Skip to main content

Full text of "Proceedings of the Indiana Academy of Science"

See other formats


Proceedings 

of the 
Indiana Academy of Science 




VOLUME 116 



2007 



NUMBER 2 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 

The PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE is a journal dedicated to 
promoting scientific research and the diffusion of scientific information, to encouraging communi- 
cation and cooperation among scientists, and to improving education in the sciences. 

EDITOR: James W. Berry, Department of Biological Sciences, Butler University, Indianapolis, 
Indiana 46208. Telephone: (317)-940-9344; FAX: (317)-940-9519; e-mail: jwberry@butler.edu 

EDITORIAL BOARD: Hans O. Anderson (Indiana University, Bloomington); Robert F. Dale 
(Purdue University, West Lafayette); Rebecca Dolan (Butler University); Kara W. Eberly (St. 
Mary's College); Uwe J. Hansen (Indiana State University); Daryl R. Karns (Hanover College); 
Gene Kritsky (College of Mt. St. Joseph); Stephen A. Perrill (Butler University); Paul Rothrock 
(Taylor University); Thomas P. Simon (U.S. Fish & Wildlife Service, Bloomington); Michael 
R. Tansey (Indiana University, Bloomington); Robert D. Waltz (Indiana Department of Natural 
Resources, Indianapolis); J. Dan Webster (Hanover College); Harmon Weeks (Purdue University, 
West Lafayette); John O. Whitaker, Jr. (Indiana State University) 

THE INDIANA ACADEMY OF SCIENCE 

PRESIDENT: John Robert Schutt (Taylor University - Ft. Wayne) 

PRESIDENT-ELECT: Nils I. Johansen (University of Southern Indiana) 

TREASURER: Edward L. Frazier (Indianapolis) 

SECRETARY: Heather Bruns (Ball State University) 

EXECUTIVE DIRECTOR: Nelson R. Shaffer (Indiana Geological Survey) 

ARCHIVIST: Anika Williams (Indiana State Library) 

EXCHANGE ITEMS: Items sent in exchange for the Proceedings and correspondence concerning 
exchange arrangements should be sent to the Indiana Academy of Science, John S. Wright Memorial 
Library, 140 North Senate Avenue, Indianapolis, Indiana 46204. Questions regarding back issues or 
undelivered issues should be addressed to Anika Williams at awilliams@library.IN.gov. Telephone 

(317)-232-3686. 

Cover photo: The photo was taken during the archeological study of a pioneer period house site (ca. 
AD 1835) known as the Reddick Site. Under the "St. Mary's Treaty" signed in 1818, native Ameri- 
cans surrendered the land in the central region of the new state of Indiana to European- American 
settlers. This site passed through several land owners and in 1903 was purchased by the United 
States government for use as Fort Benjamin Harrison. This area became the Fort Harrison State 
Park in Marion County, Indiana in 1995. The scene is the 100% surface collection of artifacts, and 
the orange flags mark the 1 -meter units. This work was done by the Next Step Education Through 
Archaeology project administered by Martin University. In 2006 it was shown that the areas that 
were surface collected had not been plow disturbed, and may represent a natural deposit since the 
Pleistocene. (See related article on page 1 1 7 of this issue.) 

Visit the Indiana Academy of Science website at: 
www.indianaacadcmyofscience.org 

Publication date: 31 December 2007 
This paper meets the requirement of ANSI/NISO Z39.48-1992 (Permanence of Paper). 



ISSN 0073-6767 



2007. Proceedings of the Indiana Academy of Science (1 16):1 17-125 



PEOPLE, POTS, AND PROSPERITY: THE CERAMIC VALUE 
INDEX AND AN ASSUMPTION OF ECONOMIC CLASS 



James M. VanderVeen: Department of Sociology and Anthropology, Indiana 

University South Bend, 1700 Mishawaka Avenue, South Bend, Indiana 46634-71 1 1 
USA 

ABSTRACT. The ceramic value index is a powerful empirical tool used in historical archaeology to 
assess the required economic access necessary for a family or individual to accumulate specific household 
goods. The focus of this method is primarily on the status of the artifact assemblage itself, however, and 
not the people who acquired the objects. Since this measure of socioeconomic status is quantified onl\ 
through the pottery used by the site occupants, it may not take into account the various perspectives of 
the occupants have towards their domestic vessels, nor does it consider the wider social context of the 
study area. Although the formula has been used extensively in historic archaeology, this has been done 
without significant critique. Sites from 19 th -century Indiana are used here as examples of the potential 
successes and failures of a formula built on the assumption that consumers utilize archaeological objects 
for all the same reasons. 

Key words. Socioeconomic status, historic ceramic vessels 



AN EXPLANATION OF 
"CERAMIC VALUE" 

Many archaeologists have a special interest 
in ceramic objects, due in part to the durability 
and prevalence of the material in a historic 
site. Further, the form, function, and design of 
a ceramic vessel typically allow for interpre- 
tations of the passage of time, cultural affili- 
ation, and categorically distinct activities. 
Like an archaeological version of the silicon 
chip, a ceramic sherd, only a small piece of 
fired clay and temper, can hold a tremendous 
amount of data. 

There may be a great deal of additional in- 
formation, however, that can be "read" from 
that sherd, with the proper tools. More than 
20 years ago, George Miller (1980, 1991a, b) 
created a method of classification and inter- 
pretation of ceramics that continues to be em- 
ployed by archaeologists. Working with doc- 
umentary records such as bills of lading, price 
lists of manufacturers and retailers, and ship 
manifests, Miller argued that the remains of 
the domestic ceramic vessels reflect the socio- 
economic status of households from which 
they were found. For example, he used Staf- 
fordshire price-fixing agreements to suggest 
that vessel decoration related directly to its 
cost to the consumer (Miller 1980). Moreover. 
the cost of the simplest undecorated ware. 



called cream color or CC, remained relatively 
stable throughout the eighteenth and nine- 
teenth centuries (Miller 1980. 1991a). 

Using the undecorated ware as a baseline. 
Miller calculated the ratio of the price of three 
other categories of decoration to this baseline. 
He determined that the comparison of the 
number and value of each type of decoration 
category could be used to construct a propor- 
tion of expensive to less expensive wares. 
called a ceramic value index. In its most sim- 
ple operation, the ceramic value index works 
like a weighted mean. The prices for each t\ pe 
are analyzed and scaled in reference to the 
undecorated ware. An archaeologist need onl\ 
count the number of vessels in each level, 
multiply this count by the index value as- 
signed to that level, sum the products and di- 
vide by the total number of vessels recovered 
(see Miller 1991a for a full description of the 
formula). 

Ceramic vessels take many forms. Some arc 
basic utilitarian dishes needed b\ all. while 
others may be high status lu\ur\ goods. The 
four decorative classes created by Miller 
(1980) cover the majority of table, kitchen. 
and toilet wares recovered from across North 
America during the late eighteenth to early 
twentieth centuries. Since the pieces are ubiq- 
uitous and reflective of price, it was thought 



117 



118 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



that "having internal value scale for ceramics 
[was] going to increase our ability to perform 
socioeconomic analysis on archaeological col- 
lections" (Miller 1980). Indeed, the analyses 
occurred as predicted, as evidenced by the 
number of studies in the past three decades 
that dealt with the different economic classes 
suggested by recovered ceramic artifacts (e.g., 
Adams & Boling 1989; Andrews & Fenton 
2001; Stine 1990). The method was seen as 
an objective measure of the socioeconomic 
status of a historical site, and was quickly 
adopted by researchers looking for a more 
systematic, empirical procedure on which to 
base their claims. Yet it seems as if the focus 
of these studies was more on the status of the 
artifact assemblage than on the people to 
whom it belonged, as if the pottery had an 
agency of its own. The analysis was then ex- 
tended to the occupants and not the other way 
around. The formula has been used extensive- 
ly in this way without much published criti- 
cism (see Majewski & O'Brien 1987 for an 
exception). 

This paper examines a few of the assump- 
tions underlying the application of the index 
to all sites indiscriminately. It is not meant to 
refute the value of the method, nor is the in- 
tent to discredit Miller. In fact, Miller himself 
warns of relying on historic archaeological 
data without reference to documentary sourc- 
es (1980, 1991a); and he has placed other ca- 
veats on his technique (Miller & Hurry 1983). 
Instead, the objective is only to raise the yel- 
low flag of caution in order to slow the speed- 
ing analyses of socioeconomic status through 
ceramics and warn the researchers of dangers 
in the road. 

AN APPLICATION OF THE INDEX TO A 
PREVIOUSLY UNSTUDIED SITE 

Aside from the date of occupation, very lit- 
tle is usually known about a small historic pe- 
riod archaeological site. It could be argued 
that the ceramic value index is just the tool 
needed to reach conclusions about the people 
who lived there. After all, it has an ease of 
application and a precedent for interpretation. 
If it were true that the style and number of 
food serving and production dishes were an 
accurate measure of economic wealth, then 
calculating the index would allow a researcher 
to ascertain the probable socioeconomic status 
of the occupants of the site. This information, 



by extension, may assist in determining a po- 
tential occupation of the settlers, the size of 
the household structures, or even the number 
of people in the family living there. Each of 
the above is a factor that contributes to socio- 
economic status, in much the same way as the 
dollar value of the associated ceramic assem- 
blage. 

In the case of the Reddick site in Marion 
County, Indiana, however, little information 
about house or family size is known. The sys- 
tematic recovery of ceramics, and a smaller 
number of architectural and domestic materi- 
als, has allowed a date of approximately 1 845 
to be assigned to the site (VanderVeen 2001). 
Aside from some later county atlases docu- 
menting land ownership, and thus the name, 
to the first recorded European-American set- 
tlers in that area of the county, little else is 
known about the site or its occupants. Census 
data exist but cannot be reliably applied, for 
at the time it was a typical practice for an 
individual or family to squat on land owned 
by others. Further, it was not uncommon for 
an individual to purchase land speculatively 
and not establish residence for some time, if 
at all. Thus the actual identities and number 
of occupants of the site are unknown. 

What is known, however, is the history of 
the land itself. With the "New Purchase" trea- 
ty, signed in 1818, Native Americans surren- 
dered their land in the central region of the 
new state of Indiana to European-American 
settlement. The area was formally opened for 
legal purchase in 1820, and prior to then it 
may have been occupied by members of the 
Delaware, Miami, and Potawatomi nations, or 
by illegal European-American settlers. Be- 
cause the site is located is within the swampy 
eastern portion of Marion County, it was set- 
tled more slowly than the rest of the county. 
According to an early history of the area, 
many of the settlers were of Scottish, Irish, 
English, and German descent; and they pri- 
marily traveled west via the Ohio River from 
Ohio, Pennsylvania, and the North Carolina 
Piedmont (Sulgrove 1884). 

The first documented European-American 
settlers in the township were Elisha Reddick, 
his wife Elizabeth, and their infant son. In 
1832, Reddick purchased property from John 
Johnson in the southwest quarter of Section 
36 (Sulgrove 1884). Reddick and his brother 
Joshua held the property until 1848, after 



VANDERVEEN— PEOPLE, POTS, AND PROSPERITY 



119 



Table 1. — Index Values for Recovered White 
Earthenware Sherds at the Reddick Site. 







Index 


Total 


Ceramic type 


Sherds 


value 


value 


Undecorated wares 


511 


1.00 


5 1 1 .00 


Minimally decorated 








(total) 


255 


1.16 


295.80 


Annular/banded 


78 






Edge-decorated 


69 






Mocha ware 


14 






Monochrome glaze 


26 






Spattered/sponged 


68 






Painted 


74 


1.30 


96.20 


Transfer-printed 


80 


2.50 


200.00 


Total 


920 




1095.68 



which time it passed through a number of 
owners until it was purchased in 1903 by the 
United States government for use as Fort Ben- 
jamin Harrison. Land surrounding the prop- 
erty then became a state park in 1995. 

The site is situated on top of a ridge over- 
looking Fall Creek and its flood plain, within 
the present boundaries of the state park. No 
evidence of structural footings have been 
found, but the type and distribution of the ar- 
tifacts suggest a small, possibly temporary 
residence. Enough brick has been recovered 
to indicate the likely presence of a hearth or 
even a small chimney. Metal hardware and 
window glass also have been collected. Com- 
bined with the amount of household ceramics, 
the archaeology attests to a modest domestic 
structure. 

This is an ideal situation for the employ- 
ment of ceramics, by far the most common 
artifact, towards developing a picture of the 
people who lived on the site. The ceramic ma- 
terial collected from three archaeological field 
seasons were analyzed using a version of 
Miller's formula revised by McBride & 
McBride (1987) to better account for broken 
artifacts. The results of the formula as applied 
to sherd counts rather than whole vessel forms 
should be viewed cautiously, but previous re- 
search using this method found the results to 
correspond with occupational levels at a de- 
gree similar to that of Miller's formula 
(McBride & McBride 1987; Huser 1993). As 
seen in Table 1, the sherds were typed ac- 
cording to level of decoration style and the 
number within each level was multiplied by 



the scale provided by Miller (1991a). Only re- 
fined ware, typically tableware, was analyzed 
this way, keeping with the procedure. The re- 
sulting ceramic value index for the Reddick 
site was 1 .20. 

IMPLICATIONS OF INDEXING 

Naturally, there are some practical limita- 
tions to the methodology behind creating a ce- 
ramic value index, as there are with many oth- 
er quantitative measures of social phenomena. 
The mean ceramic value is based on the price 
of the vessel at the time of the initial acqui- 
sition. Yet heirloom pieces and gifts would be 
examples of traditionally more expensive 
dishes given at no cost to an individual or 
family. More formal and more expensive ce- 
ramics are also used less often and are more 
carefully curated by their owners, so they tend 
not to break as frequently as those used in 
everyday circumstances and therefore are less 
well-represented in the archaeological record. 
Of course, archaeological data are always in- 
complete and may be biased due to site for- 
mation processes such as selective discard and 
scavenging (Schiffer 1972). 

Another caveat is that some sherds may 
have been misclassified as undecorated when. 
in fact, they included a pattern that was not 
exhibited in the particular portion of the vessel 
that was recovered. Most vessels of the type 
studied either show decoration over the whole 
of the body, in which case the decoration 
would not be missed, or on the rim only. With 
regards to the latter, a body sherd from a rim- 
decorated piece may be incorrectly typed, but 
even so, the value differences between undec- 
orated and minimally decorated vessels are 
relatively small. 

Miller himself cautions about shortcoming 
of his method with regards to infrastructure, 
market access, and economic isolation (Miller 
& Hurry 1983). Nevertheless, adequate trans- 
portation systems were in place in the Mid- 
western United States by the earl) to middle 
nineteenth century, and new types of ceramics 
would appear even in the most remote areas 
within a few years of introduction in England 
(Lofstrum et al. 1976). Thus, the market ac- 
cess likely had little or no effect on the ce- 
ramic value index. A review of the Indiana 
Gazetteer, a director) of merchants and ser- 
vice providers, suggests the issue of distant 
markets was not problematic for the Reddick 



120 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



site. At least four different retailers of ceramic 
vessels, and "queensware" in particular, are 
known to have establishments in the Indian- 
apolis area by 1866 (Cowen 1866). Addition- 
ally, many of the new residents of the area 
had established connections with some of the 
other vendors operating west of the Appala- 
chian Mountains. 

The economic center of a region is usually 
wealthier and more developed than those ar- 
eas at the edges, and access to economic 
goods is typically more restricted in the pe- 
riphery (Cressey et al. 1982). If a commodity 
is not as readily available once people begin 
moving to less accessible areas, then value is 
added during the transportation of the item. 
Consequently, the price of a piece of pottery 
may increase in proportion to the distance of 
the supplier from the final point of purchase. 
Some scholars, on the other hand, suggest that 
there is no "tyranny of distance" per se (e.g., 
Baugher & Venables 1987). Rather, the access 
to a market has more to do with economic and 
political considerations than does physical 
proximity. If someone can afford to purchase 
an item, then "someone was ready, willing 
and able to ship it" (Baugher & Venables 
1987). 

COMPARATIVE SITES AND RESEARCH 
DEVELOPMENTS 

To control for the various issues, including 
the economic access present in this "border- 
land" of the period, the Reddick site is com- 
pared here with other frontier sites in Indiana. 
Unfortunately, a very small number of historic 
sites within the state have properly docu- 
mented archaeological investigations, and 
fewer still date as far back as the middle 19 th 
century. Three such sites do exist that share 
with the Reddick site a similarity of house lo- 
cations, contemporaneous time period of oc- 
cupation, and number of recovered and ana- 
lyzed ceramic sherds (see Table 2). When 
possible, only information determined to be 
solely from the occupation context dating to 
the appropriate time period was used. The de- 
scriptions below and data included in Table 2 
are revised from the work of Huser (1993) and 
Stillwell (1990). 

William Conner house: Believed to be one 
of the first brick buildings in Central Indiana, 
the Conner house was constructed in 1823. 
Originally a rural family residence, it is lo- 



Table 2. — Listing of compared sites. 







Mean ce- 




Site name 


County 


ramic date 


Sherds 


Conner 


Hamilton 


1851.1 


1281 


Godeke 


Warrick 


1845.3 


1286 


Reddick 


Marion 


1845.4 


1536 


Richardville/ 








LaFontaine 


Huntington 


1830-1870 


2051 



cated on a ridge spur above the White River 
valley in Hamilton County, immediately north 
of Marion County. The house is presently part 
of the grounds of the Conner Prairie Pioneer 
Settlement, a living history museum. 

William Conner was born in Ohio and 
around 1 800 traveled to Indiana as a fur trad- 
er. Conner married a Native American woman 
from the Delaware tribe with which he did 
business. His role in the government's rela- 
tions with the Native American people re- 
mains subject to some debate. After his wife's 
departure from Indiana with the rest of her 
tribe, Conner married a European-American 
woman recently arrived from New York, and 
eventually established a distillery and several 
mills. Finally, he served as a state represen- 
tative for several terms. 

Conner's many descendants who occupied 
the house made their fortunes in medicine, 
politics, business, and the military. Conse- 
quently, the Conner family, at least during the 
period examined here, were of high economic 
wealth and social status (Stine 1990). 

Godeke site: There was no standing struc- 
ture present at the Godeke site at the time of 
investigation, but several subsurface cultural 
features were found during excavation. The 
area sits on a low hill about a kilometer east 
of Bluegrass Creek in Warrick County. Al- 
though little is known about the history of the 
site, it is estimated to have been occupied be- 
tween about 1830 and 1860. Census records 
list the various owners of the property as 
farmers and as a store clerk, but the property 
owners might have leased the land to others. 
Because of a number of assessments, includ- 
ing the lack of permanent structural remains 
and occupational activities, the socioeconomic 
status of the inhabitants of the Godeke site is 
evaluated as rather low (Stine 1990). 

Richardville/LaFontaine site: Currently, 
the house at the Richardville/LaFontaine site 



VANDERVEEN— PEOPLE, POTS, AND PROSPERITY 



121 



Table 3. — Comparison of Ceramic Value Indices. 







Documented 


Site name 


CVI 


occupation 


Conner 


1.41 


Physician; politician 


Godeke 


1.15 


Farmer; clerk 


Reddick 


1.20 


Unknown 


Richardville/ 






LaFontaine 


1.39 


Politician; merchant 



is a large, two-story wood-frame residence. 
The site is located at the confluence of the 
Wabash River and the Little River, on a flat 
plain in Huntington County. Some documen- 
tary evidence suggests that the house was 
built just after the Miami Chief John Richard- 
ville moved his tribal council to the Forks of 
the Wabash in 1831. Chief Richardville died 
in 1841 and, while he possessed political pow- 
er in his position with the Miami, his trade 
business had been greatly depressed by the 
time of his death. Francis LaFontaine, Ri- 
chardville's son-in-law, assumed the duties of 
tribal chief and inherited the section of land 
on which the house stood. LaFontaine's own 
descendents continuously occupied the site 
until the property passed out of tribal owner- 
ship early in the 20 th century. 

The inclusion of the Richardville/La- 
Fontaine site may be in some ways problem- 
atic. In strictly economic terms, chiefs Ri- 
chardville and LaFontaine should certainly be 
seen as part of the upper middle class (Stine 
1990). The ceramic value index of the site is 
not significantly different than that of the Con- 
ner site, and both houses were at one time 
owned by politicians. Given their membership 
in an ethnic group different than that of the 
dominant society, however, the social status of 
the Richardville/LaFontaine families is uncer- 
tain. Because of their positions of power and 
respect within that minority, and the access 
and means to acquire expensive material 
goods, some may treat them as individuals of 
high status. Still, the prestige given to them 
from members of the majority may be reduced 
due to their ethnicity, thus also reducing that 
high status. 

As seen in Table 3, the ceramic value index 
of the sites appears to correlate with other 
models of determining household wealth, such 
as occupation and house size (Powers 1982). 
Previous research using the Miller analysis on 



sherd counts indicates that ceramic values of 
1.20 to 1.30 can be interpreted as "middle in- 
come level," while values above or below 
may be seen as "upper" and "lower class," 
respectively (McBride & McBride 1987). 

AN EXAMINATION OF THE 
ASSUMPTIONS WITHIN "VALUE" 

But what of the people who lived at the 
sites? The process of assigning a ceramic val- 
ue seems to be rather deterministic in that a 
collection of dishes defines the household. 
Through Miller's formula, the application of 
an easy and effective quantitative measure re- 
moves the power of consumer choice and di- 
minishes the effect of other, more intangible. 
variables. For example, it could be argued that 
there is a "saturation point" of some sort with 
regards to the presence of ceramic vessels, as 
richer families will not continue to buy a new 
suite of place settings each year. After all. 
there is certainly a decreasing marginal utility 
for each additional ceramic dish purchased 
within a single household. Artifacts do not 
simply consume themselves (Cook et al. 
1996). There are consumers apart from the 
commodities that must be given agency. Peo- 
ple make choices, and are motivated to make 
those decisions for several reasons, deter- 
mined by issues other than that of available 
economic wealth, or even class ideals. Fur- 
thermore, although some tasks, demographics, 
and behaviors may be assigned to all of the 
members of a household, ever} individual in 
that group does not necessarily share the same 
status (Wall 1999). The assignment of one lev- 
el of "class" to the household is in man) 
ways a simplification on the order of that 
made by the once-a-decade census, and equal- 
ly as given to errors. 

The role of class is given considerable pow- 
er by researchers. Many proponents of objec- 
tive measures of class report a strong relation- 
ship exists between economic roles 
(occupation), social stratigraphy, and the ma- 
terial culture recovered from a site (e.g.. Xiek- 
olai 2003; Spencer- Wood & Heberling 198""). 
but they refrain from discussing what consti- 
tutes these various elements. Since this paper 
is meant to critique the assumptions of the ce- 
ramic value index and its accepted relation- 
ship with class, socioeconomic status, and 
household wealth, the concepts of those terms 
must be defined, at least eenerallv. 



122 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Human beings appear to crave categories, 
and they create classifications whenever pos- 
sible to better understand the world. While 
these categories are not natural, they are often 
rationalized as such in order to justify the re- 
stricted access to resources afforded to some 
groups of people (Beteille 1981). As for class, 
it is used to reflect a ranked social position, 
usually determined by wealth and occupation, 
but also based on prestige and family or social 
ties (Wurst & Fitts 1999). Class distinctions 
are relative and may be linked further with 
gender, race, and ethnicity. Regardless, when 
a group shares a similar lifestyle based on 
their economic position, they likely see them- 
selves as members of the same class (Powers 
1982). 

It is not necessary for income to dictate 
class. Families with the same levels of wealth 
may choose to spend money in different ways, 
for money alone does not equate with com- 
parable tastes. Some people are simply not in- 
terested in the approval of their peers and will 
behave in a manner entirely to their own lik- 
ing. What scholars typically put forth as 
"class" seems to be a manifestation of a 
Bourdieu-like process of socialization, in 
which class members "learn" what posses- 
sions are held to be desirable or improper. A 
class, in this case, is defined "as much by its 
being perceived as by its being, as by its con- 
sumption — which need not be conspicuous in 
order to be symbolic — as much by its position 
in the relations of production (even if it is true 
that the latter governs the former)" [emphasis 
in original] (Bourdieu 1984). In sum, a class 
is what a person makes of it; and, as such, an 
empirical measure meant to calculate class 
across individual actors is bound to be mis- 
leading. The important issues of perception, 
agency, and the complex symbolic nature of 
ceramics are concepts unfortunately not at- 
tended to by Miller's index; the commodities 
present in an assemblage cannot objectively 
predict something as amorphous as socioeco- 
nomic status. Historical interpretations are 
never completely neutral (Nickolai 2003). 

The ceramic value index, with all its merits, 
is constructed on a number of assumptions 
that reduce its usefulness and should temper 
its results. For example, it is supposed that 
economic wealth (or, even more accurately, 
economic access) equates directly with the 
perceived social class (Wurst & Fitts 1999). 



Yet research has shown that, in some instanc- 
es, slaves had more expensive vessels than 
their masters or many northern European- 
American farmers and business owners (e.g., 
Adams & Boling 1989). Whether the ceramics 
were given to the slaves or bought on their 
own through money earned by extra work, the 
end result is that high ceramic value index 
scores were calculated for individuals with 
known low levels of social status, at least with 
reference to the dominant society. If the slaves 
indeed purchased the ceramics, one could as- 
sume that the motivation was to convey a par- 
ticular message, and not necessarily one of 
adoption or imitation of the particular stan- 
dards of the slave owners. This symbolic 
meaning may not have been related to the 
monetary value of the ceramics at all. Ceram- 
ics may be emblems as much as everyday 
utensils; they may possess symbolic value be- 
yond that of a simple indicator of economic 
wealth (Beaudry et al. 1991; Wall 1999). 

Further, it is not always true that the ceram- 
ics assembled by a household will accurately 
reflect the socioeconomic status of each of the 
members (Garrow 1987), since males and fe- 
males within the household may be seen as 
possessing different levels of "status." Al- 
though the researcher assigns class, those be- 
ing researched may not have necessarily be- 
lieved in the same definitions, or even felt any 
pressure to behave as others did. Miller (Mill- 
er & Hurry 1983) referred to this problem, 
although not exactly in the same context as he 
intended. He used a documentary record that 
showed that a particular settler became a 
wealthy landowner in an isolated area. This 
man probably purchased the ceramics in his 
house piecemeal and not in sets, owing to the 
shortage of available commodities. Without 
knowing the circumstances, the ceramic value 
index calculated for the assemblage recovered 
at this site could lead one to believe that the 
owner was from the lower class. 

Finally, although wealth may allow one to 
purchase the "correct" symbols (a set of par- 
ticular dishes for instance), it is the manifes- 
tation of the symbolic behavior (such as table 
manners and etiquette) that more clearly de- 
notes membership in particular classes to oth- 
ers (Wurst & Fitts 1999). Status is defined by 
more than money; it is also a social construct. 

People are not the passive products of eco- 
nomic models; rather, they often tend to make 



VANDERVEEN— PEOPLE, POTS, AND PROSPERITY 



123 



unpredictable choices concerning the reallo- 
cation of resources or the reinterpretation of 
the values of a particular class. In these in- 
stances, as well as many others, a household 
may be in possession of considerable wealth, 
and even earn the respect or adoration of their 
neighbors, yet the resulting collection of ce- 
ramics could very well be below any arbitrary 
level demarcating their appropriate class. The 
right set of dishes simply may not be a pri- 
ority, and assets are instead allocated else- 
where. 

Based on its ceramics, the Reddick site was 
assigned to the level of lower middle socio- 
economic status. Without many other docu- 
mentary sources on which to base an evalua- 
tion, this site could be categorized as a simple 
farmstead with little access to wealthy goods 
and lacking in prestige, at least as viewed by 
others in the area. Yet in actuality, the scale 
of socioeconomic status on which the index 
categorizes people is quite relative and pred- 
icated on the acceptance of people whom it 
describes. Either of at least two opposing sit- 
uations may be within the realm of possibili- 
ties: the settlers of the site may have pos- 
sessed little money but were seen as wealthy 
by other, poorer, inhabitants of the area, or the 
same individuals may have had quite a sum 
of money but no contact with others and thus 
no need to present their wealth through ceram- 
ics. If it is true that the "social interaction that 
marked class affiliation called for prescribed 
behaviors, including participation in complex 
dining rituals that required expensive items of 
material culture" (Andrews & Fenton 2001), 
without the interaction there would be no need 
for the dishes. Conversely, if the dishes did 
not exist, a "dining ritual" could still occur, 
only shifted in its emphasis or alternatives for 
the dishes used. In each of the instances, the 
ceramic value index would not accurately rep- 
resent the truth of the situation. 

RECOGNIZING SELF-DEFINED 
"CLASS" THROUGH WARE RATIOS 

It appears that the employment and inter- 
pretation of the index relies heavily on the 
central tenet of all archaeology — context. 
Class values can be reflected in the choice of 
particular ceramics, just as income levels can 
also be represented in the total assemblage. 
Because it cannot understand the motivation 
behind that choice, or the circumstances sur- 



rounding that income, the index has little or 
no overall value, particularly in isolation. In- 
dividuals may choose not to participate in the 
same discourse as the larger part of society, 
or they may select instead to challenge the 
status quo. At the scale of a single person or 
household, decisions could be made to trans- 
late the accepted norms in a way that better 
reflects the needs and desires present at the 
time (Stine 1992). Both class and material cul- 
ture can be social constructions. 

Material culture may have different mean- 
ings or functions depending on its users. Pro- 
ducers may set the price of an object, but they 
cannot control how the consumer ultimately 
perceives and employs that item (Beaudry et 
al. 1991). Archaeological and documentary 
research can assist in providing context, but 
only if the interpretations are made with re- 
gards to the intrinsic distinctions meaningful 
to the consumer. Aesthetic appeal or other 
considerations are bundled with any object, 
and the choice of assigning importance to the 
different elements of a commodity resides 
with both the consumer and the surrounding 
culture; it is not fundamentally tied to the ob- 
ject's function (Marshall & Maas 1997). 

Accordingly, a modest alteration to the 
problem of reading the ceramic value index 
calculated from the Reddick site is proposed: 
an investigation into the life cycle of the 
house, the occupants, and the artifacts them- 
selves would provide the necessary context 
into which to place the index. For example. 
an investigative technique that compares ce- 
ramic utility wares to table wares could aid in 
interpretation. In the case of the Reddick site. 
the ratio of unrefined ware to the total collec- 
tion (34.8%) is much higher than that found 
in the comparison sites (ranging from approx- 
imately 9-22%) (Huser 1993: Stillwell 1990). 
While unrefined ceramics, like undecorated 
stoneware or redware. are needed to cook and 
store food, and are thus typically present at 
certain levels in all households, refined ware 
is different. It includes types of ceramic ves- 
sels that would likely be used to serve food, 
especially the types known as whiteware. 
pearlware, and yellow ware. The clay body in 
these wares is thinner, with fewer large inclu- 
sions, and the vessels tend to be much more 
highly decorated than unrefined ware. The 
more fragile and decorative serving dishes are 
then used for less practical purposes and may 



124 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



be employed to show status or reflect other 
values important to the owners. A high ratio 
of production to serving vessels could mean 
that the inhabitants of the site had few oppor- 
tunities to host their neighbors, or little incli- 
nation. 

Further research concerning the frequency 
of unrefined ware corroborates the belief that 
the site may have been relatively isolated at 
the time of occupation. Typically, there is 
more evidence of home-canned goods at res- 
idences during this period than appears in the 
archaeological record of the Reddick site 
(William Wepler pers. commun.). The abun- 
dance of redware could lead one to think that 
those within the household did much of the 
food production at the site or nearby. Either 
way, this would suggest at least one behavior 
in which there was a lack of interaction with 
others in the area. 

As for the domestic economy, it is interest- 
ing that economic status is usually inferred 
from the occupation of the male of the house- 
hold, while the ceramic tableware, at the time 
more of a woman's domain, is the feature an- 
alyzed to support the class membership (Cook 
et al. 1996; Wall 1994). Ceramics are then 
"translated" into monetary value and "thus 
converted back into a measure of the status of 
the breadwinner" (Cook et al. 1996). More- 
over, women are said to have orchestrated 
meals as rituals during this period in history 
(Klein 1991). Particular forms of behavior are 
used to create or affirm the values of the fam- 
ily (Wall 1999), as those outside the family 
often view indicators like the lack of table 
manners as an indication of poor upbringing. 
The role of women is unknown at the Reddick 
site; the rare census records show no wife for 
Elisha Reddick's brother, if that is who lived 
at the site. 

Finally, the artifactual remains at the site 
imply that the dwelling might have been 
meant only for temporary or short-term use. 
There is a diversity of domestic refuse, but a 
low number of architectural elements, and, as 
of yet, no privy or outbuildings have been 
confidently located. So, if the structure was 
used for only a short while, by individuals 
new to the area, and without the "refining" 
influence of female companionship (Worthy 
1982), the ceramic value index for the site 
may not accurately reflect the wealth or status 
of the occupants. The low proportion of dec- 



orated vessels could be from isolation, frugal- 
ity, preference, or simply convenience. Selec- 
tion of commercial goods may be based on 
more than levels of wealth. 

In conclusion, historical archaeologists 
must link consumer choices of ceramics, or 
any commodity, to individual acts as well as 
to the function of that particular good. Con- 
sumption of goods extends beyond the eco- 
nomic realm and is found within the social 
domain as well (Cook et al. 1996). The cate- 
gories of material culture constructed by in- 
dividuals constantly shifts over time and 
across space, and people often manipulate the 
meanings of artifacts while negotiating the 
concepts of class and status (Wurst & Fitts 
1999). Relying on one "objective" method, 
then, is inadequate to measure this variation, 
especially if it decontextualizes that which is 
supposed to be studied. The goal of the re- 
search instead should be to look at the pro- 
venience of any suspected prosperity, to see 
the person, as well as the pot. 

LITERATURE CITED 

Adams, W.H. & S.J. Boling. 1989. Status and ce- 
ramics for planters and slaves on three Georgia 
coastal plantations. Historical Archaeology 
23(l):69-96. 

Andrews, S.C. & J.P Fenton. 2001. Archaeology 
and the invisible man: The role of slavery in the 
production of wealth and social class in the Blue- 
grass Region of Kentucky, 1820 to 1870. World 
Archaeology 33:115-135. 

Baugher, S. & R.W. Venables. 1987. Ceramics as 
indicators of status and class in eighteenth-cen- 
tury New York. Pp. 31-54, In Consumer Choice 
in Historical Archaeology. (S.M. Spencer- Wood, 
ed.). Plenum Press, New York. 

Beaudry, M.C., L.J. Cook & S.A. Mrozowski. 
1991. Artifacts and active voices: Material cul- 
ture as social discourse. Pp. 150-191, In The Ar- 
chaeology of Inequality. (R.H. McGuire & R. 
Paynter, eds.). Basil Blackwell, Oxford, England. 

Beteille, A. 1981. The idea of natural inequality. 
Pp. 59-80, In Social Inequality: Comparative 
and Developmental Approaches. (G. Berreman & 
K. Zaretsky, eds.). Academic Press, New York. 

Bourdieu, P. 1984. A Social Critique of the Judg- 
ment of Taste. (R. Nice, transl.). Harvard Uni- 
versity Press, Cambridge. 

Cook, L.J., R. Yamin & J.P. McCarthy. 1996. 
Shopping as meaningful action: Toward a redef- 
inition of consumption in historical archaeology. 
Historical Archaeology 30(4):50-65. 

Cowen, M.V.B. 1866. The Indiana State Gazetteer 



VANDERVEEN— PEOPLE, POTS, AND PROSPERITY 



125 



and Shippers' Guide for 1866-7. Rosser, Spring 
& Cowen. Lafayette, Indiana. 

