Skip to main content

Full text of "The value of science in the smithy and forge"

See other formats

This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 
to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 
are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 
publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 

We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liability can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at |http : //books . google . com/ 


- r 





Fourth Edition, Revised. Pp. i-xv+263. With 71 Illustrations. 


Including the Author's Practical Laboratory Couhse. 
'* Just the kind of work for Students gommenoinq the study of Metallui^, or for 
Bngineerinq StudentB."— Practical Engineer. 

In Crown 8vo. Pp. i-vii-l-lOS. Handsome Cloth. Illustrated. 


"It is an excellent and handy book for its purpose, and will have a far wider 
range of usefulness than for class work alone. "—Practical Engineer. 

Second Bdition. Thoroughly Revised. Pp. i-xvii-|-313. With 196 
Photo-Micrographs, Diagrams, and Figures. 


Revised and Corrected by L. P. SYDNEY. 
"Of all the books which have dealt with this subject in its many aspects, surely 
this one remains awpreme"— Chemical World. 

Second Edition. Extra Crown 8vo. Pp. i-xii+139. With Illustrations. 


Professor of Metallurgy in the University of Birmingham. 
" Ironfounders will find much information in the book."— /ron Trade Circular 


In Crown 8vo. Pp. i-viii+46. Handsome Cloth. With Illustrations. 


Author of " Quantitative Metallurgical Tables," etc. 
*' Excellent ... It is essentially suitable for the practical man, and to the 
beginner particularly it should prove valuable."— /^ron and Coal Trades Record. 

In Crown 8vo. Pp. i-ix-|-108. Handsome Cloth. With Illustrations. 


*' May be confidently recommended." — Ironmonger. 

Third Edition, Revised. In Cloth. Pp i-xxii+25-2. With 234 Illustrations. 


By R gran JON and P. ROSEMBERG. 
Translated by D. Richardson, Wh.Ex., A.M.LM.E. 
•* A simple and clear textbook."— 7'iwi«ir. 
"We can unhesitatingly recommend this book ... it is reuUy helpful."— ITorJb. 

Other Scientific Books for Practical Men are listed at the end of this Voiume. 










DSC, d.met., f.r,s., f.i.c, f.o.s. 



D.Sc, LL.D., M.Inst. C.E. 

TMlitb 75 5llu0tratlon0, mogtlfi pbotomicroflrapba. 





[All Eights Reserved,'] 




\ A^ 


It is not usual for anyone who is constantly engaged 
in the workshop to attempt to write and lecture on the 
practical application of modern science. Mr Cattcart, 
however, who has been thoroughly trained in practical 
smith-work in the blacksmith's shop, has not only 
attempted but has succeeded in writing on the subject, 
showing clearly how much benefit blacksmiths would 
derive if they were to apply more science in the conduct 
of their everyday work. It is evident that the author 
has by patient study mastered the elements of metal- 
lography and the effect of heat on the structure of iron 
and steel, for he has in most lucid language sought to 
show how such knowledge oan be applied. Knowing 
that there is still much prejudice in the mind of the 
practical worker against theory, the author has taken 
some pains to show that the practical worker himself 
bases his practice on theory, and that "theory and 
practice are inseparable." " The practical man is always 
of necessity a theoretical man, whether he admits it or 
not." What is clearly shown is the necessity for black- 
smiths having more theory in order that their practice 
may be the more perfect. Most of the very excellent 
photomicrographs illustrating the lecture are Mr 
Cathcart's own work. As a result of his private 
research on welding iron, he has revealed the interest- 
ing fact that, when heating to welding temperature in 




a smith's forge, the iron absorbs carbon on the surface, 
and that the juxtaposed faces of finished welds, in such 
cases, may contain between 0*2 per cent, and 0*8 per 
cent, carbon, and are actually steel. As a consequence,, 
the welded portions show greater tensile strength than 
the iron on each side of the weld. 

If the little encouragement and assistance I have 
given the author has helped him in his study of metal- 
lography, and led to the better understanding of iron 
and steel, I am deeply gratified. I feel sure that Mr 
Cathcart's book will do much to lead others to see the 
value of modern science in the blacksmith's shop. 




I HAVE read Mr Cathcart's manuscript with very great 
interest. With his "thesis" — the value of science in 
the workshop — I need hardly say I am in complete 
agreement, and I would further say that he has upheld 
it admirably. 

His treatment of the subject in the early pages is 
excellent. Much of the distrust of technical education 
— or training in science, as I should prefer to call it — 
is due to writers who ought to know better, but who, 
lacking any sound knowledge of science themselves, and 
priding themselves on being "practical men," have 
endeavoured to make people believe that there are two 
classes of men connected with any craft or profession : 
those who are practical men and those who know some- 
thing of the science pertaining to the craft; — as if a 
man could not be a "practical man" if he had taken 
the trouble to learn the science that underlies the 
processes with which he deals. 

The latter part of the work has naturally been of 
special interest to me, as it contains many matters which 
I have not hitherto studied in such detail, and these 
are undoubtedly of great importance to men of his 
craft, as well as to those practising other branches of 

His "conclusion" is very excellently put and very 


I look forward to the publication of Mr Cathcart's 
work. It will be of very great interest and value to 
many outside the circle of his own great craft. One 
cannot but admire the perseverance with which he has 
laboured to achieve the results he records. His own 
photomicrographs are splendid and intensely interesting, 
and anyone who masters what he offers for study will 
know a great deal of which most smiths and engineers 
have very little conception — much less true under- 


Anniesland, Glasgow, 

24^A August 1916. 


The following pages are the outcome of a lecture 
delivered at a meeting of the Associated Foremen Smiths 
of Scotland for the purpose of interesting the members 
in the scientific training of the rising generation of 
smiths and forgers. That the object was favourably- 
received, not only by those to whom but by those in 
whose interests it was first presented, and also by dis- 
tinguished educationalists, is fully borne out by the fact 
that it has been repeated by request on several occasions ; 
not the least gratifying of which repetitions has been 
one specially requested by a deputation of smiths, who 
organised a meeting which was attended by an audience 
of about three hundred, the chair being occupied by Dr 
Henry Dyer, Chairman of the Glasgow School Board. 

On each of these occasions different practical examples 
were introduced. While these — for obvious reasons — 
could not be dealt with during a single lecture, it has 
been thought desirable to insert them for the present 
purpose. At the same time the original intention of 
dealing more or less briefly with each subject, as far as 
possible, has not been departed from. In the meantime, 
I am not so much concerned about imparting informa- 
tion to young smiths as I am about drawing their 
attention to the many different sources from which 
suitable knowledge may be obtained, and showing how 
desirable and essential it is that it should be obtained in 
order to be able to deal more effectively with the every- 
day problems of practical work. 


From several different sources we have of late heard 
of scientific training being called in question, but this is 
by no means an argument against the value of such 
training. The outstanding cause of the trouble is due 
to a lack of the right kind of education and a proper 
method of imparting it. 

It is a most surprising fact that, while almost every 
trade that one can think of is provided with a suitable 
class, there are few if any in existence for smiths and 
forgers. When one does hear of such a class, it usually 
forms part of a course for mechanical engineers. It 
would appear from the nature of a number of questions 
in a recent examination paper that these classes were 
for the benefit of engineers only. If they were intended 
for smiths and forgers, it only emphasises the fact that 
those authorities who are responsible for instituting 
trade classes evidently consider the art of smithing and 
forging as being an inferior branch of industry and of 
very little importance. 

Whatever may be the cause of this lack of interest in 
the education of smiths and forgers, it is deeply to be 
regretted that it should be so, for in this direction there 
is a field in which a vast amount of good could be done 
to the craft in question and which would reflect ad- 
vantageously upon the whole engineering industry. 

While fulfilling its primary purpose, it is hoped that 
the subject-matter contained herein may be at the same 
time a basis on which to arrange suitable classes, and also 
demonstrate the importance of and necessity for doing so. 

It will be observed that I have not dealt with methods 
of doing work. The best and only school for teaching 
smiths how to handle their tools and perform the many 
varied operations which pertain to their craft is the 
smithy. Such training does not come within the 


province of the science class, the true function of which 
is to teach the laws and principles which govern the 
practical work of the particular craft. 

I cannot conclude without expressing how deeply 
indebted I am to Dr J. E. Stead, whose kindliness of 
heart in most generously assisting in many ways has 
contributed in a large measure to the success of the 
lectures. For many years I have been studying his most 
valuable research^ in connection with iron and steel. 
From this source, and latterly from private correspond- 
ence, I have been greatly benefited by him, and have also 
had the very great privilege of receiving a few personal 
lessons in practical metallography, without all of which I 
could not have acquired the knowledge which has enabled 
me to make this attempt to help my fellow-craftsmen. 

In connection with the publication of this work I 
have still further to thank Dr Stead. Not only has he 
permitted me to reproduce in these pages a number of 
his highly instructive photomicrographs, but he has 
very kindly edited the manuscript, and by his generous 
advice and suggestion has given me much encouragement 
and inspiration in my studies connected therewith. 1 
am sure his kindness will be all the more appreciated 
when I state that he is one of the greatest authorities 
on the subject, and that by practical men and scientists 
alike he is acknowledged all over the world as being in 
a large measure responsible for having demonstrated by 
his splendid research the practical value of the science 
of metallography. 

I am also deeply indebted to Professor Archibald Barr 
for having very kindly reviewed the manuscript, and for 

1 Journals and Proceedings of the various Institutions and 
Societies, and also the Microscopic Analysis of Metals, published by 
Charles QriflBn & Co., Ltd., London. 


having given me many valuable practical suggestions of 
which I have very gladly availed myself. Not only so, 
but he has most generously communicated to me much 
information which will be of very great practical value 
to me in my further studies. 

I have also to express my thanks to Dr W. Rosenhain 
for having very kindly given me the use of several of 
his most interesting lantern slides, and for permission to 
reproduce them here ; to Sir J. A. Ewing for permission 
to reproduce four of his very instructive photomicro- 
graphs ; to Dr F. Giolitti, of Italy, for his kindness in 
assisting me in my studies on case-hardening; to 
Professor A. Campion, of the Royal Technical College, 
Glasgow, from whose classes I have derived much benefit, 
and to whom I heartily commend all who are interested 
in metallurgy ; also to many other friends who have in 
various ways contributed to the success of the lectures. 

By kind permission of the Councils of the Royal 
Society and the Institution of Mechanical Engineers, to 
whom I am much indebted, the following photomicro- 
graphs have been reproduced: figs. 27, 28, 29, and 30 
from the Philosophical Transactions and Proceedings of 
the former, and figs. 24, 25, and 26 from the Proceedings 
of the latter. 

My indebtedness is also due to the North British 
Locomotive Company, Limited, whose General Manager, 
Mr Thomas M'Gregor, has always extended to my re- 
search work his most generous permission and approval, 
without which the practical demonstration of these 
studies could not have been brought to a successful issue. 


CresswelIi, Bishopbriggs, 
December \^\b. 



Preface by Dk J. E. Stead v 

Prefatory Note by Prof. A. Barr vii 

Author's Preface ix 

Introduction. . 1-7 


The Length of Material required to make a Forging— Loss of 
Material While Heating — Alteration of Forging — Length 
Required to make a Ring — Amount of Contraction on Out- 
side and Expansion on Inside of a Ring — Advantage of 
being able to Calculate 8-1^ 


DimensioDS of Eyebolt — Dimensions of Bar for Sace — Safe Load 

for Race 20-27 




DRAWING. . . . 3i_34 




Blocking Hoops — Shrinking Fit for Hoops — Expansion and 

Contraction 36-41 


The Microscopic Examination of Iron and Steel — Fibre in Iron 
and the Crystalline Structure of Iron and Steel — Slip Bands 
and Alternating Stresses . 42-62 


Overheating — Annealing — Different Critical Points — Practical 
Example to show the Effect of Incomplete Refining-^Rect- 
angular or Stead's Brittleness — Effect of Work at the Blue 
Temperature— Distortion of Steel when Cold Worked . 63-106 


Oxidation and the Burning of Steel — Fluxes — Restoration of- 

Burnt Steel— Metallurgy— Strength of Welds . . 107-139 

CASE-HARDENING. . . 140-153 

CONCLUSION. . . . 154-166 

Appendix : Bibliooraphy 157-159 

Index 160-168 





Before attempting to prove that science is of great 
value in the smithy and forge, I think it necessary to 
explain the manner in which I propose to do so and the 
reasons for so doing. 

From time to time I have been approached by young 
smiths who, thinking that I might be able to advise 
them, have asked me many intelligent questions regard- 
ing studies suitable for smiths and forgers. 

This is a matter to which I have given a good deal 
of study and not a little labour, which has been directed 
more particularly to the needs of young men than to 
those of experience. That being so, the object of this 
book is, in the first place, to direct the attention of 
young smiths to the fact that science may be of immense 
value to them in their work in the smithy and forge ; 
andi in the second place, to try to show them how it may 
be utilised in a thoroughly practical manner. 

The whole idea has been to consider the subject from 
the point of view of its practical application to the needs 



of young smiths and forgers who are beginning work at 
their craft; so that I should like to deal with it princi- 
pally for their benefit. At the same time, the subject- 
matter is put forward with the knowledge that many 
experienced craftsmen do not realise the extent to 
which science may be applied in the workshop, and it 
is hoped that the study of these pages may lead them 
to a greater appreciation of the advantages to be derived 
from a scientific training. 

In introducing a subject of this kind to young men 
there is at the outset a great difficulty to be overcome, 
a diflSculty which ought not to exist, but which never- 
theless does exist — namely, the prejudice against that 
which is frequently though incorrectly termed technical 
education, when scientific training is really what is 
meant. It should be noted that most people — and it is 
a pity that they do — refer to technical education as if it 
were synonymous with scientific training. To avoid any 
such misconception, and at the same time to prove that 
this prejudice is really a judgment or opinion formed 
without due examination, let us consider briefly what 
is technical education and what is scientific training. 
What is the meaning of the word "technical"? It 
means "belonging to an art or profession." What 
does " education " mean ? It means " instruction." To 
educate is to instruct, to impart knowledge. Therefore 
technical education applied to our craft means " in- 
struction belonging to the art of smithing and forging," 
but it must be distinctly understood that this instruction 
is imparted only in connection with the skill exercised 
in the mechanical practice of the art. What is scientific 
training, and what is its true relationship to technical 
education? Science consists in systematic observation 
of facts and intelligent deductions therefrom, while 


art is the application of these facts. The man who has 
been scientifically trained has been taught the principles 
of science, whereas the man who has been technically 
trained has acquired skill in the actual practice of his 
art or craft. It should be interesting to reflect that 
every craftsman — although he may not think so — is a 
technician — that is, one who is skilled in the practice of 
an art, or, in other words, one who has received technical 
education. It is absurd, therefore, for anyone to say 
that he does not believe in technical education, for he 
who does so is virtually denying his own existence as a 
practical craftsman. 

It may now be stated that the prejudice mentioned 
is in reality directed against scientific training, and it 
is proposed to demonstrate in these pages that, while 
scientific training is certainly not the same thing as 
technical education, they are nevertheless inseparable — 
indeed, the latter cannot exist without the former ; and, 
further, that the practical or (as we may rightly call him) 
technical man who profitably receives scientific instruc- 
tion acquires knowledge which may be of inestimable 
value to him in actual practice. The instruction re- 
ceived consists principally in the practical application of 
* the scientific facts or truths which govern the art or craft. 
In view of this, does any intelligent smith or forger 
wish to put himself in the extremely absurd position of 
rejecting the acquisition of a knowledge of the truths 
which govern his craft? One cannot imagine any in- 
telligent smith or forger refusing to avail himself of the 
knowledge that is necessary to his existence as an efiicient 
craftsman. Since it is not in the nature of things to be 
contented with little, why should smiths or forgers be 
content with insufiicient or imperfect knowledge when 
there is such a wealth of it lying to their hands ? 


Much of this prejudice is due to a misconception 
regarding the relationship between theory and practice. 
What is " theory " ? It is the exposition of the general 
principles of a science or art. It is simply a knowledge 
of the rules that show how something should be done. 
Practice is the application of this knowledge to some 
useful end. It is the actual performance of the rules. 
Can it, then, be thought for a single moment that a 
person is able to perform anything successfully or 
economically when he does not know how to set about 
it, or when he knows nothing about the laws which 
govern it? It is not quite reasonable to expect it; 
speaking generally, success is likely to be proportionate 
to the extent of the knowledge. Such a person is only 
groping in the dark, a condition of affairs which might 
be pardoned if there were no light, but the fact remains 
that there is light, and that in abundance. 

We sometimes hear the expression, " It's practical men 
we want"; inferentially, theoretical men are evidently 
not wanted. Certainly we want practical men, but men 
who know what they are doing. Theory and practice 
are inseparable. The practical man is always of 
necessity a theoretical man, whether he admits it or not. 
The man who delights in saying that he is a practical, 
and not a theoretical, man, seems to be totally ignorant 
of the fact that the only truly practical man is he who 
has an accurate knowledge of his business. 

Perhaps, one reason why many young craftsmen look 
unfavourably on anything theoretical is that they see 
many examples of theory which are almost entirely 
speculative, and which may not be practicable when put 
to the test. Theory of this kind, however, fills its place 
too, for if men had never theorised and then experi- 
mented, many of the remarkable discoveries and in- 


ventions with which we are familiar would still be 

As an outstanding example of this we have only to 
look at the splendid achievement of Sir Henry Bessemer 
in connection with the making of mild steel. What he 
set out to do was in some respects a great failure. What 
he accomplished was a glorious success. 

The manufacture of wrought iron is an oxidising 
process. From the fact that the oxygen of the air com- 
bines with the carbon in the pig-iron for the purpose of 
removing it, Bessemer conceived the idea that by blow- 
ing air through molten pig-iron he could produce much 
better wrought iron, and that more quickly than by 
puddling. This was not a success, because when the 
complete removal of the carbon was attempted, the metal 
was rendered extremely brittle and rotten owing to the 
absorption of oxygen. It was not till Mr Robert Mushet 
suggested the use of manganese that success finally 
crowned his long and arduous labours. This was added 
in the form of ferro-manganese, a pig-iron rich in 
manganese. The oxygen in combination with the 
manganese was removed in the slag, while the addition 
of the carbon resulted in the manufacture for the first 
time of what is now known as mild steel. Thus, 
Bessemer in attempting to improve the making of 
wrought iron succeeded in making what might be called 
a new metal, which completely revolutionised engineer- 
ing practice all over the world. 

Professors and other teachers of science are sometimes 
rather lightly spoken of because they are not always 
practical men; but it does not follow, though a man 
cannot perform a certain operation himself, that he is 
incompetent to impart some information which may be 
applied in a thoroughly practical way by the practical 


man, and so prove most useful to him. The professor of 
metallurgy, for instance, tells us what takes place chemi- 
cally when we use sand or borax in the welding of iron 
or steel; and while he perhaps could not weld a rod 
although the safety of his own life were to depend on 
it, we are able to make valuable practical use of the 
scientific fact which he teaches. The professor of mathe- 
matics may not even know what we mean when we speak 
of " drawing down " and " staving up," yet all the while 
he is teaching scientific principles which may be utilised 
by us in a most beneficial manner. It is certainly to be 
regretted that they are not more fully conveVsant with 
the practical points in smith and forge work to which 
their teaching could be applied. If it were possible for 
the professors of metallurgy and mathematics, etc., to 
be at the same time practical smiths, they would make 
the acquaintance of many things of great importance 
which otherwise do not come under their notice. Con- 
sequently, their teaching could be directed along lines 
more in accordance with the needs of practical smiths 
and forgers. On" the other hand, if smiths and forgers 
with all their practical skill had added to it the know- 
ledge that these professors and scientists possess, they 
would be ideal craftsmen. 

While much of the work of scientists is of necessity 
theoretical, the scientists are at the same time in the 
highest sense practical men. Who among us would be 
likely to object to the possession of a knowledge of so 
high a practical nature as would enable us to tell the 
kind of treatment steel had received, by the examination 
of a small piece cut from it ? Yet many experts are able 
to tell whether the material has been over or under 
heated, or has been burnt, or cold-drawn or hammered 
when in the cold state, or crushed after forging. 


While it is impossible under existing circumstances 
for the professor or other teacher of science to be 
thoroughly in touch with all the points relating to the 
work of smiths and forgers, it is not impossible for the 
smith or forger to acquire as much of the professor's 
knowledge as may be practically applied by him. We 
shall consider how far this is possible by taking a 
number of practical examples in order to see if we can 
find something in any of the branches of science which 
may be applied to them. Therefore, instead of introduc- 
ing and explaining the nature of the scientific subjects 
that young craftsmen should be taught to take an 
interest in, I propose to consider these practical examples 
and to draw from the various branches of science just 
that which is necessary for each particular case. This 
method of dealing with the subject will at one and the 
same time show the need for scientific training in the 
smithy and forge ; the nature of that which should be 
taught, and also give absolute proof that it is essentially 


The Length of Material required to make 
a Forging. 

Take first the making of a forging, as shown in fig. 1. 

It is necessary to show with reasonable accuracy the 
length of material 4 inches diameter that will draw 
down to make the central part of the forging 24 inches 
long and 2 inches diameter. This is usually done by 




-4- §:- 

Fig. 1. 

guessing and trying a length. A smith of experience 
may come near to it, but even he does not know till he 
tries, and is more often wrong than right, especially if 
it is a size of work somehwat different from that which 
he is in the habit of doing. 

What then of the young smith of little or no experi- 
ence ? He has nothing to aid him in his guessing. Why 
let him grope in the dark, when by acquiring a thorough 
understanding of a few simple rules in mensuration, 



together with a suflBcient knowledge of arithmetic, — in 
order that the working of the rules may be accurately 
performed — he may always find his length with ease 
and confidence? 

The length for a forging of this kind is always found 
by dividing the volume of the forging by the area of a 
cross-section of the stock. This is best shown by stating 
it as a simple formula. 

Formida : — 

When L is the length, 

D the diameter of forging, 
d the diameter, 
and I the length of stock, 

^^^^ ^^ cZ^x7854 

_ 24x2x2 

= 6 inches. 

In this particular case we might have made use of the 
mathematical law that when the diameter of a circle is 
twice that of another, its area is four times that of the 
other. A glance at this forging will show that it comes 
imder this law, and by simply dividing the length of the 
central part of the forging by 4, the length of the stock 
is obtained. The same law applies to square or any 
other regular section. 

Loss of Material while Heating. 

In many cases an allowance will have to be made for 
loss of material due to causes which are quite familiar 
to smiths and forgers, and which may easily be provided 


for when consideration is given to the nature of the 
forging in hand. If a forging is to be made of iron, it 
is likely to be subjected to several welding heats which 
will necessitate an allowance being made for loss of 
material in the fire. Even with mild steel, which must 
not be heated to a temperature above a white heat, there 
is a loss due to oxidation. I am quite conscious that it 
is impossible to estimate this loss accurately, but at the 
same time an approximate rule may be adopted as a 
guide which should suflSce for all practical purposes. 
After careful observation of many different forgings, I 
have thought it suitable to fix the amount of loss at 
about 3 per cent, for mild steel, 6 per cent, for iron 
receiving what is termed a " wash-heat," and 12 per cent, 
for iron receiving a good welding heat. 

In the example already given, if the forging had to be 
made of iron which required a good welding heat, there 
would have to be an allowance of 12 per cent. Since 
12 per cent, of 6 is 0*72, the length would have to be 
about 6f inches instead of 6 inches. 

Alteration of Forging. 

Take another example. Someone has blundered, or 
for some reason or other it is necessary to alter the 
length of a forging. It has been discovered that a 
forging of the dimensions given in fig. 2 is about 7f 
inches too short. 

