure 57. Because of the slug-

1 1

1

A L

1 1

a a a

1 1

-

-

A C

a a b

Chloritoid

Staurolite +

+ Fluid

Almandine
+ Hercynite

+ Fluid

-

A

A A

•

i a a ■
/
/
S
/
/

\

0

/ Iron Cordierite

\ Hercynite +

0

/ + Hercynite
/ + Fluid

\ Mullite +
\ Liquid +
\ Fluid

- O O 0

XX X

\

- O 0 ,

1 1 y

1

I 1 1 1

i V i

Temperature in °C

Fig. 57. Preliminary univariant curves for the
reactions chloritoid^±iron cordierite + hercynite
+ fluid (solid lines), chloritoid^±almandine +
staurolite + hercynite + fluid (solid lines), iron
cordierite + hercynite + fluid^±staurolite + alman-
dine (long dashes), and iron cordierite -r-fluid^=±
mullite + liquid + fluid (short dashes).

Explanation of Symbols

Runs made in sealed platinum tubes and one
in an evacuated glass tube: Solid circle, chlori-
toid synthesized from mixes of oxides. Open
circle, natural chloritoid did not change. Cross,
natural chloritoid broke down to iron cordierite,
hercynite, and fluid; iron cordierite + hercynite
synthesized from oxide mixes. Solid square,
natural chloritoid broke down to almandine,
staurolite, hercynite, and fluid; almandine -f-
staurolite + hercynite synthesized from oxide
mixes.

Runs made in co-operation with Boyd and
England in squeezer apparatus built by them:
Open triangle, natural chloritoid persisted. Open
square, natural chloritoid broke down to alman-
dine, staurolite, hercynite, and fluid.

Solid rectangle, point determined by quench-
ing (Schairer and Yagi, 1952).

gishness of the reactions at pressures below
10,000 bars and the difficulty of controlling
the state of oxidation, these results must be
regarded as preliminary. Nevertheless,
they do provide an acceptable alternative
to the widely held belief that the action of
stress at any particular temperature and
composition determines the mineral asso-
ciations in regional metamorphism. This
alternative is that the differences in min-
eralogy between regionally metamorphosed
rocks and those formed in contact aureoles
may be the result of different pressures
prevailing in the two environments. Thus,
in quartz-bearing rocks containing chlori-
toid, the reaction chloritoid + quartzrfron
cordierite + fluid might be expected in con-
tact metamorphism, and the reaction
chloritoid + quartz ^ almandine + stauro-
lite + fluid in regional metamorphism.
Chloritoid-cordierite associations are known
in the Vredefort Dome area of South
Africa and the Santa Monica Mountains of
California. Chloritoid-almandine-staurolite
associations are known from many places
in the Alps and from Unst in the Shetland
Islands. At intermediate pressures the
association staurolite-cordierite might be
expected. Such associations are known
from the Lizard area of Cornwall and
from the contact aureole of the Bushveld
complex in South Africa.

The conditions under which one of the
chloritoid polymorphs will form in prefer-
ence to the other are still unknown. The
differences in energy required to produce
one in place of the other are probably very
small, and, hence, could be detected only
with very great difficulty by the methods
at present available. Natural occurrences,
however, do provide some clues. Chlori-
toids obtained from hydrothermal veins
and from hydrothermally altered rocks
have been found to be triclinic. Chloritoids
from regionally metamorphosed rocks have
been found to be monoclinic or triclinic or
both. Chloritoids associated with alman-
dine, kyanite, or staurolite are monoclinic.
This information may suggest that the

228 CARNEGIE INSTITUTION OF WASHINGTON

monoclinic polymorph forms more readily
than the triclinic under severe conditions
of pressure and temperature, or under
more moderate conditions prevailing for
a long time.

Studies of chloritoid-bearing rocks, pub-
lished reports of chloritoid occurrences,
and phase-rule considerations indicate that
chloritoid can exist under certain conditions
in equilibrium with one or more of the
following minerals: quartz, chlorite, mus-
covite, magnetite, hematite, ilmenite, rutile,
paragonite, almandine, staurolite, cordi-
erite, biotite, pyrophyllite, kaolinite, kya-
nite, andalusite, and possibly sillimanite,
corundum, diaspore, glaucophane, albite,

anorthite, epidote, zoisite, margarite, cal-
cite, and siderite. The most common min-
eral assemblage in chloritoid-bearing rocks
is chloritoid-quartz-chlorite-muscovite-ru-
tile-iron oxide, the iron oxide being one or
more of magnetite, hematite, ilmenite.
These mineral associations and the en-
vironment in which they occur — regional
rnetamorphism, contact metamorphism,
hydrothermal veins and hydrothermally
altered rocks, emery deposits — give further
support to the suggestion that stress plays
no greater part in the formation of chlori-
toid than in the formation of such minerals
as muscovite, andalusite, pyroxene, or
feldspar.

ALKALI AMPHIBOLES

W. G. Ernst

Alkali amphiboles, one of the three
principal groups of amphiboles, occur in
a wide variety of rock types. Riebeckite is
an important constituent of alkalic, silicic
igneous rocks; it has also been described
as a low-temperature authigenic mineral.
Glaucophane and crossite (an intermediate
member of the riebeckite-glaucophane se-
ries) are abundant in certain metamorphic
rocks.

Hydrothermal study of the riebeckite-
glaucophane series was initiated in 1956;
preliminary data are now available. Mag-
nesian riebeckite, Na2Mg3Fe2+3Si8022-
(OH)2, and ferrous riebeckite, Na2Fe3+2-
Fe2+3Si8022(OH)2, have been synthesized,
and work on these minerals as well as on
magnesian glaucophane, Na2Mg3Al2Si8-
022(OH)2, is in progress.

With the exception of magnesian glauco-
phane, all compositions investigated con-
tain iron capable of existing in two oxida-
tion states; therefore the physical param-
eters governing stability include the partial
pressure of oxygen as well as the tempera-
ture and total (H20) pressure. Accord-
ingly, in experiments employing iron-
bearing minerals the Po2 was fixed using
buffers as described by Eugster.

Magnesian riebeckite. Preliminary

Ph2o-T stability diagrams for Na2Mg3-
Fe2+3Sis022(OH)2 using several Po2 buff-
ers are presented in figures 58, 59, and 60.
Each diagram represents the intersections
of boundary surfaces occurring within the
Ph2o-T-Po2 volume with the surface de-
fined by a specific buffer, and projected
along the Po2 axis (compare fig. 13 with
%• H).

All three riebeckite diagrams exhibit the
same general sequence of phases. The
form of the riebeckite stability field is
similar to that of other hydrous minerals.
At low T and at a Ph2o too low to form
riebeckite, the stable assemblage consists
of one or two iron oxides, an olivine,
Na20 • 2MgO • 6Si02, Na20 ■ 5MgO • 12Si02
(compounds previously reported in the
system Na20-MgO-Si02 by Schairer,
Yoder, and Keene), and vapor, except for
the magnetite-hematite diagram, where
acmite takes the place of Na20'2MgO'
6Si02. At slightly higher temperatures, the
assemblage consists of one or two iron ox-
ides, Na20 • 5MgO • 12Si02, an olivine, liq-
uid, and vapor. In a still higher tempera-
ture range the Na20-5MgO'12Si02 melts
incongruently to orthopyroxene and liquid,
and the assemblage then becomes one or
two iron oxides, an olivine, an orthopy-
roxene, liquid, and vapor.

1000

950

900

850

800

Hematite ♦ magnetite + forsterite (?)
+ Na20- 5MgO- l2Si02+
-liquid + vapor

Hemotite + magnetite + forsterite (?)
+ enstotite + liquid + vapor

Riebeckite + vapor

^Hematite + mognetite + forsterite D(?)
+ Na20 5MgO l2Si02
+ acmite + vapor

500 1000

Pressure H20, bars

1500

2000

Fig. 58. Pn2o-T diagram for magnesian riebeckite using the hematite-magnetite buffer.

Magnetite + olivine
+ N020 5MgO-l23i02v
.+ liquid + vapor

Magnetite + olivine + hypersthene
+ liquid + vapor

a Riebeckite + vapor

a
a

Magnetite + olivine
+ Na20- 5MgO- l2Si02 +
No20 ■ 2MgO ■ 6S1O2 + vapor

Pressure (H20), bars
Fig. 59. Ph2o-T diagram for magnesian riebeckite using the magnetite-fayalite-quartz buffer.

O 850

1000

Pressure (H20). bars

2000

Fig. 60. Ph2o-T diagram for magnesian riebeckite using the magnetite-wiistite buffer.

229

230 CARNEGIE INSTITUTION OF WASHINGTON

The upper stability limit of riebeckite at
a given Ptot is elevated by increasing the
partial pressure of oxygen. The iron in
the amphibole is predominantly ferric, and
much of the iron is bivalent in the high-
temperature phases. During decomposi-
tion there is an evolution of free oxygen.
The evolution of oxygen results in a vol-
ume increase, so increased oxygen partial
pressure tends to drive the reaction to-
ward the assemblage of smaller volume
(riebeckite) .

All phases containing MgO also contain
some FeO, the amount depending on the
bulk composition and Po2 as well as on
the temperature and total pressure. Rie-
beckite contains some ferrous iron, too, and
breaks down over a small temperature
interval (approximately 20°). For this
reason, curves bounding the riebeckite
fields in figures 58, 59, and 60 define the
highest temperature at which amphibole
is stable.

Na20 • 5MgO • 12Si02 and Na20 • 2MgO •
6Si02 apparently increase their upper sta-
bility limits as the partial pressure of oxy-
gen declines. Some ferrous iron is probably
incorporated in these minerals. Oxygen in
equilibrium with these compounds is used
up during breakdown (an oxidation proc-
ess in regard to the liquid and solid

phases). Hence a diminished Po2 tends
to increase the stability field of the large-
volume assemblage (sodic silicate plus
oxygen) .