Cressey, P.J., J.F. Stephens, S.J. Shephard & B.H. 
Magrid. 1982. The core-periphery relationship 
and the archaeological record in Alexandria, Vir- 
ginia. Pp. 41-62, In Archaeology of Urban 
America: The Search for Pattern and Process. 
(R.S. Dickens, Jr., ed.). Academic Press, New 
York. 

Garrow, PH. 1987. The use of converging lines of 
evidence for determining socioeconomic status. 
Pp. 217-232, In Consumer Choice in Historical 
Archaeology. (S.M. Spencer- Wood, ed.). Plenum 
Press, New York. 

Huser, Jr. W.A. 1993. Archaeology and Socioeco- 
nomic Evaluation of the William Conner House 
Site (12H608): A 19 th Century Rural Residence 
in Hamilton County, Indiana. Master's thesis, 
Ball State University, Muncie, Indiana. 

Klein, T.H. 1991. Nineteenth-century ceramics and 
models of consumer behavior. Historical Archae- 
ology 25(2):77-91. 

Lofstrom, T, J.R Tordoff & D.C. George. 1982. A 
seriation of historic earthenware in the Midwest, 
1780-1870. Minnesota Archaeologist 41(1 ):3- 
29. 

Majewski, T & M.J. O'Brien. 1987. The use and 
misuse of nineteenth-century English and Amer- 
ican ceramics in archaeological analysis. Pp. 97- 
209, In Advances in Archaeological Method and 
Theory. (M.B. Schiffer, ed.). Academic Press, 
New York. 

Marshall, Y. & A. Maas. 1997. Dashing dishes. 
World Archaeology 28(3):275-290. 

McBride, W.S. & K.A. McBride. 1987. Socioeco- 
nomic variation in a late antebellum southern 
town: The view from archaeological and docu- 
mentary sources. Pp. 143-161, In Consumer 
Choice in Historical Archaeology. (S.M. Spen- 
cer-Wood, ed.). Plenum Press, New York. 

Miller, G.L. 1991a. A revised set of CC Index Val- 
ues for classification and economic scaling of 
English ceramics from 1787 to 1880. Historical 
Archaeology 25(1): 1-25. 

Miller, G.L. 1991b. Classification and economic 
scaling of 19 th century ceramics. Pp. 37-58, In 
Approaches to Material Culture Research for 
Historical Archaeologist. (G.L. Miller, O.R. 
Jones, L.A. Ross & T Majewski, eds.). The So- 
ciety for Historical Archaeology, Ann Arbor. 

Miller, G.L. 1980. Classification and economic 
scaling of 19 th century ceramics. Historical Ar- 
chaeology 14(1): 1-40. 

Miller, G.L. & S.D. Hurry. 1983. Ceramic supply 
in an economically isolated frontier community: 
Portage County of the Ohio Western Reserve, 
1800-1825. Historical Archaeology 1 7(2): 80- 
92. 



Nickolai, C.A. 2003. Class and gender in nine- 
teenth-century rural Michigan: The Merriman- 
Sharp hillside farm. Historical Archaeology 
37(4):69-83. 

Powers, M.G. 1982. Measures of socioeconomic 
status: An introduction. Pp. 1-28. In Measures 
of Socioeconomic Status: Current Issues. (M.G. 
Powers, ed.). AAAS Selected Symposia Series, 
vol. 81. American Association for the Advance- 
ment of Science, Washington, DC. 

Schiffer, M.B. 1972. Archaeological context and 
systematic context. American Antiquity 48:675- 
706. 

Spencer- Wood, S.M. & S.D. Heberling. 1987. 
Consumer choices in white ceramics: A compar- 
ison of eleven early nineteenth-century sites. Pp. 
55-84, In Consumer Choice In Historical Ar- 
chaeology. (S.M. Spencer- Wood, ed.). Plenum 
Press, New York. 

Stillwell, L.N. 1990. The Richardville/LaFontaine 
House and the Troyer Site: A Material Culture 
Comparison. Master's thesis, Ball State Univer- 
sity, Muncie, Indiana. 

Stine, L. 1992. Social differentiation on the farm- 
stead. Pp. 103-110, In Exploring Gender 
Through Archaeology: Selected Papers from the 
1991 Boone Conference. (C. Claassen. ed. ). Pre- 
history Press. Madison, Wisconsin. 

Stine, L. 1990. Social inequality and turn-of-the- 
century farmsteads: Issues of class, status, eth- 
nicity, and race. Historical Archaeology 24(4i: 
37-49. 

Sulgrove, B.R. 1884. History of Indianapolis and 
Marion County, Indiana. L.H. Everts. Philadel- 
phia, Pennsylvania. 

VanderVeen, J.M. 2001. Archaeological Investi- 
gation of Site 12Ma648 at Fort Harrison State 
Park, Marion County, Indiana. Report to the In- 
diana Division of Historic Preservation and Ar- 
chaeology, Indianapolis. Indiana. 

Wall, D.D. 1999. Examining gender, class, and eth- 
nicity in nineteenth-century New York City. His- 
torical Archaeology 33(1 ): 102-1 17. 

Wall, D.D. 1994. The Archaeology ot Gender: 
Separating the Spheres in Urban America. Ple- 
num, New York. 

Worthy. L.H. 1982. Classification and interpreta- 
tion of late nineteenth- and early twentieth-cen- 
tury ceramics. Pp. 329-360. /;/ Arehaeolog\ of 
Urban America: The Search for Pattern and Pro- 
cess. (R.S. Dickens. Jr.. ed.). Academic Press. 
New York. 

Wurst, L. & R.K. Fitts. 1999. Introduction: \Yh\ 
confront class'? Historical Archaeolog\ 33( 1 ): 
1-6. 

Manuscript received 16 May 2007, revised 8 Sep- 
tember 2007. 



2007. Proceedings of the Indiana Academy of Science (1 16): 126-138 



VEHICLE IMPACTS ON VEGETATION COVER AT CAMP 

ATTERBURY, INDIANA: PART 1. INITIAL IMPACTS AND 

VEGETATION RECOVERY 

Alan B. Anderson: U.S. Army Engineer Research Development Center, Champaign, 
Illinois 61822 USA 

Paul D. Ayers: University of Tennessee, Knoxville 37996-4531 USA 

Heidi Howard: U.S. Army Engineer Research Development Center, Champaign, 
Illinois 61822 USA 

Kenneth D. Newlin: Indiana National Guard, Indianapolis, Indiana 46124-5000 USA 

ABSTRACT. Geographic Positioning System (GPS) based vehicle tracking systems were installed on 
three military vehicle types (M88 tank recovery vehicle, M35A3 cargo truck, and M1009 utility cargo 
vehicle) at Camp Atterbury, Indiana, to assess the impact of vehicle traffic on vegetation. Vehicle tracking 
systems recorded the position of each vehicle every second. Instrumented vehicles were driven through 
courses of varying velocities and turning radii. GPS position data were used to calculate vehicle velocities 
and turning radii throughout the course. Vegetation damage along vehicle tracks was recorded immediately, 
5 months (end of the first growing season) and 12 months after tracking. Vegetation damage was quantified 
by both the amount of vegetation lost and the area impacted. Vehicle type, turning radius (TR), velocity 
(V), and TRV interaction were found to significantly affect all vegetation damage measures. The tracked 
M88 tank recovery vehicle caused more vegetation loss than either of the wheeled vehicles (M35A3 cargo 
truck and Ml 009 utility cargo vehicle). Decreasing turning radius increased vegetation loss for all vehicles. 
Increased vegetation loss associated with turning was a function of both greater vegetation loss within the 
track and a wider tracked area. Power equations using only turning radius and vehicle type as independent 
variables predicted vegetation damage measures with R 2 values ranging from 0.822 to 0.933. A critical 
turning radius between 15-20 m differentiated turning radii with relatively high vegetation loss as com- 
pared to straight-line tracking. Recovery of vegetation cover to pretreatment levels ranged from approx- 
imately 6-12 months, depending on impact treatment. 

Keywords: Vehicle impacts, off-road, vegetation impact, impact assessment 



The Department of Defense is responsible mation, decreased macropore space, restricted 
for administering more than 10 million hect- water movement, reduced soil strength and 
ares of federally-owned land in the United structure, and physical damage to root sys- 
States. Military training, especially vehicular terns. The immediate physical disturbance af- 
training, is an intensive land use that can neg- fects not only vigor and mortality of current 
atively impact soil and vegetation (Goran et vegetation but also the rate of vegetation re- 
al. 1983; Demarais et al. 1999). Numerous covery (Thurow et al. 1995; Prosser et al. 
studies have investigated the effects of vehicle 2000; Lovich & Bainbridge 1999). 



traffic on soil and vegetation (Johnson 1982 
Payne et al. 1983; Webb & Wilshire 1983 
Prose 1985; Braunack 1986; Wilson 1988 



The amount of vegetation damage resulting 
from vehicle traffic is determined by vehicle 
characteristics and site conditions. Site con- 



Shaw & Diersing 1990; Ayers 1994; Trumbull ditions that are important in determining veg- 

et al. 1994; Demarais et al. 1999; Ayers et al. etation damage include soil type, soil mois- 

2000; Hirst et al. 2000; Milchunas et al. 2000; ture, slope, vegetation type, and plant growth 

Hirst et al. 2003). Potential consequences of stage (Payne et al. 1983; Wilson 1988; Thu- 

vehicle traffic are loss of vegetation, exposed row et al. 1995). Static vehicle characteristics 

soil, increased erosion, soil compaction, soil important in determining vegetation damage 

puddling, displaced surface horizons, rut for- include surface contact area, surface pressure, 

126 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 



127 



total weight, and track design (Ayers et al. 
1994; Ayers et al. 2000). Dynamic vehicle 
properties important in determining vegetation 
damage include speed, turning radius, and 
driving pattern (Braunack 1986; Ayers et al. 
2000; Halvorson et al. 2001). 

A number of vehicle impact studies have 
assessed the impact of vehicle traffic on veg- 
etation without characterizing the vehicles or 
activities that caused the disturbance (Johnson 
1982; Shaw & Diersing 1990; Milchunas et 
al. 1999; Milchunas et al. 2000). Typically 
these studies compared vegetation on relative- 
ly large tracked and untracked sites. Usually 
vegetation impacts were assessed after an un- 
known period of time and/or by an unknown 
combination of vehicles using the study sites. 
Goran et al. (1983) reported a sequence of ve- 
hicle-induced effects on vegetation ranging 
from minor vegetation disturbance from ap- 
parent one-time only traffic, to increased bare 
ground and loss of sensitive plant species for 
occasional to frequent use, to complete vege- 
tation loss and soil movement for frequently 
and intensely used areas. The authors also re- 
ported local site damage resulting from single 
turns as similar to intensely-used areas. While 
these studies are useful for quantifying the cu- 
mulative impact of tracking on vegetation, 
they provide little quantitative information 
that relates type and level of vehicle use to the 
amount of vegetation damage. 

A number of studies related the impact of 
specific vehicles to a specified level of use 
(Payne et al. 1983; Wilson 1988; Thurow et 
al. 1995; Prosser et al. 2000; Grantham et al. 
2001). Typically these replicated studies in- 
volved repeated tracking of study plots with a 
specific vehicle. Thurow et al. (1995) assessed 
the impact of 1,4, and 10 straight-line passes 
of a 22.5 metric ton (t) tracked M2 Bradley 
Infantry fighting vehicle on vegetation during 
wet and dry soil conditions. Similarly, Payne 
et al. (1983) assessed the impact of 2, 8, and 
32 straight-line passes of a 2.2 metric ton (t) 
wheeled Chevy Blazer on vegetation over a 
period of time to quantify temporal effects of 
tracking on vegetation. While dynamic vehi- 
cle properties like velocity were not always 
reported in these studies, the vehicle's dynam- 
ic properties generally maintained constant 
throughout the disturbance regime. However, 
it is not clear if the dynamic vehicle properties 
used in these studies are representative of ac- 



tual site use or include the most damaging ve- 
hicle activities. For example, Grantham et al. 
(2001) quantified the impact of 1, 2. 4, and 8 
straight-line passes of a 62.6 metric ton ftj 
tracked M 1 A2 Abrams combat tank operating 
at 48 km per hr on vegetation. While turns 
were not included in the study design, the au- 
thors observed that single turns caused more 
site damage than the maximum 8 straight-line 
passes used in the study. Similarly, Prosser et 
al. (2000) assess the impact of 0. 37. and 74 
straight-line passes of a 10.9 metric ton (t) 
tracked Ml 13 personnel carrier on vegetation. 
While only straight-line tracking was included 
in the study, the authors also noted that turns 
caused substantially more damage than an\ 
straight-line tracking treatments. Belcher and 
Wilson (1989), while studying leafy spurge 
infestations on military lands, found the ma- 
jority of infestations associated with vehicle 
turns rather than straight-line tracking because 
of the amount of bare soil exposed during 
turns. Wilson (1988), while assessing the im- 
pact of 4 to 35 straight-line passes of a combat 
tank on vegetation, noted a single vehicle turn 
had an obvious and immediate impact on veg- 
etation by exposing bare ground and was 
much more severe than straight-line tracking. 

Another group of vehicle impact studies in- 
cluded both straight-line tracking and turning 
in the study design (Braunack 1986; Watts 
1998; Halvorson et al. 2001). Halvorson et al. 
(2001) assessed the impact of 1. 2. 4. and S 
straight-line passes and turns of a M1A2 
Abrams combat tank on vegetation. Watts 
(1998) assessed the impact of a 60.8 t tracked 
Ml combat tank during straight-line tracking 
and turns on vegetation. In both studies, turns 
caused substantially more vegetation damage 
within the track than straight-line tracking. 
Braunack (1986) measured rut width o\ single 
pass straight-line and turning tracking of an 
Ml 13. Turns damaged almost twice the area 
of straight-line tracking. These studies clearly 
quantified the impacts of both straight-line 
tracking and turns. However, these studies did 
not specify the type o\' turns included in the 
study. 

Ayers (1994) assessed the impact of an 
Ml 13 turning at radii o\' 30, 12. 8, and 4 m 
on vegetation damage. Sharper turns caused 
greater vegetation loss. However, the number 
of turning radii treatment was insufficient to 
accurately determine the shape of the relation- 



128 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



ship. Specifically, the data were insufficient to 
determine if there was a threshold turning-ra- 
dius such that sharper turns disproportionately 
cause greater vegetation loss and if this 
threshold turning-radius was vehicle specific. 
While a number of studies have assessed 
the impact of vehicle type, turning-radius, and 
number of vehicle passes on vegetation dam- 
age, little is know about the relative impact of 
these factors on vegetation loss. This lack of 
understanding results from studies that ex- 
amine unique impact factors at different sites. 
The objective of our study was to quantify the 
impact of three military vehicles (M88 tank 
recovery vehicle, M35A3 cargo truck, and 
Ml 009 utility cargo vehicle) on vegetation 
cover loss and determine the relative impact 
of vehicle type, turning-radius, and velocity 
on vegetation cover loss. A secondary objec- 
tive of this study was to estimate vegetation 
recovery times for the study site. 

METHODS 

Study site. — The study was conducted at 
Camp Atterbury, Indiana, an Army National 
Guard training facility that encompasses 144 
km 2 in central Indiana (Tetra Tech 2000). Prior 
to establishment in 1942 as a military instal- 
lation, historic land use consisted of inter- 
mixed farms and woodlands. The terrain rang- 
es from fairly flat historically-agricultural land 
forms on the north, rolling hills in the central 
portion, to steep hills and valleys in extreme 
southern portion. Elevations range from 195— 
297 m above sea level. Temperatures range 
from -29° C in winter to 43° C in summer. 
The last killing frost averages 27 April and 
first killing frost averages 10 October. Sum- 
mer months are characterized as hot, with pro- 
longed dry conditions. Precipitation is distrib- 
uted fairly evenly throughout the year, so there 
is no pronounced wet or dry season. Annual 
precipitation averages 104 cm, with 43 cm as 
snow. Vegetation ranges from open grasslands 
to hardwood forests. 

The study site is located in training area 3A 
(39.33° N, 85.99° W). The study site was se- 
lected because it is representative of many ar- 
eas on the installation used by vehicles. Study 
site soils were classified as a Genesee (Wig- 
ginton & Marshall 2004). The Genesee series 
is a fine-loamy, mixed, superactive, mesic 
Fluventic Eutrudept. Vegetation at the study 
site consists of native and introduced forbs 



and grasses. Common plant species occupying 
the site include annual ragweed (Ambrosia ar- 
temisiifolia L.), giant ragweed (Ambrosia tri- 
fida L.), common milkweed (Asclepias syria- 
ca L.), mustards (Brassica spp.), trumpet 
creeper (Campsis radicans (L.) Seem, ex Bu- 
reau), thistle (Cirsium spp.), Queen Anne's 
lace (Daucus carota L.), blue boneset (Eu- 
patorium coelestinum L.), morning-glory (Ip- 
omoea spp.), fescue (Lolium spp.), white 
sweet clover (Melilotus alba Medikus), timo- 
thy (Phleum pratense L.), raspberry/blackber- 
ry (Rubus spp.), cereal rye (Secale cereale L.), 
Canada goldenrod (Solidago canadensis L.), 
eastern poison ivy (Toxicodendron radicans 
(L.) Kuntze), red clover (Trifolium pratense 
L.), and wild grape (Vitis spp.). Plant nomen- 
clature follows USDA, NRCS (2004). 

Study design. — A field study was conduct- 
ed on 24 July 2001 using three vehicles: M88 
tank recovery vehicle, M35A3 cargo truck, 
and Ml 009 utility cargo vehicle (Figs. 1-3). 
The M88 tank recovery vehicle is a tracked 
vehicle that is 8.52 m long, 3.40 m wide, 3.10 
m high with a weight of 50.8 t. The M88 track 
width is 71.1 cm with pads that are 26.7 cm 
wide by 16.5 cm long. The M35A3 cargo 
truck is a six-wheeled vehicle with a 4.50 m 
total wheelbase, 2.40 m outside to outside 
width, and weight of 3.5 t. The M35A3 tire 
height is 107.0 cm with a tread width of 23.5 
cm. Tire pressure was 345 kPa during the 
study. The Ml 009 utility cargo vehicle (sim- 
ilar to a Chevrolet CD- 10506) is a four- 
wheeled vehicle with a 2.70 m wheelbase, 
1.40 m width, and a weight of 2.4 t. The 
Ml 009 tire height is 78.7 cm with a tread 
width of 26.7 cm. Tire pressure was 345 kPa 
during the study. 

Each vehicle drove a systematically 
planned course (spiral) within a randomly lo- 
cated treatment plot. Each vehicle tracked 
three treatment plots. Each spiral course with- 
in a treatment plot consisted of a section of 
straight-line travel followed by a section of 
constantly decreasing turning radius. The spi- 
ral was complete after reaching the vehicle's 
minimum turning radius. One spiral for each 
vehicle was traversed at a slow, medium, or 
fast velocity. A preliminary spiral path was 
marked in each treatment plot. However, ve- 
hicle drivers were allowed to deviate from the 
marked path to maintain a constant velocity. 
The fast velocity spiral represents the fastest 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 1 



129 




M3 5A3 



3 

i j 






K \ *' 








**;.-:> 



M1009 

Figure 1-3. — Three vehicle types used in tracking study include the MSS tank recovers vehicle t Fie. 
1), the M35A3 cargo truck (Fig. 2), and the Ml 009 utility cargo vehicle (Fig. 3). 



velocity the vehicle could safely be driven for varied between vehicle types due to vehicle 

the site conditions. Maximum velocities used design capabilities. 

in this study are faster than a vehicle would Each vehicle was equipped with a vehicle 

typically be driven during a training event un- tracking system (Avers et al. 2000). The ve- 

der similar conditions. Maximum velocities hide tracking system consisted of a 12 chan- 



130 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



nel Trimble® GPS receiver with Omnistar® 
differential correction that logged location in- 
formation, a data storage device and a power 
source. Vehicle position was recorded every 
second. Vehicle dynamic properties (velocity, 
turning radius) were calculated from the ve- 
hicle tracking system position data using the 
methods of Ayers et al. (2000). 

Sampling methodologies. — Initial vehicle 
impacts were measured immediately after 
tracking as disturbed width and impact sever- 
ity. All measurements were made along the 
inner track of each spiral. The first sample lo- 
cation was randomly located within the first 
10 m of the straight-line tracking portion of 
each spiral. Subsequent samples were system- 
atically located every 5 m along the vehicle 
track resulting in approximately 20 sample 
points per spiral. Each sample point consisted 
of a paired subplot. One subplot was located 
within the track and the other subplot was 0.5 
m adjacent to and on the inside of the track 
in undisturbed vegetation. 

Disturbed width (DW) was measured per- 
pendicular to the vehicle track and encom- 
passed the area where soil and/or vegetation 
were impacted by the vehicle tire/track. The 
disturbed width included areas where vegeta- 
tion was flattened but not killed and areas 
where soil was removed or piled up. 

Vegetation cover was estimated using a line 
transect established perpendicular to the track 
(same measurement line as the disturbed 
width measurement). A second line transect 
was established perpendicular to the track and 
a 0.5 m from the track. Each undisturbed 
paired plot was located to the inside of the 
spiral in untracked vegetation. For each line 
transect (within track and adjacent to the 
track), bare ground was visually estimated and 
reported as a percent of plot length. One ob- 
server estimated all sample plots. 

Impact severity (IS) was defined as the per- 
cent vegetation cover within the disturbed ve- 
hicle track that was removed by the vehicle 
resulting in exposed bare soil. Impact severity 
was calculated by subtracting the disturbed 
vehicle track subplot vegetation cover esti- 
mates from the non-tracked subplot vegetation 
cover estimates. Impact severity ranged from 
(no vegetation loss) to 100 (complete veg- 
etation loss). 

Payne et al. (1983) noted that single pass 
straight-line tracking by light vehicles often 



resulted in some crushed vegetation laying 
horizontal to the soil surface. They noted that 
some of this horizontal vegetation was dead 
while other vegetation was still viable and re- 
covered within a few weeks. To help interpret 
recovery data from our study, we recorded the 
types of damage observed at each measure- 
ment location. 

Vegetation impacts were measured a second 
time at the end of the growing season during 
which tracking treatments occurred. Due to an 
extremely warm fall and late winter, 14 Dec 
2001 represented the end of the growing sea- 
son. Impact severity was measured using the 
same methods as described for the original 
sampling. Impact severity was estimated for 
the original disturbed track width. 

Vegetation impacts were measured a third 
time on 17 July 2002. Impact severity was 
measured using the same methods as de- 
scribed for the original sampling. Impact se- 
verity was estimated for the original disturbed 
track width. In addition, percent ground cover 
of forbs, grasses and total vegetation was re- 
corded in the disturbed track and adjacent to 
the track. Cover of forbs, grasses, and total 
vegetation was visually estimated for a plot 
centered in the vehicle track and 0.5 m from 
the outside edge of the track. The dimension 
of each square plot was the track width. Cover 
estimates were made independent of other 
vegetation components and were not intended 
to sum up to total vegetation cover. 

Soil moisture was determined gravimetri- 
cally on the day of tracking for the to 10.16 
cm depth using methods of Gardner (1986). 
Soil samples were dried at 105° C for 48 h in 
a conventional oven. Water content was cal- 
culated on a mass basis as a percentage of dry 
soil. Air temperature was recorded 1 m above 
the soil surface at the beginning and end of 
tracking treatments. 

Statistical analysis. — Vehicle dynamic 
properties (velocity, turning radius) were cal- 
culated from the GPS vehicle tracking system 
position data using the methods of Ayers et 
al. (2000). Vehicle dynamic properties were 
calculated for each field data sample location 
within each tracking course. 

Vehicles impact a larger area (disturbed 
width) during turns (Ayers 1994; Ayers et al. 
2000). A cumulative impact (CI) measure was 
used to quantify overall vehicle impacts that 
incorporated both severity of vegetation dam- 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 1 



131 



age and area affected. Cumulative impact was 
calculated as the product of the impact width 
and impact severity on a sample plot basis. 

Vegetation impact data were analyzed using 
the Proc Reg procedure with the stepwise op- 
tion of SAS® (SAS Institute Inc., Cary, North 
Carolina) to determine which factors signifi- 
cantly contributed to vegetation damage. Ve- 
hicle type, turning radius, velocity, and all in- 
teractions were included in the stepwise 
regression analysis as independent variables. 
Impact severity, impact width, and cumulative 
impact were each used as the dependent var- 
iable of the stepwise analysis. All dependent 
variables were transformed using a log trans- 
formation. 

After significant model variables were de- 
termined from the stepwise regression analy- 
sis, nonlinear regression analyses were con- 
ducted using raw dependent and independent 
variables. Impact severity, impact width, and 
cumulative impact were each used as the de- 
pendent variable. Independent variables in- 
cluded in the model were variables found to 
be significant in the stepwise regression anal- 
ysis and that also accounted for a meaningful 
amount of the variation in dependent vari- 
ables. Independent variables included in the 
model were turning radius and vehicle type. 
Vehicle type was included in the model by the 
use of dummy variables such that d x = and 
d 2 = for the M88 tank recovery vehicle, d x 
= 1 and d 2 = for the Ml 009 utility cargo 
vehicle and d x = and d 2 = 1 for the M35A3 
cargo truck Model parameters were estimated 
using SAS® Proc Model. The general model 
has the form 

y = [a + (a^dO + (a 2 -d 2 )].x^ +{b ^ + ^- d ^ 

where y is the dependent impact variable (IS, 
DW, or CI), x is the independent variable for 
vehicle dynamic property, d x is the dummy 
variable to indicate the Ml 009 utility cargo 
vehicle, d 2 is the dummy variable to indicate 
the M35A3 cargo truck, a is the intercept co- 
efficient for the M88 tank recovery vehicle, a { 
is the intercept shift coefficient for the Ml 009 
utility cargo vehicle, a 2 is the intercept shift 
coefficient for the M35A3 cargo truck, b is 
the slope coefficient for the M88 tank recov- 
ery vehicle, b x is the slope shift coefficient for 
the M35A3 cargo truck, b 2 is the slope shift 
coefficient for the Ml 009 utility cargo vehi- 
cle. 



Model parameters and R 2 values were esti- 
mated with all model terms included in the 
model. Model terms were then systematically 
removed in a stepwise manner to determine 
the simplest model that adequately character- 
ized the relationship between impact measure 
and vehicle dynamic properties. The adjusted 
R 2 value of the model was used as the criteria 
for model selection. 

Site recovery times were estimated as the 
number of months required for vegetation 
cover to in the disturbed subplots to reach un- 
disturbed subplot vegetation cover levels. Re- 
covery times are based solely on total vege- 
tation cover. 

Differences in forb and grass cover between 
tracked and untracked paired plots were cal- 
culated. An analysis of variance for forb and 
grass cover was conducted using track and ve- 
hicle type. Track types are curved (<20 m ra- 
dius) and straight (>20 m radius). Vehicle 
types are M88 tank recovery vehicle, M35A3 
cargo truck, and Ml 009 utility cargo vehicle. 
Tukey's Honestly Different Test was used to 
test differences among means. 

RESULTS 

Initial vehicle impacts. — Vegetation cover 
for undisturbed sample points averaged 100 
indicating a densely vegetated site. Soil water 
content in the to 10.16 cm layer, at the time 
of tracking, averaged 59c. This moisture con- 
tent represents a relatively dry soil condition 
typical for many maneuver activities at this 
site. Air temperature ranged from 34—36° C 
during tracking treatments. Average M88 tank 
recovery vehicle course velocities ranged 
from 5.04 km per hr for the slow course to 
16.92 km per hr for the fast course. Average 
M35A3 cargo truck course velocities ranged 
from 8.28 km per hr for the slow course to 
15.48 km per hr for the fast course. Average 
M1009 utility cargo vehicle course velocities 
ranged from 8.28 km per hr for the slow 
course to 24.12 km per hr for the fast course. 

For all vehicles, straight-line tracking gen- 
erally resulted in flattened vegetation with lit- 
tle to no rutting or exposed soil. Shearing of 
vegetation and horizontal movement of soil 
primarily occurred for tracked vehicles at 
smaller turning radii. 

Impact severity, disturbed width, and cu- 
mulative impact increased exponentialh with 
decreasing turning radii (Fies. 4-6). The M88 



132 



s 



125 



100 



75 



50 



25 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 

-0.281 



IS = (182.7827 -(145.219*d,)-(139. 67* d, ))TR 




O M88 observed 
□ M35A3 observed 
A M1 009 observed 

M88 predicted 

M35A3 predicted 

M1 009 predicted 




AA nigy^ A D AA D 



25 



50 



75 



100 



125 



150 



Turning radius (m) 



Figure 4. — Impact Severity (IS) as a function of turning radius (TR) for the M88 tank recovery vehicle, 
M35A3 cargo truck and Ml 009 utility cargo vehicle. Parameterized power equation for fitted lines pro- 
vided at the top of the graph. Variables d x and d 2 are dummy variables that account for vehicle type. The 
dummy variables d ] and d 2 have the value of J, = d 2 = for the M88, d x = and d 2 = 1 for the M35A3, 
and d { = 1 and d 2 = for the Ml 009. R 2 fit for the equation is 0.924. 



caused substantially more damage than either 
of the other two vehicles. For wheeled vehi- 
cles (M35A3 cargo truck, and Ml 009 utility 
cargo vehicle), disturbed width increased as 
turning radii decreased because rear wheels 
did not track directly behind front wheels dur- 
ing turns. For the tracked vehicle (M88 tank 
recovery vehicle), disturbed width increased 
with decreasing turning radii because the ve- 
hicle would pivot on a portion of the track 
causing the rear portion of the track to slide 
outward. This sliding action resulted in veg- 
etation and soil being scraped out of the 
tracked area. 

Impact measures increased suddenly at 
turning radii less than approximately 15-20 m 
for all vehicle types. Despite drastically dif- 
ferent static vehicle design characteristics, the 
critical turning radii for site damage were very 
similar among vehicles. 

Natural logarithm transformations of all im- 
pact measures (impact severity, disturbed 
width, and cumulative impact) and vehicle dy- 



namic properties (turning radius and velocity) 
resulted in linear relationships between depen- 
dent and independent variables. Figure 7 
shows a typical relationship between trans- 
formed impact measure and vehicle dynamic 
property. 

In the stepwise regression, vehicle type, 
turning radius, velocity, and turning radius by 
velocity interaction were found to significant- 
ly (P < 0.10) affect impact severity, disturbed 
width, and cumulative impact. Model R 2 val- 
ues were 0.789, 0.743, and 0.853 for impact 
severity, disturbed width, and cumulative im- 
pact, respectively. Even though velocity and 
turning radius by velocity interaction model 
terms were significant, they accounted for lit- 
tle variation in impact measures after vehicle 
type and turning radius were already included 
in the model. Partial R 2 values for velocity and 
turning radius by velocity interaction model 
terms never exceeded 0.044 after vehicle type 
and turning radius were already included in 
the model. 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 1 



133 



250 



200 



150 



■a 

T3 
<D 

■g 100 



Q 



50 



DW = (324.0867 - (219.108 * d } ) - (156.59 * d 2 ))TR 



-0.3326 



O 
O 




O M88 observed 
□ M35A3 observed 
A M1 009 observed 

M88 predicted 

M35A3 predicted 

M1009 predicted 



25 



50 75 1 00 

Turning radius (m) 



125 



150 



Figure 5. — Disturbed Width (DW) as a function of turning radius (TR) for the M88 tank recovery 
vehicle, M35A3 cargo truck and Ml 009 utility cargo vehicle. Parameterized power equation for fitted 
lines provided at the top of the graph. Variables d x and d 2 are dummy variables that account for vehicle 
type. The dummy variables J, and d 2 have the value of d x = d 2 = for the M88, d x = and d z = 1 for 
the M35A3, and d x = 1 and d 2 = for the Ml 009. R 2 fit for the equation is 0.822. 



When analyzing only straight-line tracking 
data (straight-line tracking is defined as turn- 
ing radius greater than 20 m), velocity did not 
significantly (P < 0.05) affect impact severity, 
disturbed width, or cumulative impact. Only 
vehicle type significantly affected impact 
measures. 

Variation in vegetation loss in straight-line 
portions of the vehicle course for individual 
vehicles appears to be due to small-scale var- 
iation in surface roughness. Most straight-line 
tracking resulted in compression type vege- 
tation damage and little exposed soil for all 
vehicle types. Where bare ground was ex- 
posed, the track or wheel frequently passed 
over small depressions or hills. Hills and de- 
pressions approximately !/ 8 the height or depth 
of the vehicle track or wheel were associated 
with increased bare ground. At these locations 
the track or wheel came in contact with the 
soil surface at an angle rather than parallel to 
the soil surface. Though not quantified in this 



study, observations indicate that sites with 
greater surface roughness resulted in greater 
vegetation loss during straight-line tracking. 

Modeling vehicle impacts. — Natural loga- 
rithm transformations of impact measures and 
vehicle dynamic properties resulted in Linear 
relationships between the variables. For all 
impact measures, vehicle type and turning ra- 
dius were the first variables included in mod- 
els during stepwise regression analysis. Ad- 
ditional model terms, though significant. 
accounted for little additional variation in the 
dependent variables. Based on these findings. 
we chose to develop models to explain the 
relationship between vehicle dynamic prop- 
erties and impact measures using a power 
equation with vehicle type and turning radius 
as model terms. Table 1 shows alternative 
forms of the power equation evaluated and 
corresponding R : values for each of the three 
impact measures. Model I represents a model 
that accounts for turning radius but does not 



134 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



250 



200 



£ 

co 150 

Q. 

£ 

> 

H 100 

3 

E 



o 



50 



CI = (538 .2323 - (50 1 .464 * d x ) - (473 .774 * d, ))TR 



-0.5764 




o 


M88 observed 


□ 


M35A3 observed 


A 


M1 009 observed 




— M88 predicted 







- M35A3 predicted 




-M1 009 predicted 





25 



50 



75 



100 



125 



150 



Turning radius (m 



Figure 6. — Cumulative impact (CI) as a function of turning radius (TR) for the M88 tank recovery 
vehicle, M35A3 cargo truck and Ml 009 utility cargo vehicle. Cumulative impact was calculated as the 
product of the impact width and impact severity. Parameterized power equation for fitted lines provided 
at the top of the graph. Variables J, and d 2 are dummy variables that account for vehicle type. The dummy 



variables d } and d 2 have the value of d ] = d 2 = for the M88, d x 
d, = 1 and d 2 = for the Ml 009. R 2 fit for the equation is 0.924. 



and d 2 



1 for the M35A3, and 



differentiate between vehicle types. Model 4 
represents a model that accounts for vehicle 
type in both the "a and '£>' terms of the power 
model. Models 2 and 3 represent models that 
account for vehicle type in only the '#' or '&' 
term of the model, respectively. Model 2, 
which only accounts for vehicle type in the 
'<:/' term of the model was found to best fit 
the data for all impact measures or fit the data 
nearly as well as more complex models. Table 
2 shows parameter estimates, standard errors, 
and R 2 values for the selected power models 
used to characterize the relationship between 
vehicle dynamic properties and impact mea- 
sures. Model R 2 values exceeded 0.82 for all 
impact measures using only vehicle type and 
turning radius as independent variables. Fig- 
ures 4-6 show model predictions and field ob- 
servations for each vegetation impact mea- 
sure. 

Vegetation recovery. — Recovery times for 
total vegetation canopy cover ranged from 



less than 5 months to about 12 months for all 
vehicle and impact types (Fig. 8). Recovery 
times for the Ml 009 utility cargo vehicle 
where shorter than the M35A3 cargo truck, 
which were shorter than the M88 tank recov- 
ery vehicle. Recovery times were similar for 
straight-line tracks and curves despite differ- 
ences in initial impacts. 