We will assume that it has been made of mild steel 
and that welding is prohibited. A certain amount, say 
\ inch, has been added to the diameter for turning. It 
is desirable to know if there will still be sufficient 
material for the purpose if it be drawn to the altered 
length. The important question is — What will the new 


diameter be ? Of course it could be tried, but the job is 
in a hurry, and after spending valuable time in trying, 
perhaps it will not do, and if that had been known at 
the outset a new one could have been well on the way. 

On the other hand, the forging may be thrown un- 
necessarily to scrap because of the apparent unlikelihood 
of it coming out to the altered length, and the decision 
having been come to that it would only be waste of time 
to try. Young smiths and forgers should be taught 
that there is a method which would enable them quickly 

^3-^ 4,^' ^3 f, 

Fig. 2. 

to decide whether to commence at once to draw such a 
forging or to make a new one. 

This forging being of steel, will lose about 3 per cent. ; 
3 per cent, of 48 is 1'44. 

Therefore the material equivalent to a length of 
48 -1-44 = 46-56 inches. 

Fomiula : — 

When L is the length, 

d the diameter of forging, 
I the new length, 
and D the required diameter, 

then D = J^^^ 




= 77^5673 = 2-75 inches. 


This forging, if drawn, would be 2f inches in diameter, 
the knowledge of which fact would enable those concerned 
at once to decide what course to adopt. 

Length required to make a Ring. 

Consider next the making of a ring. A ring of the 
dimensions given in fig. 3 is required. The length of 
material must first be found. There are several ways of 

Fig. 3. 

doing this, but, like many other things, there is only one 
correct way, and that is by means of a very simple rule 
in mensuration. 

In every circle the ratio of the diameter to the 
circumference is as 1 is to 3*1416, or 3|, which means 
that the diameter must be multiplied by 3^ in order to 
find the circumference. 

In a ring, however, there are two diameters, the 
outer and the inner. How is the young inexperienced 
smith to proceed ? This causes us to notice a most 
important feature of the whole subject. It is such a 


case as this that proves how desirable it is to have 
mathematics taught in a more practical way. The 
science teacher cannot be expected to be familiar with 
all the points in practice to which science could be 
advantageously applied. What he teaches is excellent 
in its way, but more good could be done if he were 
able, as in this particular case, to teach the theory 
of what actually takes place when a bar is bent to 
make a ring. 

The breadth of bar must be taken into account, 
because in bending a bar the outside material stretches 
and decreases in depth ; the inside material is compressed 


^ -^ f^" ^ 


Fig. 4. 

and increases in depth ; from which it naturally follows 
that a part near the centre must remain constant, as it is 
neither in tension nor in compression. This part is called 
the neutral axis, and if the elastic limit be not exceeded, 
it coincides with the centroid of a cross-section of the 
bar, which in this case is a line drawn through the 
centre ; but if the elastic limit be exceeded, the neutral 
axis will leave the centre line, and the smaller the radius 
of the ring, the nearer will it approach to the compres- 
sion edge. This may be demonstrated by taking two 
thin strips of steel of the same length, and fixing them 
as shown in fig. 4, one on the top and the other on 
the bottom of a steel bar. After having drawn lines 
across the bar joining the ends of the strips, let the bar 
be bent as in fig. 5. It will be observed that the line 


joining the free ends of the strips has now altered 
its position in relation to them, and that it is almost 
midway between them. This proves that the inside 
of the bar has been shortened almost as much as the 
outside has been lengthened, fronp which fact it may 
safely be concluded that it will serve the purpose 
quite well to regard the centre line as having remained 

That being so, it follows that it is the mean diameter 

^-'-— Z^"— --^^ 

Fig. 5. 

which must be multiplied by 3| in order to find the 
length of material required to make the ring. 

The mean diameter of this particular ring is 14 inches. 

Formula : — 

When D is the mean diameter, 
and C the length of the bar, 

then C = Dx3f 


" 7 

= 44 inches. 

A piece of bar 1 inch square, 44 inches long, will 
make the ring in question. An allowance will have 


to be made for staving up and for loss of material iii 
the fire. 

The same result may be obtained by adding the 
thickness of the bar to the inside diameter of the ring ; 
but the method given is preferable, because it explains 
the true reason for using the rule, and also proves that 
it must be correct. 

Amount of Contraction on Outside and 
Expansion on Inside of a Ring. 

Jt is quite possible to find approximately, by means of 
a very simple rule, the thickness that the outside edge 
of a ring will be after the bar is bent. 

It is required to make a ring as in fig. 6. What will 
be the thickness of the outside edge ? 

ForTYiula : — 

When D is the mean diameter, 
d the outside diameter, 
t the depth of bar, 
and T the outside thickness after bending, 

then T = ^^ 

2 X 2 X 24 
= 1^1 inch. 

The practical value of this will be made more apparent 
by showing that a slightly different formula will give 
the thickness of bar necessary to make a ring having 
a given thickness on the outside edge. 

Suppose that it is necessary to order bars for the 


purpose of making a large number of rings to dimensions 
as shown in fig. 7. They have to be turned all over, 

Fig. 6. 

Fig. 7. 

for which \ inch has been added to the breadth and 
\ inch to the thickness. 



: — 


D is the mean diameter, 

d the outside diameter, 

t the depth of bar, 


T the thickness required to make the ring, 

_ dxt 

then T = ^- 


~ 32 





= -97 inch. 

Bars 3J inches broad, 1 inch thick, will be required 
to make the rings, and when bent there will be about 
\ inch for turning on the flat. 

Advantage of being able to Calculate. 

Mathematical calculations for smiths and forgers have 
been objected to on the grounds that it is impossible to 
work to such exactitude in making a forging. That to 
some extent is true, but those smiths who speak in this 
way seem to overlook the fact that if fault there be, it 
lies with themselves. No craftsman has ever yet been 
able to make anything with mathematical accuracy even 
though possessed of the finest tools ; but looking at the 
matter from a common-sense point of view, surely it is 
far better for a smith to know that if everything could 
be performed perfectly, the forging would work out 
exactly true to the calculation, say, 6f inches, as in 
fig. 1, p. 8, than to be in doubt as to whether 5, 7, or 


even 8 inches is the correct amount. The man who has 
a definite starting-point, even although he may not hit 
the mark with mathematical nicety, will undoubtedly 
meet with a larger measure of success than the man 
who has no such definite knowledge to aid him. 

At the same time it should be pointed out that while 
it is absolutely necessary to understand the fundamental 
principles and to be able to calculate accurately, it is not 
always necessary to calculate minutely. For example, 
if the mean diameter of a ring is 14J inches, by 
multiplying by 3|, the length of bar required is 45*5714 
inches ; but it would be quite absurd to expect any smith 
to attempt to work to such a minute calculation. It 
has already been shown that this method of calculating 
for this particular example is scientific and practical, 
and it may now be said that it does not become one 
whit less so if the smith, while practically applying it, 
exercises common-sense by stopping his calculation at 
two figures or even one figure beyond the decimal point, 
and taking his result as 45yV or ^^h inches. This 
question of minuteness of statement in calculations will 
be referred to again in connection with certain other 

Some smiths find lengths for drawing down and 
staving up by means of a table of weights. The scope 
of this method is limited, because all forgings are not of 
regular shape, such as are shown in these tables. Besides, 
there are many problems in smith-work with which 
drawing and staving have no connection, and which 
can only be solved by means of some simple rule in 
mensuration. Moreover, the smith who depends on a 
table of weights is at a loss if it be not at hand, whereas 
the smith with a knowledge of mensuration carries it 
with him as a part of himself, always ready when 


required. In many cases he may have obtained his 
length before the other has had time to turn up his 
table of weights, which process after all may involve a 
calculation far more elaborate than that required by the 
method adopted by the one who is a mathematician. 
I do not wish to condemn entirely any tables which 
may be helpful to smiths and forgers, but this I do say, 
that they do not teach them anything which in any 
way makes them less dependent on others, and more on 
themselves. What self-respecting smith or forger would 
care to have part of his work given to another because 
he was unable to do it ? In the same way, it ought to 
be as much his business to be able to calculate the 
amount of material as it is to make the forging. 

Meantime we must leave this most important study, 
but I may say that at present I am engaged on subject 
matter for another book devoted entirely to calculations 
for smiths and forgers. 


The practical examples already given prove the value 
of a knowledge of mensuration; the value of some 
knowledge of the strength of materials will be more 
apparent when we consider that many craftsmen are 
not in touch with a drawing office, and they have, 
therefore, to do a little on their own account in the 
way of determining the sizes that certain forgings 
ought to be. 

Many years ago I had the need for such a knowledge 
much impressed upon me, by having brought to my 
notice the fact that the lack of it on the part of a 
smith and an engineer resulted in a fatal accident. 
They had arranged between them the dimensions of an 
attachment made by the smith for a " kettle," as it is 
called, which was used for lowering men and material 
in a mine shaft in course of sinking. The attachment 
broke, not because there was anything wrong with the 
material or the workmanship, but because the dimen- 
sions were totally inadequate to support the load that 
it was required to carry. 

Dimensions of Eyebolt. 

Take a practical example. A smith is asked to make 
an eyebolt which he is told must carry a load of 8 tons. 


It is to be made of mild steel, the ultimate breaking 
stress of which is 30 tons per square inch. Since the 
eyebolt has to be in constant use, 5 may be taken as a 
factor of safety. Dividing 30 by 5 gives 6 as the safe 
load in tons per square inch. What must the diameter 
of the bolt and of the eye be ? 

Fonnula : — 

When d is the required diameter of bolt, 
L the given load in tons, 
and I the safe load in tons, 

then cZ = y^--V__ 

V 6 X -7854 

= Vl-6976 

= 1*3 or 1 j^^ inch approx. 

The diameter of the eye may be found by formula. 

FoTTnula : — 

When d is the diameter of bolt, 
and D the diameter of eye, 

then T>=./- — ^^^^■ 
V 2 X -7854 



V 1-57 

= vro7 

= 1-034 or IjjV or IJg- inch. 

The forging of the eyebolt may now be proceeded 
with in accordance with the dimensions shown in fig. 8. 


The question of unnecessary minute calculation may 
well be considered in connection with this problem ; it is 
at once conceded that with such problems as this it 
is not necessary to calculate so minutely as I have done 
in this instance. This should be quite apparent after 
briefly considering a few points. Although it is stated 




Fig. 8. 

that the eyebolt is to carry a load of 8 tons, this is 
probably quite a rough estimate, and at any rate is not 
at all likely to have been calculated to a fraction of a 
pound. Again, the fact that the load may not be 
perfectly steady will affect the question, and may do so 
very considerably. Then the 30 tons taken as the 
ultimate breaking stress of the material is variable even 
in the same bar, and the tensile strength of a forging is 
not the same as that of the bar from which it is forged. 


Further, the factor of safety of 5 is very indefinite; 
some designers allow 4, while others allow 6. 

For these reasons, in problems of this kind, one may 
be quite justified in using | as a good enough approxima- 
tion to '7854. Dr A. Barr, the distinguished Professor 
of Engineering, has very kindly communicated to me, 
along with other valuable information, his exceedingly 
interesting opinion on this question of minuteness of 
statement in calculations, which is well worth quoting 

He says : " Even the most scientific writers occasionally 
— I might say often — state constants, etc., to a quite 
unscientific degree of minuteness ; and most people who 
deal with calculations almost invariably do so. For 
example, if the breadth of a table top is given as 1 metre, 
some people will say that the breadth is 39'370113 inches. 
The two statements are not the same. It is true that 
1 metre is 39-370113 inches, but when we say a table is 
1 metre wide, we do not mean that it is not 1*00000001 
metre wide. To state that it is 100001 metre wide would 
be a very unscientific statement, since the breadth is not 
definite to anything like a ^a^^aa metre. To say that 
the table is 39*370113 inches broad means that it is 
definite in breadth and has been measured to one- 
millionth part of an inch. 

"In practical calculations we ought only to state 
constants, or results, to something like the degree of 
minuteness that we can trust them to. Thus, for 
example, I have always objected to the strength of a 
steel plate being given as 28*37 tons per square inch. 
There is no definite strength of a given steel plate. 
No two test specimens cut from the plate will give 
results in agreement to a y^^ part of a ton. Besides 
this, it is true to say that there is no definite strength 


even of a given test specimen. If you could test it 
rapidly and again slowly — at 40° F., and again at 60° F., 
— or test it continuously and again with halts in the 
application of the load, you would get markedly different 
results. There may be some justification — and therefore 
some truth — in saying that the strength is 28 tons per 
square inch, because we understand in a rough way by 
what kind of test we refer to, but it is untrue to say — 
even for a given specimen of plate — that the strength is 
28*37 tons per square inch. 

"Again, in all calculations regarding the strength 
required for a given piece we should deal in round 
figures. Take, for example, the connecting rod for a 
locomotive. We have only a very rough idea, not only 
of what the strength will be when the thing is made, 
but a still rougher idea of what duty it will be called 
upon to perform. The steam pressure that may come 
upon the piston is not accurately known. The inertia 
stresses are not known. If there is water in the cylinder 
the connecting rod may be subjected to very great 
extra stress, and so on and so on. Again, the alternating 
stresses of push, pull, bending, etc., are very different 
from the testing machine stresses in their effects. It is 
for such reasons that a factor of safety of 4, 5, 6 or 
more is allowed, which only means that we do not 
know within some hundreds per cent, what the condi- 
tions are." 

For these reasons, therefore, Dr Barr states that for 
problems such as the above, | is a good enough approxi- 
mation to "7854. 

Before passing on it will be of much value and interest 
to work out the first part of the last problem by a method 
recommended by Professor Barr, a method which will 
at once be considered as being much easier and much 


more rational — for such problems — than the methods 
taught by some scientists, who have not the advantage 
— as he has — of a thorough practical experience in com- 
bination with a deep scientific knowledge. 

XTST 1 . , 30 tons per sq. inch 
Workmg stress = ^-^ — -* 

= 6 tons per sq. inch. 
Load = 8 tons. 

Area required = — sq. inch. 

As the bolt is to be of circular section and the area of 
a circle is about f diameter squared, the square of the 
diameter required will be 

8 4_16 
6^3"" 9 ' 

and the square root of -^ = — ^ 

= g = 1^ or ly«^ or If inch. 

It will be observed that the result in each case is the 
same, since — for the reasons given — it was not thought 
necessary to calculate to more than one decimal place in 
the first case. 

Dimensions of Bax for Race. 

A bar of wrought iron is required to be fixed up as 
a rstce at a smith's fire ; the ends are to be supported 
10 feet 2 inches apart. If the greatest load likely to 
be carried is 20 cwt., what must be the breadth and 
depth of the bar ? 


The depth of the bar may be taken as four times the 
thickness, and is found by the following formula. 

FoTTThula : — 

When D is the depth in inches, 
L the length in feet, 
and W the load in tons, 

then D=./^j^ 


12 X 106 

^ y 122 
V 1-272 
= 5^95^ 
= 4'5 inches or 4^ inches. 

The thickness of the' bar is one-quarter of the depth. 

Therefore the thickness = -r 




= 1-125 or IJinch. 

A bar of wrought iron 4 J inches deep and 1^ inch 
thick is required to suit the purpose. 

In problems of this kind, the cube root may be ob- 
tained sufficiently near the truth for practical purposes, 
by trial. Thus, 4^ = 64; 5^=125. The required cube 
root is between these ; try 4J ; the cube is 91*125 or 91^. 
The answer therefore is fully 4J inches. 

Safe Load for Race. 

It is necessary sometimes to have some idea of the 
load which may be safely placed on an existing race. 


What load may be carried by a bar 6 inches deep, 
1^ inch thick, supported at the ends 12 feet apart ? 

Formula : — 

When W is the load in tons, 
D the depth in inches, 
T the thickness in inches, 
and L the length in feet, 

then w=-^?ii^TxD* 


_ -424x3x36 

= 1*9 or about 2 tons. 

The practical application of these problems shows that 
smiths may profit by having a little knowledge of the 
strength of materials ; the working out of the problems 
shows the value of a little knowledge of algebra. 



A PIECE of J-inch plate is bent in the form of a truncated 
cone, as shown in fig. 9. It is necessary to know the 
shape of the plate before bending, as in fig. 10. 

fl \ \ 

4 W' ' Y\ 


1 : 

I ,' i 

i. i2y2- 

Fio. 9. 
Produce the centre line of the two sides till they meet 
at the point O. With O as centre and OB as radius, 
describe the arc BF. With the same centre and OD as 




radius, describe the arc DE. Step off on the large arc 
a length equal to the mean circumference of the bottom 
of the cone. This length terminates at F. Join OF, 
cutting DE at E. The figure BDEF represents the 
form of plate required to make the cone-shaped article. 

Fig. 10. 

This is a typical example of the development of sur- 
faces, a subject with which all general smiths should have 
some acquaintance because of the varied nature* of their 
work. This particular case has been chosen because 
smiths have often to make bevelled rings, which require 
to be bent on edge before being bent to form a ring. 
The radius to which the bar must first be bent on edge 
may be found by means of the method just described, 


but this is not always convenient. It may quite easily 
be found by means of a simple formula. Suppose that 
a small stiffening ring 1 inch deep and \ inch thick is 
to be put inside the bottom of the last example, as 
shown in fig. 9. 

Forviula : — 

When D is the large mean diameter, 
d the small mean diameter, 
B the depth of bar, 
and R the outside radius of curve, 


then R = 


_ 47x2 

= 23^ inches. 



The drawing of the example shown in fig. 9 (p. 28) 
necessitates a knowledge of practical geometry, the 
value of which may further be shown by the following 

Two rods, each IJ inch in diameter, have been used 
to suspend a load, but it has been found necessary to 
alter the arrangement in such a way that one rod may 
replace the two if it be made equal to the two in strength. 
One rod will require to have an area equal to the sum of 
the area of the two rods. The diameter of the rod may 
be found by calculation, but it is very easily arrived at 
by geometry. 

Describe a circle 1^ inch in diameter, as in fig. 11, 
and draw two diameters at right angles to each other, 
producing them beyond the circle. Bisect the angle 
AOB with the line OC. At C draw a line perpen- 
dicular to 00, cutting OB at B. With OB as radius 
describe the large circle, which has an area double 
that of the small one. This large diameter measures 
barely 2^ inches, which is the size of a rod required to 
support the same load as two rods, each 1| inch in 




The converse of this may be illustrated by another 
example. A flange 9 inches in diameter has to be staved 

Fig., 11. 


Bottom Die 

Fig. 12. 

up on the end of a long round bar 4 inches in diameter, 
but it is discovered that the hydraulic press at which the 



work has to be done is not powerful enough to do it. A 
plan is conceived whereby it may be done in two 
operations. A loose disc with a tapered point on it, as 
shown in fig. 12, is placed in front of the 9-inch ram, the 
point being half the area of the large ram. This is 
pushed up to the correct thickness and then withdrawn. 
The 9-inch ram is then brought to bear on the flange. 

Fig. 13. 

and of course it has only to act on the outer circular 
part. The diameter of the point of the disc may also in 
this case be found by calculation, but again it may easily 
be obtained geometrically. 

Bisect the radius of the large circle at C, fig. 13. With 
C as centre and CD as radius describe a semicircle. 
Erect a perpendicular CA at C. With OA as radius 
describe a circle which will have an area equal to half 
the area of the large circle. The diameter is about 
6| inches. 



Mechanical Drawing. 

It is an easy step from geometrical to mechanical 
drawing. Drawing is the language o£ the workshop, 
and the smith or forger who is not conversant with it 
has to contend against a very serious drawback. It is 
imperative that all smiths and forgers should be able to 
read a working drawing, and the best way to learn to do 
so is to learn to draw one. A knowledge of this kind 
makes it possible to impart and receive information 
intelligibly regarding work to be done; lack of this 
knowledge not only hinders a man at his work, but is a 
fruitful source of mistakes. 



It has many times given me considerable amusement to 
see the astonishment that was expressed at the idea of 
a smith studying mechanics. If there were time it 
would be a matter of no great difficulty to prove that 
there is as much scope for a knowledge of this most 
important subject in the smithy and forge as in any 
other department of engineering. 

Blocking Hoops. 

The following is a good practical example. A smith is 
engaged making a number of large hoops ; after having 
welded them, it is necessary to reheat them for the 
purpose of rounding them up to the proper diameter on 
what is called a segment block. This block is usually 
cast in three parts, which are expanded by means of 
tapered wedges. Now these wedges are of more import- 
ance than one would think ; the smith, before making 
them, must consider what taper he will give them. This 
he may determine by thinking only of the amount by 
which he may require to expand his block, but he may 
make the taper so large that the wedges will keep 
jumping out instead of driving in. He may know from 



experience that if he makes the taper very small, this 
will not be the case, which fact naturally leads to 
the correct conclusion that if the taper be gradually 
decreased it will come to a point where the wedge will 
cease to jump out. By studying practical mechanics the 
smith would become acquainted with what is called the 
limiting angle of friction, the angle at which the wedge 
ceases to jump. He would also learn the law which 
governs it, which would enable him to make use of the 
following formula : 

Formula : — 

When t is the thickness of the point, 
L the length of the wedge, 
0*2 the coefficient of friction, 
and T the thickness of head of wedge, 

then T = ^+(Lx-2). 

In this case the wedge requires to be 12 J inches long 
and \ inch thick at the point. What is the thickness 
of the head ? 

Formula : — 

T = ^ + (Lx-2) 
= -5 + (12-5 X -2) 
= •5 + 2-5 
= 3 inches. 

It is essential that the head be made a little less than 
3 inches in order to keep it within the limiting angle 
of friction. The wedge is shown in fig. 14. 

The law of friction is made quite clear by means of a 
diagram. The inclined plane in fig. 15 is arranged so that 
it may be raised or lowered, increasing or diminishing the 
angle ABC. When the angle is very large the body placed 



on the plane will slide, but if the plane be lowered it will 
arrive at a particular angle where the sliding movement 

Fig. 14. 

will cease, even though extra weight be added. Accord- 
ing to the ordinary laws extra weight would produce 
no effect. The angle ABC is the Limiting Angle of 

Fig. 15. 

Friction. If the wedges are made of mild steel and the 
segment blocks of cast iron, as in this example, the 
angle may be about 11 or 12 degrees and the coefficient 


of friction about 0*2, but it is impossible to state these 
as definite constants. The angle of course varies with 
different substances, and for these the coefficients 
usually given are different, but it also varies with the 
same substance under certain conditions. For instance, 
in the practical example under consideration, much may 
depend on the kind of mild steel or cast iron, and the 
coefficients depend very greatly on the degree of rough- 
ness or smoothness, and any slight application of a 
lubricant would also greatly alter it. For these reasons, 
in making such a calculation, due consideration must be 
given to the particular conditions, and even then it is 
essential that some allowance be made for any possible 

For wrought iron in contact with cast iron the angle 
is usually stated as being about 10 degrees, and the 
coefficient of friction about 0*18. For mild steel in 
contact with mild steel the angle is about 8 degrees, 
and the coefficient of friction about 0*14. 

The law of friction is frequently seen in operation in 
connection with tapered sets for use under the steam 
hammer, and it is no uncommon sight to see these 
forcibly projected if the angle has been too great. 

To those who may still consider it absurd to 
recommend teaching of this kind to smiths or forgers, 
let me say that an accident came under my notice many 
years ago in connection with the blocking of a hoop on 
a segment block. 