The experimental formation of riebeck-
ite in equilibrium with a melt rich in
Na20 and Si02 agrees with natural oc-
currences where riebeckite appears as a
magmatic mineral in certain alkalic, silicic
intrusives.

Ferrous riebeckite. Exploratory runs on
the composition Na2Fe3+2Fe2+3Si8022-
(OH)2 using a fayalite-magnetite-quartz
buffer indicate that ferrous riebeckite
breaks down at a temperature approxi-
mately 150° lower than the breakdown
temperature of its magnesian analogue.
Decomposition products below about 1500
bars Ph2o are fayalite, acmite, magnetite,
quartz, and vapor; above this pressure
riebeckite melts incongruently to fayalite,
magnetite, quartz, liquid, and vapor.

Magnesian glaucophane. Reconnaissance
runs on the composition Na2Mg3Al2Si8-
022(OH)2 have yielded an amphibole
whose upper stability limit is 20° to 80°
lower than that of magnesian riebeckite.
The high-temperature assemblages include
forsterite, albite, enstatite (?), liquid, and
vapor.

RECONNAISSANCE IN THE SYSTEM FeO-Fe203-Si02-H20

/. R. Smith

The problem of the origin and subse-
quent enrichment and metamorphism of
iron ores of the Lake Superior type con-
tinues to be of great importance to eco-
nomic geologists and petrologists. The
bulk compositions of many of the ores
and related rocks lie close to the quater-
nary system FeO-Fe203-Si02-H20. Ex-
perimental studies in the FeO-Fe203 bi-
nary system have already contributed to
our understanding of the conditions under
which the iron oxides might have formed,
but little is known of the stability relations
of ternary and quaternary compounds in
the quaternary system, largely because of
the experimental difficulties involved in

controlling the partial pressures of oxygen
in the presence of water under pressure.
The technique developed by Eugster of
using as buffers polyphase assemblages for
which the temperature versus partial pres-
sure of oxygen equilibrium relations are
known offered a means of investigating
portions of the system. It was planned to
investigate first the stability fields of min-
nesotaite (3FeO'4Si02*H20) and green-
alite (3FeO-2Si02-2H20), both of which
occur with the iron ores of the Lake Su-
perior region. Knowledge of the stability
fields of these minerals in terms of partial
pressure of oxygen as well as of tempera-
ture and water pressure would provide

GEOPHYSICAL LABORATORY 231

further evidence of the conditions existing
during the formation and later history of
the ores.

In attempts to synthesize minnesotaite
and greenalite, over 100 experiments have
been made at temperatures from 250° to
600° C, at total water pressures from 500
to 30,000 psi, and at various partial pres-
sures of oxygen. In the experiments it was
found that, at temperatures above 400° C
and at a partial pressure of oxygen in
equilibrium with iron and magnetite,
fayalite forms readily from finely ground
mixtures of silica glass and any of the fol-
lowing: ferrous oxalate, hematite, magnet-
ite, wiistite, or native iron. Silica in ex-
cess of that required by the fayalite formula
crystallizes as quartz. Small amounts of
phases other than fayalite and quartz ap-
peared in the products of these experi-
ments: In charges which were initially
ferrous oxalate and silica glass in minne-
sotaite proportions, a phase with an atomic
spacing of 10.2 A was formed at tempera-
tures between 450° and 600° C; when the
charge was held for longer periods under
the same conditions, this phase was re-
placed by another with an atomic spacing
of 7.2 A, which is similar to the basal spac-
ing of greenalite. Still longer treatment of
the same material failed to increase the
amount of the phase with the 7.2-A spac-
ing relative to the amount of fayalite and
quartz, which made up the bulk of the
products of the experiment. Similarly, in
experiments starting with native iron and
silica glass in either minnesotaite or greena-
lite proportions, at temperatures above
400° C, a total water pressure of 30,000 psi,
and a partial pressure of oxygen in equi-
librium with iron and magnetite, small
amounts of a phase with an atomic spac-
ing of 13.1 A form with predominant
fayalite and quartz; these small amounts
persist after further treatment under the

same conditions, but do not grow at the
expense of fayalite and quartz.

In experiments of long duration at tem-
peratures between 300° C and 400° C,
mixtures of wiistite and silica glass with
an Fe : Si ratio of 3 : 4 react to give a green
isotropic material which has a refractive
index near that of minnesotaite, but which
gives no X-ray diffraction pattern, even
with long exposures on the powder camera.
When this material was held for 6 weeks
at 350° C, 30,000 psi water pressure, and
a partial pressure of oxygen in equilibrium
with iron and magnetite, fayalite and silica
glass were obtained. Fayalite forms from
wiistite and silica glass mixed in greenalite
proportions (Fe:Si = 3:2) at temperatures
as low as 315° C, in spite of the high water
pressures. In experiments with other start-
ing materials, such as ferrosilicon alloys,
ethyl orthosilicate, and colloidal silica,
fayalite was the only identifiable iron sili-
cate formed. Natural fayalite held for 46
days at 250° C, 30,000 psi water pressure,
and a partial pressure of oxygen in equi-
librium with iron and magnetite showed
no sign of decomposition.

Under the conditions of these experi-
ments, it therefore appears either that
hydrous silicates of iron are unstable or
that fayalite forms metastably and there-
after fails to react with silica and/or water.
In nature, fayalite is absent from rocks
whose bulk compositions fall in or near
the system FeO-Fe203-Si02-H20, except
in the highest grades of metamorphism.
This makes it highly probable that in the
experiments described above, especially in
those below about 500° C, fayalite formed
metastably, and that equilibrium was not
attained. Investigation of this geologically
important system under controlled pres-
sures of oxygen therefore remains a chal-
lenging problem.

232 CARNEGIE INSTITUTION OF WASHINGTON

ISOGRAD PROBLEMS IN METAMORPHOSED
IRON-RICH SEDIMENTS

H. S. Yoder, Jr.

The iron-rich sediments are sensitive to
changes in pressure and temperature; their
mineral assemblages are, therefore, very
useful as indicators of metamorphic grade.
Ferruginous sediments, however, are not
common rocks, and their metamorphosed
equivalents are even less common on a
world-wide basis. For this reason the fer-
ruginous metamorphic rocks have not been
studied in the field as intensively as those
derived from the dominantly magnesian
and aluminous sediments. The principles
obtained from field investigations of the
magnesium- and aluminum-rich rocks,
however, and the physicochemical princi-
ples derived from laboratory studies of
analogous systems, are directly applicable
to the problems in the iron-rich meta-
morphic rocks.

The metamorphism of two critical types
of iron-rich sediments, those containing
principally greenalite and those consisting
dominantly of the chamosites, will be ex-
amined in the light of the established
principles. For the most part the rocks
containing greenalite can be represented in
the system FeO-Si02-H20, and those con-
taining chamosite, in FeO-Al203-Si02-
H20. The members of the two principal
mineral groups have a kaolinite-like struc-
ture, and it is probable that a complete
series of solid solutions exists between the
end members greenalite, Feti+2Si4Oio-
(OH)8, and Fe4+2Ai4Si2Oio(OH)8, as yet
not named. Substitutions of the type Mg— >
Fe+2, Fe+3->Al+3, and Fe+3-+Fe+2H+1 are
known, but their extent has not been de-
lineated. In addition, more complex sub-
stitutions trending from trioctahedral to-
ward dioctahedral character are possible.

The minerals in the FeO-Si02-H20 sys-
tem are plotted in figure 61, and the tie
lines are those believed to exist at room
temperature, for example. Two important
concepts may be obtained from a study of
this diagram. All the possible phases are

present at the lowest temperatures, yet only
those assemblages greenalite + quartz +
water (analogous to a sediment) or green-
alite + fayalite + water (analogous to a
partly serpentinized dunite) are commonly
observed. In the sediments, for example,
the phases are in equilibrium with water,
and, therefore, only those assemblages in
which water can occur as a phase are per-
missible. The remaining assemblages may
occur in environments where water does
not exist as a phase.

The second concept arises from a con-
sideration of the corner (FeO) . The com-
pound wiistite, FeO, which has not been
found occurring naturally, is not stable be-
low approximately 570° C, and its bulk
composition is represented by magnetite +
iron. In nature most rocks contain magnet-
ite; on the other hand, native iron is ex-
ceedingly rare. For the present purposes,
therefore, only those assemblages in equi-
librium with magnetite will be considered.
Oxygen, an important component in the
iron-bearing rocks, is itself worthy of a
detailed discussion. For the present prob-
lems, however, only a few brief remarks
are necessary.

In the tetrahedron in the upper left cor-
ner of figure 61 are plotted the possible
phases in the system Fe-Si-02-H2. The
evolution of the Fe-Si-O face of the tetra-
hedron may be followed from the sche-
matic diagrams in figure 62, which is based
on the work of Darken and Gurry (1946).
The total pressures in nature far exceed
those represented by the top curve in fig-
ure 62, which is less than approximately
10~5 atm. It is seen in figure 61 (inset)
that the bulk compositions of rocks con-
taining the silicates and magnetite must
lie in the tetrahedron magnetite-fayalite-
quartz-water. Here "water" represents the
homogeneous gas phase, the composition
of which, although close to H2O, may be
enriched in the component oxygen or the

GEOPHYSICAL LABORATORY 233

component hydrogen as well as iron and
silica. If the component oxygen is in ex-
cess of that of the saturated gas in equi-
librium with magnetite and silicates, all
the silicates and some or all of the magnet-
ite would be oxidized to an assemblage
of magnetite + hematite + quartz + gas or
hematite + quartz + gas.