Though total vegetative cover returned to 
pretreatment levels after approximately one 
year for the most severely impacted areas, 
other measures of site condition might not 
have fully recovered in this period. To eval- 
uate if differences in species composition ex- 
isted between impacted areas and adjacent ar- 
eas, we measured percent grass and forb 
cover. No significant difference (P < 0.05) in 
grass cover between tracked and untracked ar- 
eas was evident one year after tracking for any 
vehicle type. No significant difference (P < 
0.05) in forb cover between tracked and un- 
tracked areas was evident for the Ml 009 util- 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 1 



135 



O M88 

D M35A3 

A M1009 
M88 linear 
M35A3 linear 
M1009 linear 




ln(turning radius) 

Figure 7. — Relationship between logarithmic transformation of both cumulative impact and turning 
radius data by vehicle type. Lines show best fit for linear relationship between variables. R 2 fit for M88 
tank recovery vehicle, M35A3 cargo truck and Ml 009 utility cargo vehicle are 0.90, 0.40. and 0.49. 
respectively. 



ity cargo vehicle or M35A3 cargo truck. How- 
ever, the M88 tank recovery vehicle tracks 
with turning radii less than 20 m had 17.7% 
less forb cover than untracked areas (signifi- 
cant at P < 0.05). Visual observation of plots 
indicated that in compressed vegetation areas 
(primarily straight-line tracking) many of the 
same individual plants regenerated. In turns 
where the track or wheel exposed soil and cre- 
ated piles of soil adjacent to the track, new 
plants colonized the disturbed area. New 
plants were from seed, rhizomes, or stolons 
remaining in the soil. 

DISCUSSION 
Results of this study are consistent with 
previously published studies. Vehicle turns 



caused more damage than straight-line travel 
(Braunack 1986; Belcher & Wilson 1989: Av- 
ers 1994; Watts 1998; Prosser et al. 2000: Hal- 
vorson et al. 2001). Smaller vehicle turning 
radii caused more vegetation loss than larger 
turning radii (Ayers 1994). We also found that 
velocity and turning radius by velocity inter- 
action significantly affects vegetation loss that 
has not previously been reported in the liter- 
ature. 

The vehicles tested in our study had critical 
turning radii between 15 and 20 m where veg- 
etation loss dramatically increased at smaller 
turning radii. This critical turning radius is 
similar to those reported for other sites (Hau- 
gen et al. 2003). 



Table 1. — Adjusted R 2 values for alternative models considered in the model selection process. 



Model 



Model parameters 



Impact 
severity 



Disturbed 
width 



Cumulatn e 

impact 



y = [a + (a v d x ) + (a 2 -d 2 )]TR^ 

y = a TR^ + ^<>onb 2 -j 2 )i 

v = [a + (a v d x ) + (a y d 2 )]TR^ + ^ 



0.027 


0.120 


0.054 


0.924 


0.S22 


0.933 


0.000 


0.763 


0.924 


0.924 


0.S31 


0.932 



136 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Table 2. — Independent model variables, parameter estimates, and R 2 values for selected nonlinear re- 
gression models. 



Model 


Impact 


severity 


Disturbed 


width 


Cumulative 


impact 


parameter 


Estimate 


Std error 


Estimate 


Std error 


Estimate 


Std error 


a 
Model R 2 


182.7833 8.930 
-145.219 7.949 
-139.670 7.663 

-0.281 0.017 
0.924 


324.087 21.029 

-219.108 17.102 

-156.590 13.779 

-0.333 0.024 

0.822 


538.232 
-501.464 
-473.774 

-0.576 

0.933 


38.489 

37.076 

35.241 

0.030 



Vegetation recovery times were relatively 
short (less than or equal to one year) on our 
site compared to more arid ecosystems that 
had recovery times ranging from a few years 
to a hundreds of years (Thurow et al. 1995; 
Lovich & Bainbridge 1999; Prosser et al. 
2000). Recovery times for our study site were 
comparable to other grassland sites (Payne et 
al. 1983; Prosser et al. 2000). Payne et al. 
(1983) observed recovery times of about one 
year for an upland grassland prairie in Mon- 
tana. Similarly, Prosser et al. (2000) observed 
recovery times for grasslands in North Dakota 
to be less than two years. The vegetation, soil 



and climate found at our study site help ex- 
plain the short recovery times observed. Our 
study site has relatively fertile soils, sufficient 
moisture, and primarily early succession plant 
species adapted to colonizing disturbed areas. 

In our study, vehicle impacts were quanti- 
fied when the soil was relatively dry. Althoff 
& Thein (2005) demonstrated that vegetation 
loss by vehicle traffic increases with soil 
moisture. Vegetation recovery rates may have 
been longer if we conducted our study when 
the soils were wetter. 

Carrying capacity models currently used by 
the military to estimate the capacity of lands 



100 



80 






_ 60 

s 

0) 

(/> 

3 40 

a 
E 
20 




o 



M1009 



B. 



M35A3 



Turn 



D Initial impact 

□ 5 Month recovery 

□ 12 Month recovery 



fa 



M88 



M1009 




Straight 



Vehicle by impact type 



Figure 8. — Impact severity by vehicle type (M88 tank recovery vehicle, M35A3 cargo truck and Ml 009 
utility cargo vehicle) for straight-line and turn tracking immediately, 5 months and 12 months following 
tracking. Straight-line tracking includes all data for turning radii greater than or equal to 20 m. Turn 
tracking includes all data for turning radii less than 20 m. Error bars represent ± one standard error. 
Impact severity at 5 and 12 months for Ml 009 is zero for both turn and straight tracking. Impact severity 
at 12 months for M35A3 is zero for both turn and straight tracking by 12 months. 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 



137 



to support vehicle-training activities incorpo- 
rate vegetation impact and recovery time es- 
timates from tracking studies (Diersing et al. 
1988; Wilson 1988; Shaw & Diersing 1989; 
Anderson et al. 1996; Concepts Analysis 
Agency 1996). Typically, straight-line track- 
ing study data have been used to estimate veg- 
etation loss in these carrying capacity models. 
Haugen et al. (2003) found that approximately 
16% of vehicle activities during real training 
exercises are at turning radii less than the crit- 
ical turning radii determined in our study. As 
a result, carrying capacity estimates based on 
straight-line tracking data may over estimate 
training land capacity by underestimating ve- 
hicle impacts. The magnitude of the overes- 
timate depends on the manner in which ve- 
hicles are used. 

ACKNOWLEDGMENTS 

We acknowledge the Strategic Environmen- 
tal Research and Development Program 
(SERDP) and the Army Environmental Qual- 
ity Technology program for providing finan- 
cial support for this study. We gratefully ac- 
knowledge Dr. George Gertner, Dr. Guangxing 
Wang, and Dr. Shoufan Fang for their guid- 
ance on statistical analysis of the data. 

LITERATURE CITED 

Althoff, P. & S. Thein. 2005. Impact of M1A1 
main battle tank disturbance on soil quality, in- 
vertebrates, and vegetation characteristics. Jour- 
nal of Terramechanics 42:159-176. 

Anderson, A.B., L. Chenkins, L. Winters, R. Hunt, 
C. Couvillon, D. Mc Ferren, S. Sekscienski & P. 
Sydelko. 1996. Evaluation of Land Value Study 
(ELVS)/Army Training and Testing Area Carry- 
ing Capacity (ATTACC), Pp. 8-17, In 5th An- 
nual LRAM/ITAM workshop proceedings. Au- 
gust 27-29, 1996. LaCrosse, Wisconsin. 

Ayers, P.D. 1994. Environmental damage from 
tracked vehicle operation. Journal of Terrame- 
chanics 31:173-183. 

Ayers, P., M. Vance, L. Haugen & A.B. Anderson. 
2000. An evaluation of DGPS-based continu- 
ously operating vehicle monitoring systems to 
determine site-specific vehicle event severity fac- 
tors. United States Army Engineer Research De- 
velopment Center, Technical Report TR-00-43. 
46 p. 

Belcher, J. & S. Wilson. 1989. Leafy spurge and 
the species composition of mixed-grass prairie. 
Journal of Range Management 42:172-175. 

Braunack, M. 1986. Changes in physical properties 
of two dry soils during tracked vehicle passage. 
Journal of Terramechanics 23:141-151. 



Concepts Analysis Agency. 1996. Evaluation of 
Land Value Study (ELVSj. United States Army 
Concepts Analysis Agency, Study Report CAA- 
SR-96-5, Bethesda, Maryland. 

Demarais, S., D. Tazik, P. Guertin & E. Jorgensen. 
1999. Disturbance associated with military ex- 
ercises, Pp. 385-395, In Ecosystems of the 
World 16, Ecosystems of Disturbed Ground. (L. 
Walker, ed.). Elsevier, New York. 

Diersing, V, R. Shaw, S. Warren & E. Novak. 
1988. A user's guide for estimating allowable 
use of tracked vehicles on non wooded military 
training lands. Journal of Soil and Water Con- 
servation 43:191-195. 

Gardner, WH. 1986. Water content. Pp. 493-544. 
In Methods of Soil Analysis, Part 1 : Physical and 
Mineralogical Methods, Agronomy Monogram 
9. (A. Klute, ed.). American Society of Agron- 
omy/Soil Science Society of America. 

Goran, W, L. Radke & W Severinghaus. 1983. An 
overview of the ecological effects of tracked ve- 
hicles on major U.S. Army installations. Con- 
struction Engineering Research Laboratory, 
Technical Report N-142. 74 pp. 

Grantham, W.P., E.F. Redente, C.F. Bagley & M.W. 
Paschke. 2001. Tracked vehicle impacts to veg- 
etation structure and soil erodibility. Journal of 
Range Management 54:71 1-716. 

Halvorson, J.J., D.K. McCool. L.G King & L.W. 
Gatto. 2001. Soil compaction and over-winter 
changes to tracked-vehicle ruts. Yakima Training 
Center, Washington. Journal of Terramechanics 
38:133-151. 

Haugen, L.B., P.D. Ayers & A.B. Anderson. 2003. 
Vehicle movement patterns and impact during 
military training exercises. Journal of Terrame- 
chanics 40:83-95. 

Hirst, R.A., R.F. Pywell, R.H. Mans & P.D. Put- 
wain. 2003. The resistance of a chalk grassland 
to disturbance. Journal of Applied Ecolog} 40: 
368-379. 

Hirst, R.A., R.F. Pywell & P.D. Putwain. 2000. As- 
sessing habitat disturbance using an historical 
perspective: The case of Salisbury Plain military 
training area. Journal of Environmental Manage- 
ment 60:181-193. 

Johnson, F.L. 1982. Effects of tank training activ- 
ities on botanical features at Fort Hood. Texas. 
The Southwestern Naturalist 27:309-314. 

Lovich. J.E. & D. Bainbridge. 1 C ) C ) L ). Anthropogen- 
ic degradation of the southern California desert 
ecosystem and prospects for natural reco\ er\ and 
restoration. Environmental Management 24:309- 
326. 

Milchunas. D.G.. K.A. Shultz £ R.B. Shaw. 1999. 
Plant community responses to disturbance b\ 
mechanized military maneuvers. Journal of En- 
vironmental Quality 28:1533—1547. 

Milchunas. D.G.. K.A. Shult/ £ R.B. Shaw. 2000. 



138 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Plant community structure in relation to long- 
term disturbance by mechanized military maneu- 
vers in a semiarid region. Environmental Man- 
agement 25:525-539. 

Payne, G., J. Foster & W. Leininger. 1983. Vehicle 
impacts on northern Great Plains range vegeta- 
tion. Journal of Range Management 36:327-331. 

Prose, D.V. 1985. Persisting effects of armored 
military maneuvers on some soils of the Mojave 
Desert. Environmental Geology and Water Sci- 
ence 7:163-170. 

Prosser, C.W., K.K. Sedivec & W.T. Barker. 2000. 
Tracked vehicle effects on vegetation and soil 
characteristics. Journal of Range Management 
53:666-670. 

Shaw, R. & V. Diersing. 1989. Allowable use es- 
timates for tracked vehicular training on Pinon 
Canyon Maneuver Site, Colorado, USA. Envi- 
ronmental Management 13:773-782. 

Shaw, R. & V. Diersing. 1990. Tracked vehicle im- 
pacts on vegetation at the Pinon Canyon Maneu- 
ver Site, Colorado. Journal of Environmental 
Quality 19:234-243. 

Tetra Tech. 2000. Final integrated natural resources 
management plan (INRMP) for Camp Atterbury, 
Indiana. Tetra Tech, Inc., Duluth, Georgia 30096. 



Thurow, T.L., S.D. Warren & D. H. Carleson. 1995. 
Tracked vehicle traffic effects on the hydrologic 
characteristics of central Texas rangeland. Trans- 
actions of the American Society of Agricultural 
Engineers 36:1645-1650. 

Trumbull, V.L., PC. Dubois, R.J. Brozka & R. Guy- 
ette. 1994. Military camping impacts on vege- 
tation and soils of the Ozark Plateau. Journal of 
Environmental Management 40:329-339. 

USD A, NRCS. 2004. The PLANTS Database, Ver- 
sion 3.5 (http://plants.usda.gov). National Plant 
Data Center, Baton Rouge, Louisiana 70874- 
4490 USA. 

Watts, E. 1998. Short-term influence of tank tracks 
on vegetation and microphytic crusts in shrub- 
steppe habitat. Environmental Management 22: 
611-616. 

Webb, R.H., & H.G. Wilshire (eds.). 1983. Envi- 
ronmental Effects of Off- road Vehicles: Impacts 
and Management in Arid Lands. Springer- Verlag, 
New York. 

Wigginton, M. & D. Marshall. 2004. Soil survey 
of Bartholomew County, Indiana. U.S. Dept. Ag- 
riculture, Soil Conservation Service. 753 pp. 

Wilson, S. 1988. The effects of military tank traffic 
on prairie: A management model. Environmental 
Management 12:397-403. 



2007. Proceedings of the Indiana Academy of Science (1 16): 139-147 



VEHICLE IMPACTS ON VEGETATION COVER AT CAMP 
ATTERBURY, INDIANA: PART 2. PREDICTING IMPACTS OF 

UNTESTED VEHICLES 

Alan B. Anderson: U.S. Army Engineer Research Development Center, Champaign, 
Illinois 61822 USA 

Paul D. Ayers: University of Tennessee, Knoxville 37996-4531 USA 

Patricia Sullivan: U.S. Army Engineer Research Development Center, Vicksburg. 
Mississippi 39180 USA 

William R. Ochsner: Camp Atterbury, Indiana 46124-5000 USA 

ABSTRACT. Vehicle tracking systems were installed on four military vehicles (M813 cargo truck. 
M998 utility vehicle, M548A cargo carrier, Ml 025 utility vehicle) at Camp Atterbury, Indiana to assess 
the impact of tracking by these vehicle types on vegetation loss. Study data were used to estimate param- 
eters for models previously reported in the literature and to validate model results. Instrumented vehicles 
were driven through courses of varying velocities and turning radii. Vegetation loss was recorded im- 
mediately after tracking. The tracked M548A cargo carrier caused the most site damage. The wheeled 
M813 cargo truck caused more vegetation loss than either of the other wheeled vehicles (M998 utility 
vehicle or M1025 utility vehicle). Power equations using vehicle type and turning radius as the independent 
variables predicted vegetation loss with an R 2 value of 0.845. Using only straight-line tracking data from 
our study to estimate parameters of a model proposed in an earlier reported study, we were able to predict 
vegetation loss for a range of turning radii almost as effectively as using the complete data set (R 2 = 
0.843). Using vehicle weights combined with impact models proposed in an earlier study, we were able 
to predict vegetation loss for untested vehicles almost as well as with field data (R 2 = 0.810). Results 
from our study indicate that vehicle impact data and models can be applied to untested vehicles and 
reasonably estimate vegetation loss at Camp Atterbury. The ability to estimate site impacts of untested 
vehicles allows installation natural resources personnel to more accurately assess proposed land manage- 
ment actions in a timely and economical manner. 

Keywords: Vehicle impacts, off-road, vegetation impact, impact assessment 



The Department of Defense is responsible and vegetation has been extensively studied 
for administering more than 10 million hect- (Anderson et al. 2005). However, the effective 
ares of federally-owned land in the United use of this information in environmental im- 
States. Much of this land is used for vehicle- pact assessments has been limited by a num- 
based training activities. Continued manage- ber of factors (Morrison-Saunders & Bailey 
ment of these lands requires assessing the im- 2003). Factors that limit the utility of impact 
pact of vehicles on installation natural study data in assessments include vehicles 
resources. These assessments are often man- having multiple configurations, assessments 
dated by the National Environmental Policy required before vehicles are available for test- 
Act of 1969 (NEPA) PL. 91-190, which re- ing, and assessments that involve multiple ve- 
quires analysis and documentation of potential hide types. 

environmental effects associated with all ma- Impact assessments may be required before 
jor federal decisions. The fielding of new vehicles are physically available for testing. In 
weapon systems or the relocation of military these situations, impact studies have been con- 
units and their vehicles to new locations are ducted using vehicles with similar static prop- 
activities that require assessments of potential erties (Haugen et al. 2003). In these cases im- 
vehicle impacts. pact data must be inferred from substituted 

The impact of off-road vehicle use on soil test vehicles to the vehicle being fielded. Stat- 

139 



140 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



ic vehicle properties important in determining 
vegetation damage include contact area, sur- 
face pressure, total weight, and track design 
(Ayers et al. 1994). 

The range of static vehicle properties that 
represents a vehicle type complicates the in- 
tegration of vehicle impact study data into im- 
pact assessments. Individual weapons systems 
are often fielded in more than one configura- 
tion, each with unique static vehicle proper- 
ties. Static vehicle properties like tire pressure 
can also be modified during use. As an ex- 
ample, the eight-wheeled Stryker armored 
combat vehicle comes in eight configurations. 
Depending on configuration and payload, in- 
dividual vehicles can vary in weight from 
12.7-18.6 metric tons. Tire pressure for the 
Stryker vehicle can also be varied during op- 
eration using the central tire inflation system. 
Field studies that quantify vehicle impacts 
typically use a single vehicle type configura- 
tion (Anderson et al. 2005). 

A new weapon system is not fielded inde- 
pendently of other vehicles. Multiple vehicle 
types make up military units. Assessing new 
weapon systems or relocation of existing units 
requires comparison between different units, 
each made up of varying vehicle types. Insuf- 
ficient funding, study area, and time limit the 
number of vehicles that can be studied. Typ- 
ically, field studies quantify the impacts of a 
dominant vehicle where the dominant vehicle 
represents the most common or most poten- 
tially damaging vehicle (Anderson et al. 
2005). 

Current environmental impact assessments 
lack the ability to objectively predict impacts 
of untested vehicles (or alternative configura- 
tions) using vehicle static properties and data 
from existing vehicle impact studies. Predict- 
ing vehicle impacts based on static vehicle 
properties is important because static proper- 
ties are known for currently-fielded vehicle 
types and are approximately known during de- 
sign and development phases of fielding new 
vehicles. 

Sullivan & Anderson (2000) proposed the 
use of vehicle static properties to estimate 
vegetation damage as a percentage of the 
damage caused by a baseline vehicle. Models 
used by the authors incorporated vehicle 
weight and properties used to estimate vehicle 
ground contact area. A preliminarily valida- 
tion of the methodology was conducted using 



subject matter expert opinion. Results of the 
proposed methodology correlated reasonably 
well with predicted subject matter expert 
opinions (R 2 = 0.77) using 37 vehicles that 
varied widely in vehicle static properties. 
However, the authors were not able to validate 
their methodology more rigorously with field 
data because of the lack of data representing 
a range of vehicles tested under similar con- 
ditions. 

Vehicle impact factors based on Sullivan & 
Anderson (2000) have been used to assess ve- 
hicle impacts on installation resources as part 
of environmental impact assessments of new 
weapon systems (Tetra Tech Inc. 2003; Col- 
orado State University 2004; Tetra Tech Inc. 
2004; Shoop et al. 2005). The vehicle impact 
factors also have been used to develop land 
repair funding requirements based on pro- 
posed training schedules (Anderson et al. 
1996; Concepts Analysis Agency 1996). 
While these approaches for evaluating vehicle 
impacts are easily used in decision support 
processes, the relative impacts of vehicles 
have not been thoroughly validated using field 
data. 

Anderson et al. (2007) quantified and mod- 
eled the impact of vehicles on vegetation loss 
using vehicle type and vehicle dynamic prop- 
erties (speed and turning radius). The study 
indicated that a common model form was ap- 
plicable for all vehicles tested at the study area 
and that turning radius was the critical vehicle 
dynamic property required to predict vegeta- 
tion loss. While the authors were able to mod- 
el vegetation loss using vehicle type, they did 
not directly use static vehicle properties that 
would have allowed the models to be applied 
to other vehicle types. 

The overall objective of our study was to 
evaluate the use of vehicle static properties 
and existing vehicle impact study data to pre- 
dict vegetation impacts of unstudied vehicles 
as a function of vehicle dynamic properties. 
The first objective was to validate that models 
proposed by Anderson et al. (2007) were gen- 
erally applicable for similar sites and environ- 
mental conditions. The second objective was 
to predict vegetation impacts using only 
straight-line vehicle tracking data and equa- 
tions that incorporate vehicle dynamic prop- 
erties. In this objective, we used models de- 
veloped by Anderson et al. (2007) as the 
foundation for extrapolating straight-line 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 2 



141 




M813 



M998 




M548A 



Ml 025 



HHHHHl 



Figures 1-4. — Four vehicle types used in the tracking study include the M813 cargo truck, M998 utility 
vehicle, M548A cargo carrier and Ml 025 utility vehicle. See Anderson et al. (2007) for pictures of the 
M88 base reference vehicle and additional vehicles used in model development. 



tracking data. The third objective was to es- 
timate vegetation damage using only vehicle 
static properties and models proposed by An- 
derson et al. (2007). 

METHODS 

Study site. — The study was conducted at 
Camp Atterbury, Indiana, an Army National 
Guard training facility. A more detailed de- 
scription of terrain, soil and vegetation typical 
of the installation can be found in Anderson 
et al. (2007). The study site is located in train- 
ing area 2A (39.70° N, 86.33° W). The study 
site was selected because it was representative 
of many vehicle use sites found on the instal- 
lation. Study site soils were classified as a Sto- 
nelick (Wigginton & Marshall 2004). The Sto- 
nelick series is a coarse-loamy, mixed, 
superactive, calcareous, mesic Typic Udiflu- 
vent. Vegetation at the study site consisted pri- 



marily of introduced grasses with a smaller 
component of native and introduced forbs. 

Study design. — -A field study was conduct- 
ed on 28 May 2003 using four vehicles: 
M548A cargo earner. M813 cargo truck. 
Ml 025 utility vehicle and M998 utility vehi- 
cle (Figs. 1-4). The M548A cargo carrier is a 
tracked vehicle that is 4.86 m long. 2.69 m 
wide, 2.50 m high that weighs 12.832 kg. The 
M548A track width is 38.1 cm with pads that 
are 15.2 cm wide by 8.9 cm long. The M813 
cargo truck is a three-axle ten-wheeled vehicle 
with a 5.23 m total wheelbase. 2.49 m width, 
and weight of 10.037 kg. The MS 13 tire 
height is 106.7 cm with a tread width of 190.5 
cm front and 221.0 cm rear. Tire pressure at 
the time of tracking was 391 to 514 kPa dur- 
ing the study. The Ml 025 utility vehicle and 
M998 utility vehicle are the same vehicle type 
but represent different fielding configurations. 



142 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



The Ml 025 is a four-wheeled vehicle with a 
3.30 m wheelbase, 2.13 m width, and a weight 
of 3720 kg. The M1025 tire height is 93.3 cm 
with a tread width of 182.9 cm. Tire pressure 
at the time of tracking was 176 to 194 kPa 
during the study. The M998 is a four-wheeled 
vehicle with a 3.28 m wheelbase, 2.13 m 
width, and a weight of 3493 kg. The M998 
tire height is 88.2 cm with a tread width of 
182.9 cm. Tire pressure at the time of tracking 
was 95 to 132 kPa during the study. 

The field study design is based on methods 
described in Anderson et al. (2007). Each ve- 
hicle drove a systematically planned course 
(spiral) within four randomly located treat- 
ment plots. Each spiral course within a treat- 
ment plot consisted of a section of straight- 
line travel followed by a section of constantly 
decreasing turning radius. Two spirals for 
each vehicle were traversed at a slower or 
faster velocity. The fast velocity spiral repre- 
sents the fastest velocity the vehicle could 
safely be driven for the site conditions. 

Each vehicle was equipped with a vehicle 
tracking system that allowed monitoring of 
vehicle velocity and turning radius. See Ayers 
et al. (2000) and Anderson et al. (2007) for 
details about the tracking systems. 

Sampling methodologies. — Sample points 
were randomly located approximately every 5 
m along the inner vehicle track of each spiral 
resulting in approximately 20 sample points 
per spiral per vehicle. Each sample point con- 
sisted of a paired subplot. One subplot was 
located within the track and the other subplot 
was 0.5 m adjacent to and on the inside of the 
track in undisturbed vegetation. 

Immediately after tracking, vehicle impacts 
were assessed as disturbed width (DW), im- 
pact severity (IS) and cumulative impact (CI). 
Disturbed width was measured perpendicular 
to the vehicle track and encompassed the area 
where soil and/or vegetation were impacted 
by the vehicle tire or track. Vegetation cover 
was visually estimated within the disturbed 
vehicle track and control subplots. Impact se- 
verity was calculated as the percent cover in 
the undisturbed subplot minus the percent 
cover in the disturbed subplot. Cumulative im- 
pact was calculated as the product of disturbed 
width and impact severity. 

Soil moisture was determined gravimetri- 
cally on the day of tracking for the 0-10.16 
cm depth using methods of Gardner (1986). 



Soil samples were dried at 105° C for 48 h in 
a conventional oven. Water content was cal- 
culated on a mass basis as a percentage of dry 
soil. Air temperature was recorded one meter 
above the soil surface at the beginning and 
end of tracking treatments. 

Statistical analysis. — Nonlinear regression 
analysis methods were used to determine best- 
fit model parameter values. Nonlinear regres- 
sion analyses were conducted using raw de- 
pendent and independent variables. 
Cumulative impact was the dependent vari- 
able. Independent variables included in the 
model were variables found by Anderson et 
al. (2007) to be significant in quantifying veg- 
etation loss by vehicles. Independent variables 
included in the model were turning radius and 
vehicle type. Vehicle type was included in the 
model as dummy variables such that d t = 
except for d l = 1 for the M813 cargo truck, 
d 2 = 1 for the M998 utility vehicle, d 3 = 1 
for the Ml 025 utility vehicle and d 4 = 1 for 
the M548A cargo carrier. The a model term 
was set to 538.2323 to make the model pa- 
rameter values directly comparable to results 
from Anderson et al. (2007). The b model 
term was first estimated directly from the data 
and then set to —0.5764 (obtained from An- 
derson et al. 2007) in two separate analyses. 
Estimating the b term from the data was to 
determine the best overall fit of the model. 
Setting the b term to a value obtained from 
Anderson et al. (2007) was to assess the utility 
of their proposed model to our site. Model 
parameters were estimated using the Proc 
Model of SAS® (SAS Institute Inc., Cary, 
North Carolina). The general form of the 
model employed is 

CI = [a + {a x -d x ) + (a 2 -d 2 ) + (a r d 3 ) 

+ (a 4 -d 4 )]TR h 

where CI is the dependent variable cumulative 
impact, TR is the independent variable turning 
radius, d x is the dummy variable to indicate 
the M813 cargo truck, d 2 is the dummy vari- 
able to indicate the M998 utility vehicle, d 3 is 
the dummy variable to indicate the Ml 025 
utility vehicle, d 4 is the dummy variable to 
indicate the M548A cargo carrier, a is the in- 
tercept coefficient, a x is the intercept shift co- 
efficient for the M813 cargo truck, a 2 is the 
intercept shift coefficient for the M998 utility 
vehicle, a 3 is the intercept shift coefficient for 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 2 



143 



the Ml 025 utility vehicle, a A is the intercept 
shift coefficient for the M548A cargo carrier, 
b is the slope coefficient. 

Most published studies quantify vehicle im- 
pacts using only straight-line tracking (An- 
derson et al. 2005). However, several studies 
have demonstrated that turning can cause 
more damage than straight-line tracking (Ay- 
ers 1994; Anderson et al. 2007). To determine 
if we can extrapolate straight-line tracking 
data for tested vehicles to vehicle tracking of 
any turning radius, we calculated the average 
cumulative impact value for turning radii 
greater than 40 m. This subset of the field data 
was used to estimate straight-line tracking be- 
cause 40 m is well above the critical turning 
radius of 15-20 m reported by Anderson et al. 
(2007). Critical turning radii derived from oth- 
er studies are also well below 40 m (Haugen 
et al. 2003). Equations obtained from Ander- 
son et al. (2007) were then used to calculate 
vehicle specific a n parameter values for a turn- 
ing radius of 17.5 m. A turning radius of 17.5 
m was used because this was the midpoint of 
the 15 to 20 m critical turning radius reported 
by Anderson et al. (2007). The resulting mod- 
el was then used to estimate vegetation impact 
values for each measured point. Model R 2 val- 
ues were calculated to quantify how well the 
model fit the complete data set that included 
straight-line tracking and turning at varying 
radii. 

To determine if we can predict vegetation 
loss for untested vehicles, we estimated ve- 
hicle specific a n parameter values for the equa- 
tion obtained from Anderson et al. (2007). Pa- 
rameter values were estimated using vehicle 
weight as a surrogate variable for vehicle type. 
For vehicles used in Anderson et al. (2007) 
we developed a linear regression model to de- 
scribe the relationship between published a n 
parameter values and vehicle weights. The re- 
sulting regression model was used along with 
the weights of vehicles used in our study to 
predict a n parameter values for the vehicles. 
The resulting model was then used to estimate 
vegetation impact values for each measured 
data point. Model R 2 values were calculated 
to quantify how well the model fit the com- 
plete field data set. 

RESULTS 

Vehicle impacts. — At the time of tracking, 
soil water content, averaged 23.7%. This rep- 



resents a soil condition typical for many ma- 
neuver activities at this site and is wetter than 
soil conditions tested by Anderson et al. 
(2007). Air temperature ranged from 12-26° 
C during tracking treatments. Vegetation cov- 
er for undisturbed sample points averaged 100 
indicating a densely vegetated site. 

The tracked M548A cargo carrier caused 
more damage than any of the wheeled vehi- 
cles. For wheeled vehicles the M813 cargo 
truck caused more damage than the M998 util- 
ity vehicle and Ml 025 utility vehicle. 

Cumulative impact increased exponentially 
with decreasing turning radii (Fig. 5). Cumu- 
lative impact increased suddenly at turning ra- 
dii less than approximately 15-20 m for all 
vehicle types. This critical turning radius 
where increase site damage occurred is similar 
to that reported by Anderson et al. (2007) 

Modeling vehicle impact using all field 
data. — Nonlinear regression analyses indicat- 
ed that a model of the form proposed by An- 
derson et al. (2007) reasonably describes the 
data in our study. The model using a and b 
parameter values from Anderson et al. (2007) 
had an R 2 value of 0.845. Figure 5 shows pre- 
dicted values plotted against field observa- 
tions. The model R 2 remained 0.845 when 
model parameter values were not restricted to 
values obtained from other studies. 

Log transformations of the dependent and 
independent variables did not result in com- 
pletely linear relationships between the data as 
found in Anderson et al. (2007). This was 
most apparent for the M548A cargo carrier. 
The lack of a linear relationship indicates that 
other model forms may better describe the 
data for this specific study site. In our study. 
the models tended to over estimate impacts for 
turning radii between 15-25 m. The dramatic 
increase in vegetation loss at turning radii less 
than 25 m may be due to the dense above and 
below-ground vegetation typical of this site. 
Foster et al. (2006) found similar results in 
their study of the impact oi' turning vehicles 
on vegetation loss. 

Predicting vehicle impact using only 
straight-line tracking data. — Figure 6 show s 
measured and predicted cumulative impact as 
a function of turning radius for the MS 13 ear- 
go truck. M998 utility vehicle, M548A cargo 
earner and M1025 utility vehicle. Cumulative 
impact was predicted using straight-line track- 
ing data from our study to estimate a,, param- 



144 PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 

CI = (538.2323 - (451 .441 * d, ) - (495.137 * d\) - (500.829 * d~ ) - (423.741 * d, ))77r 05764 



70 



60 




o 


M81 3 observed 


- 


M998 observed 


A 


M1025 observed 


□ 


M548A observed 




-M81 3 predicted 




- - - 


- M998 predicted 




- M1025 predicted 




— M548A predicted 



25 



50 75 1 00 

Turning radius (m) 



125 



150 



Figure 5. — Cumulative Impact (CI) as a function of turning radius (TR) for the M813 cargo truck, 
M998 utility vehicle, M548A cargo carrier and M1025 utility vehicle. The parameters of the power 
equation for fitted lines are provided at the top of the graph. Variables d { , d 2 , d 3 and d 4 are dummy 
variables that account for vehicle type. Dummy variables d l = except that d x = 1 for the M813, d 2 — 
1 for the M998, d 3 = 1 for the Ml 025, and d 4 = 1 for the M548. Parameter values were estimated using 
all field data and the equation from Anderson et al. (2007). The R 2 fit for the equation is 0.845. 



eter values for individual vehicle types for use 
in the equation from Anderson et al. (2007). 
The R 2 fit for the model containing all four 
vehicles is 0.843. The R 2 value for a model 
based only on straight-line tracking data is 
only slightly less than for a model using all 
the field data. R 2 values for individual vehicles 
ranged from 0.486 for the Ml 025 utility ve- 
hicle to 0.843 for the M813 cargo truck. High- 
er correlations were obtained for the heavier 
vehicles. The higher correlations for the 
heavier vehicles resulted from a larger range 
of vegetation loss across turning radii relative 
to the variation in vegetation loss at any spe- 
cific turning radius. 

Predicting vehicle impacts using static 
vehicle properties. — Figure 7 shows mea- 
sured and predicted cumulative impact as a 
function of turning radius for the M813 cargo 
truck. M998 utility vehicle, M548A cargo car- 
rier and Ml 025 utility vehicle. Cumulative 
impact was predicted using vehicle weight to 
estimate vehicle type specific a n parameter 
values for inclusion into a model obtained 
from Anderson et al. (2007). The R 2 fit for the 
model containing all four vehicles is 0.810. 



The R 2 value for a model based only on ve- 
hicle weight is only slightly less than for a 
model using only straight-line tracking data or 
all the field data. R 2 values for individual ve- 
hicles range from 0.486 to 0.843. Higher cor- 
relations were obtained for the heavier vehi- 
cles. 