The hoop having been made rather small, the wedges 
were driven in as far as they would go, and were left 
there with the object of stretching the hoop as it cooled. 
It is quite obvious that an enormous pressure would be 
brought to bear on the wedges by the ring during 
contraction. The angle of one of the wedges had been 


very close to the limiting angle of friction, because, on 
receiving a side blow for the purpose of loosening it, 
it immediately shot up into the air to a height of about 
60 feet, and before anyone could realise what had 
happened it dropped back with fatal result. If that 
wedge had been made to an angle less than the limit- 
ing angle of friction, no power whatever could have 
moved it. 

Shrinking Pit for Hoops. 

This last very striking proof of the value of a know- 
ledge of mechanics leads to the consideration of another 
example in connection with the strength of materials, 
and also introduces the study of heat. It is most 
important that a hoop which has to be shrunk on some 
article should have the proper allowance left in turning 
in order that it may just be heated sufficiently to permit 
of it passing over the article, and then when cold giving 
a compression which is adequate for the desired purpose. 

Take as an example a pulley 31 J inches in diameter 
on which a hoop has to be shrunk. It is to be made 
of mild steel 2 inches broad and \ inch thick. To what 
diameter should the hoop be turned inside in order to 
secure a proper shrinking fit ? 

Formula : — 

When d is the diameter of the pulley, 
and D the required diameter of hoop, 

then D = d-iooo 

-31-25 ^^ 
= 31-218 inches. 


Expansion and Contraction. 

In connection with the expansion and contraction of 
metal we may consider another example. A die is 
required to punch out round discs 12 inches in diameter. 
As the metal will be hot, the die and punch must be 
made larger than 12 inches to allow for contraction. 
The plates are likely to be heated to a bright red, about 
900° C. The diameter of the die may be found by the 
following formula. 

Formula: — 

when D is the diameter of disc, 
T the given temperature, 
t the atmospheric temperature, 
d the required diameter of disc, 
and 0*000012 the coefficient of expansion for 
mild steel, 

then d = D + {D X •000012(T - 1)) 
= 12 + 12 X 000012x885 
= 12-127 or fully 12^^ inches. 

In this instance the atmospheric temperature is taken 
at about 15° C. 

Smiths are in the habit of allowing \ inch to the foot 
for contraction, but a little reflection will show that this 
can only apply to a particular temperature, which is a 
bright red or about 900° C. 

It is not strictly true, but it will suffice for all practical 
purposes to say that for every degree Centigrade of 
increase, a forging made of mild steel will increase 
•000012 of its own length. This constant is called the 
coefficient of expansion. 

An intelligent stamper will take care that he finishes 
all his stampings as nearly as possible, at the same tem- 


perature. Suppose he were making levers with a boss 
on each end in which a centre is stamped in order to 
facilitate drilling, and he were to finish one at a bright 
red heat, it would, as we have already seen, contract 
about \ inch if the centres were 12 inches apart ; but if 
he were to finish another at a yellow heat — about 
1100° C. — it may easily be proved by the formula given 
that it would contract another -^^ inch. 



One of the most instructive and fascinating studies is 
that of metallography, the science which describes the 
structure of metals and of their alloys. M. F. Osmond and 
Dr J. E. Stead in their excellent work, The Microscopic 
Analysis of Metals, have in a most interesting manner 
compared metallography with other sciences, and have 
shown that the study of metals may be divided into 
several subdivisions quite analogous to those of medical 
science. These subdivisions are anatomical, biological, 
and pathological. Briefly stated, the first defines the 
physical and chemical constitution of metals ; the second 
shows how the constitution may be infiuenced by the 
various forms of necessary treatment, whether thermal 
or mechanical, to which the metal is legitimately 
subjected while being manufactured or used ; the third 
deals with the diseases of metals, which may be of two 
kinds : those which are due to impurities in the metal 
or to defects in the process of manufacture, and those 
which are due to subsequent improper treatment. By a 
careful study of these subdivisions much may be learned 
regarding the prevention and cure of diseases in metals. 
Metallography, and particularly that branch of it 
which deals with the microscopic examination of iron 
and steel, has gone far to add to our knowledge of the 



nature of steel and its behaviour under thermal and 
mechanical treatment. The mysteries are being gradu- 
ally explained and the erroneous ideas dispelled. 

With two of the latter we may deal, since they are of 
special interest to smiths, many of whom have formed a 
wrong conception of what fibre in iron really is, and who 
do not know that iron and steel are initially crystalline, 
but are under the impression (and they are not alone 
in this) that they only become crystalline through 
being subjected to vibrations or alternating stresses. 

Before dealing with these points, however, it is 
necessary to explain the method of examining metals 

The Microscopic Examination of Iron and Steel. 

For the benefit of those who are not familiar with the 
examination of iron and steel under the microscope let 
it be briefiy stated here that small pieces are cut from 
the metal in question, and after polishing and etching 
are ready for examination. The polishing is usually 
done by means of increasingly fine emery papers fixed 
on a small revolving disc, finishing off with fine diaman- 
tine powder placed between two pieces of selvyt cloth 
soaked in water. 

The specimens are polished until all mechanical 
scratches have disappeared, and on occasion are examined 
at this stage for cracks, slag inclusions or other features 
which are not so readily observed after etching. The 
information obtained by suitable etching is of the 
greatest importance, for not only are the constituents 
of the specimens of metal revealed, but also the history 
of the last heat or other treatment. The effect of the 
etching reagent is to eat out the soft parts and leave the 
hard parts in relief. 


There are various etching reagents. The specimens 
from which my own photomicrographs were taken were 
immersed, some of them in a 1 per cent, solution of 
nitric acid in alcohol, and others in a 4 per cent, solution 
of nitric acid in iso-amyl alcohol. At present it is not 
necessary to consider other reagents, each of which is of 
course made use of for different metals and for the same 
metals under different conditions. 

In general, the method of application is to immerse 
the specimen in the reagent until the brightness on the 
surface begins to disappear. On removal, it is washed 
in alcohol and dried, preferably in an air blast. Person- 
ally, I prefer to dip the specimen into a small vessel 
containing the reagent in such a way that the polished 
surface only is immersed momentarily, and then to 
observe it while gently moving it to ensure the even flow 
of the reagent all over the surface. If the first application 
is not sufficient to develop the desired structure it may 
be repeated. It is much easier to do so than to repolish 
because of having over-etched at the outset. The 
specimen is finally mounted quite level on a small piece 
of glass by means of a small piece of plasticine. 

In the meantime, it is proposed to describe only those 
constituents appearing in the structure of the various 
specimens of iron and steel under consideration. The 
points to be dealt with in connection with iron and steel 
and the photomicrographs which illustrate them will 
suffice at present to give some idea of the science and 
its application to practical work. After one has studied 
the subject sufficiently to become familiar with the 
different constituents, their characteristics and the con- 
ditions under which they are modified and altered, it is 
astonishing what may be learned and how much more 
easily some of the difficult problems may be solved. 


Quite recently I was told by a friend that he thought 
it a mistake to show these photomicrographs to smiths 
and forgers, because, he said, " They do not understand 
them." If my friend had been trying to find an argu- 
ment in favour of showing them, he could not possibly 
have found a better one. The mere fact of smiths 
being ignorant of the valuable information to be obtained 
from this source is an all-sufficient reason for bringing 
it under their notice, and I hope to see the day when 
many, if not all, young craftsmen may be taking an 
active interest in the scientific side of their craft. 

Fibre in Iron and the Crystalline Structure 
of Iron and Steel. 

Smiths are apt to confuse fibre with lamination, which 
is an imperfection. Lamination is a source of weakness 
and is due to the method of manufacture. Oxides 
and other impurities under the general term of " cinder " 
are present between the layers of iron that make up 
the pile from which the bar is rolled, as is shown in 
the upper part of fig. 18, p. 49. A bar of iron is 
composed of a number of more or less imperfect welds. 
In the layers themselves, cinder to a certain extent is 
intermixed with the granules of pure iron. In making 
the iron, everything that can be done is tried to eliminate 
these impurities, in order to obtain as far as possible a 
homogeneous mass. 

The structure of iron is always crystalline and is as 
much so as steel. By slowly bending or pulling in the 
cold state, the crystal grains are stretched enormously, 
and these fine-drawn crystal grains are what constitute 
true fibre. If a piece of iron be nicked and broken 
suddenly it will show a more or less crystalline fracture, 


and this is due to the obvious fact that the crystal grains 
had no time to stretch. If a piece be nicked and broken 
very slowly, it will show the crystal grains in the solid 
parts between the laminations drawn out very fine with 

Fig. 16. — A piece of wrought iron showing that it is possible to obtain a 
fibrous and a crystalline fracture in the same piece of good material. 

a somewhat silky appearance. Fig. 16 shows that it 
is possible to obtain these two kinds of fracture in the 
same piece of good wrought iron. 

Dr Stead has given me some valuable evidence to 
prove this. In connection with two different pieces of 
iron, one of which broke with a fibre and the other 
without, he tells me that, " The microscopic examination 


of cut and polished sections of each of these proves that 
both are crystalline. It is the points of the drawn-out 
crystals which give the fibrous appearance of the broken 
iron : it is the act of breaking cold which really produces 
the fibre. True fibre is never present in even the best 
wrought iron unless it has been cold-drawn. In break- 
ing by bending, iron with cinder intervening breaks up 
in detail, i.e. the separating layers breaking up one 
after the other, and the fracture indicates lamination." 
He further says, " Let me finally tell you that even the 
best wrought-iron bars, if nicked and broken by dynamite, 
will give a complete crystalline fracture. This is direct 
practical proof that wrought iron is not initially fibrous, 
but crystalline, and that fibre is only produced during 
the time the iron is broken, when the crystals are being 
drawn out." 

No comment of mine is required to emphasise the 
importance of such a statement by such a distinguished 

In mild steel there is no lamination because of it 
having been in the molten state, and, like iron, it is also 
crystalline in its micro-structure, as may be seen in the 
lower part of fig. 18. In breaking pieces of steel the 
crystals behave in the same way that they do in iron, 
and unless in this sense it is entirely wrong to speak 
of fibre in either iron or steel. 

The question of the crystalline structure of iron was 
investigated in a most interesting manner by Mr David 
Kirkaldy about fifty years ago, and shortly afterwards 
Dr Percy, in dealing with the same subject, paid a high 
tribute to the splendid work done by Kirkaldy. In 
more recent years the microscope has revealed indis- 
putable evidence ; pieces have been polished, etched, and 
magnified to a high power, and have given proof that 


iron and steel are really crystalline. The micro-structure 
of wrought iron is shown in fig. 17. It is composed of 

^:^' •(*•■• ".v^i; 


v^ - 

Fig. 17. — A transverse section of a piece of Yorkshire iron. 
Magnified 80 diameters. 

a number of crystal grains, and the network appearance 
is given by the boundary lines of the crystal grains. 

While preparing several specimens to show this 
microscopically, I had a piece cut from a bar which had 
been supplied as best wrought iron. The microscope 



has revealed in fig. 18 that the bar of so-called wrought 
iron had been made up of a mixture of alternate layers 


■f '■ ^ 



of iron and soft steel. Tensile tests were subsequently 
made, giving results which could not possibly be obtained 
from any make of wrought iron. We shall assume 
that the iron maker was unconscious of this having 
taken place, and be all the more disposed to exercise 



this charitable spirit, since he has supplied us, not only 
with an excellent proof of the detective possibilities of 
metallographic methods, but at the same time with a 
specimen which suits our present purpose, in that it 
reveals in one piece the characteristic structure of 
wrought iron and mild steel, and that they are each 
without doubt initially crystalline. 

It will be observed that the two portions of the 
photomicrograph have one feature in common, that is, 
the white ground mass, which is built up, as it were, 
of small irregular areas having fine junction lines 
between. This white ground mass is iron, or f errite, as 
it is called; the''small irregular areas are the crystal 
grains of iron in the wrought iron and in the soft steel. 
In the wrought iron portion the long grey streaks are 
slag inclusions, which are the direct cause of lamination 
in iron, since they are elongated along with the iron 
itself as it is being hammered or rolled. Fig. 18 repre- 
sents a longitudinal section. Compare it with fig. 17, 
a transverse section of another piece of iron which 
shows the streaks of slag cut transversely. It also 
shows the slag inclusions in the bodies as well as at 
the junctions of the crystal grains. 

In the soft steel portion of fig. 18 the small black 
masses which are absent in the wrought iron are the 
parts which contain the carbon and are known as 

The chief distinction between wrought iron and very 
soft steel is here depicted. As has been said already, it 
is due to the method of manufacture. The wrought 
iron has been built up of a number of layers from 
between which the whole of the slag has not been 
squeezed out, whereas the soft steel has been cast direct 
from the molten state. 


On examining a piece of mild steel which shows a 
highly crystalline fracture when broken suddenly — it 
may be after years of service — we have always been of 
the opinion that it became crystalline because of the 
vibrations or shock to which it had been subjected. 
Metallography has not only proved this to be a fallacy, 
but has supplied a very simple explanation of this re- 
markable phenomenon. 

When molten steel is cooling down, the pure iron or 
ferrite begins to solidify first and forms into crystals 
which shoot out somewhat in the fdrtn of a fir tree. 
The regular formation is soon arrestel because of their 
coming in contact with each other. This retards their 
growth in a straight line, but new branches continue to 
shoot out ?it right angles to each of the earlier formed 
branches, until all the interstices are gradually filled up. 
Meantime the residue containing the carbon present has 
arrived at its freezing point, which is lower than that of 
pure iron. Consequently it, along with any other im- 
purity present, solidifies last, and in veiy mild steel it 
forms in tiny islets of pearlite surrounded by the iron 
crystal grains. 

Pearlite exists in the steel as alternate layers of car- 
bide of iron and ferrite, and any increase of carbon in 
the steel has a corresponding increase of pearlite and 
a decrease of free ferrite, until in steel containing about 
0'9 per cent, of carbon the entire structure is composed 
of pearlite. The structure of a complete series of carbon 
steels from 0*2 to 0*9 per cent, is shown perfectly in the 
case-hardening photomicrograph, fig. 71 (p. 147). The 
white portions are ferrite and the black pearlite. The 
top part is entirely composed of pearlite, and therefore 
contains about 0*9 per cent, of carbon. The gradual 
diminution of the pearlite clearly indicates a similar 


diminution of the carbon, until in this instance 0*2 per 
cent, is represented at the bottom part. 

When the carbon exceeds 0*9 per cent., carbide of iron 
is in excess, and thin lines of carbide of iron now appear 
white in a mass of black pearlite, as in fig. 61 (p. 129). 

Fig. 19. — Surface of an ingot of antimony showing the 
fir-tree-like formation of crystal. 

The nature of the formation of crystals in metals is 
shown by means of a photograph of the face of an ingot 
of antimony which I had taken for the purpose. On 
the surface the crystalline structure is more perfect than 
it can be in the interior, because of greater freedom of 
growth (see fig. 19). This is still better shown in fig. 20, 
which is a photograph (natural size) of crystallites from 



the cavity of a large steel ingot. Fig. 21 is a sulphur 
print (natural size) of a cross-section through three 

Fig. 20. — Crystallites from a cavity in a large steel ingut. 
Natural size. (Stead. ) 

crystallites from a large steel ingot, proving that the 
skeleton crystals (white) that form first are free from 
sulphur, and showing how the sulphur is trapped between 
the skeleton branches. 


Sulphur prints are obtained by carefully pressing a 
polished section of the metal on bromide of silver paper 

Fig. 21. — Sulphur print of a cross-section of three crystallites 
from a large steel ingot, showing the white skeleton 
crystals with sulphur trapped between the branches. 
Natural size. (Stead.) 

which has been soaked in a 3 per cent, solution of strong 
sulphuric acid in water. The print is then washed and 
placed like an ordinary photograph in sodium hypo- 



sulphite. After again washing it is ready for drying 
and mounting. 

The primary crystallites are also shown in fig. 22, 
which represents a cross-section of a piece of steel 
polished and developed by Dr Stead's new reagent for 

Fig. 22. — Steel rich in phosphorus. A cross-section polished and 
developed by a new reagent for revealing the position taken by 
phosphorus. The dark parts are the primary crystallites free from 
phosphorus. The light parts are rich in phosphorus. (Stead.) 

revealing the position taken up by phosphorus, The 
dark parts are the primary crystallites free from phos- 
phorus; the white portions represent the parts rich in 

In no steel does the crystalline structure show up so 
well as in silicon steel. In fig. 23 is represented the 
junction of several crystal grains, illustrating the variable 
dip or direction of the cleavages of the crystals in con- 


tiguous grains. While the crystallites are all stratified 
in the same plane and at the same angles in a single 
grain, they are in widely different planes and angles in 

Fia. 23.— Several crystal grains in iron showing the variable 
direction of the cleavages of the crystals in contiguous grains. 
The black portion shows the manner in which a fracture 
follows the lines of cleavage. (Stead.) 

the different crystal grains. This photomicrograph also 
very clearly indicates the manner in which a crack 
travels through the crystal grains of a piece of steel, 
that is, along the lines of cleavage. It supplies an 
excellent proof of the interesting fact that when good 


normal steel is broken, the strongest parts lie along 
the junction lines between the crystal grains, and the 
weakest parts lie within the body of the crystal grain, 
the lines of cleavage being really lines of weakness. 

Slip Bands and Alternating Stresses. 

When a piece of steel is strained, the elements of which 
each of the crystals is made up slip over each other. 
These are known as slip lines or slip bands, and were 
first observed by Dr Rosenhain. They may be seen 
under the microscope as fine black lines which represent 
on the surface of the metal the irregular movement of 
slip bands under stress. Three photomicrographs from 
Dr Rosenhain show very clearly the effect of this plastic 
strain. Fig. 24 shows a piece of metal before, and fig. 25 
after, it has been subjected to stress. The slip bands are 
seen very distinctly, especially in fig. 26, which represents 
a portion of fig. 25, much enlarged. 

Now, when a piece of steel is at work it may not be 
subjected to a stress sufficiently strong to cause any 
deformation as a whole, but some of the crystal grains 
may be weaker than the others, or may be placed in such 
a position as to receive undue share of the stress, which 
causes them to yield a little along the cleavage planes. 
If the stress were continuous in that direction it might 
do no harm, but if it be reversed, the now somewhat 
weakened crystal grains yield again in the opposite 
direction; and if this alternating stress be repeated a 
considerable number of times, the surfaces of the cleavage 
faces will become less and less coherent and will event- 
ually begin to separate. The piece of steel will now be 
weaker than ever. On continuing the alternate stressing 
the crack soon develops and travels across the adjacent 


crystal grains because of the extra stress being brought 
to bear on them through the yielding of the others. 

Fig. 24. 

Fig. 25. 

Fig. 26. 

A piece of iron before and after plastic strain. 

Fia. 24. — Represents the piece before straining. 

Fig. 25. — Represents it after straining. , 

Fig. 26. — Is a portion of fig. 25 at a higher magnification. (Rosenhain. ) 

Finally, the piece of steel suddenly snaps, exposing more 
or less polished surfaces in the fracture where the 


primarily separated faces had been rubbing against each 
other. The portion that suddenly breaks is not smooth ; 
it has a crystalline appearance. It should be noted that 
slip bands are what appear on the surface ; there is no 
evidence of them below the surface, hence the change of 
slip bands to cleavage. 

Vibrations or alternating stresses do not cause little 
crystals to grow into big ones, nor fibrous material to 
become crystalline. Steel is never anything else but 
crystalline from the moment that it passes from the 
molten to the solid state. The crystalline appearance of 
the fracture is entirely due to the way the steel is 

The development of a crack in a rod of Swedish iron 
which had been subjected to alternating bending stresses 
while rotating is shown in four very interesting photo- 
micrographs from Sir J. A. Ewing. These represent the 
same crystal grain on a polished and etched portion of 
the rod. Fig. 27 was taken after 1000, fig. 28 after 2000, 
fig. 29 after 10,000, and fig. 30 after 40,000 reversals of 
stress. They show the gradual development of what 
finally becomes an actual fissure. 

There are many cases in which, although there be no 
apparent actual bending, the stresses have been suSicient 
to produce internal slip, gradual separation, and final 
complete rupture. 

Fig. 31 is a photograph of a piece of a broken steam- 
hammer piston rod. The crack started at one edge and 
gradually travelled across, as indicated by the series 
of bright circular portions. The bright appearance is 
caused by the rubbing together of the two sides of each 
succeeding portion of the crack while the piston rod was 
at work. The extent of the bright parts proves that the 
rod must have been in use while only a limited portion 

Fig. 27. Fio. 28. 

^ ■ {^ 

'. - L^"^ 

Fig. 29. 

Fig. 30. 

Specimen of Swedish iron after reversals of stress. 

Fig. 27. — Specimen after 1,000 reversals. 

Fi(i. 28.— Thesame ,, 2,000 

Fig. 29.— „ ,, 10,000 „ 

Fig. 30.— „ „ 40,000 „ (Ewing.) 



was solid, a fact which often indicates — as it does in this 
case — that a broken shaft or other piece of machinery is 
not necessarily bad material because it has been broken, 
but is an excellent proof of its being good. Indeed, 
microscopic examination of the structure of the whole 
area just below the fracture may show little if any 
difference. Failure of this kind may be due to the design 

Fig. 31.- 

-A fracture of a steam-hammer piston rod which 
broke while in use. 

and want of sufficient metal to withstand the stresses to 
which the forging is subjected. The piston rod mentioned 
is a good example. The photograph shows in a remark- 
able manner the step by step progress of the fracture 
before finally passing suddenly through the last portion. 
This, in conjunction with the fact that the end of the 
piston rod which takes the hammer face was too large 
compared with the body of the rod, and the fillet at the 
neck where the fracture took place was too small, clearly 

62 SClkNCk IN THE SMltltY AND FOkC&. 

indicates that the initial cause of the fracture had to be 
looked for in another direction than that of a question 
of quality of material or its thermal treatment. 

Steam-hammer piston rods afford excellent examples 
of alternating stresses, especially when it is noted that 
the blows on a piece of metal are of necessity in many 
instances delivered more often than not by the edge of 
the hammer face instead of by the centre of it. 

The wave-like appearance of fig. 31 is typical of what 
may often be observed in fractures of shafts and even 
in certain classes of reciprocating parts of mechanism. 
It should be noted, however, that if the stresses applied 
are all in one direction and are strong enough, they will 
lead to what is known as fatigue fractures or breaking 
by degrees. 



There are none to whom the study of heat should 
appeal so much as to smiths and forgers. One cannot 
think of a smith without having heat brought very 
prominently before one's mind. Separate heat from the 
smith, and he ceases to exist. Like Othello the Moor, 
" His occupation's gone." Despite all this, it is a most 
surprising fact that in the smithy and forge the correct 
heat treatment of iron and steel is a matter that does 
not receive the careful study which should be given to 
it. It is a subject of the very greatest importance, 
and it is to be regretted that there are many things 
in connection with it of which young smiths and 
forgers are totally ignorant ; and the worst feature of it 
all is, that filmost nothing is being done to induce them 
to take an interest in it. 


It is not generally known among smiths that a 
piece of mild steel may be overheated, even although 
the temperature be lower than that at which it would 
actually "burn"; that if a piece be heated to a high 
temperature such as an ordinary welding heat, or even 



lower, and be left to cool naturally without receiving 
any mechanical work, it is a most remarkable fact that 
the crystals will be much more coarse with prolonged 
cooling than with rapid cooling. In many cases, even 
although work is put on such material, the reduction 
by forging may not be sufficiently great to break down 
the coarse structure. Evidence that finishing at too 
high a temperature is followed by crystal growth on 
cooling is to be found in connection with drop-hammer 
stampings, as the rapidity with which they are produced 
has a tendency to cause work on them to be finished 
at too high a temperature. If the same things were 
finished at a lower temperature, as they are when hand- 
forged, they would have a finer structure. 