Consider a finely banded rock of alter-
nating layers of magnetite, hematite, and

it is concluded that in these cases oxygen
does not diffuse freely from one layer to
another. In addition, James (1955) finds
as a result of his field studies in Michigan
that the oxygen content of a given layer
does not appear to change with meta-
morphism. A hematite + quartz layer per-
sists through all grades of metamorphism.
Since none of the silicates with the excep-
tion of quartz can coexist with hematite

H90

lFeO)/ v

Quartz

Fayalite

FerrosiliteC?)

MOL PER CENT

SiO,

Fig. 61. Projection of the system fayalite-quartz-water-magnetite onto the FeO-Si02-H20 plane
at approximately room temperature. The composition FeO is represented by iron + magnetite at
this temperature. The inset shows the location of the system in the tetrahedron Fe-Si-02-H2.

a mineral such as greenalite or stilpnome-
lane which may possess FeO/Fe203 in
various ratios. Only the hematite layer
could be in equilibrium with a gas phase
containing oxygen in excess of the gases
in equilibrium with magnetite and the
silicates; the others are prohibited from
occurring with such a gas. It is believed,
therefore, that each layer is in itself essen-
tially a closed system with regard to oxy-
gen. Under equilibrium conditions all the
minerals may exist at the same total pres-
sure but with varying oxygen contents of
the layers. Since such layering is common,

and a gas containing oxygen in excess of
that in equilibrium with magnetite or
the silicates, no iron-bearing silicates are
formed during metamorphism. In beds
initially consisting of magnetite + quartz,
the various iron silicates form. The gas
plays an important role because it acts as
a stabilizer even though the amount of
solids far outweighs the amount of gas
present.

Considering only those assemblages in
equilibrium with magnetite, the sequence
of metamorphic changes in beds consist-
ing dominantly of greenalite can be de-

234 CARNEGIE INSTITUTION OF WASHINGTON

duced. The basis for the deductions is the highest temperature is fayalite + quartz +

general principle that the thermal stability vapor. At all temperatures, all assemblages

of the hydrous minerals increases with de- are in equilibrium with magnetite. Each

creasing water content. The first triangle of these reactions will be marked by an

in figure 63 is the same as that given in isograd. If the initial bulk composition of

3
UJ

<r

TEMPERATURE

Fig. 62. Schematic representation of the assemblages in the Fe-Si-O system stable at various tem-
peratures and pressures. The curves are based on the work of Darken and Gurry (1946). The
stippled areas indicate those bulk compositions for which oxygen exists as a phase.

figure 61. With increasing temperature
greenalite reacts with quartz to yield min-
nesotaite + vapor. Next greenalite decom-
poses. Minnesotaite reacts with fayalite,
and grunerite now appears in the presence
of vapor. Minnesotaite breaks down at
higher temperatures, and finally grunerite
decomposes. The assemblage stable at the

the sediment were in the field greenalite
+ quartz + water + magnetite, then the in-
dex minerals to appear with progressive
metamorphism would be minnesotaite,
grunerite, and fayalite (or hypersthene) .
This sequence has been observed in north-
ern Michigan and elsewhere.
The minerals in the FeO-Al203-Si02-

GEOPHYSICAL LABORATORY 235

H2O system are projected onto the FeO- quartz is probably detrital. Magnetite

Al203-Si02 plane from H2O in figure 64. may be detrital or may be diagenic in ori-

All the phases that appeared in figure 61 gin, through the reduction of hematite by

are projected onto the side line FeO-Si02. organic means, for example. The 2M mica

The chamosites, which are critical to the arises from the IMd mica (illite or glau-

H?0

(FeO)

Fayalite

Grunerite

Quartz

(FeO)

(FeO)

(FeO)

(FeO)

(FeO)

Fig. 63. Assemblages stable at successive elevated temperatures in the system (FeO)-Fe304-
Si02-H20 projected onto the (FeO)-Si02-H20 plane. The phase "H20" or "V" represents the
homogeneous gas phase.

formation of many of these minerals, are
the low-temperature polymorphs of the
chlorites that lie along part of the join
extending from greenalite to pseudothu-
ringite. At the lowest grade of metamor-
phism, a schist containing chlorite + 2M-
muscovite + quartz + magnetite is usually
recognized first. It is of interest to examine
the source of even these minerals. The

conite) laid down in the sediment, prob-
ably passing through the stages IMd— »1M
-*2M. It is likely that when the poly-
morphs of the micas are examined more
closely an isograd may be established to
mark the 1M— »2M transition. The iron
chlorite, which has a 14-A structure, arises
from the polymorphic transition of the
7-A chamosites, marking the chlorite iso-

236 CARNEGIE INSTITUTION OF WASHINGTON

grad. It is believed that the initial schist
is produced in these ways. With increasing
temperatures, the minerals in the schist
react and form new minerals.

Most field workers agree that the first
new mineral to appear in the chamosite-
rich sediments is stilpnomelane. Although
the composition of this interesting mineral
is not known with certainty, its ferrous end
member, ferrostilpnomelane, probably lies
close to 5FeO-Al203-7Si02-2H20. Stilp-

progressive metamorphism of a ferrugi-
nous sediment. It may appear with a pure
muscovite as the result of the reaction of
glauconite with a chlorite as given in figure
65, or it may simply form as the result of
the reaction of a high-silica sericite with a
chlorite, as suggested by Lambert for the
magnesian biotites.

The iron-rich garnet almandite usually
appears at higher grades than biotite. Al-
though garnet is one of the most easily

S1O2 Quartz
Tridymite
Cristobalite

Pyrophyllife
Montmorillonite

Minnesotaite
Ferroonthophyllite
Grunerite,
Ferrosilite

Greenahte
Ferroantigonte/^'

Fayalite

FeO

Hercynite
Mole per cent

Fig. 64. Minerals in the FeO-Al203-Si02-H20 system projected onto the FeO-Al203-Si02 plane.

nomelane presumably arises from the re-
action of an iron chlorite solid solution
(along the join greenalite-daphnite) with
quartz.

Chloritoid is next to appear, but it cannot
form in rocks containing stilpnomelane.
The join chlorite-quartz, which appears to
persist to higher grades of metamorphism,
prohibits their association. For chloritoid
to appear there must be a source of AI2O3,
and diaspore, boehmite, and gibbsite are
the only possible sources. Sediments con-
taining large amounts of these minerals
are known. The reaction for the formation
of chloritoid might be, for example, chlo-
rite + diaspore + quartz.

A biotite usually appears next in the

recognized minerals, the precise nature of
the reaction leading to its formation is not
readily observed. It is believed by most
field investigators to form as a result of
the reaction chlorite (on the join daphnite-
pseudothuringite) + quartz. In the event
that the chlorite has been consumed in
previous reactions, the garnet may arise
with orthoclase from the reaction of musco-
vite + biotite + quartz. Orthoclase normally
appears at higher grades of metamorphism.
Another possibility is the reaction of biotite
+ plagioclase, yielding garnet and musco-
vite. Some observers have noted an in-
crease in the muscovite content with the
formation of garnet.
Staurolite (fig. 64) is next to appear,

GEOPHYSICAL LABORATORY 237

and, like chloritoid, its presence depends on
a high alumina content. There is some
doubt as to the exact composition of stauro-
lite. Juurinen's recent formula for stauro-
lite, H4Fe4Ali8Si8048, does not balance in
charge. Many analyses closely approach
FeO-2Al203-2Si02-H20, but the water
content is usually low and the alumina

AI203>Fe20j- (K20«Na20)

Koolinite
Pytophyllite

K20« No20

^>Greenah1e
Celoilomte Anmie Fe0-Mg0-Ti02

MOL PER CENT

Fig. 65. Projection of the reacting pairs glau-
conite and daphnite which produce biotite and
muscovite. The same products may result from
the reaction of a high-silica sericite with daphnite.

content appears to be too high. Possibly
(Fe,Al)+3 is replacing Fe+2H+1, as in the
chamosites. Staurolite may arise with gar-
net as the result of the reaction of chlori-
toid + quartz. Some workers have sug-

gested that, in those rocks in which chlori-
toid is absent, staurolite forms with quartz
at the expense of garnet + kyanite (the
reacting pairs may be visualized in fig. 64) .
The appearance of kyanite at the next
higher grade is in some cases dependent
on the reverse of this reaction.

Cordierite appears at the highest tem-
peratures in the hornfels of the contact
aureoles and is absent or rare in progressive
regional metamorphism. It is a breakdown
product of many of the hydrous iron sili-
cates.

The first appearance of an index mineral
as the result of a reaction has been taken
as the marker of an isograd. The examples
given above, unfortunately, are probably
not the only possible ways in which an
index mineral may arise. It is clear that the
specific reaction marking an isograd must
be stated. There is a great need, therefore,
for accurately identifying minerals on each
side of an isograd. It may be more im-
portant to establish the disappearance of
a mineral than its appearance, as was em-
phasized by Bowen (1940) in his discus-
sion of the progressive metamorphism of
a siliceous dolomite. Since most of the
iron minerals have now been synthesized
in the laboratory, quantitative data on
the principal reactions will doubtless be
forthcoming.