DISCUSSION 

GPS technology is currently being used to 
track vehicles during live training events 
(Haugen et al. 2003). This data is useful for 
characterizing how vehicles are used in live 
training exercises. These studies characterize 
vehicle locations, velocities, and turning radii. 
However, to quantify the overall impact of a 
training event, we must also be able to esti- 
mate vehicle impacts for the full range of ve- 
hicle operating conditions observed during 
these training events. Currently most impact 
studies have only documented the impact of 
straight-line tracking on vegetation loss, with 
straight-line tracking often being one of the 
least destructive vehicle operating conditions. 
In our study, we demonstrated the use of 
straight-line vehicle tracking data to calibrate 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 2 



145 



CI = (538.2323 
70 r 



(451 .764 *</,)- (486.799 * d?) - (496.482 *d,)- (41 1 .107 * d- ))TR 



60 



E 
3, 50 

o 

(0 

Q- 40 
£ 

0) 

> 30 

*-• 
re 

£ 20 

3 

o 

10 



■? 




o 


M8 13 observed 


- 


M998 observed 


A 


M1025 observed 


□ 


M548A observed 




-M813 predicted 




- - - 


■ M998 predicted 




•M1025 predicted 






-M548A predicted 





25 



50 75 100 

Turning radius (m) 



125 



150 



Figure 6. — Cumulative Impact (CI) as a function of turning radius (77?) for the M813 cargo truck. M998 
utility vehicle, M548A cargo carrier and Ml 025 utility vehicle. The parameters of the power equation for 
fitted lines are provided at the top of the graph. Variables d u d 2 , d 3 and d 4 are dummy variables that 
account for vehicle type. Dummy variables d { = except that d ] = 1 for the M813, d 2 = 1 for the M998. 
d 3 = 1 for the M1025, and d 4 = 1 for the M548. Parameter values estimated using straight-line tracking 
field data only, equation from Anderson et al. (2007), and calculating a n parameters using a critical turning 
radius of 17.5 meters. R 2 fit for the equation is 0.843. 



models that predict vegetation impacts of ex- 
isting vehicles for different turning radii. Our 
study also verified that models developed at 
one study site are applicable to other study 
sites with similar soil and vegetation condi- 
tions. 

In environmental impact assessments we 
often need to predict the impact of vehicles 
that have not been tested. In our study we 
demonstrated that we could predict vegetation 
loss for untested vehicles by using simple ve- 
hicle static properties like weight and data 
from field tests of other vehicles. 

Several issues are relevant when interpret- 
ing results from our study. First, our study re- 
sults are site specific. While we validated our 
approach using a different study site from the 
original study (that varied in soil type, vege- 
tation composition, and soil moisture), we do 
not have data from a wide range of sites or 
moisture conditions to quantify how far con- 
clusions or models can be extrapolated. Al- 
thoff & Thein (2005) demonstrated that soil 
moisture can significantly affect vegetation 
loss due to vehicle traffic. Soil moisture or 



vegetation characteristics could explain differ- 
ences in the shape of our response data be- 
tween the two study sites. Models extrapolat- 
ed from Anderson et al. (2007) tended to 
underestimate large turning radii impacts and 
over estimate small turning radii impacts for 
our study site. 

A second issue is that all modeling ap- 
proaches we proposed for predicting vehicle 
impacts require some existing site-specific im- 
pact data to estimate model parameters. The 
approaches differed only in the amount of 
site-specific data required. If the approach is 
to be repeated at another location, at least one 
vehicle must be tested for a range of turning 
radii and several vehicles must be tested to 
develop a relationship between static vehicle 
properties and vegetation impact. 

Application of the models we evaluated 
may be limited by the empirical nature of the 
models as they are applied to more diverse 
sites. A more productive long-term approach 
may be to model vehicle static and dynamic 
properties in ways that relate more directly to 
assessed impacts. Li et al. (200") used pro- 



146 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



CI = (538.2323 - (463.876 * d, ) - (488.564 * d 2 ) - (494.268 * d 2 ) - (381 .780 * d 2 ))7Yr 03764 
70 




o 


M8 13 observed 


- 


M998 observed 


A 


M1025 observed 


D 


M548A observed 




-M813 predicted 




- - - 


■ M998 predicted 




•M1025 predicted 




-M548A predicted 





Turning radius (m) 

Figure 7. — Cumulative Impact (CI) as a function of turning radius (77?) for the M813 cargo truck, M998 
utility vehicle, M548A cargo carrier and Ml 025 utility vehicle. The parameters of the power equation for 
fitted lines are provided at the top of the graph. Variables J,, d 2 , d 3 and d 4 are dummy variables that 
account for vehicle type. Dummy variables J, = except that d l = 1 for the M813, d 2 = 1 for the M998, 
d 3 = 1 for the Ml 025, and d 4 = 1 for the M548. Parameter values estimated using vehicle static properties 
and equation from Anderson et al. (2007). R 2 fit for the equation is 0.810. 



cess-based models to predict vegetation im- 
pacts based on vehicle static and dynamic 
properties as well as soil strength. Including 
soil strength in their model allows site con- 
ditions like soil moisture to be directly incor- 
porated into the model. The approach of Li et 
al. (2007) was demonstrated to work for a 
much more diverse range of study sites than 
evaluated in our study. 

Despite potential limitations of the models 
we evaluated, we demonstrated that vehicle 
impact study data is representative of impacts 
measured for similar sites. We also demon- 
strated that relatively simple models that re- 
late vehicle static properties to vegetation loss 
could be used to predict impacts of untested 
vehicles for a range of vehicle dynamic prop- 
erties. These models will help resource man- 
agers evaluate the impact of current and pro- 
posed vehicle use at military installations. 
Most importantly, the ability to predict im- 
pacts from vehicles before available for test- 
ing allows land managers to proactively pre- 
pare for potential impacts. 



ACKNOWLEDGMENTS 

We acknowledge the Strategic Environmen- 
tal Research and Development Program 
(SERDP) and the Army Environmental Qual- 
ity Technology program for providing finan- 
cial support for this study. We gratefully ac- 
knowledge Dr. George Z. Gertner, Dr. 
Guangxing Wang, and Dr. Shoufan Fang for 
their guidance on statistical analysis of the 
data. 

LITERATURE CITED 

Althoff, P. & S. Thein. 2005. Impact of M1A1 
main battle tank disturbance on soil quality, in- 
vertebrates, and vegetation characteristics. Jour- 
nal of Terramechanics 42:159-176. 

Anderson, A.B., P.D. Ayers, H. Howard & K.D. 
Newlin. 2007. Vehicle impacts on vegetation 
cover at Camp Atterbury, IN: Part 1. Initial im- 
pacts and vegetation recovery. Proceedings of 
the Indiana Academy of Science 116(2):000- 
000. 

Anderson, A.B., A.J. Palazzo, P.D. Ayers, J.S. Feh- 
mi & S. Shoop. 2005. Assessing the impacts of 
military vehicle traffic on natural areas: Intro- 



ANDERSON ET AL.— VEHICLE IMPACTS AT CAMP ATTERBURY: PART 2 



147 



duction to the special issue and review of the 
relevant military vehicle impact literature. Jour- 
nal of Terramechanics 42(3-4): 143-158. 

Anderson, A.B., L. Chenkins, L. Winters, R. Hunt, 
C. Couvillon, D. Mc Ferren, S. Sekscienski & P. 
Sydelko. 1996. Evaluation of Land Value Study 
(ELVS)/Army Training and Testing Area Carry- 
ing Capacity (ATTACC), Pp. 8-17, In 5th An- 
nual LRAM/ITAM workshop proceedings. Au- 
gust 27-29, 1996. LaCrosse, Wisconsin. 

Ayers, P.D. 1994. Environmental damage from 
tracked vehicle operation. Journal of Terrame- 
chanics 31:173-183. 

Ayers, P., M. Vance, L. Haugen & A.B. Anderson. 
2000. An evaluation of DGPS-based continu- 
ously operating vehicle monitoring systems to 
determine site- specific vehicle event severity fac- 
tors. United States Army Engineer Research De- 
velopment Center, Technical Report TR-00-43. 
46 pp. 

Colorado State University. 2004. Transformation 
of the U.S. Army Alaska final environmental im- 
pact statement. Center for Ecological Manage- 
ment of Military Lands, Colorado State Univer- 
sity, Fort Collins, Colorado 80523-1490, USA, 
February 2004. 

Concepts Analysis Agency. 1996. Evaluation of 
Land Value Study (ELVS). US Army Concepts 
Analysis Agency, Study Report CAA-SR-96-5, 
Bethesda, Maryland. 

Foster, J., P. Ayers, A. Lombardi-Przybylowicz & 
K. Simmons. 2006. Initial effects of light ar- 
mored vehicle use on grassland vegetation at 
Fort Lewis, Washington. Journal of Environmen- 
tal Management 81:315-322. 

Gardner, W.H. 1986. Water content, Pp. 493-544, 
In Methods of Soil Analysis, Part 1 : Physical and 



Mineralogical Methods, Agronomy Monogram 
9. (A. Klute, ed.). American Society of Agron- 
omy/Soil Science Society of America. 

Haugen, L.B., P.D. Ayers & A.B. Anderson. 2003. 
Vehicle movement patterns and impact during 
military training exercises. Journal of Terrame- 
chanics 40:83-95. 

Li, Q., P.D. Ayers & A.B. Anderson. 2007. Pre- 
diction of impacts of wheeled vehicles on terrain. 
Journal of Terramechanics, 44(2^:205-2 15. 

Morrison-Saunders, A. & J. Bailey. 2003. Practi- 
tioner perspectives on the role of science in en- 
vironmental impact assessment. Environmental 
Management 31:683-695. 

Shoop, S., R. Affleck, C. Collins, G. Larsen & P. 
Sullivan. 2005. Maneuver analysis methodology 
to assess vehicle impacts on training lands. Jour- 
nal of Terramechanics 42(3-4): 28 1-303. 

Sullivan, P. & A.B. Anderson. 2000. A methodol- 
ogy for estimating Army Training and Testing 
Area Carrying Capacity (ATTACC) vehicle se- 
verity and local condition factors. ERDC Tech- 
nical Report TR-00-02. 44 pp. 

Tetra Tech, Inc. 2003. 2 nd Armored Cavalry Regi- 
ment transformation and installation mission 
support, Joint Readiness Training Center (JRTC) 
and Fort Polk, Louisiana and long-term military 
training use of Kisatchie National Forest Lands 
draft environmental impact statement. Tetra 
Tech, Inc., Fairfax Virginia. 

Tetra Tech, Inc. 2004. Transformation of the 2nd 
Brigade of the 25th Infantry Division (Light) to 
a Stryker Brigade Combat Team (SBCT) in Ha- 
waii final environmental impact statement. Tetra 
Tech, Inc., Honolulu, Hawaii 96819. 

Wigginton, M. & D. Marshall. 2004. Soil survey 
of Bartholomew county Indiana. U.S. Dept. Ag- 
riculture, Soil Conservation Service. 753 pp. 



2007. Proceedings of the Indiana Academy of Science (1 16): 148-157 



AMENDMENTS FOR FIELD-SCALE PHYTOTREATMENT OF 
PB, CD AND ZN FROM AN INDIANA SUPERFUND SOIL 

J.R. Jacob, C.K. Hee and J. Pichtel: Dept. of Natural Resources and Environmental 
Management, Ball State University, Muncie, Indiana 47306 USA 

ABSTRACT. A field study was conducted to determine the effectiveness of a mixed grass crop, sun- 
flower (Helianthus annuus), or ragweed {Ambrosia trifida) and several amendments in revegetation and 
treatment of soil severely contaminated with lead, cadmium and zinc. Amendments included composted 
municipal solid waste, dried sewage sludge, citric acid, ethylenediaminetetraacetic acid (EDTA) (single 
and multiple applications), and control. The mixed grass crop was capable of growth on all treatments. 
Soil Pb and Zn occurred primarily in the carbonate, organic-bound and residual forms (23.1%, 31.8%, 
and 44.4%, Pb, respectively, and 11.4%, 26.5%, and 60.2% Zn, respectively) as determined by sequential 
extraction. The MSW and SS treatments resulted in greatest plant cover and dry matter production on the 
field plots. Dry matter production was significantly (P < 0.05) higher in the municipal solid waste (1.09 
MT/ha) and dried sewage sludge (95 MT/ha) treatments. The single EDTA treatment resulted in signifi- 
cantly (P < 0.05) increased Pb uptake when compared to the other treatments. The EDTA, citric acid and 
municipal solid waste significantly (P < 0.05) increased Cd uptake by plants. In a growth chamber leaching 
study, soil Pb leached most from the 0.5 M EDTA treatment; the citric acid, mixed NPK fertilizer, 
municipal solid waste, sewage sludge and control treatments resulted in negligible leaching. Phytostabil- 
ization in combination with organic amendments may be the most appropriate technology to ensure sta- 
bilization of soil metals at this site. 

Keywords: Cd, Pb, Zn, phytostabilization, revegetation, sewage sludge, municipal solid waste 



Heavy metal contamination of soils at der- 
elict industrial sites is a significant issue 
worldwide. Lead (Pb), cadmium (Cd) and zinc 
(Zn) are among the most commonly encoun- 
tered heavy metals at contaminated facilities 
(Lasat 2007); and excess concentrations can 
be detrimental to plant growth. Such soils can 
also be adversely affected by poor drainage, 
low organic matter content, and diminished 
populations of indigenous microbes that cycle 
nutrients. Revegetation is essential to limit 
soil erosion by wind and water, including run- 
off of metallic sediments. A long-term goal 
for such sites is the development of a self- 
sustaining ecosystem that can support produc- 
tive land use activities and is aesthetically ap- 
pealing. 

Plant species have been identified that have 
the capability to either immobilize or accu- 
mulate heavy metals. Recent research has ex- 
amined the use of plants as either stabilizing 
or extractive tools for metal-contaminated 
soils (Pichtel & Bradway 2007; Mills et al. 
2006; Tie et al. 2006; Datta & Sarkar 2005; 
Pichtel et al. 2000). During phytostabilization, 
mobility of contaminants is reduced by accu- 



mulation within roots, adsorption to root sur- 
faces, or conversion to immobile species with- 
in the rhizosphere (Vangronsveld et al. 1995). 
In contrast, phytoextraction involves the en- 
gineered use of plants to remove contaminants 
from the soil. 

Establishment of a long-term vegetative 
cover can retain contaminants in place, thus 
reducing dispersion to local environs (Pulford 
& Watson 2003). When revegetation is com- 
bined with application of soil amendments 
such as organic matter, the mobility of con- 
taminants in the soil can be further reduced 
(Mench et al. 2000; Madejon et al. 2006). 

For site revegetation to succeed, the degree 
of plant tolerance to metallic contaminants 
must be assessed. Recent research has docu- 
mented effective plant stabilization or extrac- 
tion of heavy metals (Wu et al. 2006; Li et al. 
2005; Wilde et al. 2005); however, little is 
known regarding revegetation and/or treat- 
ment of highly toxic and/or infertile metallif- 
erous wastes. 

Metals on weathered metalliferous sites oc- 
cur in complex forms and vary widely in sol- 
ubility and bioavailability (Tie et al. 2006; 



148 



JACOB ET AL.— PHYTOTREATMENT OF PB, CD AND ZN 



149 



Jensen et al. 2006; Selim & Kingery 2003). 
Chemical fractionation procedures have prov- 
en useful for segregating soil metals into var- 
ious reactive forms (Almas et al. 2006; Ber- 
mond et al. 2005; Chague-Goff 2005; Steele 
and Pichtel 1998). Each metal fraction is as- 
sociated with a certain degree of mobility in 
the biosphere, and hence with bioavailability 
to plants. 

An Indiana Superfund site (40°10'34"N, 
85°25'36"W) was the focus of study. As a re- 
sult of uncontrolled disposal of industrial 
wastes, the site is contaminated by Pb, Cd, Zn 
and other metals. The site is bordered on one 
side by the White River and by numerous res- 
idences on the opposite side. The soil has a 
massive structure and is classified as 'Made 
Land'. Remediation at the site as directed by 
the EPA Record of Decision (issued 2001) has 
involved isolation rather than removal of con- 
taminants; steel sheet pilings have been in- 
stalled along the riverbank. The site has sub- 
sequently been removed from the EPA 
National Priorities List (U.S. EPA 2006); 
however, soil material is still enriched with 
heavy metals. Due to proximity of the site to 
residential dwellings, loss of metals to 
groundwater or via airborne dispersal is still 
a concern. 

If revegetation of the site is to succeed, a 
cover crop must be able to withstand poten- 
tially toxic soil conditions. Selected soil 
amendments may enhance plant establishment 
and enhance ecosystem sustainability. The 
purpose of the reported study was to assess 
revegetation of the toxic metalliferous soil at 
the site, and to study the influence of soil 
amendments on plant growth and plant stabi- 
lization and/or uptake of soil metals. 

METHODS 

Field study. — Three test blocks were estab- 
lished in March 2006 at a Superfund site lo- 
cated in Delaware County, Indiana. Plots mea- 
suring 2 X 4 m were set within each block. 
The following treatments were applied to the 
plots at the initiation of the study (three rep- 
licates each): composted municipal solid 
waste (MSW); dried sewage sludge (SS); cit- 
ric acid (as Fisher-grade product); ethylene- 
diaminetetraacetic acid (EDTA) (as 
Na 2 EDTA) (two application rates, e.g.. 
EDTA1 and EDTA2); and control (no treat- 
ment applied). The MSW and SS were each 



applied at 25 MT/ha (metric tons per hectare). 
Throughout the growing season, citric acid 
and EDTA1 were applied to the plots at 
monthly intervals (May, June, July, and Au- 
gust) at 2 mmol/kg soil. The EDTA2 was pro- 
vided as a single application of 500 mmol/kg. 
These concentrations were adapted from B lay- 
lock et al. (1997) and from previous studies 
conducted in our laboratories. The source of 
the MSW was Bedminster Corp., Sevierville, 
Tennessee, and the dried sewage solids (re- 
covered from belt press) were obtained from 
the Muncie, Indiana Bureau of Water Quality. 

Seeds of a grass mixture (smooth meadow- 
grass, Poa pratensis; red fescue, Festuca rub- 
ra; and perennial ryegrass, Phleum praten.se) 
(each approx. 10 kg/ha) were sown onto tilled 
soil by broadcast seeding. Both red clover 
(Trifolium pretense) and sunflower (Helian- 
thus annuus) seeds were sown at the site, and 
ragweed (Ambrosia trifida) was transplanted 
from two- week old plants. All species except 
the grass mixture failed within the first 30-60 
days, however. In June above-ground grass 
tissue and surface soil (0-20 cm) were sam- 
pled from each plot. Tissue was cut approxi- 
mately 5 cm above the ground surface to limit 
contamination by soil material and was sub- 
sequently rinsed in deionized water to remove 
attached soil particles. Soil was sampled from 
4—5 random points from the surface 20 cm of 
each plot using a stainless steel sampling tube. 
Soil material was composited, air-dried, and 
sieved through a 2 mm mesh sieve. At the 
conclusion of the growing season grass shoots 
were again harvested and surface soil sampled 
to assess changes in heavy metals content over 
the course of the growing season. Percentage 
vegetative cover was visually determined b\ 
two of the researchers. 

Plant tissue samples were dried in a gra\ it\ 
convection oven (Baxter model DS-64) at 
80°C for 48 h and weighed to determine total 
above-ground dry matter. Tissue was then 
ground in a Wiley mill (Bel-Art. Pequannock. 
New Jersey), and digested with hot (440 O 
H 2 S0 4 and H : : in a Haeh Digesdahl di- 
gestion apparatus. The digestate was diluted 
to 100 ml with deionized water. Flame atomic 
absorption spectroscopy (FAAS) using a Per- 
kin Elmer AAnalyst 200 was utilized to de- 
termine levels of Pb. Cd. and Zn. 

Soil was analyzed for total organic carbon 
(TOO usin£ loss on ignition (LOl) (360°C for 



150 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



2 h) (Nelson & Sommers 1982). Soil pH was 
measured using an AB15 Accumet® Basic pH 
meter in a 1:1 soikwater suspension. Total N 
was measured using the method of Bremner 
& Mulvaney (1986). Soil samples were ana- 
lyzed for extractable (in IN ammonium ace- 
tate) Ca, Mg and K using the method of Lan- 
yon & Heald (1986), and total Pb, Cd and Zn 
concentrations using the Hach Digesdahl® ap- 
paratus followed by FAAS analysis as de- 
scribed above for plant tissue. 

A chemical fractionation procedure (Spo- 
sito et al. 1982) was used to determine soil 
metal fractions. The soluble fraction was de- 
termined by mixing 2.0 g (dry weight) soil 
with 25 ml DI H 2 and then shaking in a re- 
ciprocating shaker (Eberbach 6010) for 2 h. 
The soil slurry was centrifuged (International 
Centrifuge Universal Model UV) for 15 min 
at 3000 rpm and the supernatant decanted. 
The procedure was repeated three times and 
the supernatants combined. The exchangeable 
fraction was determined by mixing the soil 
residue with 25 ml of 0.5 M KNO, for 16 h. 
The solution was then centrifuged and the su- 
pernatant decanted. The organic fraction was 
assessed by mixing the soil residue with 25 
ml of 0.5 M NaOH for 16 h. The carbonate 
fraction was determined by mixing the soil 
residue with 25 ml of 0.05 M Na 2 EDTA for 
6 h. The sulfide/residual fraction was assessed 
by mixing the soil residue with 1 3 ml of 4 M 
HN0 3 and heating for 16 h with 12 ml of 4 
M HNO3 added at the end of heating. All so- 
lutions were centrifuged and the supernatant 
decanted, and all supernatants were analyzed 
for Pb, Cd, and Zn using FAAS. 

Metal mobility study. — A growth chamber 
study was conducted to determine the effec- 
tiveness of different amendments for their 
ability to mobilize soil Pb, Cd, and Zn. Sur- 
face soil from the Superfund site was packed 
into 45 cm length PVC columns (5.1 cm i.d.), 
with a final density of approx. 1.1 g/cm 3 . The 
columns (five replicates each) were exposed 
to the following treatments: EDTA (0.001 M, 
0.01 M, 0.1 M, 0.5 M); citric acid (0.1 M), 
NPK solution (2.5 mM KNO, and 0.5 mM 
KH 2 P0 4 ); MSW; SS; and control. The MSW 
and SS were applied at field-equivalent rates. 
The MSW, SS and control were leached with 
deionized H 2 only. The columns were 
leached for 20 pore volumes using a Master- 
flex® peristaltic pump and the leachate was 



Table 1 . — Selected chemical properties of the Su- 
perfund soil. TOC = Total organic carbon. 



Parameter 


Mean 


Range 


PH 


8.3 


— 


TOC, % 


5.4 


5.3-5.5 


Total N, mg/kg 


0.11 


0.07-0.17 


CEC, cmol/kg 


8.4 


— 


Extractable, mg/kg 






Ca 


3681 


3785-3578 


Mg 


241.5 


229-254 


K 


61.5 


59-64 


Total metals, mg/kg 






Pb 


39,864 


35,040-49,520 


Cd 


10.1 


8.2-11.8 


Zn 


1512 


1000-2800 


Sand, % 


64.5 


— 


Silt, % 


14.1 


— 


Clay, % 


21.5 


— 



collected and stored in Nalgene® bottles with 
two drops of concentrated nitric acid (HN0 3 ) 
added. Leachates were then analyzed for total 
Pb, Cd and Zn using FAAS. 

Statistical analysis. — Data was analyzed 
statistically with analysis of variance (ANO- 
VA) using SPSS® (SPSS, 2006). The ANOVA 
was followed by post-hoc Bonferroni f-tests. 

RESULTS AND DISCUSSION 

Soil characterization. — Soil pH measured 
8.3, and TOC and total N contents were 5.4% 
and 0.11 mg/kg, respectively (Table 1). The 
high TOC values are a result of disposal of 
hydrocarbon solvents and used oil to the site. 
Total soil Pb, Cd, and Zn concentrations av- 
eraged 39,860, 10.1, and 1512 mg/kg, respec- 
tively. Concentration ranges of Pb, Cd, and Zn 
in uncontaminated soil are approximately 10- 
84 mg/kg, 0.06-1.1 mg/kg, and 10-80 mg/kg, 
respectively (Sigel et al. 2005; McBride 
1994). Metal levels in the current study varied 
widely, a result of the heterogeneity of soil 
materials at the site. Pichtel et al. (2000) mea- 
sured soil Pb concentrations of 29,400 mg/kg 
at the site, and Hee (2005) measured from 
1,900 to 6,050 mg/kg. Pichtel et al. (2000) 
measured average soil Cd concentrations up 
to 7.8 mg/kg at the site. 

Soil metal fractionation. — The majority of 
the control soil Pb occurred in the residual 
(45.7%) and carbonate (37.2%) fractions (Ta- 
ble 3). Soil Pb occurring in the soluble and 
exchangeable fractions was negligible (0.5 



JACOB ET AL.— PHYTOTREATMENT OF PB, CD AND ZN 



151 



Table 2. — Selected chemical properties of the or- 
ganic amendments used in this study. MSW = mu- 
nicipal solid waste, SS = dried sewage sludge; 
TOC = total organic carbon; 'extracted by 1 N am- 
monium acetate; 2 Pichtel & Anderson 1997. 



Parameter 


MSW 


SS 


pH 


6.8 


7.2 


TOC, g/kg 


390 


605 


Total P, g/kg 


6.2 


8.2 


Extractable, mg/kg 1 






Ca 


6850 


1460 


Mg 


608 


690 


K 


1620 


3480 


Total metal content, mg/kg 2 






Cr 


21 


57 


Cu 


28 


40 


Pb 


210 


340 


Zn 


655 


770 



and 0.2%, respectively). Using SEM-EDAX 
and x-ray diffraction analysis, Pichtel et al. 
(2001) identified both PbS0 4 (anglesite) and 
metallic Pb in soil from this site. Steele & 
Pichtel (1998) found a majority of soil Pb to 
occur in the organic (31%), carbonate (31%), 
and residual (35%) fractions of a Superfund 



soil. Zinati et al. (2004), Chlopecka et al. 
(1996), and Heil et al. (1996) all found soil 
Pb from contaminated sites to occur primarily 
in the carbonate and residual fractions. 

A substantial portion of control soil Cd oc- 
curred in the soluble (21.0%) and exchange- 
able (11.3%) fractions; however, the majority 
occurred in the less-available carbonate and 
residual fractions (48.7%) (Table 3). Jaradet et 
al. (2006) found 33% of soil Cd to occur in 
the exchangeable form, and Sanchez et al. 
(1999) found the greatest proportion of soil 
Cd to occur in the exchangeable fraction. 
Steele and Pichtel (1998) found a majority of 
soil Cd to occur in the residual (54.6%) and 
carbonate (36.9%) fractions of a Superfund 
soil. 

Soil Zn occurred predominantly in the re- 
sidual (60.4%) and carbonate (29.6%) frac- 
tions (Table 3). Soil Zn in the soluble and ex- 
changeable fractions measured 1.5 and 0.1%. 
respectively. Li et al. (2005) found a prepon- 
derance of soil Zn in residual and organic- 
bound fractions. Soil Zn speciation is strongly 
influenced by pH; as pH increases the relative 
proportions of Zn in the exchangeable fraction 
will decrease. The pH of the soil (8.3. Table 



Table 3. — Chemical fractions of Pb, Cd, and Zn in the Superfund soil. Values shown are mean values 
standard deviation. 



Treatment 


Soluble 


Exchangeable 


Organic 


Carbonate 


Resid 


ual 


Pb, % 




























EDTA 


0.7 


-+- 


0.1 


0.3 


± 0.05 


27.2 


+ 


9.8 


27.4 


+ 


2.9 


44.4 = 


13.8 


Citric acid 


0.8 


+ 


0.1 


0.2 


± 0.05 


18.9 


+ 


5.2 


31.2 


-i- 


7.6 


49.0 ± 


10.1 


MSW 


0.2 


+ 


0.02 


0.2 


± 0.01 


23.5 


+ 


11.0 


29.0 


+ 


12.4 


47.1 ± 


10.7 


SS 


0.4 


+ 


0.04 


0.3 


± 0.02 


29.5 


+ 


8.6 


34.1 


+ 


7.1 


35.7 ± 


10.6 


Control 


0.6 


+ 


0.1 


0.2 


± 0.04 


16.3 


+ 


4.6 


37.2 


± 


11.8 


45.7 = 


12.1 


Mean 


0.5 


+ 


0.07 


0.25 


± 0.03 


23.1 


■+■ 


7.8 


31.8 


-+■ 


9.9 


44.4 = 


1 1.5 


Cd, % 




























EDTA 


20.0 


+ 


2.1 


11.9 


± 2.9 


20.4 


+ 


1.0 


21.9 


± 


3.3 


25.8 ± 


3 . 3 


Citric acid 


22.3 


+ 


5.7 


8.1 


± 1.9 


20.3 


+ 


3.1 


18.9 


± 


5.7 


30.5 ± 


1.5 


MSW 


24.1 


+ 


2.4 


11.6 


± 1.1 


16.0 


+ 


2.1 


22.5 


r 


3.4 


25.8 = 


3.1 


SS 


21.4 


+ 


1.5 


17.3 


± 2.9 


13.9 


+ 


2.0 


19.5 


± 


> j 


27.9 ± 


2.0 


Control 


21.0 


+ 


2.2 


11.3 


± 1.7 


16.8 


+ 


0.4 


23.8 


± 


3.5 


27J ± 


0.2 


Mean 


21.8 


■+■ 


2.8 


12.0 


± 2.1 


17.5 


+ 


1.7 


21.3 


± 


3.8 


2^.4 ± 


2.0 


Zn, % 




























EDTA 


1.7 


■+■ 


0.2 


0.2 


± 0.04 


12.4 


+ 


2.0 


28.7 


± 


6.6 


5~.0 = 


15.2 


Citric acid 


1.7 


■+• 


0.1 


0.1 


± 0.01 


11.3 


+■ 


1.8 


23.9 


± 


7.0 


63.0 ± 


15.4 


MSW 


1.5 


+ 


0.2 


0.3 


± 0.1 


10.6 


± 


2.4 


•yi n 


± 


7.8 


64.3 = 


18.6 


SS 


1.7 


■+■ 


0.2 


0.3 


± 0.1 


14.5 


± 


3.7 


27.0 


± 


7.2 


56.5 ± 


16.0 


Control 


1.5 


-+- 


0.2 


0.1 


± 0.02 


8.4 


-I- 


2.0 


29.6 


± 


6.3 


60.4 = 


13.2 


Mean 


1.3 


H- 


0.2 


0.2 


± 0.1 


11.4 


± 


2.4 


26.5 


± 


7.0 


60.2 = 


15." 



152 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Table 4. — Dry matter yields and percent vegeta- 
tive cover on the treated plots. MSW = municipal 
solid waste; SS = dried sewage sludge; MT/ha = 
metric tons per hectare. 





Dry matter 


Coverage 


Treatment 


(MT/ha) 


(%) 


EDTA1 


0.42 ± 0.07 


70 


EDTA2 


0.13 ± 0.03 


85 


Citric acid 


0.58 ± 0.07 


70 


MSW 


1.09 ± 0.20 


90 


SS 


0.95 ± 0.34 


90 


Control 


0.54 ±0.15 


75 



1) will promote Zn precipitation as carbonates 
and other insoluble minerals (Lindsay 1979); 
likewise, high soil pH will render humic ma- 
terials more reactive with soil metals (Brady 
2000). 

Field study. — The grass mixture was the 
only plant treatment capable of survival on the 
site. Red clover, sunflower and ragweed grew 
early in the field experiment, but soon disap- 
peared from all plots. The loss is explained 
primarily by the inability of these species to 
tolerate the toxic conditions of the site, com- 
bined, to a lesser extent, with the massive soil 
structure and poor drainage conditions. 

Plant tissue dry matter yields ranged from 
1.09 MT/ha (MSW) to 0.13 MT/ha (EDTA2) 
(Table 4). Dry matter yields were significantly 
highest (P < 0.05) for the MSW and SS treat- 
ments. The MSW and SS treatments resulted 
in the greatest vegetative cover (both approx- 
imately 90%). Cover was 70% on the EDTA1 
and citric acid plots, 85% on the EDTA2 
plots, and 75% on control plots. The increased 
growth on the MSW and SS treatments is at- 



tributed in part to the added nutrient supply 
from both amendments, including Ca, Mg and 
K (Table 2). Conversely, the relatively lower 
yields in the EDTA and citric acid treatments 
may be a result of excess metal (including nu- 
trient base) solubilization. 

Amendment application did not significant- 
ly (P < 0.05) increase concentrations of soil 
Pb, Cd, or Zn in the soluble and exchangeable 
fractions when compared with the control (Ta- 
ble 3). Soil Pb was low in the soluble and 
exchangeable fractions (0.5% and 0.25% av- 
eraged over all treatments). These proportions 
are not significantly (P < 0.05) different from 
the control fractions (0.6% and 0.2%, respec- 
tively). Amendments did not significantly in- 
crease the presence of Cd in the soluble or 
exchangeable fractions; mean values for the 
amended plots were 21.8 and 12.0%, respec- 
tively, and those of the control averaged 21.0 
and 1 1.3%, respectively. Soluble or exchange- 
able soil Zn fractions did not change signifi- 
cantly (P < 0.05) with amendment addition; 
mean values for the amended plots were 1.6% 
and 0.2%, respectively, whereas those of the 
control were 1.5% and 0.1%, respectively. 

Metal uptake: Tissue Pb concentrations in- 
creased (P < 0.10) from June to October (31.7 
to 51.6 mg/kg, respectively) (Table 5) for all 
treatments. The EDTA, citric acid and MSW 
treatments resulted in increased Pb uptake 
when compared to the control; however, by 
the October sampling only the EDTA2 data 
was significantly (P < 0.05) higher than for 
the other treatments. Addition of amendments 
did not significantly increase the amount of 
plant-available (e.g., soluble and exchange- 
able) soil Pb fractions (Table 3). The EDTA2 



Table 5. — Grass shoot concentrations of Pb, Cd, and Zn, June and October samplings. Values shown 
are mean values ± standard deviation, n.d. = not determined; MSW = municipal solid waste, SS = dried 
sewage sludge. 





Pb (mg/kg) 


Cd i 


; mg/kg) 


Zn (i 


mg/kg) 


Treatment 


June 


Oct. 


June 


Oct. 


June 


Oct. 


EDTA1 


29.3 ± 19.2 


46.0 ± 21.8 


17.7 ± 1.5 


21.9 ± 5.1 


220.5 ± 60.8 


373.1 ± 188.8 


EDTA2 


62.0 ± 19.4 


108.5 ± 14.2 


8.6 ± 1.1 


9.9 ± 0.9 


n.d. 


n.d. 


Citric acid 


25.1 ± 14.6 


45.2 ±17.1 


21.5 ± 3.9 


20.3 ± 1 .9 


147.3 ±81.1 


358.8 ± 116.4 


MSW 


36.9 ±31.8 


59.3 ± 33.8 


16.7 ±5.7 


23.1 ±2.6 


118.3 ±38.8 


437.1 ± 24.7 


SS 


24.2 ± 23.4 


21.7 ± 17.1 


18.7 ± 3.4 


17.8 ± 2.3 


182.7 ± 104.1 


370.5 ±213.9 


Control 


12.9 ± 11.0 


29.0 ± 15.5 


10.2 ± 3.2 


13.7 ± 2.9 


117.5 ± 37.0 


387.3 ± 68.7 


All treatments 


31.7 


51.6 


15.6 


17.8 


157.3 


385.4 



JACOB ET AL.— PHYTOTREATMENT OF PB, CD AND ZN 



153 



treatment increased tissue Pb from 62.0 mg/ 
kg (June) to 108.5 mg/kg (October); the single 
EDTA application (EDTA2) apparently solu- 
bilized soil Pb and enhanced uptake and trans- 
port of Pb from roots to shoots. In contrast, 
the EDTA1 (multiple doses over the growing 
season) accumulated 46.0 mg/kg by October. 
The smaller EDTA applications (EDTA1) 
may have non-selectively reacted with soil Ca, 
Fe and other cationic metals instead of with 
Pb. Control tissue accumulated 12.9 mg/kg in 
June and 29.0 mg/kg in October. 