Mild steel may with safety be raised to a welding 
heat if the subsequent hammering be continued right 
down till the temperature has fallen to a red heat, about 
700° C. If continued further, it should certainly not be 
lower than a dark red heat, about 550° C, as at lower 
temperatures internal strains are likely to be set up. 

A good example of the evil effect of stopping work at 
too high a temperature is seen in welded pieces that 
break at a point outside the actual welded part. This 
part has been heated to a temperature almost, if not as 
great as, the parts welded, but has received little if any 
work, and its structure has become extremely coarse and 
the metal brittle as a consequence. 


This leads to the consideration of another most 
interesting point in heat treatment. It was demon- 
strated many years ago that a most remarkable change 
takes place in the micro-structure of mild steel — such 


as maj^ be used for welding purposes — when heated to 
870° C. or above. Mild steel which has been heated too 
greatly, or finished at too high or too low a temperature 
or which may have been by any means rendered grossly 
crystalline, may be restored to its softest, toughest, and 
finest condition by reheating to 900° C, keeping it at 
that temperature for a short time, and then withdrawing 
it from the fire or furnace to cool naturally in the 
atmosphere. For mild steel to be used for structural 
purposes, this treatment is true annealing ; this condition 
is the desired object of annealing. Forging of all kinds 
should be subjected to it; it has been experimented 
with, highly approved of, and adopted by many eminent 
practical men. 

To show the practical value of annealing, I prepared 
three pieces of mild steel which were nicked and broken 
after being treated in diff*erent ways. These were 
portions of a piece which had been drawn down and 
intentionally finished at about 1100° C. The first 
piece received no treatment after forging. It broke at 
the second blow, showing a crystalline structure less fine 
than it would have been if work on the forging had 
been continued to a lower temperature. The second and 
third pieces were overheated at above 1300° C. When 
cold the second piece was broken at the first blow, 
revealing a grossly crystalline structure. The third 
piece was annealed for a few minutes at 900° C, and 
when cold it was broken after four blows, showing a 
somewhat fibrous fracture, more particularly in the 
upper portion, which had been in tension and had 
drawn out considerably. This was entirely due to the 
correct heat treatment having increased the ductility of 
the metal. 

An interesting example of annealing is shown in 



fig. 32. This is a piece of dead soft sheet steel, the 
entire surface of which was grossly crystalline. One 

*" 1 


■^HI^^Bt' >^ 


^ ■- '\' --A 



^A '^v'^^JM 





■i ■ "';..-" \H 


Fig. 32. — A i>iece of grossly crystalline soft steel sheet completely 
refined on the side which was reheated for a few minutes to about 
1000' C. (Full size.) 

side was reheated for a short time to about 1000° C, 
while the other side was kept under a red heat. The 
structure has been completely changed from a coarse to 
a fine one, and the important point to notice is that the 


change is not gradual — as might be expected — but is 
quite sharply defined, which fact leads to the definite 
conclusion that no refinement of structure takes place 
until immediately after a particular temperature has 
been arrived at. 

In order to demonstrate this in another practical way 
and at the. same time to prove that photomicrographs 
without doubt reveal the truth, I prepared several pieces, 
the results of which show that what is seen in fig. 33 
is undoubtedly confirmed in figs. 34 and 35. The ex- 
amination of an actual fracture is certainly of value, but 
a much clearer conception of the reasons for a finely 
crystalline steel being tough and thoroughly reliable, 
and a grossly crystalline steel being extremely brittle 
and unreliable, may be obtained by making use of the 

A piece of steel containing about 0*49 per cent, carbon, 
7 inches long, 3| inches wide, and J inch thick, was 
heated to over 1300° C. and buried in ashes. When cold 
it was divided longitudinally into three portions by 
machining two notches on each side, and on suddenlj^ 
breaking off* one of the portions the fracture showed 
extremely coarse throughout (fig. 33a). The remaining 
double portion was then reheated in such a way that 
one end was at a bright yellow, while the other was at a 
black heat, with a gradually diminishing temperature 
between. When cold this was broken along the line of 
the notch, revealing the differential fracture as shown in 
fig. 33b. Photomicrographs were afterwards taken at 
the points marked A and B, fig. 34 being the overheated, 
and fig. 35 the annealed, portion. 

These examples prove three things : that overheating 
causes steel to become coarsely crystalline ; that anneal- 
ing causes such steel to become fine in structure again ; 



and that this restoration can only be obtained when the 
metal is heated to the proper temperature. 



Fig. 33a. Fig. 33b. 

Fracturen in miKl steel. 

Fig. 33a.— Coarse fracture due to overheating. 

Fig. 33b.— Coarse and tine fracture due to re- 
heating the lower half only of tlie initially 
overheated steel to above the critical point. 

The above experiment proves this last statement by 
the fact that a portion of the coarse part near to the 
point A, fig. 33b, was heated to above a red heat, but it 


is quite evident that this was not sufficient to effect any 
change of structure in that portion. 

Fig. 34. 

Fig. 35. 

Photomicrographs from the points A and B in fig. 33b. 

Fig. 34 represents the overheated, and Fig. 35 the annealed, 
portion. Magnified 40 diameters. 

DiflFerent Critical Points. 

It ought to be more widely known that the proper 
temperature for annealing varies with the amount of 


carbon contained in the steel. The critical point at 
which the microstructure becomes refined is not the 
same for all steels. While it is correct to anneal very 
low carbon steel at about 900° C. (a bright red), and 
dead soft steel and wrought iron even higher, this 
temperature is much too high for steel containing, say, 
0*9 per cent, carbon, which should be heated to about 
730° C, a full red heat. 

Dr Stead has quite recently prepared a heat-treatment 
chart (fig. 36) which should prove of great value to all 
practical workers in iron and steel. Personally, since 
it supplies a long-felt want, I have welcomed it most 
heartily. It shows how keenly alive Dr Stead is to the 
important fact that if scientific research is to be of any 
real value, it must have a direct bearing on the problems 
which daily beset the practical worker. 

Smiths and forgers are not familiar with charts of this 
kind, but instead of that being a deterrent, it oujght to 
be an incentive to the study of them. 

In studying a chart of this kind for the first time it 
should be observed at the outset that the bottom 
horizontal line is divided by equidistant vertical lines 
which represent the amount of carbon in different steels ; 
the left-hand vertical line is similarly divided by equi- 
distant horizontal lines which represent the temperature 
in degrees centigrade. 

The whole of the chart is divided in this way by lines 
drawn parallel to the two already mentioned, and any 
point of intersection of these lines indicates a steel having 
a certain amount of carbon at a definite temperature. 

The curved lines on the chart will be better under- 
stood when it is stated that steel which has just solidified 
from the molten state is composed of a solid solution of 
carbide of iron in iron and is thoroughly homogeneous. 


It continues in this condition while cooling until a point 
is reached on the chart which is intersected by the lines 






— " 




( z 
























* 4 










y . 







^ .< 



4 i 



^ 4( 
















■ C 








# < 

















































— 1 
















Ci - 
























. — • 

.— ■ 

^ • 









■ ' 








r * 






r i 










•0 -I -2 -3 -4 -5 -6 -7 -8 -9 10 II 1-2 1-3 H 1-5 1-6 1-7 1-8 1-9 20 
Fio. 36.— Heat treatment chart, by Dr Stead. 

representing temperature and carbon, and by the curved 
line GSE. 


On cooling below this line f errite and carbide of iron 
— as the case may be — separate out ; f errite in steel with 
less than, and carbide of iron in steel with more than, 
about 09 per cent, of carbon. It has already been pointed 
out that the former is now composed of f errite and pearlite, 
and the latter of carbide of iron and pearlite. As the 
temperature falls still further, it arrives at another 
critical change point which is represented by the line 

It is important to note that if steel which is cooling 
be quenched when just above this line it will be hardened, 
but if the temperature has fallen below the line it will 
not be hardened. 

This brings us to the more important part of the 
chart, in so far as it concerns the practical worker, who 
requires to improve and restore the properties of steel 
which has become coarsely crystalline and brittle 
through overheating or some other cause. 

On reheating steel, corresponding changes take place 
in the structure, but at a higher temperature than during 
the process of cooling, and Dr Stead has represented this 
on the chart by dotted lines above the full lines GSE 
and PSK. 

The utility of the chart may now be better demon- 
strated by taking several cases in which steel requires 
to have its fine structure restored. Since mention has 
been made of annealing steel at 900° C, we shall deal 
with an example of this kind. 

Take a steel containing 0-1 per cent, of carbon. Find 
the point on the bottom horizontal line which represents 
that percentage of carbon. Follow the vertical line 
upwards till it intersects the dotted curved line GS. 
Turn at right angles to the left and follow the horizontal 
line till it meets the temperature scale. In this instance 


it is a subdivision between horizontal lines, and the 
required temperature is about 895° C. ; but since most 
commercial steels contain a little phosphorus, which 
hinders the free diffusion of the carbon in the iron, it is 
advisable to heat a little higher. If this steel be kept 
for a short time at a little above 900° C, the whole of 
the carbides will be- thoroughly diffused; the steel will 
be homogeneous, and on cooling in the atmosphere the 
coarse brittle structure will have entirely changed and 
the steel will be tough and reliable. 

The structure of the steel, however, will be composed 
of little islets of pearlite — which contain the carbides — 
embedded in a mass of pure iron ; but if it be desirable, 
it is possible to obtain the steel in a more homogeneous 
condition when cold by quenching in water from a 
temperature above the dotted line GS. This treatment 
retains the steel in a very homogeneous state because of 
there being no time to permit of the separating out of 
the ferrite and the pearlite. The steel will be very 
hard, but this may be rectified by annealing at 500° C. ; 
and if necessary, by reheating to just below the line 
PS, the steel will be obtained very soft as well as 

Ordinary structural steels up to about 0*5 per cent, 
carbon may be quenched in water, but for steels above 
that, up to about 1'2 per cent., it is advisable to quench 
in oil, as the water may cause such steel to crack. The 
tendency to crack is lessened for the simple reason that 
the rate of cooling is retarded because of the fact that 
the heat conductivity of oil is lower than that of water, 
and also that a film of charred oil forms on the surface 
of the metal. The slow cooling reduces the hardness, 
modifies the variation in the change of volume, and 
tends to minimise the consequent strains. 


The opinion is held by some practical men who really 
ought to know better that the beneficial effects of oil 
quenching are due to the oil penetrating the metal. It 
should be evident to all who reflect for a moment that 
if it were possible for this to take place the steel would 
be weakened, or rather it would indicate inherent weak- 
ness in the metal. The toughening effect is not due to 
the transmission of any special virtue from the oil to 
the steel, but to the more or less complete retention of 
the carbides of iron in the diffused state. To attain this 
as nearly as is practicable the quenching must be carried 
out as quickly as possible. Cold water therefore is the 
best quenching medium for this purpose, but for reasons 
already explained it is essential that oil be used instead 
for the higher carbon steels. Although the hardness 
obtained is less than by quenching in water, it is too 
hard for ordinary structural purposes, but by reheating 
to a temperature just below the Ar^ point (the line PS 
on the chart), the homogeneity is retained, while the 
hardness is eliminated to an extent depending on how 
closely this temperature approaches to the line PS. 

The following tests show the effect of different forms 
of heat treatment on steel containing about 0*3 per cent, 
of carbon. Five pieces were forged to 1 inch square ; 
four of these were heated to 850° C, a temperature about 
20° above that indicated on the chart. 

No. 1. As forged. 

No. 2. Quenched from 850° C, reheated to 600° C, 
and allowed to cool in the atmosphere. 

No. 3. Quenched from 850° C, reheated to 680° C, 
and allowed to cool in the atmosphere. 

No. 4. Allowed to cool in the atmosphere from 
850° C. 

No. 5. Allowed to cool in the furnace from 850° C. 


The tensile tests were taken on a length of 3 inches, 
•798 inch diameter. 


Yield Stress. 

Breaking Stress. 



Tons per sq. in. 

Tons per sq. in. 

per cent. 

per cent. 


























Compared with No. 1 the tenacity and ductility of 
No. 2 has been increased. The real object and practical 
value of the oil treatment of structural steel is fully 
demonstrated by comparing Nos. 1 and 3. In No. 3 the 
hardness due to the initial quenching is reduced to such 
an extent that the tenacity is practically identical with 
that of the original forging, while the ductility has been 
considerably increased. The strength of the material is 
retained, with the addition of much greater toughness 
and ability to withstand shock. It is interesting to 
note that whereas No. i5 is much weaker than No. 3, the 
two are almost equal in ductility. No. 5 is even weaker 
and also less ductile than No. 4, which was allowed to 
cool in the atmosphere. 

Some practical men think that the best annealing 
treatment for mild steel is that to which No. 5 was sub- 
jected, but this is not the case ; as a matter of fact*, by 
such treatment the steel is really weakened. This is 
due to the prolonged cooling giving time for the greater 
separating out of the ferrite and pearlite, thereby 
destroying the homogeneity of the material and reduc- 


ing it to a condition entirely opposite to that obtained 
by oil quenching. 

Before passing on it may be of interest to compare the 
above results with a similarly treated series of chrome 
nickel steel containing about the same amount of carbon. 
This steel contained 0*32 per cent, carbon, 3*6 per cent, 
nickel, and 0*7 per cent, chromium. 

Five pieces of this steel were forged down to 1 inch 
square, four of them being reheated to a temperature of 

No. 1a. As forged. 

No. 2a. Quenched from 800° C, reheated to 600° C, 

and allowed to cool in the atmosphere. 
No. 3a. Quenched from 800° C, reheated to 680° C, 

and allowed to cool in the atmosphere. 
No. 4a. Allowed to cool in the atmosphere from 

800° C. 
No. 5a. Allowed to cool in the furnace from 800° C. 
The tensile tests were taken on a length of 3 inches, 
•625 inch diameter. 


Yield Stress. 

Breaking stress. 



Tons per sq. in. 

Tons per sq. in. 


per cent. 


per cent. 
























Nickel chrome steels are highly susceptible to double 
heat treatment and oil quenching; they are of special 
value when it is necessary to have material which has 


greater shock-resisting power than ordinary carbon steel. 
While possessing high elastic limit, they are accompanied 
by the equally vali^able property of high ductility. 
Compared with carbon steels they have the special 
qualities of greater hardness combined with resilience, 
from which they have the power to wear and to resist 
shock in a much higher degree. 

The best results may be obtained from nickel chrome 
steel by heating for an hour at 750° C, quenching in 
oil, reheating to 620° C. and allowing to cool slowly in 
the furnace. The following result is excellent; it was 
obtained from the same steel as the last results with 
which it should be compared. 

Yield stress. 

Breaking Stress. 



Tons per sq. in. 

Tons per sq. in. 

per cent. 

per cent. 





That science in general, and this heat-treatment chart 
(fig. 36, p. 71) in particular, is an invaluable aid in 
practical work could not be better exemplified than by 
the behaviour of certain tools which had to withstand 
a considerable amount of shock. These tools were 
breaking at so alarming a rate that a new process for 
doing a certain class of work had to be abandoned. 
Steel from different makers was afterwards tried, but 
even with their own recommended treatment the results 
were no more satisfactory. 

On looking into the matter I found that the men 
who were hardening the tools were without exception 
taking great care to allow them to cool down very 
slowly on the part where the fracture invariably 
occurred. The steel consequently, although very soft, 
was far from being so tough as it might have been. 
Some of the tools were made of 0*7 per cent., and some 


of 0*85 per dent., carbon steel. Let us take the 0*7 
per cent, steel and see how our chart will help us. As 
with the 0*1 per cent, carbon steel, find the point 
on the bottom horizontal line representing 0*7 per cent, 
carbon. Follow the vertical line upwards till it inter- 
sects the dotted curved line GS. From the point of 
intersection turn at right angles to the left along the 
horizontal line till it meets the temperature at about 
750° C. 

The tools which came under my immediate notice 
were quenched in oil at 770° C. to 780° C, 20° to 30° 
above the critical point ; were then reheated at the weak 
part to 300° C. (blue), and finally quenched in . oil. In 
some instances, especially with the lower carbon steel, 
the working parts were found to be somewhat softer 
than was desirable, but this was easily remedied by 
immersing these parts only in cold water until the 
temperature was reduced to about a visible red, and 
then quickly quenching right out in oil and tempering 
to a blue colour as before. The result has been that a 
broken tool is seldom heard of. Instead of breaking, 
as at first, after performing a few hundred operations, 
and in many cases considerably less than one hundred, 
they now simply wear down till they are too short for 
use, and that after performing many thousands of opera- 
tions. It is of interest to note that these results are 
obtained from tools made of the same so-called bad 
steel as was used at the outset. 

Since this treatment is really a question of relatively 
rapid cooling in order to obtain the structure described, 
modifications of it may be with advantage resorted 
to, such as, quenching in warm water, or in some 
cases boiling water and even boiling oil ; or, when the 
shape of the tool and the situation of the weak part 


demands it, by sprinkling more or less slowly with 
water at a still lower temperature, this of course 
always from the correct temperature, as indicated by 
the chart. 

As in the case of the 0-1 per cent, carbon steel, that of 
0*7 or 085 per cent, may be softened after quenching by 
reheating to 500° C, and, if desired, maximum softness 
with homogeneity may be obtained by reheating to just 
below the line PSK. 

Another practical example of the utility of the chart 
may be given. 

Certain thin steel plates, yV ^"^^ thick, in the form of 
valves working in a special kind of engine, at the rate 
of 600 movements per minute, could not be made to 
withstand the great stresses to which they were con- 
tinuously subjected. When the matter was brought to 
my notice I found that the same mistake was made in 
this case as was made in connection with the tools 
already mentioned. The maker of the valves, without 
any regard for the critical point of the steel, had been 
heating them to a somewhat high temperature and 
allowing them to cool very slowly, with the result that 
it was no uncommon experience to have to replace two 
or three in the course of the same day ; the average life 
of a valve was a matter of a very few hours. As usual, 
the steel maker was blamed ; and while it is true that 
he sometimes does give us a little trouble, it was again 
proved that several of the steels which had previously 
failed were entirely suitable for the purpose when they 
were properly treated. 

The method adopted for each of the diflFerent jsteels 
was simply that of heating to a little above the tempera- 
ture on the chart corresponding to the carbon of the 
particular steel; quenching out in oil; reheating to a 


little below the line PSK, say about 600° C, and 
again quenching in oil. Since very thin articles of this 
kind cool somewhat rapidly in the air, modifications 
of this treatment may be adopted with advantage. 
Instead of quenching in oil, the valves may be allowed 
to cool freely in the air, or they may be cooled by placing 
them between two level metal blocks which may be 
slightly heated in certain cases: this method has the 
additional advantage of keeping such thin pieces level. 
A number of valves were treated according to the above 
methods, with the gratifying result that at the time of 
writing they have been continuously at work for a 
period, not of six hours, but of six months, and are still 
behaving in every way satisfactorily. 

While the above treatment is suitable for steel contain- 
ing up to about 1*2 per cent, carbon, it does not give the 
best results for steel with from 1-2 to 1*5 per cent. The 
only way to obtain the finest and best structure for steel 
within this range of carbon is to forge and then heat- 
treat it. 

We have already seen that in slowly cooled steel with 
more than about 0*9 per cent, carbon, the carbides separate 
out, and in high carbon steel such as this under con- 
sideration (1*2 to 1*5 per cent, carbon) these carbides 
form in m/eshes or envelopes round the crystal grains. 
This is a most brittle condition, and although the 
treatment which is suitable for lower carbon steel 
reduces the size of these envelopes and obtains the steel 
fairly soft and fine, yet the structure is not so good as 
when the steel is forged. 

The correct treatment for such steel as, say, 1*5 per 
cent, carbon, is as f olio ws :— Find 1*5 on the bottom 
horizontal line, as in the other examples. Follow the 
vertical line upwards till it intersects the dotted curved 


line SE. From the point of intersection turn at right 
angles to the left along the horizontal line till it meets 
the temperature scale. The required temperature is 
940° C, and by heating to above this temperature the 
carbides are dissolved and the whole mass is a homo- 
geneous solid solution. By forging the «teel as it cools 
from this point these envelopes of carbide are broken up. 
Care must be taken to continue the hammering, as the 
temperature gradually -tails to a point just below the 
line PSK, which is about a dull red heat. If this be 
done the steel will be of a very fine structure. Reheat 
to 750° C. to 780° C, but not higher than 780° C, and 
quench in either oil or boiling water. The steel will, 
of course, be hard, and the structure will consist of 
independent particles of carbide not connected together, 
embedded in what is called hardenite. To obtain the 
steel soft as well as fine in structure, reheat to just below 
the line PSK. 

To show the critical point as represented by the line 
PSK, and also that steel does not harden until heated 
and quenched from above it, take a piece of steel, say 
6 inches long, \ inch broad, and \ inch thick ; hold it at 
one end in a pair of tongs and heat it carefully, so that 
one end is almost white and the other black, with a 
gradually lessening temperature between. Plunge into 
cold water ; polish one of the surfaces, and with a sharp 
draw-point draw a line along the surface, beginning at 
the end which was at the lower temperature. It will be 
scratched up to a certain point, where the draw-point 
will immediately skid without cutting in the least degree, 
despite the fact that the part just below the critical 
point, — the line of junction between the hard and soft 
portions — which is scratched quite readily, was actually 
visibly red-hot before quenching. 



The critical point may be clearly revealed on the 
surface of the quenched bar by grinding off the skin and 
placing in dilute nitric acid. The hardened portion will 
become black, the unhardened portion lighter. The 
junction between the two is the critical point or the 
point where the hard-pointed instrument begins to skid. 

The heat-treatment chart may be more fully under- 
stood by studying briefly^ihe complementary diagram- 
matic chart (fig. 37), which is really a series of pictures 
of the changes that take place on heating and cooling a 
number of steels containing varying amounts of carbon. 
Each of these was heated at one end to above 1000° C, 
while the other end was kept comparatively cold. The 
parts below the sloping Ar 1-2-3 line represent soft 
annealed steel, while the wholly black parts above 
represent solid solutions of carbide of iron in iron. The 
white areas on the left of 0*9 per cent, carbon are 
ferrite or nearly pure iron; those on the right are 
cementite or carbide of iron. The top of the diagram 
represents the hot ends of the steels ; the bottom repre- 
sents the cold ends. To obtain the best structure in 
any of the steels it is necessary to heat up to the 
temperatures where the white areas completely dis- 
appear ; these temperatures — as has already been pointed 
out — may be found by referring to the other chart 
(fig. 36, p. 71). It is of interest to tool smiths to note 
that what is indicated very clearly is the fact that 
steel containing free cementite requires to be heated to 
a higher temperature than steel containing 0*9 per cent, 
carbon. For example, a tool containing 1'2 per cent, 
carbon ought to be heated to above 850° C. and 
quenched; reheated to above 740° C. and quenched 
again, in order to obtain the best results. For these 
reasons it is absolutely necessary that the carbon of 


.7; 0) o 

? ^ t^'S « 

• -jj; t, ^M « o 

^£ 5 o ^ « 

u J i^ ■ S g g 

i at o o ^ 

= .5-2 n SI 

" ? ^ " 2 ?? 

"2 S b if S „, . 

€ ^i:- 

o [^ ^ ;::3 w 2^ 
o « c^ J^ Q • 

.3 S>i I » S 

5 fl > 3:3 ^ 

J8 « O ^"^ JH 



0.2 ..2^ 


each class of steel should be known to the smith 
who is treating it. 


Practical Example to show the Effect of 
Incomplete Refining. 