CRYSTALLOGRAPHY

G. Donnay

SYNTHETIC NEPHELINES

The substitution solid solution Nas-^Kar
AlsSisC^ is of particular interest because
of the two transitions that occur at two
definite compositions, namely at x=^Y^
and # = 2.00, first found by }. V. Smith
(Year Book 52, pp. 53-56). They were
characterized as high-order transitions in
a report to the Third International Meeting
on Reactivity of Solids (Geophysical Lab-
oratory Paper 1264), which is summarized
in the present Year Book. Such transitions
differ from the familiar high-order transi-
tions found in metallurgical systems in
that they are not associated with changes

in temperature but accompany composi-
tional changes. Only preliminary data had
been obtained to locate the transition com-
position in the neighborhood of x=l/4. In
view of the interest of such high-order
transitions in mineralogical systems, addi-
tional work was undertaken (jointly with
J. F. Schairer and J. D. H. Donnay) in an
effort to check and refine previous results.
Pure sodium nepheline glass was synthe-
sized and crystallized by being held, for
various lengths of time, at different tem-
peratures (table 24) . Cell dimensions were
determined by the method of least squares
from X-ray diffraction patterns, obtained

238 CARNEGIE INSTITUTION OF WASHINGTON

TABLE 24. X-Ray Data for Pure Sodium Nepheline

(Synthetic samples with different thermal histories)

NaAlSKL

Thermal History, ° C

a, A

c, A

V,A}

c/a

1100°

900°

900°

1000°

1000°

1050°

1050°

1200°

1200°

540°

6 hr
20 days
34 days

6 days
34 days

6 days
34 days

2 days
10 days

(sy
(s)
(s)
(s)
(s)
(s)
(s)

(S)

(s)

2000 bars, hydrothermal, 1 week (B)

9.971f

8.362f

720.0f

0.8386

9.984

8.333

719.3

0.8346

9.988

8.333

719.9

0.8343

9.986

8.331

719.5

0.8343

9.986

8.328

719.2

0.8340

9.984

8.333

719.3

0.8346

9.991

8.331

720.2

0.8339

9.984

8.328

719.1

0.8341

9.984

8.328

719.1

0.8341

9.989

8.328

719.6

0.8337

# (B) Boyd, (S) Schairer.

t Quenched high-temperature form.

Weight per cent Kp

Fig. 66. Change of cell dimensions (c, a in A; V in A3) in solid solution Nag^K^AlgSigOga-
The lower abscissa scale gives the number x of potassium atoms per cell. The upper scale gives
the weight percentage of Kp in the system Ne-Kp, where Ne stands for NaAlSi04 and Kp for
KAlSi04. Black circlets, quenched high-temperature form; white circlets, low-temperature form.
For comparison previous data (Geophysical Laboratory Paper 1267) are shown by double rings.

GEOPHYSICAL LABORATORY

239

with the Norelco instrument, usually for
two samples crystallized at the same tem-
perature but for different lengths of time.
The patterns yielded by the samples that
were held at 800° C for at least 27 days
show one faint nepheline line; they were
identified as low-carnegieite patterns. The
samples held at 900° C for 20 days did not
contain detectable carnegieite; those pre-

cause in one sample a high-temperature
form was obtained by quenching; its cell
dimensions are # = 9.971, c — 8.362 A, V =
720.0 A3. Within the limits of accuracy
claimed here for the method of cell-dimen-
sion determination, we therefore have evi-
dence for the existence of two nepheline
forms in the temperature range considered.
In contradistinction to albite, which

ex-

TABLE 25. X-Ray Data for Low-Potassium Nephelines
(Synthetic samples with different thermal histories)
Nafi JCAloSLO

l8-X1^'X

8W32

Wt.%

KAlSiQ4

Thermal History, ° C

a, A

c, A

V,A3

2.00 0.144

2.50 0.180

3.00 0.2165

3.75 0.270

4.00 0.289

5.00 0.361

6.00 0.434

1200°

8 days (

1200°:

5 days (

1200°

8 days (

1060°.

29 days (

1200°.

6 days (

1200°

, 72 hr (

1200°

7 days (

1050°

12 days (

1100°

10 days (

1200°

12 days (

1100°

10 days (

1200°

5 days (

1200°

8 days (

1100°

7 days (

1200°

5 days (

1200°

8 days (

s)
s)
s)

S)

s)

s)
s)

S)
S)
S)

s)

S)
S)
B)

s)
s)

9.971f

8.356

719.4

9.989

8.333

719.7 (a)t

9.980

8.333

719.2 (b)

9.973t

8.358

719.9

9.977

8.341

719.1

9.970

8.332

719.2

9.977

8.337

718.6

9.985

8.335

719.7

9.966

8.356

718.7

9.971f

8.358

719.6

9.975f

8.360

720.4

9.975f

8.358

720.2

9.969

8.354

719.0

9.963

8.354

718.3

9.971

8.355

719.4

9.968

8.353

718.8

'9.963

8.347

717.2 (a)

9.971

8.354

718.8 (b)

9.976

8.361

720.6

* (B) Bowen, (S) Schairer.

t Quenched high-temperature form.

t(a)1/47minute, (b) ^/minute.

pared at 1000° and 1050° C for 5 days or at
1200° C for 2 days showed only nepheline.
On all the runs but one the cell dimen-
sions (fig. 66) were found to be: a— 9.986
±0.005, <r= 8.331 ±0.004 A, V= 719.5 A3.
The numerical value of a ranges from
9.984 to 9.991; that of c, from 8.328 to 8.333.
A sample of pure sodium nepheline crystal-
lized (by F. R. Boyd) hydrothermally at
540° C and 2000 bars for 1 week gave cell
dimensions in good agreement with the
above values. The above samples prepared
at high temperature must have inverted to
a low-temperature form on cooling, be-

hibits a unique stable crystalline form for
each temperature (MacKenzie, Year Book
55, 1955-1956, p. 188), nepheline shows
only two forms.

New samples were also prepared by
Schairer with compositions in the low-
potassium region of the NaAlSiOi-KAl-
Si04 solid solution (table 25). Most of
the samples transformed to the low-tem-
perature form, for which the following
data were obtained by X rays. For 2.50
weight per cent KAlSiO^, that is # = 0.180,
a decreases slightly whereas c remains con-
stant. For 3.00 weight per cent, or x —

240 CARNEGIE INSTITUTION OF WASHINGTON

0.2165, large changes are suddenly observed
in the samples studied. From this point
on, to 27.07 weight per cent, that is x =
2.00, the cell dimensions are found to in-
crease linearly, in agreement with previous
results of Smith and Tuttle (Geophysical
Laboratory Paper 1267). The composi-
tional transition that takes place near x =
0.20 is thus marked by discontinuities in
the curves of a and c; the curve of cell
volume V , on the other hand, shows small
variation; it consists of two straight-line
portions, one with nearly zero slope, the
other with positive slope. The compen-
sating effect of the opposite variations in
a and c accounts for the smallness of the
offset between the two line segments of
the V curve in the region in which the
transition composition must lie. It is im-
possible to decide whether the V curve
shows a discontinuity, which would indi-
cate a first-order transition, or a singular-
ity, which would be the sign of a second-
order transition. Since we have no evidence
in favor of the two-phase region, near x=
0.20, that would be required by a first-order
transition, the hypothesis of a second-order
transition appears to be the more reason-
able one.

The problem of determining the order
of a transition on the basis of cell volume
is likely to lead to the kind of difficulty
encountered here, as it did in the study
of high-temperature alkali feldspars (Geo-
physical Laboratory Paper 1179), in view
of the uncertainty of the composition and
the limited accuracy (0.05 per cent) with
which cell dimensions can be determined
by the usual X-ray method.

Some of the synthetic samples of low-
potassium nephelines (table 25) gave the
high-temperature form on quenching,
namely: two samples out of three at i=
0.144 (2.00 weight per cent KAlSi04) and
all three samples at # = 0.289 (4.00 weight
per cent KAlSi04) . Their cell dimensions
are shown in figure 66y where it may also
be noted that the upper limit of potassium
content of the high-temperature form is
probably x = 03 or 0.4, where the cell di-

mensions of the high-temperature form
merge into those of low-temperature neph-
eline (beyond the discontinuity of x=
-0.20).

In view of the above results, it was de-
cided to take powder patterns at varying
temperature. The Gordon Davis heating
sample-holder, recently rebuilt by F. R.
Boyd, was used. A sufficiently large
amount of synthetic pure sodium nepheline
could be gathered from the samples avail-
able. The chief difficulty came from the
lack of standard material with interplanar
spacings calibrated at different tempera-
tures. Inasmuch as the purpose of the
experiment was to ascertain the existence
of a transition, already suggested by pre-
vious results, rather than to make absolute
measurements of spacings or to determine
the transition point with accuracy, un-
corrected 20 readings would be satisfactory.

The range of 20 extending from 26° to
30° C was covered twice, that is, with
increasing and decreasing 20, for each of
the following temperatures: 240°, 410°,
610°, 810°, 1050°, and 1150° C. The av-
erage value of 20_was plotted (fig. 67) for

two spacings: 2022 and 2130. The curve
of 20 (2022) shows a break in the neighbor-
hood of 810° C; it is difficult to confirm
the break by means of the curve of 20
(2130), which by itself can be regarded as
a straight line within the limits of error.
We note that the temperature scale is
only roughly calibrated, so that the values
given here may be in error by about ±50°
C, but this fact does not invalidate the
conclusion that a transition exists. Smith
and Tuttle (Geophysical Laboratory Paper
1267) place the transition in the neighbor-
hood of 900° C.

A SODIUM NEPHELINE IN NATURE

In the course of a study of nepheline
solid solutions (jointly with }. F. Schairer
and J. D. H. Donnay), we had the oppor-
tunity to gather from the literature a large
number of chemical analyses of nephelines,
together with corresponding cell dimen-

GEOPHYSICAL LABORATORY 241

sions determined by X rays. Among
them one sample stands out as abnormal,
in that its cell dimensions are quite dif-
ferent from those of all the others. It is
a sample of nepheline from Monte Somma
(British Museum No. 51495), described
by Bannister (1931). The chemical anal-
ysis can be recast in the following formula

Kl.l8Na4.46Cao.9l[Zll.45Al7.780l8.3o032

on the basis of the 32 oxygen atoms con-
tained in the cell. The cell dimensions,

-30 00c

-29.50c

...20(2022)

uncorrected

CD

(\J
o

- 27.50°

- 2700

20(2130)
*'cx-. uncorrected

"XX.