Tissue Pb concentrations in the MSW treat- 
ment increased from 36.9 mg/kg (June) to 
59.3 mg/kg (October). In the citric acid treat- 
ment Pb increased from 25.1 mg/kg (June) to 
45.2 mg/kg (October). Elevated tissue Pb con- 
centrations for both treatments (Table 5) could 
be explained by chelating effects. Zaccheo et 
al. (2002) and Chefet et al. (1998) determined 
the presence of a wide range of humic com- 
pounds in MSW compost, some of which may 
chelate metals and be of sufficiently low mo- 
lecular size to be taken up by roots. The in- 
crease in tissue Pb for the citric acid treatment 
is further explained by its acidifying effects: 
decreasing soil pH will increase the propor- 
tion of Pb in soil solution. Shen et al. (2002) 
reported an increase in cabbage (Brassica 
rapa) tissue Pb concentration with application 
of citric acid to soil. Correlation coefficients 
(r 2 ) for tissue Pb versus Pb chemical fractions 
ranged from 0.007 (exchangeable) to 0.52 (or- 
ganic), none of which were statistically sig- 
nificant. 

Cadmium uptake: The grasses accumulated 
15.6 mg/kg Cd in June and 17.8 mg/kg in Oc- 
tober, averaged over all treatments (Table 5). 
Effectiveness of amendments for plant Cd up- 
take (combined data for June and October) 
followed the order: citric acid = MSW = 
EDTA1 > SS > control > EDTA2. Addition 
of citric acid resulted in slight soil acidifica- 
tion (pH 7.9 compared to 8.5 in control plots) 
which may have increased the soil Cd avail- 
able for plant uptake. Soil pH is the single 
most important factor relating Cd mobility in 
soil and Cd plant-availability (McBride 2002). 
Citric acid additionally serves as an effective 
chelating agent (Patel and Subramanian 
2006). The MSW and EDTA1 resulted in sim- 
ilar tissue Cd accumulation. The MSW may 
have increased available Cd concentrations 
due to possible chelating effects. Cadmium 



accumulation in the EDTA2 treatment was 
comparable to the control; this is explained by 
excessive losses from the profile; a substantial 
proportion of soil Cd initially occurred in the 
soluble fraction (Table 3). High EDTA con- 
centrations may have accelerated Cd loss from 
the root zone. Correlation coefficients (r 2 ) for 
tissue Cd in relation to Cd chemical fractions 
ranged from 0.18 (soluble) to 0.36 (residual) 
thus showing no statistical relationship (data 
not shown). 

Zinc uptake: Averaged over all treatments, 
grasses at the site accumulated 157.3 mg/kg 
Zn (June 2005) and 385.4 mg/kg (October 
2005) (Table 5), a significant (P < 0.01) in- 
crease. There was no significant effect of 
amendments on Zn accumulation, however. 
By the October sampling the greatest Zn ac- 
cumulation occurred in the MSW treatment. 
This treatment also produced the highest bio- 
mass (Table 4), the highest October tissue Cd 
and the second highest October Pb tissue con- 
centrations (Table 5). The ability of MSW T to 
enhance Zn accumulation may be a result of 
possible chelating effects. Zinc availability in- 
creases with the addition of chelates (Fergus- 
son, 1990). Correlation coefficients for tissue 
Zn in relation to Zn chemical fractions ranged 
from 0.55 (exchangeable) to 0.65 (residual). 
however, with no significant statistical rela- 
tionship (data not shown). 

Leaching study: With increase in EDTA 
concentration the amount of Pb leached in- 
creased (P < 0.01). A total of 557 mg Pb was 
leached in the 0.5 M treatment, compared with 
203 mg for the 0.001 M treatment (Fig. I). 
Luo et al. (2006) and Blaylock et al. (1997) 
found that EDTA addition to soil greath in- 
creased Pb mobility. EDTA application in- 
creased soluble Pb concentrations from non- 
detectable to 4000 mg/1 (Huang et al. l L H)~t. 
Negligible Pb was leached with citric acid. 
NPK. MSW and SS treatments (Fig. 1). The 
complex organic molecules within amend- 
ments such as MSW and SS act to complex 
and chelate Pb; likewise, the pH values of the 
MSW and SS (6.8 and 7.2. respectively), ren- 
dered soil Pb immobile. The citric acid was 
apparently neutralized by the calcareous soil 
(pH 8.3, Table 1). 

Increased EDTA concentration significant 1\ 
(P < 0.01) increased the amount of Cd 
leached (Fig. 2). A total of 1 . 1 mg was 
leached with 0.5 M EDTA. compared with 



154 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



600.0 



500.0 



400.0 



£ 300.0 
E 

200.0 



100.0 - 



0.0 



u 



«? 



///// + * # v 



Figure 1. — Leaching of Pb from the Superfund soil as affected by extracting solution. 



0.04 mg at 0.001 M EDTA. These data are 
consistent with Elkhatib et al. (2001), who 
found that EDTA was effective for increasing 
Cd mobility in soil. Cajuste & Laird (2001) 
found that increasing EDTA concentration 
(0.01 M to 0.1 M) increased Cd leached from 
1.4 to 2.2 ug/g. The ability of EDTA to en- 
hance Pb availability more than Cd was noted 
by Bucheli-Witschel & Egli (2001) and Blay- 
lock et al. (1997). The citric acid, NPK, MSW, 



SS and control did not result in marked Cd 
leaching. 

EDTA leached greater amounts of Zn com- 
pared with any other amendment (Fig. 3). The 
0. 1 and 0.5 M rates of EDTA resulted in sim- 
ilar Zn leaching rates (960 and 997 mg, re- 
spectively). 

These data are consistent with those of 
Novillo et al. (2002), Grcman et al. (2001), 
and Alvarez et al. (1996), who found that ad- 




EDTA EDTA EDTA EDTA Citric NPK SS MSW Control 
0.001 0.01 0.1 0.5 Acid 
Figure 2. — Leaching of Cd from the Superfund soil as affected by extracting solution. 



JACOB ET AL.— PHYTOTREATMENT OF PB, CD AND ZN 

1200.00 



155 



1000.00 

800.00 

c 

o) 600.00 

£ 

400.00 

200.00 

0.00 



<? N , 






^ «? <s> # 



*r 



* / 



a- 



Figure 3. — Leaching of Zn from the Superfund soil as affected by extracting solution. 



dition of EDTA was effective for increasing 
Zn mobility in soil. The citric acid, NPK, 
MSW, SS and control resulted in minimal Zn 
leaching. 

LITERATURE CITED 

Almas A.R., J. Mulder, P. Lombnaes & T.A. Sogn. 
2006. Speciation of Cd and Zn in contaminated 
soils assessed by DGT-DIFS, and WHAM/Model 
VI in relation to uptake by spinach and ryegrass. 
Chemosphere 62:1647-1655. 

Alvarez, J.M., M.I. Rico & A. Obrador. 1996. Lix- 
iviation and extraction of zinc in a calcareous 
soil treated with zinc-chelated fertilizers. Journal 
of Agricultural and Food Chemistry 44:3383- 
3387. 

Bermond A., G. Varrault, V. Sappin-Didier & M. 
Mench. 2005. A kinetic approach to predict soil 
trace metal bioavailability: Preliminary results. 
Plant and Soil 275:21-29. 

Blaylock, M.J., D.E. Salt, S. Dushenkov, O. Zak- 
harova, C. Gussman, Y. Kapulnik, B.D. Ensley 
& I. Raskin. 1997. Enhanced accumulation of 
Pb in Indian mustard by soil applied chelating 
agents. Environmental Science and Technology 
31:860-865. 

Brady, N.C. 2000. The Nature and Properties of 
Soils. MacMillian, New York, New York. 

Bremner, J.M. & C.S. Mulvaney. 1986. Nitrogen- 
Total. Pp. 595-624, /// Methods of Soil Analysis, 
Part 2. Chemical and Microbiological Properties. 
(A.L. Page, R.H. Miller, & D.R. Keeney, eds.). 
Agronomy Monograph no. 9 (2nd ed.). ASA. 
Madison, Wisconsin, USA. 

Bucheli-Witschel M. & T Egli. 2001. Environ- 
mental fate and microbial degradation of ami- 



nopolycarboxylic acids. FEMS Microbiology 25: 
69-106. 

Cajuste, L.J. & R.J. Laird. 2001. The relationship 
between the phytoavailability and extractability 
of heavy metals in contaminated soil. Pp. 59-64. 
In Environmental Restoration of Metals Contam- 
inated Soils. (I.K. Iskandar. ed.). CRC Press 
LLC, Boca Raton, Florida. 

Chague-Goff C. 2005. Assessing the removal ef- 
ficiency of Zn, Cu, Fe and Pb in a treatment wet- 
land using selective sequential extraction: A case 
study. Water, Air, and Soil Pollution 160:161- 
179. 

Chefet, B., Y. Hadar & Y Chen. 1998. Dissolved 
organic carbon fractions formed during com- 
posting of municipal solid waste: Properties and 
significance. Acta Hydrochimiea et Hydrobiolo- 
gica 26:172-179. 

Chlopecka. A.. J.R. Bacon. M.J. Wilson & J. Kay. 
1996. Forms of cadmium, lead, and zinc in con- 
taminated soils from southwest Poland. Journal 
of Environmental Quality 25:69-79. 

Datta, R. & D. Sarkar. 2005. Phytoextractioii of Zn 
and Cd from Soils Using H\ per- Accumulator 
Plants. John Wiley and Sons. New York. New 
York. 

Elkhatib. E.A.. A.G. Thabet £ A.M. Mahdy. 2001. 
Phytoremediation o\' cadmium contaminated 
soils: Role o\' organic complexing agents in cad- 
mium phytoextraction. Land Contamination and 
Reclamation 9:147-154. 

Ernst. W. 2005. Phytoextraction of mine wastes — 
Options and impossibilities. Chemie der Hide 05: 
29-42. 

Fergusson. J.E. 1990. The Hea\ \ Elements. Per- 
simmon Press. Oxford. UK. 



156 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Grcman H., D. Lestan, S. Velikonja-Bolta, D. Vod- 
nik & B. Kos. 2001. EDTA enhanced heavy 
metal phytoextraction: Metal accumulation, 
leaching and toxicity. Plant and Soil 235:105- 
114. 

Hee, C. 2005. Field Scale Phytoremediation Trials 
of Lead and Cadmium Contaminated Soil. Mas- 
ter's thesis, Ball State University, Muncie, Indi- 
ana. 

Heil, D., A. Hanson & S. Zohrab. 1996. The com- 
petitive binding of peat by EDTA in soils and 
implications for heap leaching remediation. Ra- 
dioactive Waste Management and Environmental 
Restoration 20:111-127. 

Huang, W.J., J. Chen, W.R. Berti & S.D. Cunning- 
ham. 1997. Phytoremediation of lead-contami- 
nated soils: Role of synthetic chelates in lead 
phytoextraction. Environmental Science and 
Technology 31:800-805. 

Jaradat, Q.M., A.M. Massadeh, M.A. Zaitoun & 
B.M. Maitah. 2006. Fractionation and sequential 
extraction of heavy metals in the soil of scrap- 
yard of discarded vehicles. Environmental Mon- 
itoring and Assessment 112:197-210. 

Jensen, RE., L.M. Ottosen & A.J. Pedersen. 2006. 
Speciation of Pb in industrially polluted soils. 
Water, Air, and Soil Pollution 170:359-382. 

Lanyon, L.E.S. & W.R. Heald. 1986. Magnesium, 
calcium, strontium and barium. Pp. 247-262, In 
Methods of Soil Analysis, Part 2. (A.L. Page, 
R.H. Miller & D.R. Keeney, eds.) American So- 
ciety of Agronomy, Madison, Wisconsin. 

Lasat, M.M. The use of plants for the removal of 
toxic metals from contaminated soil. See: http:// 
www.clu-in.org/download/remed/lasat.pdf. Veri- 
fied 6/7/07. 

Li, J., Z.M. Xie, Y.G. Zhu & R. Naidu. 2005. Risk 
assessment of heavy metal contaminated soil in 
the vicinity of a Pb/Zn mine. Journal of Envi- 
ronmental Sciences 17:881-885. 

Lindsay, W.L. 1979. Chemical Equilibria in Soils. 
Wiley, New York, New York. 

Luo, C, X. Li, Z. Shen & L. Lou. 2006. EDDS 
and EDTA-enhanced phytoextraction of metals 
from artificially contaminated soil and residual 
effects of chelant compounds. Environmental 
Pollution 144:862-871. 

Madejon, E., A. P. de Mora, E. Felipe, P. Burgos & 
F. Cabrera. 2006. Soil amendments reduce trace 
element solubility in a contaminated soil and al- 
low regrowth of natural vegetation. Environmen- 
tal Pollution 139:40-52. 

McBride, M.B. 1994. Environmental Chemistry of 
Soils. Oxford University Press, Oxford, UK. 

McBride. M.B. 2002. Cadmium uptake by crops 
estimated from soil total Cd and pH. Soil Sci- 
entist 167:62-67. 

Mench, M., H. Vangronsveld, N. Clisters, W. Lepp 
& R. Edwards. 2000. In situ metal immobili- 



zation and phytostabilization of contaminated 
soils. Pp. 112-139, In Phytoremediation of Con- 
taminated Soil and Water (N. Terry & G. Ban- 
uelos, eds.) Lewis Publishers, Boca Raton, Flor- 
ida. 

Mills T, G. Northcott, I. Vogeler, B. Robinson, C. 
Norling, D. Leonil, B. Arnold & S. Sivakumaran. 
2006. Phytoremediation and long-term site man- 
agement of soil contaminated with pentachloro- 
phenol (PCP) and heavy metals. Journal of En- 
vironmental Management 79:232-241. 

Nelson, D.W. & L.E. Sommers. 1982. Total car- 
bon, organic carbon, and organic matter. Pp. 
539-579, In Methods of Soil Analysis. Part 2. 
(A.L. Page, R.H. Miller and D.R. Keeney, eds.). 
American Society of Agronomy, Madison, Wis- 
consin. 

Novillo, J. A. Obrador, L.M. Lopez- Valdivia & J.M. 
Alvarez. 2002. Mobility and distribution of zinc 
forms in columns of an acid, a neutral, and a 
calcareous soil treated with three organic zinc 
complexes under laboratory conditions. Austra- 
lian Journal of Soil Research 40:791-803. 

Patel, M. & R.B. Subramanian. 2006. Effect of a 
chelating agent on lead uptake by Spinacia oler- 
acea. Pollution Research 25:77-79. 

Pichtel, J. & D. Bradway. In press. Conventional 
crops and organic amendments for Pb, Cd and 
Zn treatment at a severely contaminated site. 
Bioresource Technology. 

Pichtel, J., K. Kuroiwa & H.T. Sawyerr. 2000. Dis- 
tribution of Pb, Cd, and Ba in soils and plants of 
two contaminated sites. Environmental Pollution 
110:171-178. 

Pichtel, J., B. Vine, P. Kuula-Vaisanen & P. Niska- 
nen. 2001. Lead extraction from soils as affected 
by lead chemical and mineral forms. Environ- 
mental Engineering Science 18:91-98. 

Pulford, I.D. & C. Watson. 2003. Phytoremediation 
of heavy metal-contaminated land by trees — A 
review. Environment International 29:529-540. 

Sanchez, A.G., A. Moyano & C. Munez. 1999. 
Forms of cadmium, lead and zinc in polluted 
mining soils and uptake by plants (Soria prov- 
ince, Spain). Communications in Soil Science 
and Plant Analysis 30:1385-1402. 

Selim, H.M. & W.L. Kingery. 2003. Geochemical 
and Hydrological Reactivity of Heavy Metals in 
Soils. CRC Press, Boca Raton, Florida. 

Shen, Z., X. Li, C. Wang, H. Chen & H. Chua. 
2002. Lead phytoextraction from contaminated 
soil with high biomass plant species. Journal of 
Environmental Quality 31:1893-1900. 

Sigel, A., H. Sigel & R. Sigel (eds.). 2005. Metal 
Ions in Biological Systems. Taylor and Francis. 
Boca Raton, Florida. 

Sposito, G., L.J. Lund & A.C. Chang. 1982. Trace 
metal chemistry in arid-zone field soils amended 
with sewage sludge: I. Fractionation of Ni, Cu, 



JACOB ET AL.— PHYTOTREATMENT OF PB, CD AND ZN 



157 



Zn, Cd and Pd in solid phases. Soil Science So- 
ciety of America Journal 46:260-264. 

SPSS Inc. 2006. SPSS Guide to Data Analysis. 
See: http://www.norusis.com/ 

Steele, M.C. & J. Pichtel. 1998. Ex-situ remedia- 
tion of a metal-contaminated superfund soil us- 
ing selective extractants. Journal of Environ- 
mental Engineering 124:639-645. 

Tie, M., T. Sun, H. Li, Y. Liang, S. Zang & W. Pan. 
2006. Chemical forms of cadmium in industrial 
contaminated soil and its phytoremediation. Chi- 
nese Journal of Applied Ecology 17:348-350. 

U.S. Environmental Protection Agency. 2007. Su- 
perfund (CERCLIS). See: http://oaspub.epa.gov/ 
enviro/cerclis_web.report?pgm_sys_id=IN00018 
99269. Verified 6/8/07. 

Vangronsveld, J., J. Colpaert & V.K. Tichelen. 
1995. Reclamation of a bare industrial area con- 
taminated by non-ferrous metals: In situ metal 
immobilization and revegetation. Environmental 
Pollution 87:51-59. 



Wilde, E.W., M.A. Heitkamp, D.C. Dagnan. R.L. 
Brigmon & D.L. Dunn. 2005. Phytoextraction 
of lead from firing range soil by Vetiver grass. 
Chemosphere 61:1451-1457. 

Wu, S.C., M.H. Wong, K.C. Cheung & Y.M. Luo. 
2006. Effects of inoculation of plant growth-pro- 
moting rhizobacteria on metal uptake by Bras- 
sica juncea. Environmental Pollution 140:124- 
135. 

Zaccheo, P., G. Ricca & L. Crippa. 2002. Organic- 
matter characterization of composts from differ- 
ent feedstocks. Compost Science and Utilization 
10:29-38. 

Zinati, G.M., M. Codallo, Y. Li, H.H. Bryan & R.S. 
Mylavarapu. 2004. Distribution and fraction- 
ation of phosphorous, cadmium, nickel, and lead 
in calcareous soils amended with composts. Jour- 
nal of Environmental Science and Health — Part 
B Pesticides, Food Contaminants, and Agricul- 
tural Wastes 39:209-223. 



2007. Proceedings of the Indiana Academy of Science (1 16): 158-172 



LAND TYPE ASSOCIATION DELINEATION AND SPATIAL 

ANALYSIS FOR THE HOOSIER NATIONAL FOREST IN 

SOUTHERN INDIANA 

Andriy V. Zhalnin and George R. Parker: Department of Forestry and Natural 
Resources, Purdue University, West Lafayette, Indiana 47907 USA 

ABSTRACT. A GIS approach was used to delineate Land Type Association (LTA) map units for the 
Hoosier National Forest (HNF) area. It was assumed that the spatial distribution pattern of Ecological 
Land Type (ELT) and Ecological Land Type Phase (ELTP) map units provide a theoretical foundation for 
LTA unit delineation. A semi-automated approach using visual detection of areas of different ELT patterns 
followed by multivariate statistical analysis and clustering was used for LTA delineation. This resulted in 
four LTAs for the Brown County Hills subsection (Pleasant Run unit of HNF) and six LTAs for the 
Crawford Upland subsection (Lost River, Patoka River and Tell City units of the HNF). Differentiating 
criteria included, in the general order of most frequent use, patterning of ELTPs and soil survey units, 
landforms, bedrock type, dominant tree species occurrence, and disturbance processes. LTA boundary 
identification was based on physiographic boundaries such as stream channels or watershed boundaries 
(ridges). All units are nested within boundaries of subsequent upper and lower hierarchical units (Subsec- 
tion <-> LTA <-» ELT <-> ELTP). Spatial statistics on ELT, ELTP, soils, erosion, and elevation are reported 
to highlight differences between LTAs. Mapped LTA units will help to apply effectively those management 
activities that require a spatially specific application, such as controlled fire and selection cuttings, or 
recreational planning. Map units also will provide a valuable tool for a researcher allowing to improve 
sampling strategy and have a solid ecological foundation in interpretation of results. Maps can outline 
areas with different biological and/or ecological potentials. 

Keywords: Ecological classification, GIS, landscape analysis, land type mapping, forest ecosystems 



One of the critical components of decision 
making in natural resource management is 
ecological information. An ecological classi- 
fication (ECS) framework allows identifica- 
tion of land areas with similar properties at 
different scales for the purpose of manage- 
ment, research and education. This framework 
is hierarchical: the smaller map units compose 
larger units ("bottom-up ,, approach) or are 
created by subdividing the next larger unit 
("top-down ,, approach), and nested: bound- 
aries of smaller units do not cross those of 
larger units. The USDA Forest Service adopt- 
ed a policy of ecosystem management on 4 
June 1992 that applied to national forests and 
grasslands research programs. Later, an Eco- 
logical Classification and Mapping Task Team 
(ECOMAP) was formed to develop a consis- 
tent approach to ecosystem classification and 
mapping at multiple geographic scales. Other 
agencies such as the USDA Soil Conservation 
Service and The Nature Conservancy also 
contributed to the development of the frame- 
work, and it was adopted as the National Hi- 



erarchical Framework of Ecological Units 
(NHFEU) in 1993 (ECOMAP 1993). Bailey's 
classification of US ecoregions (Bailey 1980) 
was accepted as upper levels of ECS at the 
global scale (domains, divisions and provinc- 
es) and mapped (Keys et al. 1995). Regional 
scale (sections) was described and mapped in 
1994 (McNab & Avers 1994). 

Subsections in Indiana that represent next 
lower level in ECS follow Homoya's natural 
regions of Indiana (Homoya et al. 1984). In 
1993, a multifactor ecological classification 
described ecological land types (ELT) and 
ecological land type phases (ELTP) for the 
Brown County Hills (BCH) and Crawford Up- 
land (CU) subsections of the Hoosier National 
Forest (HNF) in southern Indiana (Van Kley 
1993), based on vegetation, soils and physi- 
ography. The leading environmental factors 
that were correlated to variation in vegetation 
composition were aspect, soil A-horizon 
depth, slope position, and soil pH. Thus, veg- 
etation is believed to be related to a moisture 
and nutrient gradient which is influenced by 



158 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



159 



the above mentioned factors. Van Kley (1993) 
identified 12 ELTPs for the BCH and 15 
ELTPs for the CU subsection. He also de- 
scribed seven ecological plant species groups 
for the Brown County Hills and eight groups 
for the Crawford Upland subsection. 

An important step in bringing ecological 
theory into application is mapping ecological 
classification units on the landscape. This 
study defines and describes Land Type Asso- 
ciation units (local level) for the BCH and CU 
subsections within the HNF purchase bound- 
ary. When mapped, LTA units reflect infor- 
mation on medium-scale influences of local 
disturbance regimes, biological productivity 
and resiliency, hydrologic patterns and func- 
tions within each unit. Such information is 
readily available for a manager to be used in 
planning, and managing natural resources at 
the local scale. Using map units in managerial 
activities and research will help to build fur- 
ther knowledge of local ecosystems and im- 
prove their management. Mapping LTA units 
will help to apply effectively those manage- 
ment activities that require a spatially specific 
application, such as controlled fire or selection 
cuttings. Recreational planning is another ex- 
ample of possible use of mapped ecological 
units at fine to medium scale. Parker & Whit- 
comb (2002) demonstrated an application of 
mapped LTAs in studying patterns of dis- 
persed campsites on the Chippewa National 
Forest. 

Map units also provide a valuable tool for 
a researcher allowing to improve sampling 
strategy and have a solid ecological founda- 
tion in interpretation of results. Maps can out- 
line areas with different biological and/or eco- 
logical potentials. LTA map units can be used 
for example to assess and rank the quality of 
Indiana bat habitats (DeMeo 2002). 

Since LTA boundaries are independent of 
political or property lines, in most cases these 
LTA units will extend beyond the purchase 
boundary and should be considered as an in- 
termediate step to mapping all LTAs within 
respective subsection boundaries. 

METHODS 

Study area. — The study area, located in 
south-central Indiana, included four units of 
the Hoosier National Forest situated within 
the Brown County Hills and Crawford Upland 
subsections according to the eastern United 



Hoosier National Forest 
ECS subsections 

222De Crawford Upland 



}] 222Df Escarpment 

222Em Brown County Hills 




Figure 1. — A map of study area. Four Hoosier 
National Forest units and Eeologieal Classification 
System subsections are shown on the map. 



States classification (McNab & Avers W4> 
(Fig. 1). The area is underlain by Paleozoic 
sedimentary bedrock (Gutshick 1966). Parent 
material of this area is early-to-middle Mis- 
sissippian age siltstones and shales of the Bor- 
den group (Schneider 1966). Prevailing soils 
of this area are acid silt Loams formed from 
weathered bedrock and small areas of loess 
(Homoya et al. 1984). Typical relief consists 
of uplands dissected by creeks with steep 
slopes and narrow hollows. 

Van Kiev (1993) developed an ecological 



160 



PROCEEDINGS OF T -IE INDIANA ACADEMY OF SCIENCE 




Figure 2. — LTA subunit boundaries for the Hoosier National Forest area. Numbers refer to temporary 
unit labels in analysis. 



classification for the forest that includes 12 
ELTPs for the Brown County Hills subsection 
and 15 ELTPs for the Crawford Upland sub- 
section. Important factors affecting the clas- 
sification were landscape physiography and 
soil parental material at the ELT level and 
physiography, A-horizon depth, and vegeta- 
tion at ELTP level. ELT map units for the 
HNF were delineated using GIS tools (Shao 
et al. 2004). 

Methodology. — LTA delineation was made 
using ELTP maps produced by Zhalnin 
(2004). It was assumed that natural features of 
the landscape such as watershed boundaries or 
streams are appropriate LTA boundaries. 
Therefore, at the first stage we visually ana- 
lyzed the ELTP map and defined areas differ- 
ent in ELTP spatial pattern and separated by 
natural landscape boundaries (streams or ridg- 
es) to establish boundaries of LTA subunits 
(Fig. 2). Areas covered with large water bod- 
ies were excluded. The next step was to ana- 
lyze each subLTA with the Patch Analyst ex- 
tension for ArcView GIS (Rempel & Carr 
2003) for differences in ELTP spatial pattern. 
Mean Patch Size (MPS) and Mean Proximity 
Index (MPI) were selected from a variety of 
spatial metrics suggested by the program as 
metrics that most reflected spatial differences 
between subLTAs according to our previous 



study (Zhalnin et al. 2002). Mean Patch Size 
is a mean area of each ELTP unit within each 
LTA. Mean Proximity Index uses the nearest 
neighbor statistics and is a measure of the de- 
gree of isolation and fragmentation of map 
units within each ELTP class. Each metric was 
calculated for each ELTP class separately. 
Next, we used multivariate statistics analysis 
(Principal component analysis, PCA, and De- 
trended Correspondence Analysis, DCA) to 
group subLTAs into final LTA units. In addi- 
tion, other sources of information (GIS layers) 
were used to identify specific areas and in- 
corporate them into the LTA classification, 
such as a bedrock map (USGS data layer), the 
STATSGO and SSURGO soil maps (NRCS 
Soil Survey data layers), and the map of dis- 
tribution of chestnut oak (Que reus prinus) in 
the HNF area, interpolated from 5 1 1 sample 
points collected in the field. More detailed de- 
scription of delineation testing procedure can 
also be found in USDA General Technical Re- 
port NE-294 (Zhalnin et al. 2002). The entire 
LTA map was converted into polygon version 
and grid GIS layer, and projected in the Uni- 
versal Transverse Mercator (UTM) coordinate 
system and North American Datum of 1927 
(NAD 27) to correspond to other GIS data 
layers within USGS Forest Service database. 
All GIS work was done in AcView GIS 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



161 



3.3(ESRI 2002). PC-ORD statistical package 
was used for multivariate statistical analysis 
of metrics (McCune & Mefford 1999). 

Naming of LTA units. — Names of LTA 
consist of a numerical designation of a re- 
spective subsection (e.g., 222Em for the BCH 
subsection), subsequent number of LTA, and 
a verbal description. LTAs were described ac- 
cording to canopy species prevailing within an 
LTA, prevailing moisture conditions (based on 
moisture gradient within an LTA), and land- 
scape forms typical for the LTA. 

Spatial metrics within LTAs. — Spatial 
metrics were calculated for ELTPs and ELTs 
within each LTA unit using the Patch Analyst 
application in ArcView to characterize indi- 
vidual units. Shannon's Diversity Index (SDI) 
and Shannon's Evenness Index (SEI) were 
used to describe ELTP and ELT diversity of 
the landscape. They are useful for estimating 
landscape value and comparing different 
LTAs. SDI and SEI are calculated using the 
following formulas: 



SDI = -2 (PflnPd 



(1) 



-E (PfinPt) 

SEI = -^ , (2) 

In m 

where P { is a proportion of landscape occupied 
by ELTP or ELT /, and m is a number of 
ELTPs or ELTs present in the landscape. 

SDI index is sensitive to occurrence of rare 
land types: the higher SDI value, the more 
unique ELTPs or ELTs occur within the LTA. 
SEI measures the other aspect of landscape 
diversity — the distribution of area among 
ELTP or ELT patches. As the evenness index 
approaches "1," the observed diversity ap- 
proaches perfect evenness, when LTA is char- 
acterized by environmentally homogeneous 
landscape with equally sized ELTPs or ELTs. 
Soil and erosion information was obtained 
from respective GIS layers (SSURGO data 
layers) for each LTA map unit. An Area 
Weighted Erosion index (AWEi) was calcu- 
lated to estimate the degree of soil erosion 
within LTAs using the following formula: 



AWEi 

EIA1 + E2-A2 + 



+ En An 



A\ + A2 + 



+ An 



(3) 



where En is an Erosion Class defined by the 
Soil Survey Manual (1993) and An is the area 
occupied by a corresponding soil Erosion 
Class. Higher values represent higher degree 
of erosion. 

RESULTS 

LTA delineation. — Delineation procedure 
is explained on the example from the Pleasant 
Run Unit of the HNF within the Brown Coun- 
ty Hill subsection. We defined 20 subunits that 
were naturally separated within the landscape 
(subLTAs, Fig. 2) in the first stage. Subunits 
5, 6, 8, 13, 16, 17 and 18 were excluded from 
the further analysis for the following reasons: 
units 5 and 6 were added to unit 7, since they 
are small parts of larger units that lie outside 
the HNF boundary and visually resemble unit 
7. Unit 8 has a unique pattern of ELTP spatial 
distribution. Units 13 and 16, as well as 17 
and 18 represent two LTAs that have a dis- 
tinctive difference from the rest of the area 
due to the pattern of limestone soil occur- 
rence. 

The ultimate reasonable number of LTAs 
for the Pleasant Run unit of the Hoosier Na- 
tional Forest is 4-5 units according to the 
LTA unit size suggested in National Hierarchy 
(Cleland et al. 1997). Three of the subunits 
were reserved for areas that have distinctive 
features. The delineation of remaining two 
LTAs was based on multivariate statistics re- 
sults. PCA of Mean Patch Size revealed that 
in general subunit variability forms two clus- 
ters: first— subLTAs 1. 2. 9, 10. 12. 14. and 
15; second— subLTAs 3. 4. 7. 11. 19, and 20. 
Subunit 19 was an "outlier" on the graph, but 
still can be considered closer to group 2 than 
to group 1 (Fig. 3, top). Groups from DCA 
were less distinctive: group one — 2. 3, 11. 12. 
15. and 20; group two — 7. 9, 10. 14. and 19. 
Subunits 1 and 4 were not closely associated 
with the rest of the subunits (Fig. 3. bottom). 
Results of multivariate analysis of Mean Prox- 
imity Index using PCA statistics suggest two 
clusters: group one — 9, 10, and 14; group 
two — 1, 2. 3, 4. 7. 11. 15, and 20. Subunit 19 
was again an "outlier"" on the graph (Fig. 4. 
top). DCA method shows two clusters: group 
one— 1. 2. 3, 4. 7. 11. 12. 15, L9, and 20: 
group two — 9, 10. and 14 (Fig. 4. bottom). 
Groups determined by these analyses were 
used to delineate two additional LTAs. The 



162 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 




DCA 




Axisl 

1 



Figure 3. — Results of PC A (top) and DCA (bot- 
tom) statistics for LTA subunits within the Pleasant 
Run Unit of the Hoosier National Forest, the Brown 
County Hills subsection. Axes represent mean area 
size variability of 12 ELTPs among subunits. 



1 .*> — [ 


Axis2 






DCA 


1.2- 










0.9- 


I 1 9 

1 A ' 


12 \ 
▲ \ 






0.6- 










O.d- 


A 
A ' A 3 


7 \ 
▲ \ 

20 £/ 

— -\- I 


I 


/ 9 

/ A 
/ ▲ 




Axisl 

1 1 






03 06 


09 


12 15 



Figure 4.— Results of PCA (top) and DCA (bot- 
tom) statistics for LTAs within the Pleasant Run 
Unit of the Hoosier National Forest, the Brown 
County Hills subsection. Axes represent mean prox- 
imity index variability of 12 ELTPs among sub- 
units. 



same approach was used for delineation of 
LTAs within the Crawford Upland subsection. 
As a result of visual and statistical analysis, 
four LTAs are described for the BCH subsec- 
tion (the Pleasant Run unit of HNF, Fig. 5) 
and six LTAs are described for the CU sub- 
section (Lost River, Patoka River and Tell 
City units of HNF, Fig. 6) as follows: Brown 
County Hills subsection: 1) LTA 222Em01, 
Mixed Oak Dry-Mesic Upland Hills; 2) LTA 
222Em02, Chestnut Oak Dry-Mesic Upland 
Hills; 3) LTA 222Em03, Oak-Maple Mesic 
Upland Plateau; 4) LTA 222Em04, Oak-Ma- 
ple Calcareous Mesic Upland Hills. Crawford 
Upland subsection: 1) LTA 222De01, White 
Oak Dry-Mesic Upland Hills; 2) LTA 
222De02, Chestnut Oak Dry Upland Hills; 3) 
LTA 222De03, Oak-Maple Calcareous Upland 
Hills; 4) LTA 222De04, Oak-Maple Wet-Me- 



sic Dissected Plateau; 5) LTA 222De05, 
Mixed Oak Dry Upland Hills; 6) LTA 
222De06, Post Oak Dry Upland Hills. 

LTA descriptions for the Brown County 
Hills subsection. — LTA 222Em01, Mixed Oak 
Dry-Mesic Upland Hills: This LTA is located 
in central part of the Pleasant Run (PRN) unit, 
extending from the northern to the southern 
purchase boundary, and is the largest LTA in 
the BCH subsection (Fig. 5). Dry ridge and 
dry slope ELTPs 10 and 20 of this LTA are 
dominated in some parts by Quercus alba 
(northwestern corner of PRN unit above the 
Monroe Lake) or Quercus prinus or both. In 
the southwestern part, this LTA borders LTA 
222Em04, Oak-Maple Calcareous Mesic Up- 
land Hills, and a few calcareous ELTPs 13 and 
23 may occur along that border. In the south- 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



163 



222Em 
n Brown County Hills 
Subsection 




LTA222£m01 

:ii!!]]]LTA222Eni02 

LTA 222Em03 

_ LTA 222Em04 

Figure 5. — Land Type Associations (LTA) of the Brown County Hills subsection (Pleasant Run unit of 
the Hoosier National Forest). LTA boundaries are restricted to purchase boundaries of the Hoosier National 
Forest. 



eastern part, this LTA borders LTA 222Em02 
(Chestnut Oak Dry-Mesic Upland Hills). The 
southern part of this LTA may justify sepa- 
rating from the northern in the area of Hickory 
Ridge Road as more information is obtained 
on characteristics of the region to the south of 
PRN Unit purchase boundary. 