Some time ago I had a broken lathe centre placed in 
my hands; the appearance of the fracture is most 
peculiar (fig. 38); it has a very coarse structure, sur- 
rounded by an outer zone which is very much finer : 

Fig. 38. — Broken lathe centre showing refined outer zone, coarse 
centre portion, and blue circular part at right-hand side. 

indeed one could be pardoned for taking it for a case- 
hardened specimen. Although the colour is not shown 
on the photograph, the half -moon portion is decidedly blue. 

This was declared to be a very mysterious fracture, 
and consequently — and this is too often the case — the 
usual subterfuge was resorted to, that it must have 
been bad steel. 

By applying to the problem what we have been con- 
sidering about heat treatment and the conditions ob- 
tained above and below the critical point, I came to the 
following conclusions: that the method of making the 


tool having been to cut it from the bar without any 
subsequent forging, it had, while being heated for that 
purpose, been overheated, which treatment had caused 
it to become very coarse throughout; that after being 
turned, and on heating for hardening, the point of the 
tool had been heated somewhat rapidly and quenched 
before the central portion had been raised to a tempera- 
ture above the critical point, the temperature of the 
outer zone only being thereby raised suflSciently high to 
restore the structure to a fine condition, or, in other 
words, to admit of the carbide and iron of the pearlite 
interdiffusing, and that during this heating and quench- 
ing the sudden expansion and contraction of the coarse 
brittle metal had caused a crack to develop where it 
shows blue, the remaining portion breaking while the 
tool was at work, this being shown by the fact that in 
drawing the temper to the " blue " the walls of the crack 
had taken on the same colour, evidence of air having 
passed into the fissure. 

Now, if it be argued that the foregoing explanation is 
wrong, it means that it would be impossible to obtain by 
working on these lines a similar fracture. One would 
be quite justified in claiming a fracture of a like nature 
as an undoubted proof that these conclusions are correct, 
if one could reproduce it at will. 

For the purpose of supplying this proof, several ex- 
periments were carried out. A piece of tool steel was 
overheated; when cold it was reheated rapidly to its 
critical point on the outside and quenched before the 
heat had penetrated much beyond the surface. This 
piece was not tempered. On breaking it, the centre 
showed a very coarsely crystalline structure, while the 
outer zone was very fine, demonstrating the effect of 
correct heat treatment on that part alone. 



Another piece was treated in the same way, but on 
quenching it split longitudinally, the fracture, however, 
not extending to the ends ; this piece was polished, and 
on close examination there were no signs of fracture 
except the longitudinal one. It was then tempered to 
a blue colour, quenched and broken, revealing the re- 

FiG. 39. — Fracture in cast tool steel. The dark portion at 
the top became blue in colour during tempering, thereby 
revealing the presence of an initial crack. The lower 
portion and the outer edge all round the top finally 
broke out of the solid. 

markable fracture shown in fig. 39. Let it be said here 
that this photograph and that of the broken centre are 
of course not exactly similar in appearance, but the 
characteristics are undoubtedly identical, although their 
magnitude and disposition are somewhat different. In 
fig. 39 the difference between the restored outer zone 
and the brittle interior is clearly indicated. The in- 
teresting feature of this fracture is that when the steel 


split longitudinally it had also, while being quenched in 
the water, cracked from that split outwards, but without 
breaking through the outer shell. The extent of this 
crack has been fully located by the tempering to the 
blue colour, which shows black on the photograph. The 
fact that the forces which burst the steel longitudinally 
and at the same time outwards from the centre were not 
sufficient to break through the thin outer shell, proves 
conclusively how much more tough and strong it must 
have been than the whole of the internal area. 

These fractures have not been presented as curiosities, 
but because they teach several valuable practical lessons, 
chief of which are: that overheating renders steel 
extremely brittle and easily broken ; that proper anneal- 
ing makes the structure of such overheated steel fine 
again and restores its tenacity and ductility; that 
correct annealing is not a mere reheating to an in- 
definite temperature; it is more than that: it is a 
reheating to above the critical temperature of the 
particular steel, throughout the mass and as uni- 
formly as possible ; that although the steel maker does 
occasionally supply us with bad steel, it is always wise 
to examine and study every fracture carefully for the 
purpose of making sure whether or not we ourselves are 
to blame before asserting thit it is bad material, as was 
not the case in the problem we have been considering. 

Rectangular or Stead's Brittleness. 

That the careful study of annealing is of the highest 
importance is clearly demonstrated by the interesting 
fact that after prolonged heating at temperatures 
between 600** C. and 760** C, iron and very soft steel, 
if initially subjected to work when cold or at tempera- 


tures below 500° C, are liable to become coarsely 

This should be of very great interest to those crafts- 
men who, being unaware of the existence of the critical 
points and the fact that steel must be heated to above 
these temperatures to effect complete refining of structure, 
have hitherto deemed it correct practice to reheat all 
steel to an ordinary red heat for a more or less prolonged 
period. It is a fact that many smiths are of the opinion 
that this is the proper method, but it has been proved 
to be a very dangerous practice if the material has been 
finished at too low a temperature in the smithy or forge. 

In the case of initially overstrained or cold- worked 
iron or soft steel, it will be evident from our study of 
heat treatment that this annealing may not serve any 
good purpose, and instead of improving the material, it 
will be evident from what follows that this reheating 
to temperatures below the critical point, about 900° C, 
will be liable to cause the crystal grains to undergo 
considerable growth. This growth of crystal grains and 
a peculiar condition of rectangular brittleness which 
accompanies it was first studied by Dr Stead, and has 
been called "Stead's brittleness" by Professor Howe, 
but Dr Stead himself describes it as "rectangular 

When iron or dead soft steel is rolled into thin sheets, 
the rolling finished at a low temperature, and the sheets 
afterwards annealed at about 750° C, Dr Stead found 
that the material sometimes crystallises with the 
cleavages at angles 45° to the direction of rolling and 
at 90° to the surface. These sheets are readily fractured 
along the three cleavages parallel to the sides of a cube 
— that is, in a direction^at right angles to each other. It 
follows that if the crystals in a piece of iron are so 


oriented that their cleavages are approximately in the 
same planes, the metal will be exceedingly liable to 
fracture if the stresses are applied at right angles to any 
of the three cleavages. 

Photograph fig. 40 freely explains what cleavage 

Fig. 40. — Piece of sheet steel showing rectangular brittleness. 

brittleness means. It represents a piece of sheet steel 
which, when placed over a dished block and struck 
with a hammer, broke up in three directions, one verti- 
cally downwards and the other two at right angles to 
each other. This photograph shows another interest- 
ing peculiarity of this rectangular brittle condition. 


Although the steel fractured very readily along the lines 
of weakness which had developed in the cleavages of the 
crystal grains, yet the steel could be bent and hammered 
close without fracture when the bending stress was 
applied at an angle of 45° to these lines of weakness. 

While it is of interest to note that all of this cold- 
worked or initially strained material became coarsely 
crystalline and brittle because of heating to between 
600° C. and 750° C, it is of special interest to note 
further that the same material on reheating to above 
900° C. was invariably refined in structure and 

A few years ago I was approached by an engineer, 
who asked my opinion about fractures which very 
frequently occurred in iron and soft steel pipes, — which 
are often cold-drawn — and large numbers of which were 
manufactured under his supervision. The fractures 
were of a cast-iron nature, irregular pieces falling out of 
the pipe. Each of these pipes had a flange brazed on 
the end, and on ascertaining that the fracture invariably 
occurred at a point where the pipe was heated to about 
a red heat, I felt thoroughly justified in describing it 
as a clear case of "Stead's brittleness." This opinion 
was fully borne out by the fact that pipes made of the 
same material, but having flanges screwed on instead of 
brazed, gave not the slightest trouble. 

Many practical examples of cold work will recur to 
the practical man, and will emphasise the importance of 
this question, in view of what has already been said, and 
also of more recent research carried out by several 
investigators who have shown that plastic deformation 
is conducive of large crystal growth on annealing at 
low temperatures. 

Professor Sauveur has added to our knowledge of the 


subject by his discovery that if cold iron is distorted 
very much or very little, large crystals do not form on 
annealing at low temperatures ; that there is a critical 
range of deformation which is favourable to development 
of large crystals. If the deformation is more or less 
outside this range, large crystals do not form. Along 
with other experiments he pressed the ball of a Brinell 
testing machine into the surface of small steel blocks 
and then annealed them at a low temperature. The 
examination of a section through the indentation 
showed that while the outer surface and a strained 
portion in the interior were unchanged, a region lying 
between had developed large crystal grains, which fact 
justified him in stating that there is a critical range 
inside of which deformation causes large crystal growth 
at low temperatures. 

Another interesting example which freely illustrates 
the conclusions of Professor Sauveur was discovered by 
Mr H. S. Kipling, one of Dr Stead's students, previous to 
the publication of Sauveur's results. It is shown in 
fig. 41, which represents a soft iron rod bent slightly when 
cold and then case-hardened at a temperature below 
900° C. There is a neutral zone in the middle of the 
bar where the crystals are small and have not grown, 
and where the cold deformation has been outside the 
critical range of strain, or, in other words, was less than 
that required to lead to the growth of large crystals. 
The metal on each side of this has been distorted within 
the critical range and the crystals have grown to a large 
size. On the outside the metal is carburised and the 
crystals are fine. 

The straightening of soft steel plates by stretching 
them until the buckled parts have become quite flat is 
liable to cause the growth of large crystals in the cold- 



strained parts if the plates are subsequently annealed at 
a low temperature. Dr Stead very kindly gave me such 

Fio. 41. — Piece of iron rod showing growth of large crystals due to 
cold-straining and then heating to a low temperature. (Kipling.) 

a piece of plate. There are large grossly crystalline 
portions separated from each other by portions which 
are very finely crystalline. The fine portions are the 


buckled parts, which had not been cold - strained 
sufficiently to cause large growth; the coarse portions 
are the other parts, which had been stretched consider- 
ably and had developed large crystals on annealing at 
too low a temperature. Fig. 32 (p. 66) represents a 
portion cut from one of the grossly crystalline patches 
with one half restored, as described, by reheating to 
1000° C. for a few minutes. 

While studying this most interesting plate it occurred 
to me that the same conditions might exist to some 
extent in chains or other lifting tackle which had been 
annealed — as they very often are — at a low temperature 
after having been initially cold-strained during service. 
To investigate this, a small ring 3 J inches inside diameter 
was made from a |-inch round bar of soft steel. When 
cold, it was strained. until two of the sides straightened 
out and closed together a little, the ring then having 
taken much the same shape as an ordinary link. Four 
pieces were sawn from the straight sides. No. 1 received 
no treatment. After No. 2 had been heated for six hours 
in a small furnace at a temperature of about 700° C, 
No. 3 was placed beside it. An hour later the gas was 
turned off and the two pieces were left to cool down 
with the furnace. No. 4 was heated for a few minutes 
to about 950° C. and then allowed to cool in the atmos- 
phere. On examining the pieces microscopically it was 
observed that the crystal grains in No. 1 were fine and 
fairly uniform in size. On the inner edge of No. 2, 
which is represented by the upper portion of fig. 42, 
very large crystals had grown. A similar growth was ' 
observed in No. 3, although not to the same extent. In 
each piece, the crystal grains beyond the inner edge, 
extending right across to the outer edge, were fine. 
Evidently the deformation of the outer edge had not 


come within the critical range at which cold-straining 
causes the growth of large crystals during annealing at 
low temperatures. In No. 4 no large crystals had grown 
while heating at 950° C. ; as anticipated, the structure of 
this piece was very fine. 

In view of the fact that it is quite a common practice 
to anneal chains and other lifting tackle by heating 
them to an ordinary red heat and allowing them to cool 
out in the furnace, it should be of great interest to note 
that No. 3, after heating for only one hour at this low 
temperature, developed large crystal grains. 

The above experiments are of value in proving that 
the growth of large crystal grains may be produced at 
will. It still remained, however, to find if chains which 
had been in service were in a similar condition. A link 
was taken from a wrought-iron chain which had every 
appearance of having been cold-strained. Pieces were 
treated in the same manner as the pieces of the soft 
steel ring, the results obtained being of a like nature. 

It might be said with truth that the investigation was 
even yet not quite complete, since a chain annealed by 
someone else in the course of ordinary practice had not 
been examined ; the opportunity for doing so was soon 
forthcoming. A large chain suddenly broke while 
carrying a load about 70 per cent, under the ultimate 
breaking load of the chain, and about 30 per cent, under 
the load to which it had been tested some time previous 
to its having been annealed at the usual low tempera- 
ture. A piece examined near to the fracture — without 
any treatment — revealed very large crystal grains. At 
the same magnification, these were much larger than 
those shown in fig. 42. An examination of the entire 
area along the line of fracture was most interesting and 
instructive; several of the large crystal grains were 


much deformed, showing the effect of still further cold- 
straining subsequent to the annealing at a low tempera- 



\^^'p^, 4^' 


^?ir ^ 



Fig. 42. — Portion of soft steel ring which was annealed at a low 
temperature after being cold-strained. The top portion, 
which represents the edge of the ring, shows growth of large 
crystals. Magnified 60 diameters. 

ture. All the characteristics observed in the other cases 
recorded were present in this wrought-iron link; the 
large crystal grains had grown near to the edge in areas 
which had been deformed within the critical range, 


whereas in other portions the structure was almost 
normal. A piece taken from the area in which large 
crystals had grown was annealed for a few minutes at 
950° C, and again these were broken up into very fine 

Now the microscopic examination of the above 
material proved conclusively that under certain condi- 
tions large crystal growth takes place, but it is necessary 
to supply evidence to prove that such material is really 
brittle and unreliable. 

Several pieces of wrought iron, \ inch broad and 
yV inch thick, after being annealed together at 950° C. 
were treated and tested as follows : — 
No. 1. Original bar. 
No. 2. Stretched \ inch on a length of 4 inches 

when cold. 
No. 3. Same as No. 2, but annealed at 700° C. for 

six hours. 
No. 4. Same as No. 2, but annealed at 700° C. for 

four hours. 
No. 5. Same as No. 3, but reheated to 950° C. for a 

few minutes. 
No. 6. Bent cold through an angle of 30° and an- 
nealed at 700° C. for six hours. 
No. 7. Same as No. 6, but reheated to 950° C. for a 

few minutes. 
No. 8. Bent cold through an angle of 90° and an- 
nealed at 700° C. for six hours. 
No. 9. Same as No. 8, but reheated to 950° C. for a 
few minutes. 
These pieces were tested to destruction in the follow- 
ing manner: —Each piece was gripped between the faces 
of a steam hammer and bent by blows from a hand 
hammer to a right angle, as shown by dotted lines in 


fig. 43 ; it was then straightened to dotted lines, as shown 
in fig. 44 These operations were repeated until the 
pieces broke. In order to measure the relative value* of 
the piece, each separate operation of bending or of 
straightening was reckoned as one. 

No. 1 broke at 17; No.2atl5; No. 3at4; No.4at5; 
No. 5 at 15; No. 6 at 10; No. 7 at 16; No. 8 at 8; and 
No. 9 at 15. The appearance of the fractures indicated 

Fio. 43. 

Fig. 44. 

very clearly the effect of the treatment; Nos. 3 and 4 
were grossly crystalline, while all the others were more 
or less fibrous in appearance. 

As the result of this series of investigations I am of 
the opinion that sufficient evidence was obtained to 
justify the following conclusions: — That the broken 
chain link in question had been subjected — as is the case 
more or less with all chains — to strain during service ; 
that the subsequent annealing had been carried out at 
too low a temperature ; that instead of doing good, much 



harm had been done; that the material had become 
brittle owing to the growth of large crystal grains ; that 
in .initially cold-strained material this growth takes 
place at an ordinary red heat, a temperature that is 
commonly thought to be correct ; that the chain ought 
to have been annealed at 950° C. to 1000° C, and finally, 
if it had been inadvertently or ignorantly annealed at 
the low temperature, the fine structure could have been 
restored and the material thoroughly toughened by re- 
heating to a temperature of 950° C. 

In a recent paper to the Iron and Steel Institute, 
Mr C. Chappell deals with the question of large crystal 
growth. One point which may prove of great practical 
value is that he demonstrates the fact that even in 
material strained while hot, large growth had developed. 
This is most interesting in view of the fact that it may 
offer an explanation of many cases of failure while 
working hot material. For example, many years ago I 
observed the need for the exercising of great care while 
straightening pieces of iron which were repeatedly 
bending in course of upsetting (staving up) uiider the 
steam hammer. When failure occurred in such material 
it was invariably while at a red heat, and on the part 
which had been repeatedly straightened. 

Effect of Work at the Blue Temperature. 

It is not very well known that to put work on mild 
steel at about 300° C. to 400° C. — the blue temperature, 
as it is called — is a very dangerous practice. At this 
temperature mild steel appears to be in a very critical 
condition, and if worked at that temperature is liable to 
become exceedingly brittle. It may, however, be sub- 
jected to more work when cold than at a blue heat with- 


out making it brittle. From the fact that mild steel may 
be straightened and even bent cold, arises the erroneous 
idea in the minds of young craftsmen that this may the 
more easily be performed the more highly the material 
is heated. A few tests will supply sufficient proof. 

Several pieces of mild steel, 6 inches long, f inch 
broad, and \ inch thick, were treated and then tested to 
destruction in the manner shown in figs. 43 and 44 and 
described on p. 96, each operation of bending and of 
straightening being again reckoned as one. 

A piece bent and straightened when cold broke at 29. 
Another piece at the blue temperature broke at 7. The 
practical value of a knowledge of this being wrong treat- 
ment was better exemplified by a third piece, which, after 
being bent to a right angle at the blue temperature, was 
allowed to cool, and when cold the operations were con- 
tinued until it broke at 18. This clearly proved that 
the brittleness was produced at the blue temperature, 
and that it remained in the steel after cooling. A fourth 
piece bent to a right angle at the blue temperature was 
allowed to cool, and was then annealed for a short time 
at 1000^ C. When cold it was operated on, breaking at 
32, affording an excellent practical example of the need 
for annealing material which has been subjected to work 
at this dangerous temperature. 

To prove still further that the metal is at this tempera- 
ture in its least plastic condition, several tensile tests 
were made. These were made on a length of 6J inches 
and '798 inch diameter. The local blue ones were heated 
on a part about 2 J inches long ; the remaining parts were 
just a little above the atmospheric temperature. These 
tests show thoroughly the unyielding nature of the 
material at this temperature, and it is worthy of note 
that while being pulled, and more particularly in the 


case of the "all blue" one, the testing machine kept 
jumping in a most remarkable manner, accompanied by 
a series of sharp cracking noises. 

Tons per 



sq. 111. 

per cent. 

per cent. 

Original bar 




Broke at 

Local blue at 




Broke at 

one end 

57-5 at 
broken end, 

cold end. 


at two 

Local blue at 



13-9 at 



other end, 
6-8 at 


of them. 

All blue . . 




Broke at 

A piece of cold angle bar, 3 inches long, was placed 
with the edges of the two flanges touching the face of 
a steam hammer. It was flattened right out until the 
root of the angle was equal in thickness to the two 
flanges ; there was not the least appearance of fracture. 
An attempt was made to flatten a second piece while 
at the blue temperature, but on receiving the second 
blow it broke in two pieces right through the root of 
the angle. 

Many instances of the effect of this treatment in actual 
practice might be mentioned, but it will be sufficient in 
the meantime to take notice of one only. A boiler-smith 
had flanged a plate and laid it down on the floor. In 
passing shortly afterwards, he observed that a part of 


the edge of the flange required to be set in a little ; he 
thereupon struck the edge of the plate one blow with 
a sledge-hammer, with the result that a semicircular 
portion broke right out. An examination of the fracture 
showed that it had become quite blue in colour, proving 
beyond doubt that the plate was at the temperature at 
which work rendered the metal extremely brittle. 

Distortion of Steel when Cold Worked. 

The hardness and elastic limit of steel which has been 
severely crushed are always increased, while its ductility 
is decreased. The tenacity of such material decreases 
in the direction in which the crushing stress is applied, 
and at the same time increases at right angles to 
that direction. I have heard the opinion expressed 
that even the most severe cold working will not injure 
mild steel, but this is not the case. The tenacity of 
severely cold-drawn steel has been jreduced to such an 
extent transversely that it has been possible to make it 
behave like a piece of cane by splitting it up into long 

The cold working of steel is often the cause of failure, 
and should always be regarded as liable to render the 
material more or less treacherous. Although it is true 
that some steels are less susceptible to cold working 
than others, it should always be avoided as far as 

A well-known example of the effect of crushing is the 
case of steel plates which have been cut by shears, and 
are always to some extent rendered brittle and liable to 
fail during service. Pieces of steel plate which have 
been shorn and then bent double very often crack right 
across, whereas pieces of the same material have bent 


double without any signs of cracking when the shorn 
edges have previously been machined. 

The practice of machining the shorn edges of plates 
is a good one and should not be discontinued in spite of 
the fact that test pieces not machined do sometimes bend 
double without fracture. It may be that the injury — 
the presence of which is beyond dispute — is not sufficient 
at the moment to cause trouble, but at some future date 

Fig. 45.— Surface of a forced drill hole. (Stead. ) 

during service it may develop in such a manner as to 
cause failure. 

The metal surrounding a punched hole is always 
brittle, and much less ductile than the remainder of the 
plate. Punched holes should always be made small and 
be drilled out afterwards to the finished dimensions; 
but even then we may not be out of the wood, as it 
has been proved that a hluni drill causes distortion of 
the metal, which may subsequently develop into cracks. 
Fig. 45 is a photograph of a forced drill hole showing 
that the blunt drill has crushed the walls of the bole. 


Bad drilling is probably the cause sometimes of cracking 
at the rivet holes in boiler plates. The side of a drill 
hole in a boiler plate that cracked during use is shown 
in fig. 46; the surface showed indication of crushing 
—note the multitude of fine cracks that developed 
while the plate was in use. Fig. 47 represents a section 
of a crushed hole in a boiler plate that cracked at this 
point when being tested — note the fracture due to 

Fig. 46. — Cracks on the surface of a drill hole in a boiler plate which 
cracked while in use. (Stead. ) 

crushing. This specimen was polished, but not etched ; 
the dark lines are cracks. 

The further development of distortion by crushing 
may depend on the nature of the conditions under 
which the material serves its purpose, and in certain 
cases it may be subjected to very great stresses. For 
example, one can quite readily conceive, in connection 
with boiler plates especially, if there be any incipient 
weakness of the above kind, that the alternate expansion 
and contraction and the repeated stresses and vibrations 
may gradually develop strains and eventually cause 


An interesting example of the effect of cold work is 
shown in fig. 48, which represents a vertical section 
through the head of a smith's set-hammer showing con- 
tortions produced by the sledge-hammer during use. 
It indicates very clearly why portions fly off, and also 

Fig. 47. — Cracks in a boiler plate which gave way when being tested 
under water pressure. This section was cut close to the point 
where the plate failed at a crushed and distorted rivet nole. 

shows the necessity for occasionally annealing the heads 
of such tools. The steel was originally forged at a 
dull red heat which left the ferrite and pearlite in 
straight lines. 

The effect of cold crushing may be demonstrated in 
a most interesting manner. Make a bend test of a 
piece of square steel, hammering the metal quite close ; 


on attempting to open out the test piece again, it will 
immediately break right through the end. On examin- 
ing the fracture it will be observed that the steel has 
been crushed; the crystals appear to flow and form 
plates at the inside, while the extended part at the out- 
side of the bend is granular. The platey arrangement of 
the crystals makes the steel exceedingly weak ; hence on 
trying to bend back again fracture always follows. 