J 1__L

500

Temperature °C

1000

Fig. 67. Uncorrected 28 angles obtained with
CuKcc radiation for spacings 2022 and 2130 of
synthetic pure sodium nepheline at varying tem-
peratures (° C, ±50°). The transition between
low-temperature and high-temperature forms is

marked by a break in the curve of 20(2022).

determined by Bannister by the oscillating
crystal method, are: a = 9.96, c = 8.33 (A
from kX), which lead to a calculated den-
sity of 2.645 g/cm3. The reported observed
density, 2.576 g/cm3, is unusually low. No
other natural nepheline has so low a value
for c.

Dr. M. Hey (Mineral Department of
the British Museum) kindly sent us part
of the original sample, consisting of frag-
ments, among which could be found a

small euhedral crystal suitable for X-ray
work. Its cell dimensions were determined
on the precession camera: a— 9.988, c —
8.328 A, both ±0.3 per cent, c/a- 0.8338,
in agreement with Bannister's results.

Another portion of the sample was used
for powder work on the Norelco diffrac-
tometer; it gave: a" = 9.985, c" = 8.372 A,
both ±0.05 per cent, c/a = 0.8385. Even
though d = a" within the limits of error,
it can be concluded that the sample con-
tains two phases, for the difference in c
exceeds the uncertainty of the measure-
ments. The phase detected as a single
crystal must exist in the sample in very
small quantity, as its effect on the powder
pattern is not noticeable. Its cell dimen-
sions are those obtained for samples of
sodium nepheline synthesized by Schairer
(see above), with less than 0.20 potassium
atom per cell (*<0.20).

It is interesting to note that Cesaro
(1920) made a careful goniometric study
of small perfect crystals from Monte
Somma, in which he was able to measure
the angle (1010) : (1011) to 2 or 3 minutes
of arc. The lowest c/a ratio he records is
0.8358. This value is lower than the c/a
ratio 0.8381, which is the smallest value
found by X rays in the low-potassium re-
gion (0.20 <#< 2.00). Cesaro too seems
to have encountered a specimen of sodium
nepheline in his Monte Somma material.

In conclusion, the chemical formula
given above cannot represent any one of
the specimens used for single-crystal work,
although it may approximate the composi-
tion of the powder sample.

Bannister had already shown, by means
of optical measurements, that the chemical
composition of the material occasionally
changes from grain to grain in one hand
specimen. Such variability in composition
has, of course, long been recognized in
mineralogy; it must be reckoned with in
any attempt to determine chemical com-
position from cell dimensions. It can only
be hoped that careful sampling and grind-
ing of the sample will yield a representa-
tive powder, whose X-ray pattern will

242 CARNEGIE INSTITUTION OF WASHINGTON

show broadened peaks, each peak being
smeared over an angular range that cor-
responds to the range in composition.

SOLID SOLUTION

Under this heading last year's report
(Year Book 55, p. 205) gave an account
of the variation of cell dimensions of syn-
thetic nephelines with different kinds of
solid solutions. One of the conclusions
was that neither omission solid solution,
Nas-yUlyAls-ySis+yOsz, nor substitution-
omission solid solution, Nas^CazDzAls-
Si8032, changes the cell dimensions in
either a or c. Further work has shown that
this conclusion must be amended.

When the work was repeated on newly
prepared samples (see Synthetic Nephe-
lines, above), the a value had to be cor-
rected from 9.971 to 9.986 A, and the c
value from 8.362 to 8.331 A; the cell vol-
ume, however, changed only slightly, from
720.0 to 719.5 A3. The numerical values
given in last year's report are indeed the
cell dimensions of crystals of the high-
temperature form obtained by quenching.
The new data pertain to the low-tempera-
ture form, into which the samples inverted
on cooling. By a curious coincidence these
same cell dimensions are found on crystals
which represent the limit of solid solution
by substitution-omission, and they also ap-
proximate those of the low-potassium sub-
stitution solid solution near #=0.3.

PHOSPHATES

The question of stereoisomerism of tetra-
metaphosphate has been raised by Drs. R. J.
Gross and J. W. Gryder (Johns Hopkins
University). The existence of two stereo-
isomers of P4O12, namely a ring form and
a boat form, was postulated by them on
the basis of chemical evidence. The two
forms can be isolated only in the solid
state; by single-crystal work they were
proved to be distinct crystalline species.
Preliminary results were reported jointly
to the American Chemical Society.

Work on the crystallography of alkali
phosphates has been continued with Drs.

J. W. Gryder and Helen M. Ondik. A
compound reported in the literature as
Na2H2P40i2 was found to be surprisingly
insoluble in water for a metaphosphate,
and its fibrous habit was intriguing (Grif-
fith, 1956) . The compound was synthesized
by Griffith's procedure, and its identity
was checked by comparing its powder pat-
tern with that of material kindly furnished
by Dr. Griffith. The transparent, colorless
product is found to consist of two distinct
crystalline forms intimately intergrown.

Form I comprises approximately 10 per
cent by weight of the sample, as estimated
from the relative intensities of powder
lines. It consists of crystals of thick tabular
habit and average dimensions 0.3 X 0.05 X
0.02 mm. They are monoclinic, elongated
[010], with cell dimensions # = 30.7, b =
6.77, c=7.12 A, all ±0.5 per cent, (3 = 94° 6'
±l(f, V=W6 A3. A pronounced pseudo-
repeat a=a/2 is evident. Because it proved
impossible to separate a sufficient amount
of phase I for accurate density determina-
tion, we can report only that its density
is 2.62 g/cm3 or greater as determined by
the flotation method using a mixture of
bromoform and toluene as the inert liquid.
For the same reason a chemical analysis of
this phase was not possible. The space
group is uniquely determined as P2i/a by
the systematic absences hOl with h odd and
0^0 with \ odd. Only the forms {100}
and {001} are observed, the larger faces
being those of {100}. As cleavage is ex-
tremely fibrous along b, good single crys-
tals are difficult to obtain. No further
work on this form is contemplated.

Form II, the bulk of the material, con-
sists of thin to thick rectangular plates
which grow up to 5 mm in length. Only
cleavage fragments can be removed from
the matrix of the melt, and therefore
morphology gives no clue concerning the
point-group symmetry. Cell dimensions
were determined on precession films taken
with MoKa radiation: #=18.74, £ = 14.79,
<r = 7.03 A, all ±0.3 per cent, 3 = 90° 0'±5',
F = 1948 A3. Parallel to {100} the cleavage
is very easy and very good; parallel to

GEOPHYSICAL LABORATORY 243

{010} it is easy and fairly good; and paral-
lel to {110} it is fairly easy and good. The
four cleavage directions result in platy to
fibrous fragments.

The cell is monoclinic but markedly
pseudo-orthorhombic. Although the zero-
level net (010) # shows symmetry 111, paral-
lel upper levels show that a* and c* are
not symmetry directions. The pseudo-
orthorhombic character is emphasized by
the optical orientation — the principal axes
of the index ellipsoid lie along the crystallo-
graphic axes within experimental error.
The optical character is biaxial positive
with the plane of the optic axes parallel
to (100) and the acute bisectrix along
[001]. The indices are na= 1.485 ±0.005,
^=1.510 ±0.005, ny= 1.545 ±0.001 (deter-
mined by J. D. H. Donnay). Since the sys-
tematic absences are h\l with h + 1 odd and
hOl with h and / both odd, the space group
is Blja or Ba. The 5-centered lattice is
used to bring out the pronounced pseudo-
symmetry with pseudo space group Bmam,
Bma2, or Blam. The test for pyroelec-
tricity with liquid nitrogen gave negative
results.

Because form II is present in much larger
amounts than form I, its density could
be determined more accurately. The low-
est density obtained by the pycnometric
method using toluene as the inert liquid
is 2.34 ±0.02 g/cm3. The flotation method
using a mixture of bromoform and toluene
indicates that the density of fragments of
the crystal intergrowth varies from 2.34
to 2.62 g/cm3.

Using 2.34 g/cm3 as the density and
Griffith's formula, we calculate 7.54 ±0.08
formula units per cell. The space groups
Blja and Ba permit only an even number
of Na2H2P40i2 units per cell; consequently
the observed density must be compared to
the calculated densities of 1.86 for 6 and
2.48 for 8 units per cell. These values lie
well outside the limits of experimental er-
ror. If one molecule of water is subtracted
from the empirical formula, the result is
Na2P40n. For 8 such formula units per
cell the calculated density is 2.36 g/cm3,

within the limits of error of the experi-
mental value. Knowing that the phos-
phorus atom surrounds itself tetrahedrally
by oxygen atoms, we conclude from the
formula that some of the P04 tetrahedra
must share more than two corners. Fol-
lowing van Wazer's terminology (1955),
the material is therefore an ultraphosphate.
It is the first crystalline ultraphosphate on
record.

DISORDER IN CRYSTALS

Present-day interest in crystal structures
has shifted from the regularity to the im-
perfections of the interatomic arrangement.
Considerable attention has already been
devoted to the study of mistakes that occur
during crystal growth or form the very
mechanism of growth, such as screw dis-
locations. Order-disorder studies in recent
years have had a profound impact on
mineralogy, particularly on our knowledge
of layer minerals and of feldspars. A study
of disorder by means of optical diffraction
has been initiated in this laboratory
(Chayes). In view of the growing im-
portance of this field, new types of crystal
disorder are of special interest.