LTA 222EmO 1 has a mean elevation of 208 
m with the highest point at 284 m and the 
lowest at 164 m (Table 1). It has the largest 
area (21,116 ha) and the highest proportion of 
dry slope ELTP 20 (32.5%) among all LTAs 
of the BCH subsection (Table 2). In general, 
the dominate ELTPs were dry slope ELTP 20, 
mesic slope ELTP 22, and mesic ridge ELTP 
12 (32.5, 27.9, and 9.3%, respectively). The 
dry slope ELTP 20 has the largest mean area 
among all BCH subsection LTAs (6.66 ha). 
Shannon's Index of Diversity was 1.89 for 
ELTPs and 1.21 for ELTs (the latter is highest 
in the BCH subsection). 

This LTA has the largest percentage of soils 
in Erosion Class 1 and least percentage in 
Erosion Class 2 (85.7 and 6.8%, respectively. 
Table 3) indicating relatively slight distur- 
bance of soils in this area. The Area Weighted 
Erosion index (AWEi) is 1.07, the smallest 
among the BCH LTAs. The major soil survey 



map units within this LTA are Brownstown- 
Trevlac-Kurtz silt loams. 20-70% slopes: 
Brownstown-Gilwood silt loams. 25-75% 
slopes; Wellrock-Brownstown-Trevlac silt 
loams, 6-20% slopes which cover 27.4. 16.7 
and 9.6% of the area, respectively. 

LTA 222Em02, Chestnut Oak Dry-Mesic 
Upland Hills: This LTA is located in south- 
eastern part of the PRN Unit and characterized 
by deeply dissected uplands underlain by silt- 
stone, shale, and sandstone. From the south 
this area borders the outwash plain of the East 
Fork of the White River. The typical topog- 
raphy is represented by exposed hills, mesic 
ravines, and river floodplains. Dry ridge and 
dry slope ELTPs 10 and 20 of this LTA are 
dominated by Quercus prinus. and sites with 
Q. alba as a dominant species are uncommon. 
The current spatial pattern of these two spe- 
cies may be in differences of their response to 
past disturbance or differences in relation to 
ecological factors, namely soil moisture and 
nutrient content. In general, this area is known 
to be on the border o\' the O. prinus natural 
range, apparently due to climatic conditions. 
and also may contribute to the intricate spatial 
variation of these species. 

LTA 222Em02 has a mean elevation of 222 



164 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



LTA222De01 
LTA 222De02 
] LTA222De03 
LTA 222De04 
LTA 222De05 
fflVW LTA222De06 




Figure 6. — Land Type Associations of the Craw- 
ford Uplands subsection (Lost River, Patoka River 
and Tell City units of the Hoosier National Forest). 
LTA boundaries are restricted to purchase bound- 
aries of the Hoosier National Forest. 



m with the highest point at 273 m and the 
lowest at 173 m. It has the second largest area 
( 1 1 ,627 ha) and the highest proportion of 
broad floodplain ELTP 42 and ELTP 43 (13.3 
and 8.2%, respectively) among all LTAs of the 
BCH subsection due to the adjacent East Fork 
of the White River. The dominate ELTPs 
within this LTA are mesic slope ELTP 22, dry 
slope ELTP 20 and mesic broad bottomland 
ELTP 42 with 28.0, 25.8 and 13.3% of the 
area, respectively). The mesic slope ELTP 22, 
wet-mesic bottomland ELTP 41, and flood- 
plain ELTPs 42 have the largest mean area 
among all the BCH subsection LTAs with 
4.95, 5.80, and 35.96, respectively (Table 2). 



This LTA has the largest proportion of area 
and the largest mean size of bottomland ELT 
4 (29.5% and 31.75 ha, respectively). Shan- 
non's Index of Diversity was 1.86 for ELTPs 
and 1.13 for ELTs (Table 1). 

This LTA has the largest percentage of soils 
in Erosion Class 3 (5%). AWEi equals 1.25 
which is the second highest among BCH 
LTAs. The dominate soil survey map units 
within this LTA are Brownstown channery silt 
loam (25-75% slopes), Gilwood-Wrays silt 
loams (10-25% slopes) Gnawbone silt loam 
(25-55% slopes) covering 26.8, 18.0 and 
15.2% of the area, respectively). 

LTA 222Em03, Oak-Maple Mesic Upland 
Plateau: This LTA is located in north-eastern 
part of the PRN Unit and characterized by 
wide and relatively level mesic ridges under- 
lain by soils formed in 0.6-1 m of loess. His- 
torically, this area was and still is heavily dis- 
turbed by agricultural practices and has the 
highest percentage of eroded soils among all 
LTAs of the PRN unit. Most exposed sites 
within ELTPs 10 and 20 of this LTA are dom- 
inated by Q. prinus; however, the majority of 
sites are dominated by Q. alba. 

LTA 222Em03 has a mean elevation of 217 
m with the highest point at 292 m and the 
lowest at 164 m. It has the smallest area of 
all LTAs (4735 ha) and the highest proportion 
of mesic ridge ELTP 12 (51.2%) among all 
LTAs of the BCH subsection. The dominate 
ELTPs within this LTA were mesic ridge 
ELTP 12, mesic slope ELTP 22, and mesic 
broad bottomland ELTP 42 with 51.2, 12.7, 
and 12.0% of the area, respectively). The dry- 
mesic ridge ELTP 11, mesic ridge ELTP 12, 
and mesic bottomland ELTP 40 have the larg- 
est mean area among all the BCH subsection 
LTAs (1.11, 34.14, and 2.23 ha, respectively). 
This LTA has the largest area proportion and 
largest mean size of ridge ELT 1 (51.8% and 
29.56 ha). Shannon's Index of Diversity was 
1.93 for ELTPs and 1.17 for ELTs (Table 3). 

This LTA has the largest percentage of soils 
in Erosion Class 2 (50%), second highest per- 
centage in Erosion Class 3 (4.7%). This LTA 
is the only one among those described that has 
5.2% of the area covered by gullied soils (Ero- 
sion Class 5). AWEi is 1.81 and is the highest 
among the BCH LTAs. High percentages of 
erosion in this LTA are due to broad mesic 
plateaus that have been used extensively for 
agriculture. The dominate soil survey map 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



165 



Table 1. — Elevation (meters) and diversity statistics of Land Type Association map units of the Brown 
County Hills subsection (the Pleasant Run unit of the Hoosier National Forest). SDI — Shannon's Diversity 
Index; SEI — Shannon's Evenness Index. 



LTA 222Em01 


LTA 222Em02 


LTA 


222Em()3 


LTA 


222Em04 


Elevation, m 
















Mode 


Max 


Mode 


Max 


Mode 


Max 


Mode 


Max 


(Mean) 


(Min) 


(Mean) 


(Min) 


(Mean) 


(Min) 


(Mean) 


(Mini 


164 


284 


229 


273 


174 


292 


164 


275 


(208) 


(164) 


(222) 


(173) 


(217) 


(164) 


(210) 


(164) 


Landscape diversity 
















SDI 


SEI 


SDI 


SEI 


SDI 


SEI 


SDI 


SEI 


ELTP 1.89 


0.74 


1.86 


0.78 


1.93 


0.80 


1.95 


0.76 


ELT 1.21 


0.88 


1.13 


0.81 


1.17 


0.84 


1.16 


0.84 



units within this LTA are Brownstown-Trev- 
lac-Kurtz silt loams (20-70% slopes), Stone- 
head-Trevlac silt loams (10-20%, eroded 
slopes), Stonehead silt loam (4-12% eroded 
slopes) covering 16.9, 16.2, and 14.0% of the 
area, respectively. 

LTA 222Em04 Oak-Maple Calcareous Me- 
sic Upland Hills: This LTA is located in east- 
ern part of the PRN Unit and characterized by 
calcareous ELTPs. The eastern boundary of 
this LTA lies on the eastern edge of the Frog 
Pond Ridge then follows Hickory Ridge Road 
and goes south along Hunter Creek and Little 
Salt Creek. The majority of calcareous ELTPs 
occur in the area of the Frog Pond Ridge and 



Little Salt Creek. However, several ELTPs 
were detected in Aliens Creek and Hardin 
Ridge State Recreational Areas on the slopes 
adjacent to Monroe Lake. Other locations may 
occur sporadically within the area in locations 
associated with Corydon or Crider soil series. 
Geologically, this area is within the transition- 
al zone extending from the Norman Upland to 
the Mitchell Karst Plain. It may be similar to 
areas across Monroe Lake within the Mitchell 
Karst Plain subsection. 

LTA 222Em04 has a mean elevation of 2 1 
m with the highest point at 275 m and the 
lowest at 164 m. It has the area of 8498 ha 
and the highest proportion of calcareous mesic 



Table 2. — Spatial metrics of Ecological Land Type Phase map units (area and c 'c) by Land Type As- 
sociation map units of the Brown County Hills subsection (the Pleasant Run unit of the Hoosier National 
Forest). 





LTA 222Em01 


LTA 222Em02 


LTA 


222Em03 


LTA 


222Em04 




Area, 




Mean 


Area, 




Mean 


Area, 




Mean 


Area. 




Mean 


ELTP 


ha 


% 


size, ha 


ha 


% 


size, ha 


ha 


% 


size, ha 


ha 


% 


si/e. ha 


10 


757 


3.6 


1.12 


332 


2.9 


1.12 


18 


0.4 


0.73 


305 


3.6 


1.15 


11 


60 


0.3 


0.91 


39 


0.3 


0.95 


12 


0.3 


1.1 1 


4 





0.S5 


12 


1,955 


9.3 


4.95 


1,481 


12.7 


6.38 


2.424 


51.2 


34.14 


77^ 


9.1 


7.09 


13 


15 


0.1 


3.81 














454 


S j 


16.8 


20 


6,869 


32.5 


6.66 


2,998 


25.8 


4.12 


378 


S 


2. OS 


2.604 


31." 


s s 5 


21 


148 


0.7 


1.07 


47 


0.4 


0.83 


58 


1.2 


0.8 


90 


1.1 


1.64 


22 


5,884 


27.9 


4.33 


3,252 


28 


4.95 


602 


12.7 


j « 


2,296 


2~ 


4.~4 


23 


1 





0.02 














298 


^ s 


[ 1.91 


40 


412 


2 


1.81 


357 


3.1 


2.15 


171 


3.6 


5 oo 


138 


1.6 


in 


41 


926 


4.4 


4.35 


580 


5 


5.8 


373 


7.9 


4.15 


415 


4» 


4.23 


42 


1,119 


5.3 


4.1 


1,546 


13.3 


35.96 


566 


12 


21.78 


386 


4.5 


3.00 


43 


1,304 


6.2 


5.64 


950 


8.2 


25 


103 


-> -> 


L2.9 


ISO 


24 


$9." 


Water 


1,665 


7.9 


3.57 


46 


0.4 


0.21 


29 


0.6 


0.24 


466 


s 5 


4.96 


Total 


21,116 


100 


4.13 


11,627 


100 


4.52 


4.735 


100 


5.04 


S.4 L )S 


100 


4.5S 



166 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Table 3. — Soil erosion statistics of Land Type Association map units of the Brown County Hills sub- 
section (the Pleasant Run unit of the Hoosier National Forest). AWEi — Area Weighted Erosion index. 
Erosion class descriptions: 1 — soils that have lost on the average less than 25% of the original A and/or 
E horizons; 2 — soils that have lost, on the average 25 to 75% of the original A and/or E horizons; 3 — 
soils that have lost, on the average 75% of the original A and/or E horizons; 5 — gullied soils. 





LTA 222 


EmOl 


LTA 222Em02 


LTA 222Em03 


LTA 222Em04 


Erosion class 


Area, ha 


% 


Area, ha 


% 


Area, ha 


% 


Area, ha 




% 


1 


18.100 


85.7 


9,243 


79.4 


1,885 


39.8 


6,648 




78.2 


2 


1 .434 


6.8 


1,790 


15.4 


2,370 


50.0 


1 ,345 




15.8 


3 


5 


0.0 


578 


5.0 


224 


4.7 


— 




— 


5 


— 


— 


— 


— 


247 


5.2 


— 




— 


Water 


1,583 


7.5 


23 


0.2 


11 


0.2 


510 




6.0 


Total 


21,122 


100 


11,634 


100 


4,738 


100 


8,503 




100 


AWEi 


1.07 




1.25 




1.81 




1. 


17 





ridge ELTP 13 and mesic slope ELTP 23 (5.3 
and 3.5%, respectively) among all LTAs of 
BCH subsection. The dominate ELTPs were 
mesic slope ELTP 22, dry slope ELTP 20, and 
mesic ridge ELTP 12 (with 31.7, 27.0, and 
9.1% of the area, respectively). The dry ridge 
ELTP 10, calcareous mesic ridge ELTP 13, 
and floodplain ELTP 43 have the largest mean 
area among all BCH subsection LTAs (1.15, 
16.80 and 89.77 ha, respectively). This LTA 
has the largest area proportion and largest 
mean size of slope ELT 2 (63.3% and 122.29 
ha). Shannon's Index of Diversity was 1.95 
for ELTPs (highest in BCH subsection) and 
1.16 for ELTs. 

The eroded area proportions of this LTA are 
similar to those of LTA 222Em02: 78.2% in 
Erosion Class 1 and 15.8% in Erosion Class 
2 (Table 3). AWEi equals 1.17 and is the sec- 
ond smallest among the BCH LTAs. Dominate 
soil survey map units within this LTA are 
Brownstown-Gilwood silt loams (25-75% 
slopes), Wrays-Gilwood silt loams (6-20% 
slope), Crider silt loam (6-12% slopes, erod- 
ed) covering 41.5, 15.4, and 11.1% of the 
area, respectively. Other soils series, such as 
Caneyville and Corydon, derived from calcar- 
eous parent material also occur. 

LTA descriptions for the Crawford Up- 
land Subsection.— LTA 222De()l White Oak 
Dry-Mesic Upland Hills: This LTA occupies 
the central part of the Lost River unit, the en- 
tire Patoka River Unit and northwestern cor- 
ner of the Tell City unit to the border between 
Crawford and Perry counties (Fig. 6). It is 
characterized by Q. alba dominated dry ridge 
and dry slope ELTPs 1 1 and 22. Quercus pri- 



nus was not found in this area during sam- 
pling. Wet-mesic slope ELTP 25 occurs com- 
monly on northeastern nose slopes adjacent to 
wide wet-mesic floodplains within this LTA. 
Calcareous mesic ELTP 26 is scattered rather 
scarcely throughout this LTA and associated 
with patches of Crider-Caneyville soil series. 

LTA 222De01 has a mean elevation of 209 
m with the highest point at 297 m and the 
lowest at 133 m (Table 5). It is the largest 
(53,257 ha) and has the highest proportion of 
mesic ridge ELTP 13 and wet-mesic bottom- 
land ELTP 41 (35.2 and 6.8%, respectively) 
among all LTAs of the CU subsection (Table 
4). In general, the dominate ELTPs were me- 
sic ridge ELTP 13, dry slope ELTP 22, and 
mesic slope ELTP 24 with 35.2, 18.6, and 
14.3% of the area, respectively). Post oak 
dominated ELTPs 10, 20, and 21 as well as 
cliff ELTP 30 are absent from this LTA. The 
mesic ridge ELTP 1 3 of this LTA has the larg- 
est mean area among all the LTAs (17.54 ha). 
This LTA has the largest mean size of ridge 
ELT 1 (16.46 ha). Shannon's Index of Diver- 
sity was 2.12 for ELTPs and 1.13 for ELTs 
(Table 5). 

This LTA has the largest percentage of soils 
in Erosion Class 2 (36.5%), and it is the only 
LTA that has 0.6% of the area occupied by 
gullied soils (Erosion Class 5). The Area 
Weighted Erosion Index is 1.61 (Table 6). 
Dominate soil survey map units within this 
LTA are Adyeville-Wellston silt loams (18- 
50% slopes), Wellston-Adyeville-Ebal silt 
loams (12-18% slopes, eroded), and Wellston 
silt loams (6-12%, slopes, eroded) covering 
16.1, 1 4.0, and 1 1 .6% of the area, respectively 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



167 



OL 



a. 
O 

a 
u 






C 

6 S 

S-H 
Kl O 

■3 5 



czi 
O 

o o 

I— 1 aj 
-O 5 



'3b 


3 





>> 









u 


w 





3 J 



Si 



L" 






<J~ 















c<3 



2.9 



^ 



— i oo oo — CO C 
— : d d ri o6 o 



r~ r o iri p» 



(N a 0\ rf, oo CO 
O — ; O On O ri GO 



in co o n n o to co xc 

■xt co O On 00 ro - ri O 
^ ri \D ^ vD — !r, CO 



-< o o 



-! I « 

CJ ! OS 



>n C^ — 
O IT) CM 



r-- on o co 
co ON -xt" 

\Q ,-h o 



O vC ri vC - C 



cm r-- r-- ^r 

m h « a 

O IT) ON CM 



1; h c 
d CM CO 



oo r- 't ri 
r^, x h — 



00^— (i—fOONONCM'^tCO 

■^oqoqvnpoqt^vqON 



O if, X r, — -f c\ 

p o^ _j oo p co o 
co 'xf co cm d -t 



O oo O 
CM 



ri co so r» r- o 



CO r~ co \Q |> MD ON CM LO in CM On C4 — in CO 00 

NO^t-CJ OCrcci^tNOONOOCCCCO^ 

(^^hv.0 rn \o in \o x — ri x - cm -t- 



On 00 CM 



On in 
r- on 



on t> r- 
r- r- oo 



rf cm 

O r-5 



O "xt 

CM CM 
in 



NO 00 "vt 

oo i^ in 

d d i< 



\C 


oo 


CM 


' 


° 


in 

CO 


so 
r) 
oc 


-1- 


CO 
CO 



r) no 

co rf 


X 


c 


^ 


ON 
00 


-f 


CO 


— 


r ] 


On <xf 


in 


d 


^6 


- 


r^ 


-t 


c 


in 


On in 


co 


sq 


00 


00 


i/-, 


CO 


r i 




co in 
CM 


d 

ri 


l> 


CO 


° 


o 


xC 


— 


X 


-xj- ^t 
in co 
"vT co 


C 
CO 
CM 


CO 

xC 


CI 
CO 
CM 


yz 

-f 


CO 

ON 

CO 


co 


-f 


r^ 


— 


"~ 














O 


NO NO 

On \D 




in 

ON 




in 

sq 






in 


CM 


*t ~ 


in 


in 




c 


CM 


X 

r i 


~ 


m] 


CM Tf 


SO 


xC 




r-; 


CO 


CM 


CM 




co d 


ri 


ri 


1 


c 


yQ 


= 


~ 


X 


r- O 

CO CM 


sO 

in 


in 

ri 


! 


in 

CO 


C^, 


m, 


ri 


X 


1—1 


— 














^f 


co — 

sq — • 


sq 


sC 


in 


£ 


co 


X 




X 


nc ri 


CO 


ON 


ri 


— 


n 


— 


^ 


in 


NO 


CO 


in 


ri 


m, 


00 


CO 


CM 




00 Tt 


-1-' 


00 


° 


~~ 


c 


r^ 




E 


[-- oc 

On — _ 


00 
CO 

sC 


co 
m, 


co 


t~- 


| 


CM 
X 
X 




CM 



ON (N l> ^xf 



CM co O — CM co -xt in, sC O O — rj 

-™CJ(N(N(N(N(N(Nr, \fxt\f 



^ ~ 



168 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Table 5. — Elevation and diversity statistics of Land Type Association map units of the Crawford Upland 
subsection (Lost River, Patoka River and Tell City units of the Hoosier National Forest). SDI — Shannon's 
Diversity Index; SEI — Shannon's Evenness Index. 



LTA 222De01 LTA 222De02 LTA 222De03 LTA 222De04 LTA 222De05 LTA 222De06 

Elevation, m 

Mode Max Mode Max Mode Max Mode Max Mode Max Mode Max 

(Mean) (Min) (Mean) (Min) (Mean) (Min) (Mean) (Min) (Mean) (Min) (Mean) (Min) 

207 297 140 248 220 287 189 272 201 266 129 268 

(209) (133) (179) (134) (199) (139) (182) (117) (191) (118) (171) (117) 

Landscape diversity 

SDI SEI SDI SEI SDI SEI SDI SEI SDI SEI SDI SEI 



ELTP 2.12 
ELT 1.13 



0.85 
0.82 



1.90 
1.10 



0.79 
0.79 



2.17 
1.12 



0.87 
0.81 



2.29 
1.25 



0.82 
0.78 



1.92 
1.13 



0.71 
0.70 



2.16 
1.19 



0.78 
0.74 



(Table 4). This LTA has also other soils series 
derived from calcareous parent material 
(1.2%) such as Crider, Caneyville, and Cory- 
don. 

LTA 222De02, Chestnut Oak Dry Upland 
Hills: This LTA is north of the junction of the 
White River East Fork and the Lost River 
within the Lost River unit. It is characterized 
by Q. prinus dominated dry ridge and dry 
slope ELTPs 1 1 and 22. Wet-mesic ELTP 25 
and calcareous mesic ELTP 26 are absent 
from this LTA. LTA 222De02 has a mean el- 
evation of 179 m with the highest point at 248 
m and the lowest at 134 m. It has the smallest 
area (4891 ha) and the highest proportion of 
bottomland ELTP 42 (11.2%) among all LTAs 
of the CU subsection. The dominating ELTPs 
were mesic ridge ELTP 13, dry slope ELTP 
22, and mesic slope ELTP 24 (31.2, 23.2, and 
21.6%, respectively). Post oak dominated 
ELTPs 10, 20, and 21 as well as cliff ELTP 
30 and calcareous ELTP 26 are absent from 
this LTA. This LTA has the largest area pro- 
portion of bottomland ELT 4 (18.8%). Shan- 
non's Index of Diversity was 1 .90 for ELTPs 
and 1.10 for ELTs. 

This LTA has the largest percentage of soils 
in Erosion Class 1 (78.5%) that indicates the 
slightest degree of soil disturbance among all 
the CU subsection LTAs. AWEi equals 1.24 
which is the smallest within the CU subsec- 
tion. Dominating soil survey map units within 
this LTA are Wellston-Tipsaw-Adyeville com- 
plex, 18-70% slopes; Wellston silt loam, 6- 
12% slopes, eroded; Apalona silt loam, 2-6% 
(53.0, 10.2, and 8.0% of the area, respective- 
ly). 



LTA 222De03 Oak-Maple Calcareous Up- 
land Hills: This LTA is located to the east 
from the Sams Creek within the Lost River 
unit and characterized by abundant occurrence 
of calcareous ELTP 26 on steep slopes along 
the Lost River and Sulfur Creek. This ELTP 
was closely associated with Caneyville-Crider 
rock outcrops. LTA 222De03 has a mean el- 
evation of 199 m with the highest point at 287 
m and the lowest at 139 m. It has the area of 
6071 ha and the highest proportion of dry- 
mesic ridge ELTP 12 (1.6%) among all LTAs 
of the CU subsection. The dominating ELTPs 
were mesic ridge ELTP 1 3 and calcareous me- 
sic slope ELTP 26 (22.2 and 3.8%, respec- 
tively). Post oak dominated ELTPs 10, 20, and 
21 as well as cliff ELTP 30 are absent from 
this LTA. The dry-mesic ridge ELTP 12 and 
dry slope ELTP 22 of this LTA have the larg- 
est mean area among all LTAs (0.87 and 9.32 
ha, respectively, Table 4). This LTA has the 
largest area proportion of slope ELT 2 
(61.2%). Shannon's Index of Diversity was 
2.17 for ELTPs and 1.12 for ELTs. 

This LTA has the second largest percentage 
of soils in Erosion Class 2 (34.7%). AWEi 
equals 1 .38 and is the second smallest among 
LTAs of the CU subsection. Dominating soil 
survey map units within this LTA are Adye- 
ville-Wellston silt loams, 18-50% slopes; 
Wellston-Adyeville-Ebal silt loams, 12-18% 
slopes, eroded; Wellston silt loam, 6—12% 
slopes, eroded (48.6, 20.1, and 9.4% of the 
area, respectively). This LTA also has the 
highest percentage of calcareous soils series 
such as Crider, Caneyville and Corydon (6.1% 
in total). 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



169 



Table 6. — Soil erosion statistics of Land Type Association map units of the Crawford Upland subsection 
(Lost River, Patoka River and Tell City units of the Hoosier National Forest). AWEi — Area Weighted 
Erosion index. Erosion class description: 1 — soils that have lost on the average less than 25% of the 
original A and/or E horizons; 2 — soils that have lost, on the average 25 to 75% of the original A and/or 
E horizons; 3 — soils that have lost, on the average 75% of the original A and/or E horizons: 5 — gullied 
soils. 



LTA 222De01 LTA 222De02 LTA 222De03 LTA 222De()4 LTA 222De05 LTA 222De06 



Erosion 


Area, 




Area, 




Area, 




Area, 


Area, 




Area. 


class 


ha 


% 


ha 


% 


ha 


% 


ha 


% ha 


% 


ha % 


1 


21 Ml 


51.0 


3,844 


78.5 


3,867 


63.7 


8,539 


51.7 13,444 


51.1 


11.947 52.2 


2 


19,463 


36.5 


896 


18.3 


2,106 


34.7 


3,286 


19.9 7,085 


26.9 


6.213 27.1 


3 


5,730 


10.8 


150 


3.1 


91 


1.5 


4,158 


25.2 5,229 


19.9 


4.338 19. (J 


5 


311 


0.6 


















Water 


595 


1.1 


6 


0.1 


10 


0.2 


528 


3.2 559 


2.1 


392 1.7 


Total 


53,277 


100 


4,895 


100 


6,074 


100 


16,510 


100 26,318 


100 


22,889 100 


AWEi 


1.6] 




1.24 




1.38 


1.73 1.68 




1.66 



LTA 222De04 Oak-Maple Wet-Mesic Dis- 
sected Plateau: Located in the northeastern 
corner of Tell City to the north from Mill 
Creek, this area is characterized by rugged ter- 
rain and steep slopes along the meandering 
Little Blue River and its forks. Cliff ELT and 
respective ELTP are typical for this LTA. 
Only one dry clayey ELTP 20 dominated by 
post oak was found in this LTA and dry clay- 
ey ELTPs 21 are very few and occur in south- 
ern tip next to the LTA 222De06 (Post Oak 
Dry Upland Hills) and the Ohio River. 

LTA 222De04 has a mean elevation of 182 
m with the highest point at 272 m and the 
lowest at 117 m (Table 5). It has the area of 
16,498 ha and the highest proportion of wet- 
mesic slope ELTP 25, dry-mesic slope ELTP 
23, and cliff ELTP 30 (11.3, 9.2, and 0.7%, 
respectively) among all LTAs of the CU sub- 
section. The dominating ELTPs were mesic 
ridge ELTP 13, dry slope ELTP 22 and mesic 
slope ELTP 24 (28.0, 16.0, and 16.0%, re- 
spectively). The post oak dominated dry ridge 
ELTP 10, dry slope ELTP 21, and dry-mesic 
slope ELTP 23 of this LTA have the largest 
mean area among all CU subsection LTAs 
(1.48, 16.89, and 5.64 ha, respectively. Table 
4). This LTA has the largest proportion of cliff 
ELT 3 (0.7%) and the largest mean size of 
slope ELT 2 (50.62 ha). Shannon's Index of 
Diversity was 2.29 for ELTPs and 1.25 for 
ELTs (both indices are highest for the CU sub- 
section). 

This LTA has the largest percentage of soils 
in Erosion Class 3 (25.2%) and AWEi is 1.73. 



which is the highest erosion index within 
LTAs of the CU subsection. Dominating soil 
survey map units within this LTA are Tipsaw- 
Adyeville complex, 25-75% slopes: Wellston- 
Adyeville-Ebal silt loams, 12-18% slopes. 
eroded; Apalona silt loam. 6-12% slopes, 
eroded and severely eroded; Wellston silt 
loam, 12-18% slopes, severely eroded (23.2. 
12.1, and 11.8% of the area, respectively). 
Corydon stony silt loam, with 20-60% slopes. 
is also present on this LTA (1.2% of LTA 
area). 

LTA 222De05 Mixed Oak Dry Upland 
Hills: This is a landscape type that represents 
a transition from northern part of the Craw- 
ford Upland subsection to its southern part. It 
is located in the east-central part of the Tell 
City unit between Crawford county and Perry 
county line in the north and Middle Deer 
Creek in the south. It has a wide variety of 
ELTPs including Cliff ELTP 30 along Jubm 
and Oil Creeks. Calcareous ELTP 26 and dr\ 
clayey ELTPs 10. 20 and 21 arc uncommon 
within this LTA. 

LTA 222De05 has a mean elevation of 191 
m with the highest point at 200 m and the 
lowest at 118 m. It has the area of 20.312 ha 
and the highest proportion of dry slope ELTP 
22. mesic slope ELTP 24. mesic bottomland 
ELTP 40. and dry ridge ELTP 1 1 (26.7, 20.3. 
2.7. and 2.4%. respectively) among all LTAs 
of the CV subsection. The dominating ELTPs 
were mesic ridge ELTP 13. dr\ slope ELTP 
22. and mesic slope ELTP 24 (20. S. 26.7, and 
26.3%, respectively). Post oak dominated 



170 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



ELTP 20 is absent from this LTA. The wet- 
mesic slope ELTP 25, mesic bottomland ELTP 
40, and floodplain ELTP 42 of this LTA have 
the largest mean area among all the CU sub- 
section LTAs (11.77, 2.32, and 33.05 ha, re- 
spectively. Table 4). Shannon's Index of Di- 
versity was 1 .92 for ELTPs and 1.13 for ELTs. 

AWEi is 1.68 that indicates relatively high 
proportion of eroded area in comparison with 
other LTAs. Dominating soil survey map units 
within this LTA are Adyeville-Tipsaw-Ebal 
complex, 20-50% slopes, very rocky; Ebal- 
Deuchars-Kitterman complex, 12-24% slopes, 
eroded; Apalona silt loam, 6—12% slopes, 
eroded and severely eroded (35.9, 13.2 and 
12.4% of the area, respectively). 

LTA 222De06 Post Oak Dry Upland Hills: 
It is located within approximately 8 km of the 
Ohio River and characterized by frequent oc- 
currence of post oak dominated dry clayey 
ELTPs 10, 20, and 21 as well as a few cliff 
ELTPs 30 along the Ohio River banks. LTA 
222De06 has a mean elevation of 171 m with 
the highest point at 268 m and the lowest at 
117 m. It has the area of 22,878 ha and the 
highest proportion of post oak dominated dry 
slope ELTP 21, ELTP 20, and dry ridge ELTP 
10 (2.8, 0.8, and 0.2%, respectively) among 
all LTAs of the CU subsection. The dominat- 
ing ELTPs were mesic ridge ELTP 13, mesic 
slope ELTP 24, and dry slope ELTP 22 (29.3, 
23.5, and 18.0%, respectively). The post oak 
dominated dry slope ELTP 20 and cliff ELTP 
30 of this LTA have the largest mean area 
among all the CU subsection LTAs (18.17 and 
15.80 ha, respectively, Table 4). This LTA has 
the largest mean size of cliff ELT 2 and bot- 
tomland ELT 4 (15.79 and 21.15 ha, respec- 
tively). Shannon's Index of Diversity was 2.16 
for ELTPs and 1.19 for ELTs. 

AWEi is 1.66 that is similar to LTA 
222De05. Dominating soil survey map units 
within this LTA are Adyeville-Tipsaw-Ebal 
complex, 20-50% slopes, very rocky; Ebal- 
Deuchars-Kitterman complex, 12-24% slopes, 
eroded; Ebal-Deuchars-Kitterman complex, 
12-24% slopes, severely eroded (30.5, 17.0, 
and 9.9% of the area, respectively). 

DISCUSSION 

The delineated LTA map is a first approx- 
imation and based on spatial variability 
among ELTP map units, soils, and geology 
within the study area. However, there are sev- 



eral limitations that should be considered 
when using this map. Delineated LTA map 
units were analyzed only within the purchase 
boundary of the HNF. Therefore, boundaries 
can be reconsidered in the future in relation 
to surrounding areas and the entire subsection, 
as more information is obtained. In addition, 
all spatial information represents each LTA 
only within the purchase boundary of the 
HNF and potentially can be different for the 
entire LTA area beyond those boundaries. 
Since only a small portion of the Escarpment 
subsection is occupied by the Lost River and 
Patoka River units of the HNF, this area is 
considered as part of the Crawford Upland 
subsection. The more precise delineation of 
the boundary between these subsections as 
well as LTA delineation within that part can 
be done when sufficient information on ELT 
and ELTP classification for the Escarpment 
subsection is collected. This study did not re- 
veal any significant differences in ELTP, soil 
or geology pattern that would justify separa- 
tion of that area. 

Differentiating criteria used in LTA devel- 
opment for the HNF in Indiana include, in 
general order of most frequent use, patterning 
of ELTPs and soil survey units, landforms, 
bedrock type, dominant tree species occur- 
rence, and disturbance processes. These cri- 
teria are specific for the study area and may 
be different from criteria used in other re- 
gions. For example, the differentiating criteria 
used in LTA development in the Lake States 
included, in order of most frequent use: sur- 
facial geology, composition or productivity of 
historic vegetation, hydrology, meso-climate, 
patterning of ELTs and ELTPs, bedrock type, 
hydrography, and disturbance processes (Jor- 
dan et al. 2001). Some criteria were the same 
but had a different importance, while others, 
such as glacial features or local climatic influ- 
ence ("lake effect"), were not important. In 
general, most LTA projects are taking the 
same approach with geomorphology being an 
overriding differentiating criterion that is re- 
fined using soil and vegetation information. 

As to boundary identification, the approach 
used in LTA delineation for the HNF was sim- 
ilar to approaches used in other regions and is 
based on physiographic boundaries (Mc- 
Farlane et al. 2002; Zastrow et al. 2002). All 
units are nested within the boundaries of sub- 
sequent upper and lower hierarchical units 



ZHALNIN & PARKER— LTA DELINEATION AND SPATIAL ANALYSIS 



171 



(LTA <-» ELT <-» ELTP). Normally the feature 
with the sharpest transitions between adjacent 
LTAs is used to draw boundaries. 

Most other regions have used a "top to bot- 
tom" approach in delineating LTAs (Mc- 
Farlane et al. 2002; Nigh & Shroeder 2002), 
due to the scale of delineation criteria (e.g., 
glacial geology) and also to the fact that de- 
lineated areas were already defined as units in 
other classification systems in use by various 
agencies (e.g., soil associations of USDA Nat- 
ural Resources Conservation Service). The 
approach used in this study was primarily 
based on spatial distribution of smaller eco- 
logical units with an addition of information 
from other resource maps of coarser scale. 

The current LTA maps should be consid- 
ered as an initial version that will be period- 
ically updated. Changes in understanding of 
relationships between biotic and abiotic com- 
ponents of LTAs as well as feedback from us- 
ers and availability of new GIS data will lead 
to future revisions of LTA definitions and 
boundaries. 

ACKNOWLEDGMENTS 

We would like to acknowledge Ken Day, 
supervisor of the Hoosier National Forest, 
Patrick Merchant and Dave Weigel for their 
assistance. This project was funded by USDA 
Forest Service. 