Fig. 48. — Section through the crashed head of a smith's 
set-hammer. (Stead. ) 

showing a surface that appears very bright when viewed 
under vertical rays of light. It indicates that cold 
crushing must never be practised if the material is to 
receive stresses in the reverse direction to the line of 
stress. The effect of bending and closing pieces of steel 
in this way is so severe that in several experiments 
evidence was obtained to prove that at the inside edge 
of the bend, where there is maximum crushing, fracture 
may actually take place during the process of crushing, 
before attempting to open out again. This was 
demonstrated in the following manner: — a piece after 


being closed was slowly heated to the blue temperature ; 
when cold it was opened out, revealing a blue portion, 
in some cases \ inch deep, indicating thereby the 
presence and the extent of the initial fracture. The 
bending of a piece of steel in this fashion is no doubt 
an extreme case of crushing, but it is of great value in 
showing that a practice which has the possibility of 
being so drastic in its results as to cause rupture in the 
course of performing the operation, may, though carried 
out to a less extent, still be sufficient to weaken the 
material considerably and eventually lead to disaster. 



Is chemistrj^ — or metallurgy, which is in large measure 
really a branch of applied chemistry — of any value to 
young smiths ? Most assuredly it is. Why, the smithy 
and forge are veritable chemical laboratories. Not 
one, but many, practical examples could be profitably 
dealt with. 

Young smiths should be taught the chemistry of 
welding, which would shed light on a very important 
subject, and at once settle a matter about which there is 
much unnecessary controversy among practical smiths. 
This would cease to exist if the matter were thoroughly 

.The structure of iron and steel is made up of tiny 
crystals, and the result of iron welding is the formation 
of new crystals across the line of contact of the two 
pieces welded together. Dr Stead has informed me in a 
private communication that " there can be no disputing 
the law that a skilful microscopist can determine with 
absolute certainty, by examination of a section cut 
through a welded joint, whether the weld is perfect or 
not, for in a perfect weld there is no visible joint, and 
the original line or plane of junction is occupied by 
crystals, portions of which belong to one piece of 



metal and portions of the same crystals to the other." 
" When the boundary of the crystal is coincident with 
the juxtaposed plane surfaces it is evidence of non- 
welding, which is equivalent to saying that unless the 
crystals become common to the two pieces there is no 

Several conditions must be fulfilled to obtain perfect 
welding. The pieces must be heated to a temperature 
high enough to ensure the most rapid crystallisation or 
welding of the adjacent surfaces, but well below that at 
which the metal would " burn." The faces of the pieces 
to be welded must be perfectly free from oxide of iron 
or other impurities in order to have absolutely clean 
metallic surfaces in actual contact. As a help towards 
the attainment of these conditions, some suitable sub- 
stance is used as a flux. There is much diversity of 
opinion among smiths in connection with fluxes, and, 
since welding is one of the most important operations 
that a smith is called on to perform, it is worthy of 
being considered a little to find out if science can dispel 
some of the mistaken notions that have gathered round 
it. To that end it will be necessary first to take notice 
of certain most interesting phenomena. 

Oxidation and the Burning of Steel. 

Young smiths should be taught that iron has a great 
affinity for the oxygen in the air, and when iron is 
heated an interesting chemical change takes place. A 
fixed quantity of oxygen combines chemically with a 
fixed quantity of iron to form on the outside of the iron 
or steel an entirely new substance which has none of the 
characteristics of either of the elements of which it is 
made up. Smiths are familiar with it as smithy scale ; 


chemically it is known as oxide of iron. It is common 
knowledge that it will not combine with the iron in 
welding, and if present betweien the welded faces in any 
quantity it causes an unsound weld — indeed, it is one of 
the chief causes of lamination in wrought iron. This 
cannot be avoided, even although it were possible to heat 
the pieces in a non-oxidising atmosphere, because in 
passing from the lire or furnace to the steam hammer or 
the anvil, contact with the atmosphere will bring about 
the same result, more especially if there be any delay in 
bringing the pieces together. 

The higher the temperature to which the metal is 
heated, the more rapidly does oxidation take place ; if 
the heating be continued long enough after the welding 
heat has been reached, the whole of the metal may be 
oxidised — converted into oxide of iron, or "burned away " 
as smiths say. An interesting proof of this is the fact 
that if a jet of cold air be impinged on a piece of iron 
or steel at a welding heat, the oxygen of the air will 
combine so rapidly with the iron that the temperature 
will be increased considerably, just as cold air blown on 
a piece of red-hot carbon (charcoal) causes it to glow. 
A heavy bar may be completely burned through; the 
metal as it combines with the oxygen running down as 
oxide in a molten state and falling away from the bar. 

The phenomena of burning present a most interesting 
study to smiths. A smith while engaged in heating a 
piece of iron or steel in the fire has his mind continually 
reverting to the need for care in order to avoid burning 
the metal ; he is under the influence of a natural impulse 
which keeps him ever on the alert and of which he is 
almost unconscious. Being, therefore, of peculiar interest 
to smiths, it will be profitable to investigate what burn- 
ing really is, and to consider means which are helpful in 


preventing it, since it is very liable to occur during 
heating for welding. 

Smiths who are in the habit of welding steel are 
familiar with tiny cracks which occasionally appear on 
welded bars, more particularly on the corners of flat bars. 
This is direct evidence of burning. The cracks are caused 
by thin films of molten material forming between the 
crystals, and while the metal is being forged the crystals 
separate and oxygen passes into the cracks. 

When steel is heated up to a very high temperature it 
reaches a zone in which that part of the metal "which 
wjis the last to solidify on cooling is the first to melt 
on reheating. This is the point of incipient fusion, 
when the carbon and phosphorus, with a portion of the 
iron, liquefy into globules and eventually coalesce, and 
finally pass to the crystal joints. It is not difficult for 
one to conceive that if the metal be hammered or strained 
in any way while in this condition, separation will take 
place through these meshes of molten metal between 
and around the crystal grains ; and if these cracks are 
connected to the surface, oxygen will penetrate, to form 
films of oxide of iron which will prevent the crystals 
from welding together during the subsequent forging. 
One can just as readily conceive that the internal crystal 
grains which do not connect to the surface will also 
separate if strained, but that it will be impossible for 
air to find its way between them, and if the steel be 
forged at a welding temperature these internal cracks 
will close up again and the interior of the bar become 

Several photomicrographs (figs. 49, 50, 51, and 52) 
illustrate this very clearly. These were all from the 
same burnt angle, which was shattered when attempting 
to bend it at the high temperature ; the piece was free 






S ^ 


; i II 


flS o ,2 "^ "^ 
" ^ 3 V V 

PQ flS P 09 P4 

3 P o • 

S S "^ ^ 
J .2 .2 « 

p J s s 

I U I 

« O r-4 C4 

-^ la ko ko 
d d o* d 



from excrescences, the surface quite smooth, but there 
were many cracks. Fig. 49 is the surface of the bar, 
showing no cracks in the unstrained portion as at A, 
and cracks in strained portions as at B. Fig. 50, which 
is a section through the unstrained part of fig. 49 at A, 
shows the outer surface which had not been strained 
after leaving the fire; the white ferrite bands are the 
places where the liquid had formed ; but as cracks had 
not been made, no oxide could form, and this piece is 
quite free from intergranular oxidation. Fig. 51 is a 
section through the strained part at B ; it shows a 
crack at the top which connects to the surface, and 
which is more or less filled with oxide. The zigzag 
crack at the bottom is an internal one which did not 
continue to the surface, and consequently no air could 
get in. When fractured the sides of the crack were 
found to be brightly metallic, and there was no trace 
of oxide. Fig. 51 is polished only, but fig. 62, which is 
a photomicrograph of the same section etched, shows 
very distinctly that the cracks have taken place between 
the crystal grains. 

Figs; 53 and 54 represent the same burnt steel at a 
place where there was a slight concentration of phos- 
phorus. Fig. 53 was etched by alcoholic acid solution, 
whereas fig. 54, which is exactly the same area, was 
etched by Dr Stead's new reagent for revealing only 
the portions rich in phosphorus, which remain white 
after the attack. It is certain that the white parts 
were actually liquid, due to excessive heating. 

Certain conclusions have hitherto been assumed by 
scientists in connection with the burning of steel, but, 
in view of the new light that Dr Stead has brought to 
bear on the subject as the result of his more recent 
research, these conjectures must now be discarded. 



It was formerly assumed that on heating a piece of 
steel to above the point of incipient fusion, oxygen 
penetrated to the centre and caused intergranular 

Fio. 53. Fig. 54. 

The same burnt steel as fig. 49 at a part where there was a slight 

concentration of phosphorus. 

Fio, 53. — Etched by alcoholic acid solution. 

Fio. 54. — The same area etched by new reagent for revealing the presence 

of phosphorus, which is concentrated in the white portions. Magnified 

60 diameters. (Stead. ) 

oxidation throughout the entire mass, a state in which 
the steel is very brittle and practically worthless. It 
is now known that if the steel be heated to the same 
high temperature without the presence of oxygen, such 



as may be done in molten slag, the metal will be 
rendered just as tender and brittle as before, although 
there is no evidence of any intergranular oxidation — 
indeed, the method of heating precludes any possibility 
of it. 

The conditions are of course diiferent when heating 
in a smithy fire or furnace, but even in cases of this 
kind Dr Stead has obtained evidence which enables 
him to state as a fact that it is .only in the outer 
envelopes that intergranular oxidation can take place 
when heating to well above the point of incipient fusion. 
It is only on the surface when the metal becomes liquid 
and froths up or boils that the oxygen can penetrate 
to form layers of. oxide around the crystal grains ; there 
is no more dissolved oxide contained in the interior of 
the steel than there was before it was burned. 

I have repeatedly proved this by heating bars in a 
furnace and allowing them to soak thoroughly at a 
temperature sufficiently high to cause the outside to 
become so liquid that the metal frothed and boiled up, 
and on sawing a cross section of the bars the interior 
metal showed the same appearance as steel which had 
been heated in molten slag. It was only after prolong- 
ing the heating until not only the outside metal but 
also the succeeding layers had become so thoroughly 
liquid, causing the bar to collapse gradually in the form 
of a shapeless mass, that oxygen had penetrated to the 
centre of the bar. As a matter of fact, the spongy 
nature of the interior is often not due to oxidation at 
all, but to the internal evolution of occluded gases and 
liquation of the carbides, which simply run out and 
leave gaseous cavities, that do not connect to the 
surface and therefore cannot contain any oxygen — unless 
these holes or cracks are near to the surface, and the 



surface metal in boiling up in bubbles allows the oxygen 
to pass in and the liquid metal closes over it. 

Occluded gases are those which have been absorbed 
by the molten metal during the process of manufacture 
and are enclosed in the solid ingot. 

Liquation means melting, and the term is applied to 

Fio. 55. — Froth on outside of burnt steel. The dark parts are 
cavities partly filled with oxide. (Stead. ) 

the melting of the more easily fusible parts of a metal, 
such as the carbides of iron in steel. 

In fig. 55 there is shown froth on the outside of a burnt 
angle bar ; the dark parts are cavities partly tilled with 
oxide. The steel had evidently been melted on the 

A piece of mild steel was burnt to such an extent that 
it frothed up all over the surface ; the outside metal was 


reduced by grinding until the excrescences and gaseous 
cavities were almost although not entirely removed. 
What remained is shown in fig. 56, which is a portion 



< ■!-_ 


; * * 






1 , 1 

• ■" **\'' 

4 V-* 



Fig. 56. — Surface of burnt mild steel showing gas-holes and fissures con- 
taining oxide. Polished but not etched. Magnified 80 diameters. 

polished, but not etched. It shows gas-holes, and also 
fissures containing oxide of iron. 

One may now quite readily understand why it is that 
high carbon steel is more liable to burn and consequently 
more difficult to weld than low carbon steel or iron, for 


the initial point of incipient fusion steadily falls as the 
carbon rises, therefore the crystals of high carbon steels 
become surrounded by liquid at a lower temperature 
than steels containing less carbon. As would be expected, 
and as actually is the case in practice, high carbon steel 
is far more tender and brittle at a low temperature than 
low carbon steel at a high temperature. 

It is now quite evident that in heating steel for the 
purpose of welding or forging, it must never be heated 
to the point of incipient fusion. Since it is probable 
that this high temperature may be reached while heat- 
ing for the purpose of welding, it will be advisable to 
look for some practical means of keeping the temperature 
just below the point at which internal liquation or fusion 
commences. Let me lead up to this, however, by ex- 
plaining a series of experiments which were carried out 
by Dr Stead in connection with the elimination of blow- 
holes in steel ingots. These prove conclusively that not 
only may steel of any proportion of carbon be perfectly 
welded, but the operation may be performed at a 
temperature very much lower than that with which 
smiths are familiar. This, of course, was accomplished 
under conditions where oxygen was entirely absent and 
purely metallic surfaces were in actual contact. Quite 
a number of these tests were made in duplicate with 
steel ranging from O'l per cent, to 1*4 per cent, carbon. 
We shall consider one or two of them. 

Two pieces of steel were polished, and one placed on 
the top of the other. They were placed in a porcelain 
tube and heated in an atmosphere of hydrogen gas for 
two hours at a temperature of 900° C. ; when cool they 
could not be separated and had to be burst asunder with 
a hammer and chisel. They had been welded together 
at different parts where there was actual contact, an<J 


pieces which had been held together with gentle pressure 
were perfectly welded throughout. 

A hole 7 inches deep, J inch diameter, was drilled 
down the middle of a piece of square steel (0*9 per cent, 
carbon) 8 inches long and 2J inches thick, leaving a 
solid bottom 1 inch thick. A plug of the same material 
was turned a driving fit, 1\ inch long. A few drops 
of petrol displaced the air in the hole, and the plug was 
then driven in tight. The steel was heated in a fire 
to 800'' C. and gently flattened under a 12-cwt. steam 
hammer. It was then reheated to a yellow heat — 1100° 
C. — and forged into a f -inch octagon bar, which was cut 
up and made into chisels. These were put to practical 
use and did their work admirably, giving no indication 
whatever of unsoundness. 

It was proved that on heating pieces of steel with 
their plane surfaces in contact to temperatures ranging 
between 750° C. to 950° C. for half an hour only, there 
was little or no welding, but that welding was effected 
at a temperature of 800° C. when the period of heating 
was extended to two hours. These experiments show 
the remarkable results that may be obtained when the 
conditions essential to perfect welding are fulfilled. 


As has now been demonstrated, two of the essential 
conditions for successful welding are, clean metallic 
surfaces and a temperature that, while being sufficiently 
high, would be lower than that at which the metal would 
begin to melt or " burn." 

The question now arises — Is there anything that will 
help to obtain clean metallic surfaces and also lessen the 
possibility of burning the metal ? Chemistry supplies 


the answer most readily: sand and borax, etc., which 
young smiths are accustomed to see used by the older 
men while welding iron and steel, are excellent aids 
towards the attainment of the desired condition. 

The infusible sand combines with the oxide at a 
temperature much lower, than that required for welding. 
They form a chemical compound called ferrous silicate, 
which is readily fusible, and therefore flows off very 
easily. This is a point that should be made quite 
clear to young smiths. There frequently exists a mis- 
apprehension as to what a flux and its true function 
really is. The word itself is derived from the Latin 
word "fluere," which means — to flow. In particular, 
when applied in metallurgy, it is the term given to 
the substances which are used to make fusible flowing 
mixtures with substances of an infusible character. 

Most smiths and forgers think that a flux is that 
greasy-looking substance which flows all over the 
surface of the metal after the application of sand. 
The sand itself is the flux, and the fusible compound 
which it forms with the oxide of iron is not — as is also 
thought to be the case — in any sense of the nature of 
a sticky substance which helps to bind the metal in 
welding. It is as detrimental between the welded faces 
as the oxides with which it combines, and, being only a 
means to an end, it must be got rid of after it has served 
its purpose. The washing of one's dirty hands suggests 
the analogy of the effect of the soap on that which it 
is intended to remove. Any soap, if left, would be 
just as objectionable as any other foreign matter, the 
removal of which it facilitates. 

The flux is simply a cleansing and protecting agent, 
and its chief advantage in these respects lies in the 
important fact that it permits the operation of welding 


to be conducted at a much lower temperature than if 
it were not present. 

It has been argued that iron and soft steel require 
no flux, and that satisfactory welds have been obtained 
without it. That may be so, but it is often open to 
very serious question. It is true that iron and dead 
soft steel may be heated above the temperature at 
which the oxides melt, but this is precisely how the 
danger is incurred, because any observant smith can 
testify to the difficulty with which the impurities on 
the surface of his " weld " are sometimes removed, and 
when he does not actually burn the metal he approaches 
dangerously near to it. 

The use of a suitable flux considerably lessens the 
risk of reaching the danger zone, and it ought to be 
regarded as indispensable at all times and with all kinds 
of steel. The sand or borax cleanses the metal from 
all impurities which may be present on the surface and 
also protects it from being burnt. This is effected by 
reason of the fact that in combining with them the 
oxides are melted and more readily removed at a much 
lower temperature than they otherwise would be, and 
therefore clean metallic surfaces are secured at a tem- 
perature which, while preventing the possibility of 
burning the metal, would still be sufficiently high to 
enable perfect welding to take place. The metal is 
protected, not only from actual burning, but also from 
loss due to further oxidation, since the ferrous silicate, 
while flowing all over the surface, is in reality a varnish 
which prevents the oxygen of the air from attacking 
the metal. 

Sand is not suitable for use in the welding of high 
carbon steel, for the simple reason that it does not com- 
bine with the oxide of iron at a low enough temperature. 


From the fact that borax does so at a much lower tem- 
perature than sand, it will be quite obvious that this 
is the reason for its use in welding high carbon steel. 

In making welds the scarf should be of such a shape 
as to easily permit of the discharge of the ferrous silicate 
during the subsequent hammering. Although it may 
not be always practicable, an excellent conception of 
the ideal shape of the scarf for welding may be formed 
by placing the balls of one's two thumbs together. 

This raises another point of contention among smiths ; 
there is a diversity of opinion as to the part upon which 
the sand or borax should be put. Now, while it is the 
whole of the " weld " that requires to be protected from 
oxidation, is it not more particularly the face that 
requires to be cleansed and kept clean in order to have 
perfect metallic surfaces ? It is undoubtedly. Therefore 
it is beyond dispute, both from a scientific and a practical 
point of view, that it is upon the face of the scarf that 
the flux ought to be put. When anyone uses as an 
argument against this, the fact that some broken 
" welds " reveal sand in their fracture, it only strengthens 
the argument in favour of the value of a knowledge of 
the chemistry of the process. Without this knowledge, 
the proper use and importance of fluxes for welding iron 
and steel cannot be fully understood nor appreciated, 
neither can the operation be properly performed. They 
must be used judiciously with due regard to the chemical 
action which takes place; it is not sound reason to 
condemn the use of sand or borax on the faces of pieces 
to be welded simply because some smiths, in their 
ignorance of its proper application, throw it on in large 
quantities and consequently ruin their " welds." 

Some time ago I was considerably surprised, in one 
sense, although not iu another, to se^ it stated very 


emphatically in an educational text-book that a flux 
should not be used in the welding of angle bars, the 
reason given being that the flux in many cases only 
eats away the metal. Now, it is quite evident that the 
writer of the text-book had received his information 
from a practical smith, and it is just as evident that the 
practical smith had no knowledge of the chemistry of 
welding. What he states is perfectly true when lack of 
scientific knowledge causes improper use of the sand as 
a flux ; trouble of this kind is not due to the use of sand, 
unless it be used in excess, but is entirely a question of 
too much sand. An explanation of what really takes 
place should be of considerable interest and value. 

To make the matter quite clear to practical smiths, it 
will be better to lay aside chemical terms for the time 
being, and to state simply that a definite quantity of 
sand combines with a definite quantity of oxide of 
iron to form the fusible compound — ferrous silicate — 
described. If more sand be applied than the oxide of 
iron is capable of combining with, there will be present 
a mixture of sand and the fusible compound, a mixture 
which gradually becomes thickened as the sand is in- 
creased, and consequently may not be so readily removed 
from the surface of the metal. In this somewhat thick 
pasty condition the mixture of sand and the fusible 
compound, instead of flowing, begins to boil up in the 
form of small bubbles ; that part of the surface which 
is covered by the mixture is protected, but the air 
contained within the bubbles oxidises or eats into the 
metal, causing the formation of small cavities or hollows 
on the surface. This pitted appearance of the metal is 
therefore not due to the use of sand, but to the abuse 
of it. 

It is impossible to b^at a piece of steel without ^ 


certain amount of loss due to oxidation, but it is 
interesting to note that in where sand is not used, 
the extent of the loss is likely to be much greater than 
it is in cases of pitting caused by excessive use of sand. 
It is the irregular surface of the metal that leads to the 
wrong assumption that it is only in these latter cases 
that loss of metal is incurred. When a flux is not used, 
the relatively larger loss of metal escapes notice because 
of its being more uniformly distributed over the whole 
surface. There can be no doubt whatever that sand, 
when applied in accordance with an intelligent apprecia- 
tion of th6 chemical change which takes place, instead 
of causing oxidation of the steel, helps in a very large 
measure to prevent it. As has already been stated in 
these pages, no man can hope to perform any operation 
successfully unless he has studied the fundamental laws 
which govern it. The smith who has studied these laws 
has, when heating mild structural steel of any shape and 
for any class of work, no difficulty in obtaining sufficient 
fusible ferrous silicate without any superfluous sand, 
which enables him to secure the ideal conditions already 
indicated as absolutely essential for perfect welding. 

Restoration of Burnt Steel. 

One of the most startling features of Dr Stead's re- 
search on the burning of steel is that regarding which 
he affirms, and gives evidence to prove, that steel which 
has been heated to well above the point of incipient 
fusion without being oxidised or blistered on the surface, 
and which has not been strained at that high tempera- 
ture, may have its former good properties restored by 
reheating to a suitable temperature for a short time. 
The very high temperature of heating brings the steel 


back to the original ingot structure, very coarse, and the 
pearlite and ferrite in the cold metal are not intimately 
distributed, but are in large separate portions. The re- 
heating causes the carbide and ferrite to interdiffuse and 
coincidentally causes the metal to recrystallise into small 
crystals. After proper reheating and cooling in air the 
whole mass has a fine structure equal to what it had 
before heating to above the point of incipient fusion. 

It was my very great pleasure and privilege, at the 
request of Dr Stead, to carry out certain experiments 
which added to this conclusive evidence. Three classes 
of steel — low, medium, and tool steel — were burnt and 
allowed to cool in the air. A number of pieces 2J inches 
diameter, several of which were subsequently forged, 
were reheated to just above 900° C. for half an hour ; 
pieces of the same steel were treated in exactly the same 
way, excepting that they were not burnt. Comparative 
tensile and bend tests were made, each class of steel 
being treated in the following manner : — 

1. Original bar. 

2. Burnt. 

8. Burnt and reheated to 900° C. 

4. Burnt, reheated to 900"* C, and forged. 

5. Original bar forged. 

Several photomicrogi-aphs taken from the medium 
steel, containing 0*42 per cent, carbon, show very clearly 
the effect of the treatment. The original bar as received 
from the steel maker is shown in fig. 57. Fig. 58 re- 
presents the adjacent part of the same bar after burning ; 
the large polygonal structure and the white specks 
clearly indicate burning. Fig. 59 is the same as fig. 58 
after restoring by heating to 900° C. The structure 
shown indicates that burning and reheating not only 
improve the steel, but restore it to a condition even 










10 ni 







superior to the original, as was also proved by bend 
and tensile test pieces. Pieces of the same material, 
burnt and unbumt, after being reheated to 900° C, 
forged down to 1 inch square, and then annealed, revealed 
similar structures and gave similar results when tested. 