Such a new type has been encountered
in sodium ultraphosphate form II (with
Gryder and Ondik; see Phosphates,
above). The experimental evidence is as
follows. On the precession patterns con-
taining reflections hlfl, hf([f and ^3, all
reflections with / even are sharp, while
those with / odd appear as diffuse circles.
The c-axis rotation pattern consists of odd
layer lines which are diffuse streaks and of
even layer lines which contain sharp spots,
the width of the two types of layer lines
being the same. It follows that the reflec-
tions with / odd are diffuse circular disks
oriented normal to the <r# axis. This con-
clusion is confirmed by Weissenberg photo-
graphs. The intensity distribution within
the streaky layer lines varies from crystal
to crystal. The radius of the disk was
measured for one of the specimens and
was found to be 0.021 ±0.001 A"1.

These observations indicate that the crys-

244 CARNEGIE INSTITUTION OF WASHINGTON

tal structure is disordered by random dis-
placements of structural elements through
a distance of c/2 in the z direction. The
bonds parallel to the c direction must be
very much stronger, indeed of a different
order of magnitude, than the bonds paral-
lel to (001). Such requirements are met
by a chain structure with chain axes
parallel to c.

Additional evidence leads to the hy-
pothesis that rings of four phosphate tetra-
hedra are linked into chains by the sharing
of an oxygen atom between consecutive
rings. The cell height being equal to the
height of one such ring, the disordered
crystal consists of chains displaced by half
a link with respect to one another. An ex-
pression relating the observed diffuse in-
tensities to the probability of chain dis-
placement has been derived; the theoretical
treatment is similar to that given by Wil-
son for layer displacements.

POLYMORPHISM VERSUS ISOMERISM

When a chemical compound exists in
more than one crystalline modification, it
is usually a case of polymorphism, in
which each of the polymorphic forms is
stable in a definite region of the P-T dia-
gram. If the crystal structure of the com-
pound is of the molecular type, poly-
morphic forms differ in the way identical
molecules pack in the crystal, whereas
isomeric forms differ in the molecular
configuration itself and separate on crystal-
lizing from a solution in which they co-
exist in equilibrium. To discriminate
between polymorphism and isomerism,
organic chemistry usually must come to
the rescue of crystallography unless the
crystal structure is first completely deter-
mined. The working crystallographer,
however, may have occasion to suspect
the existence of isomerism in the crystals
he studies, and can draw the chemist's at-
tention to the problem. One such oppor-
tunity came to us this year.

Dr. F. W. Barnes, of the Johns Hopkins
Hospital, had given us hollow crystals of
5,5-diethylbarbituric acid (barbital, vero-

nal), grown at low temperature in his
laboratory. These crystals turned out to
be rhombohedral, thus different from the
monoclinic pseudo-orthorhombic barbital
described in the chemical and crystallo-
graphic compendia. Mr. William Seip then
referred us to a paper by Fischer and
Kofler, that appeared in the Archiv der
Pharmazie in 1932, in which three forms
of barbital, described as polymorphic
forms, had been studied optically with the
polarizing microscope. A crystallographic
re-examination of these forms was under-
taken (jointly with J. D. H. Donnay).

New data were obtained on the rhombo-
hedral form. It occurs in elongated crystals

composed of trigonal prisms: 1120, 2110,

3030, 0330, 4150, and 1540, in order of de-
creasing size. Several easy cleavages are
parallel to the c axis. Cleavage fragments
were used for X-ray work. Cell dimensions
are: tf=26.97±0.09, r=6.85±0.02 A; 18
molecules per cell give a calculated density
of 1.276 g/cm3, as compared with the
measured value of 1.26 (by M. Crute).
The space group determined by X rays is

R3 or R3; a pyroelectric test in liquid
nitrogen was negative, but R3 is the more
probable in view of the morphological de-
velopment. A powder pattern was taken
and indexed. The indices of refraction
determined by Fischer and Kofler were
confirmed to 0.001.

Crystal data on the monoclinic form of
barbital go back to Hertel (1930, 1935).
In the pseudo-orthorhombic description
(1930), Hertel had given an incorrect
space group, Cmcm instead of Ccmm, for
0=7.11, £ = 14.4, c=9.7 kX. The morpho-
logical development of the crystals was
found to be incompatible with the space
group on record; re-examination by X
rays gave the correct space group. The
monoclinic description (1935) with space
group C2/c or Cc and interchange of a
and b was confirmed by indexing a powder
pattern. The optical data of Fischer and
Kofler (1932) were confirmed; the indices
agree within ±0.003. These two authors

GEOPHYSICAL LABORATORY 245

observed the c axis to be polar; the prob-
able space group is accordingly Cc.

Preliminary data on the triclinic form
confirm the values of the indices of refrac-
tion measured by Fischer and Kofler on
twinned crystal plates. More work is
needed before the triclinic character can
be confirmed and cell dimensions reported.

The following evidence indicates that
the three forms of barbital are not poly-
morphs, but isomers. On melting either
the monoclinic or the rhombohedral form,
and letting it recrystallize from the melt,
we always retrieve the original form.
Whereas Fischer and Kofler report the
rhombohedral form as the form stable at
high temperature, the rhombohedral crys-
tals we studied were obtained at low tem-
perature and were accompanied by only
very small amounts of the other two forms.
Fischer and Kofler report that all three
forms usually occur together in their prepa-
rations; occasionally only two forms were
found by us.

DIGENITE

The work on digenite was continued.
In view of the impossibility of producing
true single crystals, it was decided to at-
tempt a determination of the crystal struc-
ture from the data obtained from twins. It
was necessary to analyze the data with a
view to finding which reflections were con-
tributed by the several crystals of the twin
to the cubically indexed diffraction spectra
on the X-ray photographs.

A tentative crystal structure has been
arrived at, on the basis of space group

R3m. All atoms lie along the threefold
axis of the rhombohedral cell described in
last year's report (Year Book 55, p. 204).
The sulfur atoms are placed at ;r=0,±%,
±%, and lead to an S-S distance of
3.92 A. The sulfur sites correspond to the
face centers of the small cube, which was
the cell reported by Rahlfs (1936). One
copper atom is located at x=Y2, and is
octahedrally co-ordinated to sulfur atoms.
The other copper atoms are found at
#=±~0.060 in tetrahedral co-ordination,

but displaced from the center toward one
face of the sulfur tetrahedron; at x —
±0.133, in triangular co-ordination; at
#=±0.250 and ±0.350, in regular tetra-
hedral co-ordination. The Cu-S distance
ranges from 2.26 A for the triangular co-
ordination to 2.77 A for the octahedral one.
This structure leads to satisfactory agree-
ment between calculated and observed in-
tensities for the superstructure reflections,
which are the reflections that appear on a
powder pattern. For the weak reflections,
observed only on "single-crystal" patterns,
the agreement still leaves much to be
desired.

Powder patterns were taken at high tem-
perature at the National Bureau of Stand-
ards, on the heating camera designed by
Mr. F. A. Mauer, who kindly agreed to
help us with this task. The sample was
heated in a helium atmosphere to prevent
oxidation. The temperature was slowly
raised to -500° C. Between 60° and 65°
the weak reflections disappear, indicating
the existence of a transition above which
the small cubic pseudocell becomes the
true cell.

The electrical conductivity of digenite
was tested by Dr. A. Franklin, at the Na-
tional Bureau of Standards. He reports
that digenite is an unusually good con-
ductor, even at room temperature; there
is no sharp change in conductivity at the
transition temperature.

CRYSTALLOCHEMICAL ANALYSIS

Identification of small amounts of crys-
talline material is a problem that is always
with us. A number of methods are well
established; for instance, determination
may be made by means of the polarizing
microscope, by the powder technique of
X-ray diffraction, by microchemical tests,
by spectroscopic analysis. When single
crystals are available, identification may be
based on the determination of the cell di-
mensions by X rays, a method that has
become possible only recently, after the
necessary determinative tables of crystal
data were published. It was thus a chal-

246 CARNEGIE INSTITUTION OF WASHINGTON

lenge to devise a practical determinative
procedure that could be applied by scien-
tists not specially trained in X-ray crystal-
lography. Such a procedure has been
worked out for each of the crystal systems,
and for all kinds of crystals, euhedral, sub-
hedral, or anhedral. It requires only a
precession camera and a two-circle goni-

ometer. It can be applied in a routine
fashion, with a minimum of interpretation,
although some judgment remains indis-
pensable.

The method has been written up as a
chapter for the coming third edition of
Physical Methods of Organic Chemistry,
edited by A. Weissberger.

APPLICATION OF THE "MOREY-SCHREINEMAKERS'
THEOREM OF COINCIDENCE"

G. W. Morey

A large part of the experimental work
of the Geophysical Laboratory has been
the determination of the melting points
and other phase-equilibrium relations of
the oxides and their mixtures and com-
pounds which make up the earth's crust.
This work has been guided by the theo-
retical precepts developed by the great
American physicist and mathematician,
J. Willard Gibbs. The greater part of our
studies have dealt with nonvolatile oxides,
in which the only experimental variables
needing consideration are the temperature
and the compositions of matter under con-
sideration. There is, however, increasing
interest in systems in which volatile sub-
stances such as water and carbon dioxide
are introduced as components. This in-
creasing interest is manifest not only in the
work of this Laboratory but also in other
laboratories interested in experimental geo-
chemistry, and to a rapidly increasing ex-
tent in the chemical industries, which see
the possibilities of radically new methods
of manufacture resulting from high-pres-
sure reactions. Such reactions require
simultaneous consideration not only of the
composition variables but also of pressure
and temperature, and involve complica-
tions resulting from critical phenomena
and significant solubility of solids in gases.
The theoretical consideration of such sys-
tems is little known and is not ordinarily
treated in textbooks.