LITERATURE CITED 

Bailey, R.G. 1980. Description of the ecoregions 
of the United States. Miscellaneous Publication 
1391. U.S. Department of Agriculture. Washing- 
ton, DC. 77 pp. 

Cleland, D.T., RE. Avers, W.H. McNab, M.E. Jen- 
sen, R.G. Bailey, T. King & W.E. Russell. 1997. 
National Hierarchical Framework of Ecological 
Units. Pp. 181-200, In Ecosystem Management: 
Applications for Sustainable Forest and Wildlife 
Resources (M.S. Boyce & A. Haney, eds.). Yale 
University Press. New Haven & London. 

DeMeo, T. 2002. Use of Landtype Associations as 
a coarse filter for ranking quality of Indiana bat 
habitat on the Monongahela National Forest. 
West Virginia. Pp. 71-80, In Landtype Associa- 
tions Conference: Development and Use in Nat- 
ural Resources Management, Planning and Re- 
search (M.L. Smith, ed.). General Technical 
Report NE-294. U.S. Department of Agriculture. 
Forest Service, Northeastern Research Station. 
Newton Square, Pennsylvania. 

ECOMAP 1993. National hierarchical framework 



of ecological units. U.S. Department of Agricul- 
ture, Forest Service. Washington, DC. 20 pp. 

ESRI. 2002. ArcView GIS. Ver. 3.3. Environmen- 
tal Systems Research Institute. Redlands. Cali- 
fornia. 

Gutshick, R.G. 1966. Bedrock geology. Pp. 1-20. 
In Natural Features of Indiana (A. A. Lindsey, 
ed.). Indiana Academy of Science. Indianapolis. 

Homoya, M.A., D.B. Abrell, J.R. Aldnch & T.W. 
Post. 1984. The natural regions of Indiana. Pro- 
ceedings of the Indiana Academy of Science 94: 
245-268. 

Keys, J., Jr., C. Carpenter, S. Hooks. F. Koenig. 
W.H. McNab, W. Russell & M.L. Smith. 1995. 
Ecological units of the eastern United States — 
First approximation (map and booklet of map 
unit tables). U.S. Department of Agriculture. 
Forest Service. Atlanta, Georgia. 

McCune, B. & M.J. Mefford. 1999. PC-ORD. 
Multivariate analysis of ecological data. Ver. 
4.20. MJM Software. Gleneden Beach. Oregon. 

McFarlane, D.W, C. Coutrous & J.R Dunn. 2002. 
Land Type Associations using digital soil maps 
and landscape topography. Pp. 1 1-23. In Land- 
type Associations Conference: Development and 
Use in Natural Resources Management. Planning 
and Research (M.L. Smith, ed.). General Tech- 
nical Report NE-294. U.S. Department of Agri- 
culture, Forest Service, Northeastern Research 
Station. Newton Square, Pennsylvania. 

McNab, W.H. & RE. Avers. 1994. Ecological Sub- 
regions of the United States. WO-WSA-5. U.S. 
Department of Agriculture. Forest Service. 
Washington, DC. 267 pp. 

Nigh, T.A. & W Shroeder. 2002. Developing and 
application of Landtype Associations in Missou- 
ri. Pp. 103, In Landtype Associations Confer- 
ence: Development and Use in Natural Resourc- 
es Management, Planning and Research (M.L. 
Smith, ed.). General Technical Report NE-294. 
U.S. Department of Agriculture. Forest Ser\ ice, 
Northeastern Research Station. Newton Square. 
Pennsylvania. 

Parker, D. & L. Whitcomb. 2002. Human prefer- 
ence for ecological units: Patterns of dispersed 
campsites within Landtype Associations on the 
Chippewa National Forest. Pp. 62-67, In Land- 
type Associations Conference: Developmenl and 
Use in Natural Resources Management. Planning 
and Research. (M.L. Smith, ed.). General Tech- 
nical Report NE-294. U.S. Department of Agri- 
culture. Forest Service, Northeastern Research 
Station. Newton Square. Pennsylvania. 

Rempel. R.S. & A. P. Carr. 2003. Patch Analyst ex- 
tension for ArcView: version 3.3. Via the Inter- 
net. http://tlash.lakclicadu.ca - rrempel patch- 
index. html. 

Shao, G.. G.R. Parker. A.Y. Zhalnin, P. Merchant & 
D. Albright. 2004. GIS Protocols in mapping 



172 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



ecological landtypes for the Hoosier National 
Forest. Northern Journal of Applied Forestry 21: 
180-186. 

Soil Survey Division Staff. 1993. Soil survey man- 
ual. Soil Conservation Service. U.S. Department 
of Agriculture. Washington, DC. 457 pp. 

Van Kley, J.E. 1993. An ecological classification 
system for the central hardwood region: The 
Hoosier National Forest. Ph.D. thesis. Purdue 
University. West Lafayette, Indiana. 436 pp. 

Zastrow, D.E., D.J. Hvizdak & M.C. Moline. 2002. 
Development of Wisconsin's Landtype Associa- 
tions: A layer within the National Hierarchical 
Framework of Ecological Units. Soil Survey Ho- 
rizons 43:58-64. 



Zhalnin, A.V. 2004. Delineation and spatial anal- 
ysis of ecological classification units for the 
Hoosier National Forest. Ph.D. thesis. Purdue 
University, West Lafayette, Indiana. 268 pp. 

Zhalnin, A.V., G.R. Parker, G. Shao & P. Merchant. 
2002. LTA delineation for the Hoosier National 
Forest: Criteria and methods. Pp. 30-38, In 
Landtype Associations Conference: Develop- 
ment and Use in Natural Resources Management, 
Planning and Research (M.L. Smith, ed.). Gen- 
eral Technical Report NE-294. U.S. Department 
of Agriculture, Forest Service, Northeastern Re- 
search Station. Newton Square, Pennsylvania. 

Manuscript received 5 December 2006, revised 11 
April 2007. 



2007. Proceedings of the Indiana Academy of Science (1 16): 173-1 83 



ZOOPLANKTON GROWTH RESPONSES TO THE 

CYANOBACTERIA MICROCYSTIS AND ANABAENA IN 

EAGLE CREEK RESERVOIR IN INDIANA 



Annette Trier weiler 1 : 

Indiana 46240 USA 



Park Tudor School, 7200 N. College Ave., Indianapolis. 



Denise Lani Pascual: Center for Earth and Environmental Science, Department of 
Earth Sciences, Indiana University-Purdue University at Indianapolis; Indianapolis. 
Indiana 46202 USA 

ABSTRACT. Eagle Creek Reservoir is a small, eutrophic reservoir located on the northwest side of 
Indianapolis. Recent assessments by the Indiana Department of Environmental Management (IDEM) have 
shown that ECR is impaired due to persistent nuisance algal blooms (IDEM 303(d) list; 2002. 2004. and 
2006). Research presented here explored the relationship between zooplankton feeding behavior and c\ - 
anobacteria morphology, specifically, how resident ECR zooplankton growth was affected by filamentous 
versus coccid algal morphologies. In August of 2003, in situ mesocosms were deployed in the reservior 
to observe zooplankton growth given two different algal food sources: Anabaena sp., a filamentous het- 
erocyst-forming cyanobacteria, and Microcystis sp., a coccid non-heterocyst forming cyanobacteria. No 
statistical difference between the overall zooplankton growth in the enriched mesocosms was observed. 
However, a taxa-treatment effect was seen as rotifer populations grew significantly faster in the Micro- 
cystis-enriched mesocosm and the copepod populations were significantly greater in the Anabaena-enhched 
mesocosm. These taxa-specific trends show that different zooplankton taxa prefer and/or tolerate different 
phytoplankton morphologies. 

Keywords: Zooplankton, Cyanobacteria, grazing preference, physical inhibition, Eagle Creek Reservoir 



Cultural eutrophication of drinking water 
reservoirs continues to be a major threat to 
drinking water supplies. Watershed nutrient 
loading spurs greater productivity in these 
systems, changing the environment to favor 
the exponential growth of nuisance algae that 
can be a problem for both recreation and mu- 
nicipal uses. Scums of nuisance algae will de- 
ter swimmers, and high oxygen demand in the 
reservoir's bottom waters can stress fish pop- 
ulations and lead to fish kills of sport fish. In 
terms of municipal uses, nuisance algae can 
interfere with water treatment: some filamen- 
tous algae can clog filters, while other algae 
can produce secondary metabolites such as 
the taste and odor-causing compounds 
2-methylisoborneol (MIB) and geosmin ((E)- 
l,10-dimethyl-9-decalol) or toxins (e.g., ana- 



1 Current address: Department of Biology. Fur- 
man University, 3300 Poinsett Hwy.. Greenville 
South Carolina 29613 USA 



toxin-a, microcystin-LR, and cylindrosper- 
mopsin). As Eagle Creek Reservior (ECR) is 
a drinking water resource for ~ 80.000 Indi- 
anapolis residents, these nuisance algal 
blooms are a problem since several genera of 
cyanobacteria in the reservoir such as Pseu- 
danabaena, Anabaena, and Aphanizomenon 
can produce secondary metabolic compounds 
that cause taste and odor problems and/or tox- 
icity. Due to the societal importance of drink- 
ing water, understanding the prerequisite con- 
ditions necessary for exponential growth and 
the food-web checks and balances that could 
naturally control nuisance algal growth arc 
important for protecting and maintaining the 
health of these ecosystems. 

In addition to the nutrient loading and 
bloom formation research at Eagle Creek Res- 
ervoir and other lakes world-wide, ecosystem 
bloom control needs be studied as these biotic 
interactions between phytoplankton and zoo- 
plankton during blooms can give insight into 



17. 



174 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



long term solutions: understanding how pri- 
mary consumers and higher tropic levels cope 
with and/or suppress cyanobacterial blooms 
may not only aid in our understanding of the 
reciprocal influence that cyanobacteria-domi- 
nated algal communities and zooplankton 
communities exert on each other, but may also 
provide a top-down, natural control for algal 
blooms. 

Studies by Lampert et al. (1986) and Som- 
mer et al. (1986) have shown that while zoo- 
plankton herbivory can clear springtime 
blooms, causing a "Clear Water Phase," the 
use of zooplankton to control late summer 
blooms of cyanobacteria is not as promising. 
Such a clearing effect is not seen to the same 
degree in late summer when the cyanobacter- 
ial blooms occur. From this inability of the 
zooplankton community to clear a cyanobac- 
terial bloom, many studies consider cyanobac- 
teria to be a poor food quality for reasons of 
toxicity, nutrient inadequacies, and physical 
inhibitions (Haney 1987; Lampert 1987). 

While toxicity to zooplankton has been a 
major research focus in these phytoplankton- 
zooplankton interactions, DeMott & Moxter 
(1991) concluded that cyanobacteria toxicity 
is perhaps overemphasized and zooplankton 
adaptations to coexist with toxic cyanobacte- 
ria are underestimated. In Gilbert (1990), this 
is seen as the strain of Daphnia pulex that had 
coexisted with toxic Anabaena ajfinis had 
evolved resistance to toxin (Gilbert 1990). In 
another study, Bosmina longirostris was found 
to be resistant to strains of Microcystis aeru- 
ginosa (Fulton 1988). 

In the field, toxicity may not be the most 
significant factor in shaping zooplankton com- 
munity structure. Various studies have found 
filamentous morphology inhibits feeding and 
growth rates for daphnids (Arnold 1971; 
Holm et al. 1983; Gliwicz & Lampert 1990). 
In one study, Daphnia responded to the dom- 
inance of a filamentous, non-toxic strand of 
Aphanizomenon with reduced growth rates 
(Kurmayer 2001). 

This mesocosm study attempts to under- 
stand the phytoplankton-zooplankton interac- 
tions in blooms that are not necessarily toxic 
to natural zooplankton populations by exam- 
ining the impact of physical inhibition on zoo- 
plankton communities by using two morpho- 
logically different cyanobacteria present in the 
reservior, e.g., the filamentous Anabaena sp. 



and the coccid Microcystis sp. We hypothesize 
that zooplankton populations will be more ad- 
versely affected in the mesocosm enriched 
with the filamentous Anabaena sp. 

METHODS 

Study site. — ECR is a small (area of 5.0 
km 2 ), shallow reservoir (mean depth of 4.2 m) 
with an estimated reservoir volume is 
20,954,000 m 3 (Tedesco et al. 2003). Located 
northwest of Indianapolis, Indiana (39.83°N, 
86.3 1°W; 39.87°N, 86.30°W) the reservoir was 
created by impounding Big Eagle Creek. 
Originally constructed as a method for flood 
control, the reservoir became a direct drinking 
water source in 1976 when the T.W Moses 
Drinking Water Plant came on-line. The res- 
ervoir is fed by three streams in Eagle Creek 
Watershed (420 km 2 , HUC 05120201120): 
Eagle Creek, Fishback Creek, and School 
Branch with the largest flow contribution 
coming from the trunk stream, Eagle Creek. 
Eagle Creek is a small stream with a median 
daily instantaneous flow of 0.9 m 3 /s (USGS 
Stream Gage #03353200, 1957-2003). Water 
balance estimates of Eagle Creek Reservoir 
resulted in a calculated residence time of 43 
days (Pascual et al. 2006). An average water- 
shed slope of <5% and the presence of pro- 
ductive soils allowed for crop production in 
the watershed, resulting in the majority of Ea- 
gle Creek Watershed land to be cultivated as 
agricultural land, 60.1% in 2003 (Tedesco et 
al. 2005). Central Indiana Water Resources 
Partnership (CIWRP) bi-weekly monitoring 
during the growing season from 2003-2005 
showed a reservoir mean total phosphorus 
concentration of 94 |xg P/L (Pascual et al. 
2006). The 2003 assessments using the Indi- 
ana Trophic State Index categorize ECR as eu- 
trophic. 

In August 2003, the ECR phytoplankton 
community was composed of chrysophytes, 
chlorophytes, and cyanobacteria. The diatoms 
present were Asterionella, Aulacoseira, while 
Cyclotella. Ankistrodesmus, Actinastrum, 
Closterium, Coelastrum, and Pediastrum rep- 
resented the chlorophytes. Merismopedia, 
Microcystis, Anabaena, Aphanizomenon, and 
Cylindrospermopsis represented the cyano- 
bacteria. The cyanobacteria used to inoculate 
the experimental mesocosms (Anabaena sp. 
and Microcystis sp.) were naturally found in 
the reservoir at the time of the study. 



TRIERWEILER & PASCUAL— ZOOPLANKTON AND CYANOBACTERIA 



175 



( CMi J 


f CM 2 j 


f MM t J 


( MM 2 j 


( AMi j 


( AM 2 ) 



Figure 1. — Mesocosm experimental design. 



Experimental design. — Mesocosms were 
deployed off a secluded dock in Eagle Creek 
Reservoir from 7 August to 12 August 2003 
(5 days). Mesocosms were built using 45-50 
gallon husky high density clear polyethylene 
bags (Poly-America) suspended from a PVC 
frame. Each bag provided a depth of 0.8 m 
and contained approximately 165 L of water. 
Two control mesocosms (CM) were filled with 
approximately 165 L of reservoir water fil- 
tered through an 80 |Jim mesh net to remove 
zooplankton and most of the larger filamen- 
tous algae. Four treatment mesocosms were 
filled with approximately 160 L of reservoir 
water filtered through a 20 |jim mesh net to 
remove zooplankton and most of the resident 
algae. Two of the mesocosms were then in- 
oculated with 5 L of cultured Microcystis sp., 
and designated as the Microcystis Mesocosm 
(MM). The other two exposure mesocosms 
were inoculated with 5 L of cultured Anabae- 
na sp. and designated as the Anabaena Me- 
socosm (AM) (Fig. 1). Original cultures were 
obtained from Carolina Biological Supply in 
January 2003 and cultured using modified 
Guillards F/2 media. Cultures were verified as 
named genera prior to mesocosm inoculation. 

To discern if zooplankton feeding behavior 
was influenced by toxicity, Microcystis culture 
samples were sent to State University of New 
York College of Environmental Science and 
Forestry in March 2004 for toxin analysis. 



Toxin concentrations were determined to be 
103 fxg microcystin-LR per gram chlorophyll 
a. Compared to microsystin-LR levels found 
in natural systems, our Microcystis culture had 
very low toxicity (G.L. Boyer pers. commj. 
During non-bloom conditions, other studies 
have reported natural microsystin levels 100- 
fold greater than our culture (Wang et al. 
2002). Anabaena cultures were not tested for 
toxicity. 

A total of six mesocosms was deployed: 
two control, two MM, and two AM (Fig. 1 ). 
Resident ECR zooplankton were harvested us- 
ing an 80 (xm mesh plankton net, as done by 
Sterner (1989), drawn over a 2 m vertical 
depth and concentrated into a 125 ml collect- 
ing bottle at the end of the plankton net. After 
three vertical tows, the zooplankton were 
transferred into an amber 1 L glass bottle for 
transport to the mesocosm. Harvested zoo- 
plankton were introduced to their respective 
treatment mesocosm exposure within 10 min- 
utes to minimalize stress. All six mesocosms 
were each inoculated with zooplankton from 
a total of 15 tows. The tows were spaced to 
prevent re-towing the same water column. Af- 
ter all mesocosms were inoculated, an addi- 
tional tow was made in ECR to calculate ini- 
tial zooplankton densities. 

Sampling. — Zooplankton and phytoplank- 
ton communities interacted in the six meso- 
cosms between 7-12 August 2003. Each 
morning, dissolved oxygen (DO), tempera- 
ture, and pH were measured with a Hydrolab® 
multiparameter field probe (Hydrolab Corpo- 
ration, Austin, Texas) inside and outside the 
mesocosms to ensure that environmental con- 
ditions remained within ranges for optimal 
phytoplankton and zooplankton growth as rec- 
ommended by US EPA: DO of > 6 mg/L. pH 
between 6-9, and temperature ranging be- 
tween 20-25 °C (Lewis et al. L994). Secchi 
disk readings were also taken to ensure that 
the mesocosm depths were consistently in the 
photic zone. After abiotic measurements were 
taken, the mesocosm bags were stirred to 
evenly distribute the zooplankton and algal 
populations. The mixed mesocosms were 
sampled for zooplankton and phytoplankton 
abundance. 

At 0, 42. 65, and S L ) hours, a 1 L sample 
was taken from the middle of each mixed me- 
socosm. This 1 L sample was concentrated 
into 125 ml using a 20 |xm mesh plankton net 



176 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



with the filtrate being returned to the meso- 
cosm (Ferrao-Filho et al. 2002). The 125 ml 
sample was split into two 50 ml HDPE cen- 
trifuge tubes and preserved with Lugol's So- 
lution. 

Identification and enumeration. — Pre- 
served samples were refrigerated at <4 °C for 
no longer than six months prior to analysis. 
Each duplicated sample was counted first for 
zooplankton and then phytoplankton. Zoo- 
plankton were counted using a 1 ml Sedg- 
wrick-Rafter cell and identified based on these 
zooplankton groups: Calanoida, Cyclopodia, 
Bosminidae, Daphniidae, Leptodoridae, Acar- 
iformes, Certopogonidae, Chironomidae, Os- 
tracoda, and Keratella (Rotifera). Zooplankton 
taxonomic categories are similar to those used 
by Newhouse & Stahl (2000). The entire 
Sedgwrick-Rafter cell was counted at 100X. 

After zooplankton counts, preserved water 
samples were allowed to settle by gravity for 
24 h and were concentrated. Phytoplankton 
were counted using a Palmer-Malone Nano- 
plankton Counting Chamber at 400 X and 
identified to genus. In each sample, at least 
100 natural units were counted. Later mea- 
surements and counts were made to convert 
Anabaena filaments and Microcystis colonies 
to densities on a per cell basis. 

Statistical methods. — Before the experi- 
ment, the alpha level of 0.05 was chosen for 
the statistical analyses. A three-way ANOVA 
test for Treatment X Replicate Mesocosms X 
Taxa was run separately for both phytoplank- 
ton and zooplankton. Additional two-way 
ANOVA tests were run to analyze the Treat- 
ment X Day interaction for individual zoo- 
plankton taxa. Whenever appropriate, the 
Holm-Sidak (HS) post hoc tests were em- 
ployed to further analyze statistical differenc- 
es. All statistical tests were preformed on 
SYSTAT Software. 

RESULTS 

Physical parameters. — All mesocosms 
stayed within the optimum growing ranges as 
determined by EPA. The temperatures of the 
mesocosms were similar to the temperature of 
ECR, ranging from 24.9-27.6 °C. The pH and 
DO within the mesocosms ranged from 8.2— 
8.7 and 7.0-9.1 mg/L, respectively. Secchi 
disk measurements taken outside the meso- 
cosms indicated that the exposures were al- 
ways within the photic zone (as estimated by 



1.7X Secchi disk reading). While nutrient re- 
straints within the mesocosms were not 
known during the experiment, limiting nutri- 
ents such as NO~ 3 , NH + 4 and TP were abun- 
dant in a surface reservoir sample located in 
the same basin as the mesocosms: [NO~ 3 ] = 
1.4 mg N/L, [NH + 4 ] = 0.30 mg N/L and [TP] 
= 0.063 mg P/L. 

Biotic results. — Biotic data were analyzed 
for changes in both the phytoplankton and 
zooplankton community densities and in tax- 
on specific populations. In analyzing the data, 
samples from replicate mesocosms were 
pooled as the replicate mesocosms were not 
determined to be significantly different for ei- 
ther the phytoplankton (F = 0.273, P = 
0.601) or the zooplankton (F = 3.615, P = 
0.058) (Tables 1 and 2, respectively). 

Phytoplankton trends: The initial total phy- 
toplankton concentrations for the CM, AM, 
and MM were 16,100, 16,200, and 31,800 
cells/ml respectively (Fig. 2). The overall phy- 
toplankton trends for the CM and AM are 
nearly identical with means always in the oth- 
er's error bars for h to 65 h. While the total 
phytoplankton in the AM increased to 23,000 
cells/mL at 89 h, the CM phytoplankton de- 
creased from 22,500 cells/ml at 65 h to 20,000 
cells/ml by 89 h. Initial total population in the 
MM, 31,800 cells/ml, are nearly double that 
found in the CM (16,100 cells/ml) and de- 
creased over time to 17,100 cells/ml (Fig. 2). 
While the three-way ANOVA test (Treatment 
X Replicate mesocosms X Taxon) showed no 
significant difference between treatments (F = 
1.875, P = 0.154; Table 1), both the taxonom- 
ic interaction (Taxon) and Taxon X Treatment 
interactions were highly significant (F = 
142.76, P < 0.001; Table 1). 

Of the many phytoplankton taxa identified, 
this paper presents the results only for chlo- 
rophytes in general, Anabaena sp., and Micro- 
cystis sp. because of relevance to the experi- 
mental design. While Merismopedia minima 
was often the major cyanophyte on a per cell 
basis in the mesocosms, the biovolume of the 
eight celled colonies was substantially smaller 
than a cell of any other phytoplankton iden- 
tified (<5 |jim 3 ). 

The results for the Chlorophyta showed 
population changes for the majority of non- 
cyanobacteria during the experiment. In the 
CM, Chlorophyta populations steadily de- 
creased from initial populations of 900 cells/ 



TRIERWEILER & PASCUAL— ZOOPLANKTON AND CYANOBACTERIA 



177 



6000 

4500 

3000 

1500 

40000 

30000 

20000 

-i 10000 

E 

J§ 6000 

4500 

3000 

1500 

40000 

30000 

20000 

10000 





o 



Chlorophyta 



Control 



M 



Microcystis 



Hi 



Anabaeria 



W7?Z\ Anabaena 
Mm Microcystis 





nil 



ni 



ITM 



Total Phytoplankton 



qli jl 



m 



20 



40 



60 



80 



— i — 
100 



120 



Hour 

Figure 2. — Phytoplankton in different treatments. Solid bars represent the control mesocosm: diagonal l\ - 
hashed bars represent the Anabaena mesocosms; and crosshatched bars represent the Microcystis meso- 
cosms. Each value is a mean (±1 SE) of eight counts. 



Table 1. — Results of three-way ANOVA testing the effects of treatment (Control. Anabaena, and Mi- 
crocystis mesocosms), Replicate mesocosm, and time on Phytoplankton Densities. SS = sum of squares 
df = degrees of freedom; MS = mean square; F = F- value; and P = P-value. 



Source 



SS 



df 



MS 



P 



Treatment 

Replicate 

Taxon 

Treatment 

Treatment 



Replicate 
Taxon 



Replicate X Taxon 

Treatment X Replicate X Taxon 

Error 



6.20 X 10 7 


i 


3.10 \ 10 


4.52 X L0 6 


1 


4.52 \ 10 


2.13 X L0 10 


9 


2.30 \ 10 


1.95 X L0 6 


2 


9.70 \ 10 


1.22 X L0 9 


18 


0.70 \ 10 


2.22 X 10 7 


9 


2.40 \ 10 


1.25 X 10 7 


18 


6.93 \ 10 


2.25 X LO 10 


1359 


1.65 \ 10 



l.S~5 
0.273 
42.761 
0.059 
4.087 
04 40 
0.042 



601 

000 
^43 
000 
99 s 

000 



178 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Table 2. — Results of three-way ANOVA testing the effects of treatment (Control, Anabaena,, and Mi- 
crocystis mesocosms), Replicate mesocosm, and time on Zooplankton Densities. Abbreviations are the 
same as those defined in Table 1. 



Source 


SS 


df 


MS 


F 


P 


Treatment 


1.81 X 10 9 


2 


9.07 x 10 8 


3.742 


0.024 


Replicate 


8.76 X 10 8 


1 


8.76 X 10 8 


3.615 


0.058 


Taxon 


2.32 X 10 9 


3 


7.75 X 10 8 


3.196 


0.023 


Treatment X Replicate 


1.76 x 10 9 


2 


8.78 X 10 8 


3.623 


0.027 


Treatment X Taxon 


5.30 x 10 9 


6 


8.84 X 10 8 


3.648 


0.001 


Replicate X Taxon 


2.66 X 10 9 


3 


8.85 x 10 8 


3.654 


0.012 


Treatment X Replicate X Taxon 


5.36 X 10 9 


6 


8.93 x 10 8 


3.686 


0.001 


Error 


1.32 X 10" 


544 


2.42 X 10 8 







ml to the final population of 430 cells/ml (Fig. 
2). In the AM, the chlorophyte populations in- 
creased from 750 cells/ml to a peak of 1300 
cells/ml at 42 hrs (Fig. 2). Unlike in the con- 
trols, Chlorophyta populations in the MM 
steadily increased for 65 h, then slightly de- 
creased to the final population of 1730 cells/ 
ml (Fig. 2). 

Microcystis sp. and Anabaena sp. were 
prominent in the mesocosms to which they 
were inoculated. In the MM, the initial Mi- 
crocystis sp. population was 24,400 cells/ml, 
more than five times the CM concentration of 
3200 cells/ml. However, while Microcystis sp. 
dominated the phytoplankton at the start of the 
experiment, concentrations rapidly decreased 
to 3300 cells/ml by 89 h. In the CM and AM, 
Microcystis sp. remained fairly constant. 

While the Anabaena sp. population in the 
AM was not the dominant alga by cell count 
(second to Merismopedia sp.), the initial An- 
abaena density was more than five times that 
found in the CM (630 cells/ml). Starting at 
3900 cells/ml, the AM Anabaena density de- 
creased to 1400 cells/ml by 89 h. This is in 
contrast to the CM, where Anabaena sp. abun- 
dance steadily increased to 870 cells/ml at 65 
h before decreasing to 540 cells/ml by 89 h. 

Zooplankton trends: Overall zooplankton 
trends, like the overall phytoplankton trends, 
showed a broad picture that is further supple- 
mented by the taxa specific trends. At the start 
of the experiment, the total zooplankton pop- 
ulations of the various treatments were not 
statistically different (Table 2). The initial 
densities for total zooplankton in CM, AM, 
and MM were respectively: 770 ± 131, 700 
±112, and 920 ± 160 organisms/L (means ± 
1 SE). Total zooplankton in all treatments ex- 



perienced continual positive growth from to 
89 h. Results at 89 h showed a striking dif- 
ference between densities in the CM, 1380 ± 
100 organisms/L, and the enriched meso- 
cosms with 2200 ± 337 organisms/L in the 
AM and 2700 ± 318 in the MM. While all 
treatments experienced growth, there was a 
significant difference (F = 3.742, P = 0.024) 
in overall zooplankton densities between the 
treatments (Table 2). Further, Holm-Sidak/?e>5t 
hoc tests found the difference between the to- 
tal zooplankton in CM from either enriched 
mesocosms AM or MM to be significant (P 
< 0.05). However, there was no significant 
difference (F = 3.742, P = 0.24) between the 
total zooplankton populations in AM and 
MM. There was a highly significant interac- 
tion (F = 3.648, P = 0.001) among the taxa 
and mesocosm treatments (Table 2). 

Of the zooplankton taxa recorded, only the 
copepod and rotifer populations were at high 
enough densities to record distinct trends. The 
copepod trend in mesocosms was different 
from the Total Zooplankton Trend (Fig. 3). 
The initial densities for the copepods in the 
CM, AM, and MM were similar (respectively: 
380 ± 65, 480 ± 104 and 480 ± 107 organ- 
isms/L; Fig. 3). The CM copepods peaked on 
42 h at 480 ± 1 30 organisms/L, decreased to 
230 ± 37 organisms/L by 65 h, and returned 
to 380 ± 58 organisms/L by 89 h. The co- 
pepods in the AM were the only population 
to increase to 810 ± 159 organisms/L by 89 
h. The population in the MM decreased slight- 
ly to 392 ± 42 organisms/L. These fluctua- 
tions over the experimental period for any 
treatment did not show a significant change 
(F = 0.987, P = 0.403). Holm-Sidak post hoc 
tests found the copepod population in AM to 



TRIERWEILER & PASCUAL— ZOOPLANKTON AND CYANOBACTERIA 



179 



3000 H 

2000 

1000 



Cl 

o 
o 
N 




3000 

2000 

1000 

4000 
3000 
2000 
1000 




Copepods 



n 



R 



_Q. 



Rotifers 



Pi w* 



n 




Total Zooplankton 




] Control 



r 

WZZA Anabaena 
Microcystis 




20 



40 



60 



80 



Hour 



100 



20 



Figure 3. — Zooplankton in different treatments. Solid bars represent the control mesocosm; diagonal 1\ 
hashed bars represent the Anabaena mesocosms; and crosshatched bars represent the Microcystis ineso- 
cosms. Each value is a mean (±1 SE) of eight counts. 



be significantly higher than the populations in 
both CM and MM for 65 h and 89 h (P < 
0.01 for both days and treatments). No signif- 
icant difference between the copepod popu- 
lations in the CM and MM (HS P = 0.92) 
was observed. 

Unlike the copepod trends, the rotifer trends 
closely resembled the total zooplankton treat- 
ment trends. The starting rotifer populations 
were distinct from each other, but the differ- 
ence did not appear to inhibit the growth in 
the enriched exposures (Fig. 3). All treatments 
showed significant (HS P < 0.05) growth of 
rotifer populations from to 89 h. By 89 h, 
rotifer densities for CM, AM, and MM 
reached 953 ± 108, 1359 ± 211, 2277 ± 278 
organisms/L, respectively. As with the total 
zooplankton populations, rotifer populations 
in the enriched mesocosms were significantly 
different (P < 0.01 for both AM and MM) 
from the CM but not from each other (P = 
0.76) except on the last day (HS P < 0.05). 



DISCUSSION 

Inoculated mesocosms resulted in higher 
zooplankton growth than the CM. which did 
not experience a statistically significant 
growth. However, different inoculations did 
not produce a statistically different effect on 
the total zooplankton from each other. This 
contradicts our hypothesis that the AM would 
more adversely affect zooplankton growth in 
comparison to the MM. However. AM and 
MM enrichments did exhibit taxa specific 
zooplankton trends. Copepods did significant- 
ly better in the AM by 65 h than in the other 
mesocosms, and rotifers did statistically better 
in the MM. These taxa specific trends reflect 
a preference for. or a tolerance tow aids, a cer- 
tain algal morphology. 

Although cladoeerans made up such a small 
percentage of the zooplankton counted that 
they could not be accurately analyzed, this 
taxa is worth mentioning in context to this 



180 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



study as a majority of zooplankton-cyanobac- 
teria studies use cladocerans. Since cladocer- 
ans have high filtering rates and a partheno- 
genic lifecycle like the rotifers (Allan 1976), 
cladocerans have been known to quickly re- 
spond to spring algal blooms and consume 
large amounts of algae such that they visibly 
clear the water, causing what is called the 
"Clear Water Phase" (Lampert et al. 1986; 
Sommer et al. 1986). In these favorable con- 
ditions, the cladocerans can out-compete and 
suppress other zooplankton with overlapping 
niches such as the rotifers (Gilbert 1990). This 
clearing effect is unlikely to occur if the algae 
are inedible, as most filamentous cyanobac- 
teria are considered to be (Lampert et al. 
1986; Sommer et al. 1986). The literature sug- 
gests that the presence of filamentous algae 
reduces the food gathering capability of 
daphnids. This is most observed with large fil- 
aments: larger filaments entering the daph- 
nid's feeding current are rejected, disrupting 
the feeding and consumption of edible algae 
(Burns 1968; Gliwicz 1977; Burns et al. 
1989). By being unable to exhibit selection in 
what they ingest, cladocerans are considered 
generalist feeders and must reject all their 
food if something inedible such as a filament 
is ingested (Gilbert 1990). As summer blooms 
in ECR are often composed of filamentous, 
heterocyst-forming cyanobacteria, the absence 
of cladocerans during the experiment is rea- 
sonable and consistent with the literature. 

The presence of other zooplankton, mainly 
copepods and rotifers, in ECR during the ex- 
perimental period indicates a tolerance if not 
adaptation to the late summer cyanobacterial 
community not exhibited by the Cladocera. 
This is probably the reason why the zooplank- 
ton community in the Anabaena mesocosms 
was not negatively impacted by the enrich- 
ment. 

Copepods, the largest taxon seen during the 
experiment, responded contrary to expecta- 
tions since they thrived in the AM (Fig. 3). 
Copepods have a very different lifestyle from 
the cladocerans and rotifers in that they repro- 
duce sexually. Because of the short time scale 
of the experiment, the copepod response to the 
inoculation may not have fully materialized as 
they have a longer reproductive cycle than do 
either the cladocerans or the rotifers (Allan 
1976). In addition, they display a very differ- 
ent feeding behavior that shows selectivity not 



seen in cladocerans. Copepods are known to 
discriminate by size, taste, and toxicity 
(DeMott 1986, 1988). In one study, copepods 
strongly avoided consuming colonial Micro- 
cystis aeruginosa because of chemical com- 
pounds associated with these cyanobacteria 
(Fulton & Paerl 1987). This active selection 
is a mechanism that allows copepods to co- 
exist with toxic algal blooms. 

While primarily filter feeders, copepods can 
capture larger particles raptorially, thus giving 
the taxon the largest size range for food par- 
ticles: 5-100 |jLm (Allan 1976; Pennak 1989). 
This size range allows them to access other 
food sources unavailable to both cladocerans 
and rotifers, such as larger filaments like An- 
abaena filaments. Some studies suggest that 
some species of copepods have a greater se- 
lectivity and feeding efficiency on larger food 
particles (Haney & Trout 1985; Vanderploeg 
et al. 1988; DeMott 1990). While it is unclear 
how efficient copepods are in handling fila- 
mentous cyanobacteria in comparison to a fil- 
amentous Chlorophyta or Bacillariophyta, 
Burns & Xu (1990) found calanoid copepods 
significantly reduced both density and tri- 
chome length of filamentous cyanobacteria. 
Regardless of their handling efficiency, cope- 
pods clearly benefited from the Anabaena sp. 
enrichment and are the only zooplankters 
morphologically capable of exerting enough 
grazing pressure to cause Anabaena densities 
and length to decrease. 