The remarkable results of the tensile and bend tests 
enabled Dr Stead to arrive at the following conclusions, 
which are best stated in his own words : — 

" 1st — That very soft steel, after burning and cool- 
ing, is not deteriorated to more than a 
slight extent. 
" 2nd — That steel with 0*4 per cent, carbon, after 
burning, has its ductility greatly reduced. 
On reheating the burnt steel to 900° C its 
good properties are completely restored. 
On reheating and forging it is made more 
ductile than it was before burning, and 
but slightly lower in tenacity than the 
original steel after reheating. 
" 3rd — That tool steel containing above 1 per cent, 
carbon is made exceedingly brittle by 
burning and cooling in air. Reheating to 
900° C. greatly improves the burnt steel, 
but it is less ductile than the unburnt 
material. Reheating and forging to smaller 
section and annealing at 900° C. more than 
completely restores the original strength 
and ductility, as was proved by two sets of 
carefully conducted trials." 
When carrying out this work I was amazed at the 
results ; as a practical smith I should never on my own 
account have dreamt of attempting what I formerly 
regarded as being an utter impossibility. Indeed, I do 
not wonder that practical men and scientists, since 

TiiE Chemistry of weIdinG, 12; 

Dr Stead published his results, have expressed them- 
selves as being somewhat sceptical of the pieces having 
been actually burnt. For this reason I welcome this 
opportunity of testifjdng to the truth regarding this 
most remarkable phenomena of the restoration of burnt 
steel. Although it is almost inconceivable, yet it is 
an undoubted fact, and may easily be proved to be so ; 
one need only make a few trials in order to prove it to 
one's own satisfaction. 

It is well known to smiths that if a piece of tool steel 
be heated above a certain temperature and struck with 
a hammer it will break up completely, and also that if 
the end of the bar be allowed to protrude beyond the 
high temperature zone of the fire the mere act of with- 
drawing it when heated to this high temperature, will 
be sufficient to cause it to break in two, even although 
the greatest care be exercised. Steel in this condition 
is what is technically called " burnt," and if it be possible 
to obtain as good results from such steel as from the 
same steel unburnt, then no one can doubt that it is 
possible to restore it. 

Having heard it suggested that I may have been 
mistaken in thinking that the steel was actually burnt 
and that the pieces had only been grossly overheated, I 
felt it incumbent on me to disprove this, and an effort 
was made to supply some more convincing evidence. If 
it had only been a question of restoring grossly over- 
heated steel I could not, as a practical smith, have been 
so profoundly impressed with the results. What Dr Stead 
hsts already affirmed and proved may be fully understood 
by describing the nature of these further experiments 
and the manner of carrying them out. 

Several pieces of tool steel, 2J inches diameter, were 
placed in a furnace, a short period of time elapsing 


between the introduction of each piece. These were 
carefully observed for the purpose of trying to secure a 
piece which had been heated as nearly as possible to the 
point of destruction. The first piece was allowed to 
melt completely and become absolutely worthless; by 
carefully noting its behaviour while heating, the remain- 
ing pieces were secured in the desired condition. These 
were almost liquid on the surface, which frothed up all 
over, and on touching them with a rod the granules of 
iron were quite easily displaced. To prevent the possi- 
bility of straining, great care was taken while remov- 
ing them from the furnace, and when cold they were 
reheated to 900° C. for half an hour. The outer oxidised 
envelopes were all carefully cut off before forging down 
to 1 inch square. Pieces were cut off to make tensile 
tests and were annealed along with similar pieces forged 
to the same size from the original unburnt material. 
Pieces of burnt and unburnt material were also forged 
down to 6 inches long, f inch broad, and \ inch thick, 
and were bent cold. In order that the bend tests should 
be subjected to the same pressure and bent at the same 
speed, they were bent side by side in a hydraulic press 
until they broke. It is an interesting fact that the 
restored burnt pieces in both methods of testing gave 
as good results as, and in some cases better results than, 
pieces of the same steel forged to the same sizes without 
having been previously burnt. 

An excellent proof of the fact that the steel had been 
heated to above the point of incipient fusion was supplied 
by the evidence of heterogeneity, which was revealed 
while machining the burnt pieces, and which had entirely 
disappeared after reheating. Local portions of the burnt 
tool steel, after slowly cooling in ashes, were so hard 
that it was impossible to machine them. While attempt- 



ing to machine strips from the burnt steel, the tool 
followed the line of least resistance and was diverted in 
some instances very considerably from the straight line; 
this effect was obviously due to the presence of segregated 
masses rich in carbide of iron. A photomicrograph of 


9' "- ' 

Fig. 60. Fig. 61. 

Piece of tool steel, carbon 1 '2 per cent., after and before burning. 

Fig. 60.— After burning, showing patches of segregated carbide and phos- 
phide of iron located at the junctions and in the bodies of the crystals. 

Fig. 61. — The original structure before burning. Magnified 50 diameters. 

the steel by Dr Stead confirms this, and clearly shows 
the patches of segregated carbide and phosphide of iron 
at the junctions and in the bodies of the crystals (fig. 60). 
Compare this with fig. 61, which represents the original 
steel before burning. It contained 1*2 per cent, carbon. 
The dark portions are pearlite and the white portions 



are carbide of iron. In the original steel the carbide 
is intimately distributed throughout the whole mass, 
whereas in the burnt steel it is segregated in large 

Another piece of the same steel, 18 inches long, 2 J 
inches diameter, was placed in the furnace on a carefully 
prepared flat bottom. After soaking slowly, it was 
heated to so high a temperature that on attempting to 
withdraw it without subjecting it to any lateral strain, 
it broke in two pieces, the portion at the back of the 
furnace not being moved in the slightest degree from its 
position. This remaining piece also broke in two in the 
same way. Meanwhile the last remaining portion had 
become thoroughly liquid on the top, and at the risk of 
subjecting it to strain it was removed on a shovel. All 
of these pieces were treated in the same way as the 
others, the result being entirely in keeping with the 
previous tests. Two of the tensile tests will serve to 
indicate what these were like, and are of special interest, 
since the burnt test was taken from the last remaining 
portion of the piece, which repeatedly broke in the 

These tensile tests were taken on a length of 2 mches, 
'625 inch diameter. 

per sq. in. 

per cent. 

per cent. 

Original bar 
Restored burnt bar 

58-9 ' 



What is most remarkable is that a bend test, 6 inches 
long, f inch broad, and \ inch thick, forged from the 


same portion, did not break after bending to an inside 
radius of f inch with the points actually touching, a 
result which was not obtained with any of quite a 
number of unburnt pieces of the same dimensions. 
Moreover, midway between the bend and one end of 
the piece there was observed before bending a crack 
\ inch deep on the edge, which was evidently due to the 
outer oxidised envelope not having been wholly removed. 
Despite the fact that the bending was carried out in 
such a way that the pressure was applied on the extreme 
points, this crack did not develop. It will be readily 
accepted that the steel was actually burnt, and from the 
facts stated it will be as readily agreed that such burnt 
steel can be, and has been, restored, and in some instances 
actually improved. The carbon in the burnt steel was 
slightly lower than it was in the original bar, and this 
probably accounts for the lower tenacity and higher 
ductility of the restored burnt steel. 

Seeing that steel makers would be likely to maintain 
that although these remarkable mechanical tests had 
been obtained, yet good tools could not be made from 
such steel, it was arranged that this point should be 
investigated. Tools of various kinds were made and 
tested against tools made from the same steel unburnt. 
These were handed out in pairs without any information 
being supplied, but simply with a request to carefully 
note their behaviour, and report, with the interesting 
result that convincing evidence was obtained to prove 
that the steel had its properties restored. Indeed, it is 
of peculiar interest to note that in each of the different 
methods of testing, in the tensile tests, the bend tests, 
and tests with tools in actual practice, there were several 
outstanding superior results, and these were obtained 
from restored burnt steel. 



A study of the metallurgy of iron and steel would be 
useful to young smiths and forgers in making them 
familiar with the processes of the manufacture of the 
metal that they are working with. Among other things 
it would prevent the possibility of their being in the 
position of not a few who argue that a piece of a steel 
casting cannot be drawn down or welded. Why not? 
In a certain respect it is the same material as mild steel. 
Mild steel in the form of an ingot is also a steel casting, 
and is simply cast in a different shape of mould. A few 
tests put the matter beyond any doubt. Four pieces 
were sawn side by side from a steel casting : No. 1 was 
tested without any treatment ; No. 2 was forged down 
from 2 J inches broad and 1 J inch thick to 1 inch square ; 
No. ^ represents two pieces of the same dimensions as 
No. 2 welded together and forged down to 1 inch square. 

The pieces were tested on a length of 6^ inches, -798 
inch diameter. 

No. 1. Original casting . 
No. 2. Forged piece 
No. 3. Welded piece 

per sq. in. 



per cent. 


per cent. 


These tests are also interesting examples of the 
effect of work upon the material. The steel casting 
was very coarsely crystalline, and the effect of working 
was to alter the structure completely, it becoming very 
finely crystalline, which caused it to be considerably 
increased in tenacity and ductility. It is of special 


interest to note that the weld test gave better results 
in every respect than the solid drawn test, which is a 
most satisfactory proof that the material may be welded 

Strength of Welds. 

Mention has already been made of the fact that the 
point at which broken welded parts give way is very 
often situated outside the actual weld. This may be due 
to the weakening of these parts through overheating, or 
the strengthening of the actual weld by means of proper 
and sufficient hammering continued while the tempera- 
ture is falling from that of welding to about a red heat. 

Some time ago, while engaged in making certain 
welding experiments, I observed a most interesting 
microstructure, which I think points to another reason 
for the actual weld being sometimes stronger than the 
adjacent parts. 

Two pieces of mild steel, 5 inches long. If inch 
diameter, were slightly rounded on the face at one end. 
These were then heated to the welding temperature 
in a smith's fire and placed together under a steam 
hammer to form a butt weld. Hammering was con- 
tinued on the end until the temperature had fallen 
below that of welding. The weld at this stage received 
no other mechanical or thermal treatment, and, when 
cold, small pieces were sawn from the junction of the 
weld, which, when polished and etched, revealed the 
microstructure shown in fig. 62. The black portion 
is pearlite, which contains — as was previously stated — 
the carbon ; the white portion is f errite or iron. It 
should be remembered — since it has already been dis- 
cussed — that in steel containing more than 0*9 per cent. 


carbon, the white portions (unless when etched with 
certain reagents other than those described on p. 44) 
represent free carbide of iron, as in figs. 60 and 61 ; and 

Fig. 62. 

Fig. 63. 

Butt weld in mild steel before and after annealing, showing excessive 
carburisation in the region of the weld. Magnified 40 diameters. 

in steel containing less than 0*9 per cent, carbon, the 
white portions represent ferrite or iron, as in fig. 62. 
It will be observed that along the line of junction of 
the weld there is a band less rich in carbon than the 
area on each side of it ; but the remarkable feature of 



the specimen is the fact that these areas on each side 
of this band contain an amount of carbon considerably 
in excess of that which was present in the steel originally. 
The steel contains, as is seen in the photomicrograph, 
about 0*15 per cent, carbon, whereas on one side of the 
junction of the welded parts there is present about 0*5 
per cent., and on the other side about 0*9 per cent. This 
must have been taken up by the steel while heating, 
and is not difficult to understand when it is remembered 
that the fire may have been in such a condition that 
instead of free oxygen there was carbon monoxide in 
excess, which, along with the solid carbon in intimate 
contact with the steel, would cause it to take up carbon 
very rapidly at the high temperature to which it was 
subjected. It is simply a form of carburisation, as 
carbon monoxide is one of the best case-hardening 
agents. The photomicrograph agrees with this com- 
pletely, as well as with what follows. 

The welded bar was afterwards reheated to a welding 
heat and drawn down to f inch square. After anneal- 
ing it was turned down to make a tensile test piece, 
which was compared with a piece of the same bar drawn 
to the same dimensions, and also with a piece of the 
original bar not drawn. 

The tests were taken on a length of 6J inches, -798 
inch diameter. 

per sq. in. 

per cent 

per cent. 

Original bar 
Welded „ . 





Pieces from the same material welded in the same 
fashion bent double without fracture. 

The tensile test of the welded bar broke outside the 

-r-,.- ^ 

f • 


Fig. 64. Butt weld in iron. Fig. 65. ^ 

Fig. 64. — Shows carburisation at the weld, polished and etched. 

Fig. 65. — Shows a perfect weld, polished only. Magnified 40 diameters. 

weld after elongating on each side of it, leaving a heavy 
portion between, proving conclusively that the actual 
weld was stronger, though less ductile, than the adjacent 
parts by reason of its containing a greater amount of 
carbon. The nature of the deformation of the test 



pieces is just what one would expect from steel of 
varying carbon contents. 

The part of the bar adjacent to that shown in fig. 62 was 
reheated to 900° C. for a few minutes, and the effect of 
such correct heat treatment is shown in fig. 63. The struc- 
ture is completely changed from coarse to fine, and the 
special features of the specimen are very clearly revealed. 

Two pieces of Yorkshire iron containing only a trace 
of carbon were welded in the same manner. Fig. 64 
proves that a considerable amount of carbon had been 
taken up at the juncture of the weld. Fig. 65 represents 
the same part before etching, clearly indicating a perfect 
weld, and that the flux to a great extent has eliminated 
the impurities in the immediate region of the weld. 
Tensile and bend tests were made, as in the case of the 
steel welds, the tensile tests being taken on a length of 
6J inches, 798 inch diameter. 

per sq. in. 

per cent. 

per cent. 

Drawn bar . 

Welded „ . . . 




In the tensile test, the welded iron bar behaved in the 
same manner as the steel weld, breaking outside the 
actual weld after elongating on each side and leaving 
an enlarged portion between. Bend tests of similarly 
welded Yorkshire iron bent close without the least sign 
of fracture. 

These structures, as shown in the photomicrographs, 
may not always be obtained in welding, but may be 
due to the existing working conditions of the fire ; they 
would seem to point to the fact that there is a field in 
this direction for much experimental work which would 


be of great value in adding to our knowledge of this 
most important part of smith work. At some future 
date I hope to be able to make known the results of 
research — on which I am at present engaged — on the 
welding of iron and steel. 

Before leaving the subject, however, since it is not 
usual to make butt welds in this particular way, it may 
be of interest to explain one of the reasons for having 
carried out the above experiments. 

Some time ago, in a paper by Dr Stanton and Mr 
Pannell, published in the Proceedings of the Institution 
of Civil Engineers, reference was made to the inte|*est- 
ing results obtained from certain welds which had been 
made in Austria. Pieces of steel bars, If inch square, 
were welded and subsequently reheated and forged 
down to f inch square, a method which, if it were 
always practicable, I should approve of, since the metal 
has the great advantage of being well wrought after 
having been heated to a high temperature. If, however, 
welds of this kind are made for the purpose of comparing 
with welds made according to the usual practice, or 
with solid bars, then the method is open to very serious 
question, because a simple calculation will at once show 
that the two ends of the welded portion must have been 
inside the grips of the testing machine, and in that case 
the true relative value could not possibly be arrived at. 
The result of these tests showed that the welded bars 
were almost as good as the solid bars used, a result 
which was undoubtedly largely due to the abnormal 
length of the scarf. 

By means of my experimental butt welds I sought to 
prove that when the art of welding is properly under- 
stood it is quite possible to make good welds even when 
the length of the scarf is at a minimum ; and this was 



best obtained when, by butting the pieces end to end, 
there was strictly speaking no scarf at all, unless in the 
sense that after the piece was drawn down the junction 
line of the weld had in every case altered slightly from its 
original position at right angles to the length of the bar. 

It will be readily conceded by every practical smith 
that the operation of forging down these butt-welded 
pieces transversely to the line of weld was in itself a 
very severe test; if perfect welding had not been 
secured, it is highly probable that the fact would have 
been revealed during the process of forging. Since, 
however, these butt - welded pieces were evidently 
perfectly welded, they would also be considerably im- 
proved by such mechanical treatment, instead of being 
adversely affected thereby. 

It should be noted that the amount of forging down 
was less than in the case of the Austrian welds, which 
were forged from If inch square to f inch square. The 
butt welds were forged from If inch round to f inch 

That the butt welds were good, that they compared 
very favourably with those made in Austria, and that 
in several instances they were better than the original 
bar, is shown by the following results, obtained from test 
pieces 8 inches long, -798 inch diameter : — 




per sq. in. 

per cent. 

per cent 

Forged steel bar . 















Forged iron bar . 




Welded „ 






Without a knowledge of chemistry and metallography 
it is impossible to understand fully and control properly 
the process of case-hardening. Case-hardening is simply 
a modification of the cementation process of making 
steel. Iron and steel take up carbon very readily when 
heated to a sufficiently high temperature in the presence 
of some carbonaceous substance. 

The usual method of procedure is to pack carefully 
the articles to be case-hardened in a cast-iron box 
amongst wood charcoal, bone charcoal, leather, or some 
such substance. There is always present in the box or 
muffle furnace in which the articles are heated a certain 
amount of air, with the oxygen of which the carburising 
agent combines to form carbon monoxide. This, acting 
on the iron, parts with a portion of its carbon to form 
carbide of iron, while the oxygen released combines 
with another part of the carbon monoxide to form 
carbonic acid. The carbonic acid leaves the metal, meets 
the carbon of the charcoal, and is converted again to 
carbon monoxide. The process is repeated, the carbon 
being absorbed by the outer part of the metal, from 
which the inner parts receive it by diffusion, until, if 
continued sufficiently long, the metal will become highly 
charged with carbon, even as high as 1*7 per cent. 

In case-hardening the process is stopped when the 



carbon has penetrated to a sufficient distance. The 
object of this is to obtain an article which will resist 
shock and wear, having a soft, tough core surrounded 
by an outer zone of steel containing sufficient carbon 
to give a very hard surface after the metal has been 
heated and quenched. 

For the purpose of showing that carbon diffuses from 
the outer zone towards the centre I had two small discs 
of steel turned to \ inch diameter, f inch thick, having 
a |-inch hole drilled through the centre of each. Two 
small plugs of soft steel were turned a driving fit and 
driven into the holes. These were then heated in a 
furnace to 950° C, one being withdrawn immediately and 
the other remaining for eight hours at about the same 
temperature. Fig. 66 shows no change in one of them, 
whereas in the other the carbon has diffused from the 
outer ring of cast steel into the core of soft steel, much 
in the same way as in case-hardening (fig. 67). 

In ideal case-hardened material the carbon merges 
gradually from high in the outer zone to low in the 
central core. There should be no line of demarcation 
between the high and low carbon zones ; no greater mis- 
take can be made than to suppose that it is good practice 
to be able to define exactly, in a fractured piece of case- 
hardened work, the depth of carburisation obtained. 

The phenomenon of peeling or blistering, which is a 
most objectionable feature of some case-hardened work, 
is entirely due to irregular carbon penetration, and it 
is a most remarkable fact that it may be caused by con- 
ducting the process not only at too high, but at too 
low, a temperature. 

Except in certain special cases, it is inadvisable to 
have more than 0*9 per cent, carbon in the outer zone. 
As we have already seen, this is the point where the 


steel is wholly composed of pearlite, and above which 
is obtained free carbide of iron, which forms meshes 

Fig. 66. Fig. 67. 

Part of tool-steel washer with mild steel plug. 

Fig. 66.— Heated to 950" C. for 8 minutes. 

Fig. 67. — Heated to 960** C, for 8 hours, showing diflfusion of carbon from 
the washer to the plug. Magnified 40 diameters. 

round the crystals and is very hard and brittle. High 
temperature, in causing very rapid carburisation, will 


very soon cause the iron to reach this highcarburisation, 
and the higher this becomes the more plainly visible in the 
structure of the material is the bad effect of the high 
temperature which produced it. The outer zone becomes 
extremely coarse and brittle, more especially if it be 
allowed to cool slowly, for in so doing the carbide of 
iron will segregate into large meshes and the decrease 
of carbon will be irregular round the big crystals. This 
segregation may be prevented by quenching from the 
carburising temperature ; but since 09 per cent, of carbon 
gives adequate hardness, it is unnecessary to produce 
conditions which require special precautions. 

In avoiding the known dangers of overheating during 
carburisation, there is the risk of falling into the greater 
danger of underheating, greater because of its being 
seldom realised as a danger. 

We have seen that above the top critical temperature 
the structure of iron is completely changed and that 
this temperature differs with the proportion of carbon 
of different steels. Below that temperature, the critical 
point called Acg, steel takes up carbon very slowly, 
above it very rapidly, so that in carburising at a low 
temperature the process goes on very slowly at first, 
until the outer zone has absorbed an amount of carbon 
which has for its Ac critical point the temperature at 
which the process is conducted. The outer zone then 
absorbs the carbon very rapidly, while the inner zone 
absorbs it very slowly. Consequently, there is a most 
sudden change from high carbon on the outer part to 
low carbon on the part below, and it is this sudden 
variation that accounts for the phenomenon of peeling. 

To prove the truth of this I carried out the following 
experiments, which confirm the valuable research of 
Dr Giolitti. 


Pieces of soft steel 1\ inch diameter were carburised 
during a period of eighteen hours in wood charcoal at 

Fig. 68. 

Fig. 69. 

Mild steel case-hardened. 

Fig. 68.— Carburised at 960" C. during 18 hours. 

Fig. 69.— Carburised at 840° C. during 24 hours. Magnified 30 diameters. 

a temperature of about 950° C. Pieces of the same 
material were carburised for twenty-four hours at about 
840° C. The microstructures obtained by these treat- 
ments are shown in figs. 68 and 69, and by comparing 



them it will be seen that even with six hours less time 
the penetration of carbon is almost double in the piece 
carburised at 950° C. The outer zone is entirely com- 

FiG. 70.— Mild steel carburised at 800" C. during 24 hours. 
Note the crack along the line where the carbon changes 
abruptly. Magnified 30 diameters. 

posed of pearlite, and the decrease towards the central 
core is gradual. 

In fig. 70 is shown a piece of mild steel carburised 
during a period of twenty-four hours at a temperature 



of about 800° C. This piece was blistered all over, and 
the photomicrograph indicates quite distinctly that it 
is along the line where the sudden difference in the 
carbon exists that the crack has taken place. 

It is not difficult to understand why cracks take 
place between layers of steel which have different 
amounts of carbon, such as in fig. 70. There are in 
intimate contact two different steels which have widely 
different properties of various kinds ; the outer zone is 
really a tool steel, whereas the adjacent inner zone is 
a medium hard structural steel which does not harden 
to the same extent when quenched, and is very much 
more tough and able to adjust itself to the mechanical 
forces set up during quenching. At the commencement 
of cooling the outer zone quickly becomes hard, brittle, 
and rigid. Immediately thereafter the underlying 
portion changes its volume without any appreciable 
harm to itself, but the outer shell cohering to it is 
unable to follow, and, being thin, — in fig. 70 it is not 
much more than ^ inch thick — it is unable to with- 
stand the internal tension, and consequently cracks and 
peels off. 

It should be easy to understand that the tendency 
in this direction is very considerably lessened when a 
piece such as that represented by fig. 71 is being dealt 
with, which gives a very clear conception of what an 
ideal piece of carburised steel is like. It represents a 
carburised outer zone of about \ inch in depth; the 
depth of carburisation is marked in inches on one side 
of the photomicrograph, while the gradually diminish- 
ing percentage of carbon is marked on the other side. 