The fundamental considerations devel-
oped by Gibbs are competent guides to the
most complicated of these systems, but
their detailed application leads to phase-

equilibrium relations difficult to under-
stand and expound because of their com-
plexity. One aspect of such relations,
namely the course of the pressure-tempera-
ture curves of univariant equilibria in
which such P-T curves follow each other
around an invariant point, can be devel-
oped and applied in a comparatively sim-
ple manner. The treatment of this problem
is based on a theorem deduced rigorously
by Morey and Williamson, and from gen-
eral considerations by Schreinemakers, and
called by Prigogine the "Morey-Schreine-
makers' Theorem of Coincidence." This
is a powerful tool, applicable to all types
of P-T curves, and simple to apply, but
since it is not included in textbooks it is
not generally known. By application of
this theorem it is possible not only to de-
termine the sequence of P-T curves around
an invariant point in a system of many
components, but also to fix the phase as-
semblages that can have stable coexistence
in the divariant regions between the P-T
curves, even though the points in such
regions are projections on the P-T plane of
an ^-dimensional hyperprism. A paper
recently published gives a detailed exposi-
tion of how this theorem may be applied
to the various types of invariant points in
a ternary system, and of the various types
of P-T curves proceeding from these in-
variant points, using as an example the
ternary system water-Na20 ^1203*25102
(nepheline)-Na20 ■ AI2O3 * 6S1O2 (albite),
in which the compounds Na20*Al203'
4Si02*2H20 (analcite) and Na20 ■ AI2O3 ■
4Si02 (jadeite) are formed.

GEOPHYSICAL LABORATORY 247

MISCELLANEOUS ADMINISTRATION

PENOLOGISTS' CLUB

The Petrologists' Club met at the Labo-
ratory on six occasions this year. After
rather rapid growth in the previous few
years, the membership has leveled off at
about 150 active participants.

The following papers were presented:

"Gravity and continental structure," by
H. E. Tatel (Department of Terrestrial Mag-
netism).

"Equilibrium texture in rocks," by J. B.
Thompson, Jr. (Harvard University).

"Origin of spilites," by G. D. Nicholls
(University of Manchester).

"The 1955 eruption of Kilauea," by G. A.
MacDonald (U. S. Geological Survey).

"Oxidation and reduction in metamor-
phism," by H. P. Eugster, J. R. Smith, and
W. G. Ernst (Geophysical Laboratory), and
H. James (U. S. Geological Survey).

"Origin of lamprophyres associated with
granitic plutons," by C. A. Hopson (Johns
Hopkins University).

SEMINARS

The Laboratory continued its weekly
series of seminars, in which papers con-
cerned primarily with work in progress
were presented largely by staff" members.
Several talks were given by guest speakers
from outside the Laboratory, including :

"The origin of lamprophyres," by C. A.
Hopson (Johns Hopkins University).

"An attempt to limit the possible composi-
tion of the ore-forming fluid," P. B. Barton,
Jr. (U. S. Geological Survey).

"Recent developments in X-ray fluores-
cence," I. Adler (U. S. Geological Survey).

"Phase relations in the system gold-silver-
tellurium," N. Markham (University of Ade-
laide, South Australia).

"Systems involving ferrous and ferric ox-
ide," A. Muan (Pennsylvania State Univer-
sity).

"Recent studies on silica," R. Roy (Penn-
sylvania State University).

SYMPOSIUM ON HIGH PRESSURES

On June 12, 1957, a one-day symposium
on high pressures was held at the Geo-
physical Laboratory. The morning session

was devoted to formal talks by Professor
H. Tracy Hall, Director of Research,
Brigham Young University, on "Chemistry
at high pressures and high temperatures,"
and by Professor Harry G. Drickamer,
Head, Chemical Engineering Department,
University of Illinois, on "The effect of
high pressures on optical properties of ma-
terials." After a buffet luncheon, the first
session of the afternoon was devoted to
general discussion of phase changes in-
duced by pressure and their relation to
geophysical problems, with Professor Birch,
of Harvard, contributing notably. The
final portion of the day was devoted to
apparatus design, with discussion led by
Professor Hall.

The group attending the symposium in-
cluded: L. H. Adams (National Bureau
of Standards), L. T. Aldrich (Department
of Terrestrial Magnetism), Charles W.
Beckett (National Bureau of Standards),
Francis Birch (Harvard University), E. H.
Carnevale (Naval Ordnance Laboratory),
Harry G. Drickamer (University of Illi-
nois), Abraham Friedman (Atomic En-
ergy Commission), Irving Friedman (U. S.
Geological Survey), H. R. Gault (Lehigh
University), John W. Graham (Depart-
ment of Terrestrial Magnetism), H. J.
Hadow (United Kingdom Scientific Mis-
sion), H. Tracy Hall (Brigham Young
University), Joseph Hilsenrath (National
Bureau of Standards), Francis T. McClure
(Applied Physics Laboratory), Donald
Newhall (Harwood Engineering Com-
pany), Thomas B. Nolan (U. S. Geologi-
cal Survey), Donna Price (Naval Ord-
nance Laboratory), Sidney G. Reed, Jr.
(Office of Naval Research), A. E. Ring-
wood (Harvard University), Eugene Rob-
ertson (U. S. Geological Survey), William
W. Rubey (U. S. Geological Survey),
Paul A. Scherer (Carnegie Institution of
Washington), H. E. Tatel (Department
of Terrestrial Magnetism), Dudley Taylor
(Naval Ordnance Laboratory), M. A.
Tuve (Department of Terrestrial Magnet-
ism), Alvin van Valkenburg, Jr. (Na-

248 CARNEGIE INSTITUTION OF WASHINGTON

tional Bureau of Standards), Charles E. posium to the Washington Crystal Col-
Weir (National Bureau of Standards), loquium.

George W. Wetherill (Department of Ter- H. P. Eugster served as Lecturer in the

restrial Magnetism), and Samuel Zerfoss Department of Geology, Johns Hopkins

(National Bureau of Standards). University, each Friday during the aca-

In addition to the above, the following demic year 1956-1957. He also lectured at

staff members and guest investigators of the fall meeting of the National Academy

the Geophysical Laboratory were in at- of Sciences.

tendance: P. H. Abelson, R. G. Arnold, J. W. Greig served as Visiting Research
H. L. Barnes, P. B. Barton, Jr. (U. S. Geo- Associate at the College of Mineral Indus-
logical Survey), F. R. Boyd, Jr., S. P. tries, Pennsylvania State University, from
Clark, Jr., G. L. Davis, J. L. England, November 1956 through March 1957. Dur-
W. G. Ernst, H. P. Eugster, J. W. Greig, ing this period he gave a series of lectures
G. Kullerud, E. H. Roseboom, J. F. on various aspects of phase equilibria in
Schairer, J. R. Smith, D. B. Stewart (U. S. ternary systems, which were attended by
Geological Survey), D. R. Wones, H. S. staff members and graduate students of
Yoder, Jr., and E. G. Zies. five departments: Mineralogy and Petrol-
ogy, Geophysics and Geochemistry, Geol-
lectures 0gy5 Metallurgy, and Ceramic Technology.

During the report year staff members G. W. Morey gave a series of three lec-

were invited to present lectures as follows : tures before the Inorganic Chemical Di-

P. H. Abelson lectured at the National vision of Monsanto Chemical Company at

Academy of Sciences; the Annual Meeting Dayton, Ohio.

of the Trustees of the Carnegie Institution J- F- Schairer delivered the Presidential
of Washington; the New York Academy Address, "The crystallization of rock-form-
of Sciences-A.A.A.S. Symposium on Mod- in£ minerals from magmas and the nature
ern Ideas on Spontaneous Generation; a o£ the residual liquid," at the annual meet-
graduate seminar at Catholic University; mS o£ the Geological Society of Wash-

the Pittsburgh Section of the American m§;05; ™ , , , ^ ,

nu^^;^\ c^;~«- ,. „ n n • • t? .u (j. K. 1 llton lectured at the Gordon Re-

Cnemical society a Colloquium in Earth , ~ r ^ TT .

c • . .i -xr i \^ T . c search Conference; Georgetown Univer-

bciences at the Massachusetts Institute of ^ r xt i -n

rr, , i , XT. iT.-1-i sity; a Conference on Nuclear Processes in

lechnology; the National Biophysical ^ J \ . c ... _ ^ . A ^ ,•

r> r i tt o ttt i ^ Geologic Settings at Boston ; and a 1 ektite

Conference; the U. S. Weather Bureau: n r ° ° ju.-ut^--- c

, T . ' . . . . _^ ' Conrerence sponsored by the Division or

the Journal Club of the Department of Earth Sciences> National Research CoundL

Geology, Johns Hopkins University; a H. S. Yoder, Jr., lectured at the Depart-

Symposium on Isotope Separation spon- ment o£ Geol Columbia University; a

sored by the Netherlands Physical Society Basak Conference sponsored by the Di-

and the International Union of Pure and vision o£ Earth Sciences, National Re-

Apphed Physics at Amsterdam; and the search Council; and the Institute on Lake

National Bureau of Standards. Superior Geology, Michigan State Univer-

F. R. Boyd, Jr., lectured at the National sity. He also gave a series of two lectures
Academy of Sciences; the Washington each, at the Department of Geology, Uni-
Junior Academy of Sciences; and the versity of Wisconsin; the Department of
Physics Department, Howard University. Geology, University of Chicago; and the

G. Donnay addressed the Point-Group Department of Geology, University of
Seminar, Physics Department, Polytechnic Illinois.

Institute of Brooklyn. Together with G.

Kullerud and J. D. H. Donnay (of Johns The "Summary of Published Work" be-

Hopkins University) she presented a sym- low briefly describes the papers published

GEOPHYSICAL LABORATORY 249

in scientific journals during the report year.
In addition, the following papers are now
prepared for publication : H. L. Barnes and
G. Kullerud, "Relations between composi-
tion of ore minerals and ore solutions";
F. R. Boyd, Jr., "Geology of the Yellow-
stone rhyolite plateau"; F. Chayes and
W. S. MacKenzie, "Experimental error in
determining certain peak locations and dis-
tances between peaks in X-ray (powder)
dinractometer patterns"; G. Donnay, J. D.
H. Donnay, and G. Kullerud, "Crystal
and twin structure of digenite, Cu9S5";
J. W. Gryder, H. Ondik, and G. Donnay,
"Disorder in a crystalline condensed phos-
phate"; W. S. MacKenzie, "The crystal-

line modifications of NaAlSisOs"; G. W.
Morey, "The system water-nepheline-al-
bite: A theoretical discussion"; G. W.
Morey, "The transition between the low-
and the high-temperature form of sodium
tripolyphosphate"; J. F. Schairer, "Melting
relations of the common rock-forming ox-
ides"; G. R. Tilton, G. L. Davis, G. W.
Wetherill, and L. T. Aldrich, "Isotopic
ages of zircon from granites and pegma-
tites"; O.F. Tuttle and N.L. Bowen,"The
origin of granites in the light of experi-
mental studies in the system NaAlSi308-
KAlSisOs-SiOs-H^O"; H. S. Yoder and
Th. G. Sahama, "Olivine X-ray determina-
tive curve."

SUMMARY OF PUBLISHED WORK

(1255)

Variations in X-ray powder diffraction
patterns of plagioclase feldspars. J. R.
Smith and H. S. Yoder, Jr. Am. Min-
eralogist, 41, 632-647 (1956).

The angular separation between the (131)

(1256) Paleobiochemistry. P. H. Abelson.
American, 195, 83-92 (1956).

Sci.

A variety of types of organic substances
have been preserved in their original form or
in only slightly altered state for many mil-
and (131) reflection in X-ray difTractometer lions of years. Some of these occurrences are
patterns of 66 chemically analyzed natural described, and their significance to potential
plagioclases, 1 1 plagioclases synthesized in the knowledge of past living forms is pointed out.
dry way, and 4 plagioclases synthesized hy-

drothermally has been measured and plotted
against composition. By this criterion, plagio-
clases synthesized in the dry way and natural
plagioclases from thick stratiform mafic intru-

(1257)

Petrographic modal analysis. F. Chayes.
New York, John Wiley & Sons, Inc.
x+113pp. 1956.

A manual, intended for graduate students

sions constitute two distinctly different series, and advanced undergraduates, outlining the

each of which is closely defined by a single geometrical basis of the procedure, describing

curve. Natural plagioclases from volcanic experimental studies of various kinds of ana-

and hypabyssal rocks and plagioclases syn- lytical and sampling errors, and discussing a

thesized hydrothermally are intermediate be- method of estimating and controlling the ef-

tween the two series. Other natural plagio- feet of grain size on experimental error in

clases, some of which have been assumed by studies of two-feldspar granites,
other workers to belong to a "low-tempera-

ture" series, do not belong to either of the
series mentioned above, and cannot be repre-
sented by any single curve. It is concluded
that composition determinations cannot be
made by means of the available curves based
on the variation of reflection separations, be-

(1258) The Holmes effect and the lower limit
of modal analysis. F. Chayes. Mineral.
Mag., 31, 276-281 (1956).

Thin-section analysis is essentially an areal
measurement, the measurement area usually
being the upper surface of the section. If
cause there is no a priori way of knowing transmitted light is used for the measurement,
how closely a given plagioclase is represented the apparent areas of opaque grains are in
by a particular curve. Given the composition general somewhat larger than their true areas
of a plagioclase, however, the curves are use- on the measurement surface. For strictly
ful for making an estimate of its degree of spherical opaque particles in a transparent
inversion toward some undefined low-tem- matrix the expected excess of apparent over
perature state. true area is shown to be (jtr2^)/(2r + ^),

250 CARNEGIE INSTITUTION OF WASHINGTON

where r is the spherical radius and \ is the
thickness of the thin section. A table shows
the relation between true and apparent area
as a function of r/\.

(1259) Pressure-temperature curves in some
systems containing water and a salt.
G. W. Morey and W. T. Chen. /. Am.
Chem. Soc, 78, 4249-4252 (1956).

A novel method for the determination of
vapor pressures of saturated solutions has
been applied to binary systems containing
water and the salts LiF, NaF, KF, NaCl, KC1,
RbCl, CsCl, PbCl2, Li2S04, Na2S04, K2S04,
T12S04, Na2C03, K2C03, Na204B203,
Na4P207, and K4P207 at some or all of the
temperatures 374°, 400°, 500°, 600°, and
700° C.

(1260) Experimental and theoretical studies of
the mica polymorphs. J. V. Smith and
H. S. Yoder, Jr. Mineral. Mag., 31, 209-
235 (1956).

An experimental and theoretical study has
been made in order to determine the number
and the structure of the possible polymorphs,
and the structural relations between them.
The simplest structures are 1M, 2Ml5 2M2,
3T, 20, and 6H polymorphs, and more com-
plicated types can be developed. Some of the
previously described polymorphs were not
contained in the theoretical list and were re-
examined. The 6M structure was found to be
a 2M2 polymorph, the 6-layer triclinic type
was found to be a 2 Mi polymorph, and the
3M structure was shown to be a 3T type.
The 24-layer triclinic structure could be de-
scribed on a simpler 8-layer cell. This type
together with a new 12-layer monoclinic
structure, as well as other structures of higher
periodicity, presumably consists of complex
stacking and results from spiral-growth
mechanism. Two extreme types of layer-
disordered crystals may be built, and a dis-
order of individual ions may also occur. Sin-
gle stacking faults result in twinned crystals.
A new twin relation (180° rotation about
the [100] axis) has been recognized. Twenty
specimens from extreme geological environ-
ments have been examined to evaluate the
control of environment on the stacking. The
type of stacking could not be attributed solely
to the influence of pressure and temperature.
Composition seems to play a dominant role

in the type of stacking, and semiquantitative
structural arguments appear to support this
contention. The influence of growth mecha-
nism is discussed. A scheme for the identifi-
cation of the mica polymorphs by X-ray
powder and single-crystal methods is given.

(1261) The chemical formula of empressite.
G. Donnay, F. C. Kracek, and W. R.
Rowland, Jr. Am. Mineralogist, 41,
722-723 (1956).

Ag5Te3 is synthetic empressite. Ag5_a.Te3
is the formula deduced, for the mineral, from
cell dimensions and density of analyzed
crystals.

(1262) A provisional reclassification of granite.

F. Chayes. Geol. Mag., 94, 58-68
(1957).

The term "granite" could be usefully re-
served for massive or weakly oriented plu-
tonic rocks of color index less than 20 per cent
and quartz content between 20 and 40 per
cent by volume. A symbolic classification of
rocks meeting these requirements is pre-
sented. The various classes are based on the
relative proportions of plagioclase and alkali
feldspar, and subclasses based on the ternary
dominance ratio quartz-alkali feldspar-plagio-
clase are also proposed. The symbols are
simple and easily remembered. A few of the
common names now in use are retained for
class designations.

(1263) Organic constituents of fossils. P. H.
Abelson. Geol. Soc. Am. Mem. 67, pp.
87-92 (1957).

This chapter describes the occurrence of
amino acids in recent shells and in a variety
of fossils as old as 360 million years. These
findings are correlated with laboratory tests of
the thermal stability of alanine.

(1264) High-order transitions in (Na,K) AlSi04.

G. Donnay. Third International Meet-
ing on Reactivity of Solids, Madrid,
April 1956. 1957.

According to Ehrenfest's definition, a high-
order transition corresponds to a discontinuity
in any one of the derivatives of the Gibbs free
energy G. The order of the transition is set
equal to the order of the lowest derivative of
G that shows a discontinuity. High-order

GEOPHYSICAL LABORATORY 251

transitions are usually detected by observing
an anomalous change in specific heat with
temperature for a given composition. Nephe-
line, Na8_a;Ka;Al8Si8032, is an example of a
low-temperature phase in which changes of
the composition a: in a solid-solution range
result in high-order transitions at x=^lA
and x = 2.00. A singularity in the curve of
cell volume V against x corresponds to a dis-
continuity in (dV/dx)^ which is equal to
(d2G/dPdx)T. Such a transition is therefore
of the second order.

(1265) Annual report of the Director for 1955-
1956.

(1266) Optical analyzer. G. Donnay and J. D.
H. Donnay. Rev. Sci. Instr., 28, 145
(1957).

Two Polaroids, coupled for synchronous
rotation, are mounted between glass plates
and placed one at the top, the other at the
bottom of a brass cylinder, which can rotate
in a brass sleeve. The crystal on the goniom-
eter head can be introduced into the cylin-
der through a hole in the sleeve and a cor-
responding one in the cylinder. The crystal
can thus be observed between crossed Pola-
roids while it is rotated about the axis of the
goniometer head.

(1267) The nepheline-kalsilite system: I. X-
ray data for the crystalline phases. J. V.

Smith and O. F. Tuttle. Am. J. Sci.,
255, 282-305 (1957).

X-ray data are given for the following
phases: high- and low- carnegieite, high-
and low-nepheline, kalsilite, orthorhombic
KAlSi04, synthetic kaliophilite, natural kali-
ophilite, anomalous natural kaliophilite, tetra-
kalsilite, and 02. Comparison of the cell
dimensions indicates that the structures of all
the phases except carnegieite are based on a
tridymite-type framework. Synthetic natural
and anomalous natural kaliophilite are not
identical, but comparison of their X-ray prop-
erties indicates that