It should now be quite clear that the best temperature 
at which to carburise is about 950"* C. ; it should never 
be lower than 900** C. The temperature at which the 



process is to be carried out having been fixed, the 
obtaining of the wholly pearlitic zone at the outer edge 
depends on the time and the size of the articles; the 


































Fig. 71. — Piece of carburised mild steel showing complete series of steels 
from 0'9 per cent, to 0*2 per cent, carbon. Magnified 30 diameters. 

time required again depends on the carbonaceous 
material used. The above experiments were made with 
wood charcoal alone, which, although one of the slowest 
mediums, gives very good results. If any other agent 


be preferred, nothing could be easier than to conduct 
a few experiments at the temperature indicated in order 
to find out the time required to give the best results. 
Pieces should then be examined microscopically to 
show the structure obtained; for case-hardened work 
microscopic examination is preferable to chemical 
analysis, if it were for no other reason than the great 
difference in the time taken to secure the desired 
information. After the specimen has been cut from 
the material it is quite possible to polish and etch it 
and examine the structure in less than a quarter of an 
hour. Moreover, the analyses of layers can only give 
the mean composition of each, and do not in some 
instances afford sufficiently accurate information, since 
it may happen that through the centre of a layer there 
may be such an abrupt change in the concentration of 
the carbon as is shown in fig. 70, whereas the micro- 
scope, by presenting a picture of the specimen intact, 
reveals the exact arrangement of the carbon, the treat- 
ment to which the steel has been subjected, and, in 
many cases, the cause of failure when such has taken 
place. In many cases it is necessary to take advantage 
of chemical analysis in conjunction with microscopic 

Assuming, then, that the articles have been properly 
carburised and that there is no free carbide of iron in 
the outer zone, several miethods of conducting the further 
necessary treatment may be recommended ; but before 
doing so attention must be drawn to the fact that at 
this stage case-hardening presents a most interesting 
problem in connection with the critical change points 
of steel containing different amounts of carbon, and that 
on the proper interpretation of this depends the success- 
ful completion of the process. 


It will be understood that the prolonged heating at 
950° C. will have produced a very coarse structure 
throughout, which must be rendered fine again, so that 
the articles may retain the property of maximum 
tenacity and ductility ; but a glance again at fig. 71 will 
remind us that we have to contend with a piece of metal 
having an outer zone gradually varying from about 0'9 
per cent, to about 0*2 per cent, carbon, and in some cases 
less, which means that the critical point AC3 of all the 
different strata is not the same. Ideal heat treatment 
demands therefore, not one, but several reheatings of the 
metal, because by reheating only to 750° C. the structure 
will become fine only up to the part which contains 
about 0*7 per cent, carbon, and in pieces where free 
carbide of iron is in excess the outer zone will not become 
fine. On the other hand, reheating to 900° C, while 
making the whole of the central core fine in structure, 
leaves the outside in a less fine condition than it might 
be. It is quite evident that some compromise must be 
made. After having obtained a clear understanding of 
all the laws which govern the process, careful experi- 
ments may be carried out for the purpose of obtaining 
data which will serve to indicate the proper modifications 
of the following methods, which may be adopted to suit 
the requirements of the various classes of work : — 

1st. Carburise at a temperature of about 950° C. 
After allowing to cool, reheat to about 800° 
C. and quench. 

This is suitable for work of a nature which does not 
demand great ductility and is not to be subjected to any 
severe shocks. The central core will not be made fine, 
and the outside portion will be rendered somewhat 
coarser than in more correct treatment, while the region 


containing about 0*4 per cent, carbon will become fine in 

2nd. Carburise at a temperature of about 950° C. 
After allowing to cool, reheat to about 900° 
C. and allow to cool again. Reheat to about 
750° C. and quench. 

Fig. 72 shows the effect of this treatment without 
quenching. Compare with fig. 73, which is the same 
piece as carburised, and without any subsequent heat 
treatment. The central core of fig. 72 is quite fine, as is 
also the outer zone, while the region between is not 
rendered coarse to any appreciable extent. 

These two photomicrographs demonstrate very clearly 
the advantage of microscopic examination in problems of 
this kind. They indicate admirably, in a manner not 
obtained by any other means, the history of the thermal 
treatment to which the material was subjected. 

3rd. Carburise at a temperature of about 950° C. 
After allowing to cool, reheat to about 900° 
C. and quench. Reheat to about 750° C. and 

Quenching after each reheating retains the metal in 
a more homogeneous condition. If the shape of the 
articles presents no risk of warping and the work is of an 
important nature, the best results will be secured by this 
quenching after each reheating ; indeed, for the purpose 
of obviating the possibility of segregation, especially if 
there be any free carbide of iron in the outer zone, 
it is advisable when possible to quench direct from 
the carburising temperature in addition to the other 

The advantage of quenching after each reheating 



should be thoroughly appreciated by reverting to the 
two heat-treatment charts, figs. 36 and 37 (pp. 71 and 

Fig. 72. Fig. 73. 

Carburised mild steel before and after annealing. 

Fig. 72.— After annealing. 

Fig. 73. — Before annealing. Magnified 30 diameters. 

83), where it was pointed out that when steel is heated 
to above the critical point it becomes thoroughly homo- 
geneous, and by hastening the cooling by quenching no 
time is given to allow the constituents to separate out 


again, and the structure is retained in its finest pos- 
sible condition. 

Fig. 74. Fio. 75. 

Centre of carburised mild steel. 

Fio. 74.— Coarse structure due to prolonged heating. 
Fio. 75. — The same part after annealing for a few minutes. Magnified 
80 diameters. 

Except at the final quenching for hardening it is not 
necessary, indeed, it is not advisable, to allow the cooling 
to proceed beyond the point at which redness disappears. 
The metal at that temperature having passed below all 


the critical points, no further change can take place in 
the structure, and therefore no further advantage can be 
gained by continuing the cooling ; but by interrupting it 
and immediately reheating, the risk of cracks forming is 
considerably lessened, since there is a probability of these 
occurring during this reheating if the preceding quenching 
has been carried right out from the high temperature. 

As has been said in connection with high-carbon steel, 
these first quenchings from the high temperature for the 
purpose of retaining a fine structure in the central core 
may be with advantage carried out in oil or hot water, 
cooling mediums which are less drastic in their action 
than cold water. 

It should be noted that for reheating purposes it is 
always inadvisable to use an open fire or a furnace from 
which the air has not been entirely excluded. The 
oxygen which enters in the air combines with the 
carbon in the steel, decarburising the outer skin and 
thereby causing partial and sometimes complete softness 
on the surface. 

Fig. 74 represents a central core of the same piece of 
steel as is shown in fig. 71. The effect of the prolonged 
heating during carburisation is shown in fig. 74; the 
same piece is shown in fig. 75 after having been reheated 
to about 900° C. for a few minutes: there is a well- 
marked difference. These illustrations supply indisput- 
able evidence of the effect of the heat treatment necessary 
to retain a soft, tough core which, as we know, is one of 
the fundamental objects of case-hardening. 



In the practical examples which I have dealt with, 
I have endeavoured to show the advantage of having 
some knowledge of arithmetic, mensuration, strength 
of materials, geometrical and mechanical drawing, 
practical mechanics, properties of iron and steel, heat, 
chemistry, metallurgy, and metallography. I feel sure 
that it will be readily admitted that a little knowledge 
of all these subjects should prove beneficial to young 
smiths and forgers. It has been said, " A little know- 
ledge is a dangerous thing." So it is, but not in the 
sense in which I use the expression, and at any rate it is 
not half so dangerous as ignorance. The only time that 
it can do harm is when a man thinks and acts as if he 
knew everything when he only knows a very little. 
When I speak of a little knowledge I do not mean an 
imperfect knowledge of the subject, but a thorough 
understanding of the little in each subject that may be 
practically applied to smith work and forging. Of 
course, we shall always have with us the man who 
affects to despise these things, but he, more often than 
not, is the very man who stands most in need of them. 

The spirit in which these pages has been written is 
that of a desire to point out to young men the kind 
of knowledge that may add considerably to their 



efficiency as skilled craftsmen, and to obtain the pleasure 
of their company along the road in quest of it. 

There are facilities nowadays which it is to be 
regretted are not more fully taken advantage of by 
young craftsmen. A choice may be made among 
evening classes, technical colleges, and correspondence 
schools. Other countries, such as America and Germany, 
are making rapid progress in the scientific training of 
their young men, and in many respects they are doing 
more in that direction than we are in Britain. The day 
has now passed when we may treat these neighbours of 
ours with indifference, and they have now to be reckoned 
with in the race for supremacy, because they are provided 
equally as well as we are with the most modern ap- 
pliances and up-to-date machinery. These, however, are 
only tools. The knowledge and intelligence of the man 
behind the tools constitute the real element of success. 
Placed on an equal footing as regards appliances, we 
depend on the quality and ability of our craftsmen to 
enable us to retain our position at the top; but it is 
desirable that our young men should see it to be their 
duty as well as to be in their own interests to acquire 
as much as possible of all the scientific knowledge that 
has a practical bearing on their craft, in order to keep, 
not only our country in the foremost position in the 
industrial world, but also our craft in the foremost 
position in the engineering world. 

In conclusion, let it be distinctly understood that I 
do not think for a single moment that the most profound 
knowledge of the subjects which we have been consider- 
ing will ever of themselves make a man a smith. I 
trust as a practical smith that I have not yet taken 
leave of my senses. It may be true that good smiths, 
like poets, are born, not made, but after being born they 


have a long childhood to pass through in the acquisition 
of the high quality of skill in the manipulation of their 
tools, which is essential to all true sons of Vulcan. But 
what I do think, make bold to say, and strongly main- 
tain without fear of contradiction from any quarter, 
is that knowledge of this kind is invaluable, and the 
smith or forger who acquires it will most assuredly 
be a far superior craftsman to what he formerly was, 
whatever may have been the degree of his ability and 

As a last word, let me say that while I believe that 
all who read these pages will be at one with me regard- 
ing the object for which they were written, it is just 
possible that some may differ from me about some one 
or other of the practical points that I have introduced. 
What I have said I at present believe to be true, and 
in connection with some of the matters discussed, I have 
quoted as my authorities those at whose feet I may 
well be content to sit and learn. 

If we differ, let us do so in the best spirit, so that for 
our mutual benefit we may arrive at a knowledge of 
the truth, and the very fact of our differing will only 
be another and my last proof of the need for, and the 
value of, Science in the Smithy and Forge. 



Metallurgy^ Iron and Steel, John Percy, M.D.F. (John Murray, 

An IntrodtLction to the Stvdy of Metallurgy, Sir W. Roberts- 
Austen, K.C.B., D.C.L., F.R.S., A.R.S.M. (Charles Griffin 

& Co., Ltd., London.) 
The Metallurgy of Iron, Professor Thomas Turner, B.Sc, 

A.R.S.M., F.I.C. (Charles Griffin & Co., Ltd., London.) 
The Metallurgy of Steel, Professor H. M. Howe, LL.D. (Hill 

Publishing Co., New York.) 
Metallurgy, Elementary, A. Humboldt Sexton, F.LC, F.C.S. 

(Charles Griffin & Co. Ltd., London.) 
The Metallurgy of Iron and Steel, A. Humboldt Sexton, F.LC, 

F.C.S., and J. S. G. Primrose, A.G.T.C, A.LM.M. (The 

Scientific Publishing Co., Manchester.) 
The Metallwrgy of Steel, F. W. Harbord, A.R.S.M., F.I.C, and 

J. W. Hall, A.M.Inst.C.E. (Charles Griffin & Co., Ltd., 

Iron and Steel, J. H. Stansbie, B.Sc, F.I.C. (A. Constable & 

Co., London.) 
Iron and Steel, W. H. Greenwood, F.C.S., M.Inst.CE., and A. 

Humboldt Sexton, F.I.C, F.C.S. (Cassell & Co., London.) 
Cast Iron in the Light of Recent Research, W. H. Hatfield, 

B.Met., A.M.I.Mech.E. (Charles Griffin & Co., Ltd., 

Iron, Steel, and Other Alloys, Professor H. M. Howe, LL.D. 

(Hill Publishing Co., New York.) 


Structural Steely H. H. Campbell. (Hill Publishing Co., New 

The Chemistry of Iron and Steel Making, W. Mattieu Williams, 

F.C.S. (Chatto k Windus, London.) 
General Fotmdry Practice, A. M*William, A.R.S.M., and P. 

Longmuir. (Charles Grifl&n & Co., Ltd., London.) 
The Chemistry of Materials of Engineering, A. Humboldt 

Sexton, RLC, F.C.S. (The Technical Publishing Co., 

The Testing of Materials of Construction, W. Cawthorne Unwin, 

F.R.S., LL.D., M.Inst.C.E. (Longmans, Green & Co., 

A Handbook on Metallic Alloys, G. H. Gulliver, B.Sc, F.R.S.E. 

(Charles Griflan & Co., Ltd., London.) 
Introduction to Metallography, P. Goerens. (Longmans, Green 

& Co., London.) 
Elements of Metallography, Dr R. Ruer. (Wiley & Co., New 

Metallography, A. H. Hiorns. (Macmillan & Co., London^) 
Metallography, C. H. Desch, D.Sc., Ph.D. (Longmans, Green 

& Co., London.) 
Alloys and their Industrial Applications, E. F. Law, A.R.S.M. 

(Charles Griffin k Co., Ltd., London.) 
The Practical Metallography of Iron and Steel, J. S. G. Prim- 
rose, A.R.T.C., A.I.M.M. (The Scientific Publishing Co., 

Metallography of Iron and Steel, Professor A. Sauveur. (Hill 

Publishing Co., New York.) 
Microscopic Analysis of Metals, F. Osmond, J. E. Stead, D.Sc, 

D.Met., F.R.S., F.LC, F.C.S., and L. P. Sidney. (Charles 

Griffin & Co., Ltd., London.) 
An Introduction to the Study of Physical Metallurgy, W. 

Rosenhain, B.A., D.Sc, F.R.S. (Constable & Co., 

The Management of Steel, George Ede. (E. & F. N. Spon, 



The Heat Treatment of Tool Steely Harry Brearley. (Longmans, 

Green & Co., London.) 
The Gase-Hardening of Steel, Harry Brearley. (Iliffe & Sons, 

The Cementation of Iron and Steel, Dr Federico Giolitti. (Hill 

Publishing Co., New York.) 
The Iron and Steel Institute Journals, (Published by The 

Iron and Steel Institute, London.) 


Acid, nitric, 44. 

Alcohol, 44. 

Algebra, 27. 

Alteration of forging, 10. 

Alternating stresses, 57, 59. 

on locomotive rod, 24. 

Amount of contraction on outside 

of ring, 15. 
of expansion on inside of ring, 

Analysis of metals, microscopic, 42. 
Anatomical metallography, 42. 
Angle of friction, limiting, 36, 37. 
Annealing, 64, 87. 

chains, 94, 98. 

Antimony, crystals on face of ingot 

of, 52. 
Areas of circles, 31. 
Art, science and, 2. 

Bands, slip, 57. 

Barr on calculations, 23. 

Bending, theory of, 13. 

Bessemer steel, 5. 

Bevelled rings, 29. 

Biblio^phy, 157. 

Biological metallography, 42. 

Blistering, phenomena of, 141. 

Blocking hoops, 35. 

Blow-holes in ingots, 117. 

Blue temperature, effect of work at, 

Borax for welding, 120, 121. 
Breaking of tools, 77. 
Brittleness, rectangular, 87. 

Stead's, 87. 

Brittle structure owing to cold 

work, 101. 
Burning of steel, 108. 

Burnt steel, restoration of, 123. 
Butt welds, structure of, 133. 

Calculations for forgings, 8. 

for smiths and forgers, 19. 

mathematical, 17. 

Carbide of iron, 51, 52, 72. 
Carbon in iron, 51, 52, 140. 
Carburisation, correct temperature 

for, 146. 
Garburising at low temperatures, 

Case-hardening, 140-163. 
Casting, welding steel, 132. 
Chains, annealing, 93, 98. 
Chappell on crystal gi'owth, 98. 
Chart, diagrammatic, 82, 83. 

heat treatment, 70, 71. 

Chemistry of welding, 107-139. 

Chrome nickel steel, 76. 

Circle having area double that of 

another, 31. 
half that of another, 

Cleavage lines in steel, 56. 
Coefficient of expansion, 40. 

of friction, 36, 37. 

Cold strain in chains, 94. 

worked iron and steel, 88, 101. 

Cone, plate for, 28. 
Contraction and expansion, 40. 

of drop stamping, 41. 

on outside of ring, 15. 

Cooling, prolonged and rapid, 64. 
Critical points, 69. 
Crushing effect on steel, 101, 104. 
Crystalline structure of iron and 

steel, 43, 45, 47, 55. 
Crystallites, primary, 55. 




Crystals, fir-tree formation of, 61, 

growth of, 64, 88, 91. 

in chains, 93. 

Dkcarburisation, 163. 
Development of surfaces, 28. 
Different amounts of carbon in steel, 


critical points, 69. 

Diffusion of carbon in iron, 140. 
Dimensions of bar for race, 25. 

of eyebolt, 20. 

Distortion of cold-worked material, 

Drawing, mechanical, 31, 34. 

and .staving, 18. 

Drop stamping, contraction of, 41. 
stampings, effect of hot work 

on, 64. 
Ductility of heat-treated steel, 75. 

Education, technical, 2, 3. 
Effect of hot work on drop stamp- 
ings, 64. 
Etching reagents, 44. 
Ewing and alternating stresses, 59. 
Expansion and contraction, 40. 

on inside of ring, 15. 

Eyebolt, dimensions of, 20. 

Failure of boiler plate, 103. 

FeiTite, 50, 72. 

Ferrous silicate, 119. 

Fibre in iron and steel, 43, 45. 

Fir-tree crystals, 51, 62. 

Fluxes, 118. 

Forging, alteration of, 10. 

Forgings, calculations for, 8. 

Fracture in iron, 45. 

of steam hammer piston rod, 

Fractures in chains, 94. 

in steel, 65, 67, 69, 84, 86. 

Friction, coefficient of, 38. 
limiting angle of, 36, 37. 

Geometry, practical, 31. 
Grains, cracks through crystal, 56. 
Growth of crystals, 64, 88, 91. 
of crystals in chains, 93. 

Hammer piston rod, broken, 59. 
Hardening point of steel, 81. 
Heat, 35. 

treated steel, ductility of, 76. 

treatment chart, 70, 71. 

of iron and steel, 63-106. 

Hoops, blocking, 35. 
shrinking fit for, 39. 

Incomplete refining, 84. 

Ingots, blow-holes in, 117. 

Intergranular oxidation, 112. 

Iron and steel, cold-worked, 88, 

crystalline structure 

of, 43, 46, 47, 55. 

fibre in, 43, 45. 

heat treatment of, 


microscopic exami- 
nation of, 43. 

overstrained, 88. 

welding, 6. 

carbide of, 51, 52, 72. 

carbon in, 51, 52, 140. 

diffusion of carbon in, 140. 

fracture in, 45. 

lamination in, 45. 

Kipling's example of cold deforma- 
tion, 91. 

Kirkaldy on the crystalline structure 
of iron, 47. 

Lamination in iron, 45. 

Length of material for a forging, 8. 

a ring, 12. 

Limiting angle of friction, 36, 37. 
Locomotive connecting rod, 24. 
Loss of material while heating, 9. 

Manganese, 5. 

Material, loss of while heating, 9. 
Materials, strength of, 20, 27. 
Mathematical calculations, 17. 
Mathematics, 6. 
Mechanical drawing, 31, 34. 
Mechanics, practical, 35. 
Metallography, 42-62. 
Metallurgy, 6, 107, 132. 
Method of etching, 44. 



Method of polishing specimens, 43. 
Microscopic analysis of metals, 42. 

examination of iron and steel, 

Micro-structure of wrought iron, 48. 
Mild steel, 5. 
Mushet, Robert, 5. 

Nickel chrome steel, 76. 
Nitric acid, 44. 

Oil quenching, 73, 78 

Osmond on microscopic analysis of 

metals, 42. 
Overheating, 63, 87. 
Overstrained iron and steel, 88. 
Oxidation, 10, 108. 

Pathological metallography, 42. 

Pearlite, 50, 52, 72. 

Percy on the crystalline structure 

of iron, 47. 
Phenomena of peeling, 141. 
Phosphorus in steel, 55, 73. 
Pic iron, 6. 

Polishing of specimens, 43. 
Practical and theoretical men, 4. 

geometry, 31. 

mechanics, 35. 

Practice, theory and, 4. 

Prejudice against scientific training, 

Primary crystallites, 56. 
Prints, sulphur, 53, 64. 
Prolonged cooling, 64, 75. 

heating, 148, 153. 

Punched holes in plates, 102. 

Quenching carburised material, 
151, 163. 

in oil, 73, 78, 153. 

in water, 73, 78. 

Race, safe load for, 26. 

dimensions of bar for, 25. 

Reagents for etching, 44. 

Stead's new, 55. 

Rectangular brittleness, 87. 
Restoration of burnt steel, 123. 
Ring, length of material to make, 

Rings, bevelled, 29. 

Rings, expansion and contraction 

on edges of, 16. 
Rosenhain and slip bands, 67. 

Safe load for race, 26. 

Sand for welding, 119, 121. 
i Sauveur on deformation of iron, 91. 
I Science and art, 2 

value of, 1. 

Scientific training, 2, 3, 7. 
I Shrinking fit for hoops, 39. 
; Silicate, ferrous, 119. 

Silicon steel, 66. 

Slip bands, 57, 59. 

Stampings, drop-hammer, 40, 64. 

Staving and drawing, 18. 

Stead's brittleness, 87. 

conclusions regarding burnt 

steel, 126. 

heat-treatment chart, 70, 71. 

new reagent for etching, 65. 

Steel, Bessemer, 5. 

cleavage lines in, 66. 

I different amounts of carbon in, 


ductility of heat-treated, 76. 

fractures in, 65, 67, 69, 84, 86. 

hardening point of, 81. 

heat-treutment of, 63. 

microscopic examination of, 


mild, 5. 

nickel chrome, 76. 

phosphorus in, 66, 73. 

restoration of burnt, 123. 

strength of, 23. 

tenacity of heat-treated, 76. 

welding high-carbon, 116. 

Strength of materials, 20, 27. 

of steel, 23. 

of welds, 133. 

Stresses, alternating, 24, 57, 59. 

Structure of butt welds, 133. 

Sulphur prints, 53, 54. 

Surfaces, development of, 28. 

Table of weights, use of, in forging, 

Taper of wedges, 35, 38. 
Technical education, 2, 3. 



Tenacity of heat-treated steel, 75. 
Testing of material, 23. 
Theory and practice, 4. 

of bending, 13. 

Training, scientific, 2, 3, 7. 
Treatment of tools, 77, 82. 

Value of science, 1. 

Wash heat, 10. 

Water quenching, 73, 78. 

Wedges, tapered, 35, 38. 

Weights, use of table of, in forging, 

Welding at low temperatures, 117. 

chemistry of, 107. 

effect of, on adjacent structure, 


high-carbon steel, 116. 

iron and steel, 6. 

steel casting, 132. 

Welds, strength of, 133. 

butt, 133. 

Wrought iron, 6, 60. 

micro- structure of, 48. 


— -^/ 




This book Is doe on die last date stamped below, or 

on uie date to wniai renewed* 

Renewed books are subjea to immediate cecaiL 

APR 2 1977 

i 1 

S^HR/fl 6 77 

^gi^WtP w '» 

. — ti\OA 


, AFR'^s^^m 


r:s. cm f,iAY 2 7 1981 

JUL 2 9 1998 


LD21— 32m— 1,'75 

General Library 

University of California 



vn 19258 


360416 ^ 

TT 2 f^'