Rotifers, like copepods, exhibit a greater se- 
lectivity over what they can consume. This 
selectivity is better attributed to their relative 
smaller size (0.2-0.6 mm, Allan 1976) than 
any capabilities to choose what they ingest 
based on toxicity or taste (Kirk & Gilbert 
1992). Rotifers use coronal cilia for suspen- 
sion feeding, which restricts their consump- 
tion size to 1-20 |jLm (Allan 1976). As a result, 
rotifers are not impeded from feeding by hav- 
ing to reject filaments as do the cladocerans 
because most filaments cannot enter their 
mouth (Gilbert 1994). This advantage against 
physical inhibition also seems beneficial 
against toxic species that are filamentous or 
form mats, as they are less likely to ingest 
those toxins as with larger cladocerans. Gil- 
bert ( 1 990) demonstrated that the growth rates 
of five rotifer species were unaffected by high 
concentrations of toxic Anabaena ajfinis while 
growth rates of large cladocerans were re- 



TRIERWEILER & PASCUAL— ZOOPLANKTON AND CYANOBACTERIA 



181 



duced. This tolerance to Anabaena filaments 
due to their small size could explain why ro- 
tifer growth in the AM was higher than their 
growth in the CM (Fig. 3). The inability of 
rotifers to utilize the inoculated Anabaena sp. 
could lead to the conclusion that rotifer pop- 
ulations in the mesocosm inoculated with An- 
abaena sp. would closely resemble the rotifer 
populations in the CM if their feeding behav- 
ior was the only factor. However, this was not 
the case. The increase in rotifer populations in 
the AM could be explained by a lack of com- 
petition with copepods for the same food 
source as copepods were likely to consume 
the Anabaena sp. filaments. Thus, the rotifers 
would have had an ample food source of chlo- 
rophytes and Microcystis. 

The success of the rotifers in this experi- 
ment was aided in part by their highly oppor- 
tunistic nature and parthenogenic reproduc- 
tion. With an r max range of 0.2-1.5 d ] , rotifers 
are quicker in their response to the inocula- 
tions than cladocerans and copepods, whose 
maximum growth rate in ideal conditions 
(r max ) are respectively 0.2-0.6 d" 1 and 0.1-0.4 
d" 1 (Allan 1976). Therefore, rotifer popula- 
tions in the MM exploded when a food source 
within their size range was enriched to more 
than five times the natural concentrations (Fig. 
3). Fulton & Paerl (1988) observed in their 
study that a rotifer's (Brachionus calyciflorus) 
ability to handle a non-toxic strain of Micro- 
cystis aeruginosa was superior to that of a cla- 
doceran (Daphnia ambigua). 

Successful short term coexistence or even 
population growth found in these taxa specific 
trends requires behavioral or physiological ad- 
aptations to select, reject or tolerate specific 
cyanobacteria. The morphology of an algal 
bloom can change the zooplankton commu- 
nity structure and can favor certain zooplank- 
ton taxa to excel due to their feeding behav- 
iors and mechanisms. Therefore, while 
zooplankton graze upon nuisance algae in Ea- 
gle Creek Reservoir, this grazing stress may 
not have the same effect in reducing cyano- 
bacterial populations during the late summer 
as in the spring-early summer "clear water 
phase." Future mesocosm studies would ben- 
efit from a longer experimental period, a more 
efficient way to measure zooplankton and 
phytoplankton populations as phytoplankton 
and zooplankton identification and enumera- 
tion were time-consuming, and nutrient flux 



measurements. More mesocosm studies are 
needed for ECR and other lakes to fully un- 
derstand these interactions and to achieve a 
long term management plan for algal blooms. 

ACKNOWLEDGMENTS 

We are grateful to Mark L. Dewart and Dr. 
Lenore P. Tedesco for guidance and support 
during this study. We also thank Gregory L. 
Boyer of SUNY College of Environmental 
Science and Forestry for toxin analysis and 
Eagle Creek Park for use of their docks. This 
study was partially supported by a grant from 
Indiana Academy of Science and funding 
from the Center for Earth and Environmental 
Science at Indiana University-Purdue Univer- 
sity, Indianapolis. 

LITERATURE CITED 

Allan, J. D. 1976. Life history patterns in zooplank- 
ton. American Naturalist 110:165-180. 

Arnold, D.E. 1971. Ingestion, assimilation, surviv- 
al and reproduction by Daphnia pulex fed seven 
species of blue-green algae. Limnology and 
Oceanography 16:906-920. 

Burns, C.W. 1968. Direct observations of mecha- 
nisms regulating feeding behavior of Daphnia in 
lake water. Internationale Revue der gesamten 
Hydrobiologie 53:83-100. 

Burns, C.W., D.J. Forsyth, J.F. Haney. M.R. James. 
W. Lampert & R.D. Pridmore. 1989. Coexis- 
tence and exclusion of zooplankton by Anabaena 
minutissima var. attenuata in Lake Rotongaio, 
New Zealand. Archiv fur Hydrobiologie. Ergeb- 
nisse der Limnologie. Beiheft 32:63-82. 

Burns, C.W. & Z. Xu. 1990. Calanoid copepods 
feeding on algae and filamentous cyanobacteria: 
Rates of ingestion, defaecation and effects on tri- 
chome length. Journal of Plankton Research 12: 
201-213. 

DeMott, W.R. 1986. The role of taste in food se- 
lection by freshwater zooplankton. Oecologia 69: 
334-340. 

DeMott. W.R. 1988. Discrimination between algae 
and artificial particles by freshwater and marine 
copepods. Limnology and Oceanography 33: 
397-408. 

DeMott. W.R. 1990. Retention efficiency, percep- 
tual bias, and active choice as mechanism ot' 
food selection by suspension-feeding zooplank- 
ton. Pp. 569-594. /// Behavioral Mechanisms of 
Food Selection. (R.N. Hughes. edJ NATO AS1 
Series. Series G: Ecological Sciences. Springer. 
New York. New York. ISA. 

DeMott. W.R. & F. Moxter. 1991. Foraging Cya- 
nobacteria by copepods: Response to chemical 
defense and resource partitioning. Ecolog\ ~2: 
1820-1834. 



182 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Ferrao-Filho, A.S., B. Kozlowsky-Suzuki & 
S.M.F.O. Azevedo. 2002. Accumulation of mi- 
crocystins by a tropical zooplankton community. 
Aquatic Toxicology 59:201-208. 

Fulton, R.S. 1988: Resistance to blue-green algal 
toxins by Bosmina longirostris. Journal of 
Plankton Research 10:771-778. 

Fulton, R.S. & H.W. Paerl. 1987. Toxic and inhib- 
itory effects of the blue-green alga Microcystis 
aeruginosa on herbivorous zooplankton. Journal 
of Plankton Research 9(5):837-855. 

Fulton, R.S. & H.W. Paerl. 1988. Effects of the 
blue-green alga Microcystis aeruginosa on zoo- 
plankton competitive relations. Oecologia 76: 
383-389. 

Gilbert, J.J. 1990. Differential effects of Anabaena 
afflnis on cladocerans and rotifers: Mechanisms 
and implications. Ecology 71:1721-1740. 

Gilbert, J.J. 1994. Susceptibility of planktonic ro- 
tifers to a toxic strain of Anabaena flos-aquae. 
Limnology and Oceanography 39:1286-1297. 

Gliwicz, M.Z. 1977. Food size selection and sea- 
sonal succession of filter feeding zooplankton in 
an eutrophic lake. Ekologia Polska 25:179-225. 

Gliwicz, Z.M. & W. Lampert. 1990. Food thresh- 
olds in Daphnia species in the absence and pres- 
ence of blue-green filaments. Ecology 71:691- 
702. 

Haney J.F 1987. Field studies on zooplankton-cy- 
anobacteria interactions. New Zealand Journal of 
Marine and Freshwater Research 21:467-475. 

Haney, J.F. & M.A. Trout. 1985. Size selective 
grazing by zooplankton in Lake Titicaca. Archiv 
fur Hydrobiologie, Ergebnisse der Limnologie, 
Beiheft 21:147-160. 

Holm, N.P., G.G. Ganf & J. Shapiro. 1983. Feeding 
and assimilation rates of Daphnia pulex fed 
Aphanizomenon flos-aquae. Limnology and 
Oceanography 28:677-687. 

Indiana Department of Environmental Manage- 
ment. 2002. Indiana Integrated Water Monitor- 
ing and Assessment Report. IDEM/34/02/004/ 
2002. http://www.in.gov/idem/owm/planbr/wqs/ 
quality/2002integrept/2002integrept.pdf. 

Indiana Department of Environmental Manage- 
ment. 2004. Indiana Integrated Water Monitor- 
ing and Assessment Report, http://www.in.gov/ 
idem/programs/water/303d/index.html 

Indiana Department of Environmental Manage- 
ment. 2006. Indiana Integrated Water Monitor- 
ing and Assessment Report, http://www.in.gov/ 
idem/programs/water/303d/ndex.html 

Kirk, K.L. & J.J. Gilbert. 1992. Variation in her- 
bivore response to chemical defenses: Zooplank- 
ton foraging on toxic cyanobacteria. Ecology 73: 
2208-2217. 

Kurmayer, R. 2001. Competitive ability of Daph- 
nia under dominance of non-toxic filamentous 
cyanobacteria. (M. Boersma & K. Wiltshire, 



eds.). Proceedings of the V Symposium on Cla- 
docera in Plon (Germany). Hydrobiologia 442: 
279-289. 

Lampert, W 1987. Laboratory studies on zoo- 
plankton-cyanobacteria interactions. New Zea- 
land Journal of Marine and Freshwater Research 
21:483-490. 

Lambert, W, W Fleckner, H. Rai, & B.E. Taylor. 
1986. Phytoplankton control by grazing zoo- 
plankton: A study on the spring clear- water 
phase. Limnology and Oceanography 31:478- 
490. 

Lewis, P. A., D.J. Klemm, J. A. Lazorchak, T.J. Nor- 
berg-King, W.H. Peltier & M.A. Heber (eds.). 
1994. Short-term Methods for Estimating the 
Chronic Toxicity of Effluents and Receiving Wa- 
ter to Freshwater Organisms. Third Edition. 
EPA/600/4-91/002. Environmental Monitoring 
Systems Laboratory, U.S. Environmental Protec- 
tion agency, Cincinnati, Ohio. 

Newhouse, S.A. & J.R. Stahl. 2000. A comparison 
of the midwater planktonic invertebrate com- 
munities of Eagle Creek, Geist, and Morse Res- 
ervoir in central Indiana using underwater light 
trapping. IDEM 32/03/005/2000. 

Pascual, D.L. & L.P. Tedesco. 2004. Eagle Creek 
Reservoir phytoplankton ecology study. Central 
Indiana Water Resources Partnership, CEES 
Publication 2004-01, IUPUI, Indianapolis, Indi- 
ana. 39 pp. 

Pennak, R.W 1989. Fresh-water Invertebrates of 
the United States: Protozoa to Mollusca. Third 
edition. John Wiley & Sons, New York, New 
York. 628 pp. 

Sommer, U., Z.M. Gliwicz, W. Lampert & A. Dun- 
can. 1986. The PEG Model of seasonal succes- 
sion of planktonic events in freshwaters. Archiv 
fur Hydrobiologoie 106:433-471. 

Sterner, R.W. 1989. Resource competition during 
seasonal succession toward dominance by Cya- 
nobacteria. Ecology 70:229-245. 

Tedesco, L.P, E.A. Atekwana, G.F. Filippelli, K. 
Licht, L.K. Shrake, B.E. Hall, D.L. Pascual, J. 
Latimer, R. Raftis, D. Sapp, G. Lindsey, R. Ma- 
ness, D. Dershing, D. Peterson, K. Ozekin, C. 
Mysore, & M. Prevost. 2003. Water quality and 
nutrient cycling in three Indiana watersheds and 
their reservoirs: Eagle Creek/Eagle Creek Res- 
ervoir, Fall Creek/Geist Reservoir, and Cicero 
Creek/Morse Reservoir. Central Indiana Water 
Resources Partnership, CEES Publication 2003- 
01, IUPUI, Indianapolis, Indiana. 163 pp. 

Tedesco, L.P, D.L. Pascual, L.K. Shrake, L.R. Cas- 
ey, B.E. Hall, RG.F Vidon, FV. Hernly, R.C. 
Barr, J. Ulmer & D. Pershing. 2005. Eagle 
Creek Watershed Management Plan: An Inte- 
grated Approach to Improved Water Quality. Ea- 
gle Creek Watershed Alliance, CEES Publication 
2005-07, IUPUI, Indianapolis. 182 pp. 



TRIERWEILER & PASCUAL— ZOOPLANKTON AND CYANOBACTERIA 



183 



Vanderploeg, H.A., G.-A. Paffenhofer & J.R. Lie- 
big. 1988. Diaptomus vs. net phytoplankton: Ef- 
fects of algal size and morphology on selectivity 
of a behaviorally flexible, omnivorous copepod. 
Bulletin of Marine Science 43(3):377-394. 

Wang, X., P. Parkpian, N. Fujimoto, K.M. Ruchir- 
awat, R.D. DeLaune & A. Jugsujinda. 2002. En- 



vironmental conditions associating microcystins 
production to Microcystis aeruginosa in a res- 
ervoir of Thailand. Journal of Environmental 
Science and Health 37:1 181-1207. 

Manuscript received JO December 2005, revised 15 
June 2007. 



2007. Proceedings of the Indiana Academy of Science (11 6): 1 84—195 



FISHES OF THE LOWER PRAIRIE CREEK AREA, 
VIGO COUNTY, INDIANA 

John O. Whitaker, Jr., Sherry L. Veilleux, Joseph Duchamp, Matthew Griswold, 
and Douglas Rees: Department of Ecology and Organismal Biology, Indiana State 
University, Terre Haute, Indiana 47809 USA 

ABSTRACT. Sixty-five species of fish are known from the lower Prairie Creek drainage of Vigo County, 
Indiana; 51 of them had been reported prior to the present study. During the summer of 1999 we seined 
15 sites in Prairie Creek, Oxendine Bayou, and Muskrat Pond. Nine of the sites had been previously 
sampled by Gerking (1945) and/or Whitaker and Wallace (1973). Ten of the 51 previously found species 
were not found in 1999 or in any collections after 1987. They are Notropis atherinoides, Phenacobius 
mirabilis, Rhinichthys obtusus, Carpiodes cyprinus, C. carpio, Erimyzon oblongus, Etheostoma flabellare, 
E. gracile, Percina maculata, and Sander canadensis. The number of species of fish known from Vigo 
County is now 111, or 52.6% of the 211 species known from Indiana. 

Keywords: Fishes, Indiana, Vigo County, Prairie Creek 



Vigo County, Indiana, was first sampled for 
fishes by Jenkins (1887), who reported 63 spe- 
cies of fishes there. However, he did not sam- 
ple the Prairie Creek area. Prairie Creek is lo- 
cated in west central Indiana and is primarily 
a small stream (25 m), mostly sandy bottom, 
but with some gravel and rocky bottomed ar- 
eas. Collections by Jordan (1877, 1890), Ev- 
ermann & Jenkins (1888), Jordan & Ever- 
mann (1902), Hubbs & Trautman (1937), 
Blatchley (1938), and Gerking (1945) added 
20 species to the known fish fauna of Vigo 
County, making a total of 83 species. Whitak- 
er and Wallace (1973) during a study of the 
fishes of Vigo County reported an additional 
25 species, yielding 108 species known from 
Vigo County. Based on historical collecting 
through 1973, 51 species of fish were known 
to occur in the lower Prairie Creek area. The 
Prairie Creek area is particularly interesting 
because it has been little studied, it includes a 
large tract of bottomland woods, and it ap- 
pears to have high diversity at least for some 
organisms. 

The purpose of this study was to assess the 
present distribution and abundance of fish in 
the lower Prairie Creek area of Vigo County, 
Indiana, and to determine changes in the fish 
community from previous studies (1945— 
1966) of this area, as compared to the period 
1980 to 1999. 



STUDY AREA 

Prairie Creek drains much of the south- 
western portion of Vigo County (Fig. 1), flow- 
ing west into the Wabash River bottoms, 
where it flows into Negro Ditch. As Prairie 
Creek turns south it drains approximately 650 
ha (1600 acres) of forest. Oxendine Bayou 
runs west from Negro Ditch, then to the south- 
west. Muskrat Pond and Round Pond both lie 
west of the north-south portion of Prairie 
Creek and south of Oxendine Bayou. Most of 
Prairie Creek has a sandy bottom, but there 
are periodic riffles in the upper east/west por- 
tion. The stream width is about 9 m and depth 
is up to 1 meter. Oxendine Bayou is 6-10 m 
wide with a mud bottom. Except for periodic 
pools, much of Prairie Creek dries up in the 
summer, but the entire bottomland part of the 
study area (all sites except 4, 1, and 14) may 
be flooded when the Wabash River floods. 

METHODS 

Fifteen collections of fish in the lower Prai- 
rie Creek area were made during 19-30 July 
1999, including ten in Prairie Creek, three 
along Oxendine Bayou, one in Round Pond, 
and one in Muskrat Pond (Fig. 1). Four of the 
ten sites along Prairie Creek, two in Oxendine 
Bayou, and the one in Muskrat Pond had been 
sampled in previous studies. Also included are 
unpublished data from the 1990 Indiana State 
University vertebrate zoology class from Ox- 



84 



WHITAKER ET AL.— FISHES OF PRAIRIE CREEK 



185 




County Line 



Figure 1. — Collection sites of the 1999 study in the lower Prairie Creek area. 



endine Bayou, which documented the first 
Gambusia in Vigo County, and one collection 
in Prairie Creek by Leanna Smith (1988). 

Collections were made with either a 15 or 
30 foot, Va inch mesh seine. Collection sites 
along Prairie Creek and Oxendine Bayou av- 
eraged 100 m in length with the average width 
being approximately 7-8 m. Muskrat Pond 
was sampled with seine hauls throughout most 
of its length and width. Nomenclature follows 
the Revised Checklist of the Vertebrates of In- 
diana (Simon et al. 2002). 

RESULTS 

Sixty-five species of fish from 15 families 
and 40 genera have been taken in the lower 
Prairie Creek area, Vigo County, Indiana (Ta- 
ble 1). Fifty of these species were taken since 
1988. The most abundant species in Prairie 
Creek by family in the recent collection were: 
Clupeidae — Dorosoma cepedianum, 44; Cy- 
prinidae — Hybognathus nuchalis, 697, Cypri- 
nella spiloptera, 174, Notropis blennius, 70, 
Pimephales notatus, 50; Catostomidae — Ca- 
tostomus commersoni, 10; Centrarchidae — Le- 
pomis macrochirus, 25, Etheostoma caeru- 
leum, 24, and E. nigrum, 13. The most 
abundant species in Oxendine Bayou were 
Gambusia affinis, 104; D. cepedianum, 52; 
Cyprinus carpio, 44; L. macrochirus, 37; Po- 
moxis annularis, 24; P. nigromaculatus, 21; 
and Hypophthalmichthys nobilis, 21. 

Fourteen species had not been found in the 
Prairie Creek area prior to 1988 and constitute 



new locality records. They are Lepisosteus os- 
seus, Ctenopharyngodon idella, Cyprinella 

whipplei, Hypophthalmichthys nobilis, Notro- 
pis stramineus, Hypentelium nigricans, Ictal- 
urus punctatus, Gambusia affinis, Labidesthes 
sicculus, Morone chrysops, Ammocrypta pel- 
lucida, Etheostoma chlorosoma, Percina 
phoxocephala, and Aplodinotus grunniens. 

Ten species were taken before 1987. bm 
they have not been taken in any of the collec- 
tions since. Listed in order of decreasing num- 
bers taken, they are: Notropis atherinoides 
(65), Erimyzon oblongus (9). Sander canaden- 
sis (9), Phenacobius mirabilis (6), Carpiodes 
cyprinus (4), Carpiodes carpio (4). Etheosto- 
ma gracile (3), Etheostoma flabellare \2). 
Percina maculata (2). and Rhinichthyes ob- 
tusus (1). All of these are listed as common 
by Simon et al. (2002). except for Erimyzon 
oblongus and Etheostoma gracile. which are 
listed as rare. 

Nineteen species o\' cyprinids have been 
collected from Prairie Creek at nine collection 
sites sampled during the present stud) and 
four of 17 sites sampled b\ Whitaker & Wal- 
lace (1973). Gerking ( 1945) collected onl> ten 
species, but he sampled onl\ one site in this 
area. The most common species of cvprinid 
fish in Prairie Creek during the prevent stud) 
were Hybognathus nuchalis, Xotropis blen- 
nius, Crprinclla spiloptera, and Pimephales 
notatus. 

Eleven species of cyprinids were collected 



186 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Table 1 . — Species of fish collected in the lower Prairie Creek area. Gerking did not provide numbers 
of individuals collected; therefore, his collections are marked by an 'X'. Species taken since 1989 are 
marked with an asterisk. PC = Prairie Creek, OB = Oxendine Bayou, MP = Muskrat Pond, RP = Round 
Pond. 



Earlier collections (1945-1966) 



Gerking 
(1945) 

PC 



Whitaker & Wallace (1973), 
1962-66 



PC 



OB 



MP 



RP 



Lepisosteidae 
Lepisosteus osseus* 

Longnose gar 
Lepisosteus platostomus* 
Shortnose gar 

Amiidae 
Ami a calva 
Bowfin 

Clupeidae 
Dorosoma cepedianum* 
Gizzard shad 

Cyprinidae 
Campostoma anomalum* 

Stoneroller 
Ctenopharyngodon idella* 

Grass carp 
Cyprinella spiloptera* 

Spotfin shiner 
Cyprinella whipplei* 

Steelcolor shiner 
Cyprinus carpio* 

Carp 
Ericymba buccata* 

Silverjaw 
Hybognathus nuchalis* 

Mississippi silvery minnow 
Hypophthalmichthys nobilis* 

Bighead carp 
Luxilus chrysocephalus* 

Striped shiner 
Ly thrums umbratilis* 

Redfin shiner 
Notemigonus crysoleucas* 

Golden shiner 
Notropis atherinoides* 

Emerald shiner 
Notropis blennius* 

River shiner 
Notropis stramineus* 

Sand shiner 
Phenacobius mirabilis 

Suckermouth minnow 
Pimephales notatus* 

Bluntnose minnow 
Pimephales vigilax* 

Bullhead minnow 




93 

94 

300 



60 



X 


15 

























X 


24 


21 






















X 


12 


213 


24 


180 




2 











X 


132 


39 






















X 


24 











X 


47 





2 







1 


2 










63 


2 










144 


9 






















X 


6 











X 


59 


4 













2 









WHITAKER ET AL.— FISHES OF PRAIRIE CREEK 



187 



Table 1. — Extended. 



Oxendine Bayou ] 


988- 


-1998 




Present (1999) 






Smith 


Class 




Class 


Overall 


(1989) 


(1990) 




(1998) 


PC 


OB 


MP 


RP 


totals 










50 


3 











53 


8 







10 


6 


20 


1 





139 



500 



594 



400 



200 



44 



52 



100 



1165 












2 











17 








100 


5 


31 


8 





144 





12 


5 


174 





1 





237 











9 











9 


1 


15 


5 


22 


44 


19 


180 


535 











3 











5 











697 


17 


20 





905 








200 


6 


21 


10 





237 











6 











30 











2 











51 


13 


12 











1 





29 























65 





150 





70 











3-3 











3 











3 























6 











50 


5 








11s 





2 





1 











5 



Table 1. — Continued. 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 







Earlier collections (1945 


-1966) 






Gerking 




Whitaker & Wallace (1973), 






(1945) 




1962- 


66 






PC 


PC 


OB 


MP 


RP 


Rhinichthyes obtusus 












Blacknose dace 


X 


1 











Semotilus atromaculatus* 












Creek chub 


X 


46 


2 








Catostomidae 












Carpioides carpio 












River carpsucker 




3 


1 








Carpiodes cyprinus 












Quillback 




4 











Catostomus commersoni* 












White sucker 




22 


1 








Erimyzon oblongus 












Creek chubsucker 


X 


9 











Hypentelium nigricans* 












Northern hogsucker 
















Ictiobus bubalus* 












Smallmouth buffalo 


X 














Ictiobus cyprinellus* 












Bigmouth buffalo 







61 





2 


Ictiobus niger 












Black buffalo 







80 





15 


Minytrema melanops* 












Spotted sucker 


X 


1 


1 








Moxostoma erythrurum* 












Golden redhorse 


X 














Ictaluridae 












Ameiurus melas* 












Black bullhead 


X 


10 


21 








Ameiurus natalis* 












Yellow bullhead 


X 


3 


1 








Ictalurus punctatus* 












Channel catfish 
















Esocidae 












Esox americanus* 












Grass pickerel 




29 


2 








Aphredoderidae 












Aphredoderus say anus 












Pirate perch 




148 


38 








Fundulidae 












Fundulus notatus* 












Blackstripe topminnow 


X 


3 











Poeciliidae 












Gambusia affinis* 












Mosquitofish 
















Atherinidae 












Labidesthes sicculus 












Brook silverside 

















WHITAKER ET AL.— FISHES OF PRAIRIE CREEK 1 89 

Table 1. — Extended (Continued). 



Oxendine Bayou 1988- 


-1998 












Smith 


Class 


Class 




Present (1999) 




Overall 


(1989) 


(1990) 


(1998) 


PC 


OB 


MP 


RP 


totals 























1 











14 





1 





63 























4 























4 











10 











33 























9 











2 











2 





1 





9 


7 


3 





20 





1 


5 





8 


21 


2 


98 




















15 


95 











6 











8 











2 











: 


6 








1 


3 


60 





101 


5 








1 











10 


2 


6 





L5 





1 





24 



30 



89 



75 1 -9 



10 80 104 1000 1200 



190 

Table 1. — Continued. 


PROCEE 


;dings of the Indiana academy oi 


F SCIENCE 






Earlier collections (1945-1966) 






Gerking 
(1945) 


Whitaker & Wallace (1973), 
1962-66 






PC 


PC OB MP 


RP 


Moronidae 
Mo rone chrysops 
White bass 











Centrarchidae 
Lepomis cyanellus* 

Green sunfish 
Lepomis gulosus* 

Warmouth 
Lepomis humilis* 

Orangespotted sunfish 
Lepomis macrochirus* 

Bluegill 
Lepomis megalotis* 

Longear sunfish 
Lepomis microlophus* 

Redear sunfish 
Micropterus punctulatus* 

Spotted bass 
Micropterus salmoides* 

Largemouth bass 
Pomoxis annularis* 

White crappie 
Pomoxis nigromaculatus* 

Black crappie 

Percidae 
Ammocrypta pellucida* 

Eastern sand darter 
Etheostoma blennioides* 

Greenside darter 
Etheostoma caeruleum* 

Rainbow darter 
Etheostoma chlorosoma 

Bluntnose darter 
Etheostoma flabellare 

Fantail darter 
Etheostoma gracile 

Slough darter 
Etheostoma nigrum* 

Johnny darter 
Etheostoma spectabile* 

Orangethroat darter 
Percina caprodes* 

Logperch 
Percina maculata 

Blackside darter 
Percina phoxocephala* 

Slenderhead darter 
Sander canadense 

Sauger 

Sciaenidae 
Aplodinotus grunniens* 
Freshwater drum 



X 


27 


83 













1 








X 





3 








X 


20 


285 








X 


3 
















1 













1 










2 


12 













43 













55 






















X 


11 











X 


1 

























X 


2 













3 











X 


53 











X 


2 











X 


4 


2 










2 






























8 


1 






WHITAKER ET AL.— FISHES OF PRAIRIE CREEK 



191 



Table 1. — Extended (Continued). 



Oxendine Bayou 1988-1998 

Smith Class Class 

(1989) (1990) (1998) 



PC 



Present (1999) 



OB 



MP 



RP 



Overal 

totals 









5 


3 


1 


2 





121 














8 


4 





13 











2 


1 








6 


10 


25 


2 


25 


37 


52 





456 











16 











19 














4 








5 











12 


2 








15 





15 


2 


5 


1 1 





o 


47 





25 


50 


5 


24 


5 





152 


8 


50 


50 


2 


21 


59 





245 











2 











2 











1 











12 











24 


1 








26 





15 

















15 























-> 























3 











13 











66 











3 











5 











10 











16 























-> 











5 











5 























9 



20 



26 



192 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



in Oxendine Bayou. Listed by decreasing 
abundance, the most common species were 
the common carp (Cyprinus carpio), the big- 
head carp (Hypophalmichthys nobilis), the riv- 
er shiner {Notropis blennius), the grass carp 
(Ctenopharyngodon idella), and the silvery 
shiner {Hybognathus muchalis). Other species 
taken in relatively low numbers were Cypri- 
nella spiloptera, Notemigonus chrysoleucas, 
Pimephales notatus, P. vigilax, Notropis ath- 
erinoides, and Semotilus atromaculatus. No- 
tropis atherinoides, Luxilus chrysocephalus, 
S. atromaculatus, and Cyprinella spiloptera 
were not netted at all during the 1999 work. 

Of the ten species of suckers that have been 
found in Prairie Creek, the most abundant was 
the white sucker (Catostomus commersonii). 
The other nine species collected were Erimy- 
zon oblong us, Ictiobus bubalus, I. cyprinellus, 
I. niger, Minytrema melanops, Carpiodes cy- 
prinus, Carpiodes carpio, Hypentelium nigri- 
cans, and Moxostoma erythrurum. Erimyzon 
oblongus was taken earlier, but not in the re- 
cent studies. 

Ten species of percids have been collected 
from Prairie Creek. Gerking (1945) collected 
six species, Whitaker & Wallace (1973) eight, 
and seven species were collected during the 
present study. The most common percid ob- 
served in Prairie Creek was Etheostoma ni- 
grum. Whitaker & Wallace (1973) captured 53 
of the 66 individuals, whereas only 13 were 
taken in the present study (Table 1). Twenty- 
five individuals of Etheostoma caeruleum 
were netted, with 24 taken in the present study 
(Table 1). Etheostoma nigrum and E. specta- 
bile are the most common darters in the sand 
streams of Vigo County (Whitaker & Wallace 
1973). Other percids observed were Percina 
caprodes, Etheostoma blenniodes, Etheostoma 
spectabile, Percina phoxocephala, Etheosto- 
ma gracile, Percina maculata, Etheostoma 
flabellare, and Ammocrypta pellucida. Whi- 
taker & Wallace (1973) sampled only one site 
on the north-south portion of the creek but 
caught no fish at all there. Four sites were 
sampled during the present study (1999), and 
only three percids were taken. 

Few percids were collected in Oxendine 
Bayou. Not surprisingly, Etheostoma chloro- 
soma was the most abundant species (15 in- 
dividuals collected in 1990), as it is typically 
found in bottomland pools and bayous. Other 
percids collected in decreasing abundance 



were Sander canadensis, Percina caprodes, 
and Etheostoma caeruleum. 

DISCUSSION 

Species that comprised the largest number 
of individuals collected from all studies of the 
lower Prairie Creek area were: western mos- 
quitofish, gizzard shad, silvery shiner, bowfin, 
common carp, and bluegill. Three species, the 
mosquitofish (Gambusia affinis), grass carp 
(Ctenopharyngodon idella), and bighead carp 
(Hypophthalmichthys nobilis), have apparent- 
ly recently become established in Vigo Coun- 
ty- 

The bighead carp was first introduced in the 
1970s in Arkansas to be used as an aquacul- 
ture species (Henderson 1976; Jennings 
1988). The grass carp was introduced to con- 
trol plant growth in ponds and lakes (Laird & 
Page 1996). The first bighead carp and grass 
carp from Indiana were taken in northern Vigo 
County in 1988 (unpubl. data). The grass carp 
has also been collected in northwest Indiana, 
at Moss Lake (Simon et al. 2004), and in the 
Little Calumet River, Lake Michigan, and sev- 
eral other lakes in Porter and LaPorte counties 
(Simon et al. 2005). The bighead carp has also 
been collected in the Wabash River, the West 
Fork of the White River, and the lower Ohio 
River (Simon 2006). 

No mosquitofish had been taken in Vigo 
County until recently, although a great amount 
of collecting has been done there. Since 1990, 
they have occurred by the thousands. The first 
record of the western mosquitofish was in 
1990, when six individuals were taken in Ox- 
endine Bayou in Vigo County, but hundreds 
were present in 1992 and since (Clem & Whi- 
taker 1996). The original range of this species 
in the state was in extreme southwest Indiana. 
We suspect that they became established in 
Vigo County about 1989 or 1990. With no 
current (except during flooding), warm water, 
and aquatic vegetation, Oxendine Bayou pro- 
vides an excellent habitat for this species. 
Eighty mosquitofish were collected in Prairie 
Creek during the present study. By far the 
most abundant species collected in Muskrat 
Pond during the present study was the western 
mosquitofish. We collected over 1000 individ- 
uals, but probably hundreds of thousands 
more were present. It had not been taken in 
the previous collections from Muskrat Pond. 
Many young and pregnant female mosquito- 



WHITAKER ET AL.— FISHES OF PRAIRIE CREEK 



193 



fish were observed. Since Muskrat Pond re- 
ceives most of its water from the Wabash Riv- 
er, it is presumed that the mosquitofish entered 
the lower Prairie Creek area via the Wabash 
River. 

Some of the species not taken prior to 1987 
deserve special comment. Whitaker & Wal- 
lace (1973) and Smith (1988) did not observe 
the bighead carp (H. nobilis) or the grass carp 
(C. idella). These two species had not been 
taken in Vigo County or in Indiana until re- 
cently. Both species were taken at several sites 
during the present study. These are both exotic 
species introduced from Asia, and each was 
first taken in the state in 1988 in Otter Creek 
below Markle's Dam in the northern part of 
Vigo County (unpubl. data, JOW). The fish 
presumably came up Otter Creek from the 
Wabash River. Our collections include young 
individuals which, along with the large num- 
bers taken, indicates that these species are 
successfully reproducing in Vigo County. 

The eastern sand darter, Ammocrypta pel- 
lucida, is listed as special concern in the state, 
and is a federally-endangered candidate spe- 
cies (Simon et al. 2002). In addition, it has 
declined noticeably in Vigo County over re- 
cent decades, as indicated in annual samples 
collected by the Indiana State University ver- 
tebrate zoology classes. Therefore, it was of 
special interest that Ammocrypta had not been 
found earlier in Prairie Creek, but it was found 
during the present work. 

The bluntnose darter, Etheostoma chloro- 
soma, is listed as rare (Simon et al. 2002) but 
likewise was not found earlier; however, 15 
individuals were found during the present 
work. This species is fairly common in the 
sloughs and ponds of the Wabash River bot- 
toms. 

The other nine species not taken before 
1989 are listed in order of decreasing numbers 
taken: Lepisosteus osseus (53), Aplodinotus 
grunniens (26), Ictalurus punctatus (24), Cy- 
prinella whipplei (9), Morone chrysops (5), 
Percina phoxocephala (5), Notropis strami- 
neus (3), Hypentelium nigricans (2), and La- 
bidesthes sicculus (1). 

The creek chub {Semotilus atromaculatus), 
silverjaw minnow (Ericymba buccata). and 
the central stoneroller (Campostoma pullwn) 
were three of the most abundant species in 
Vigo County as a whole; but they were taken 
only in low numbers at Prairie Creek in pre- 



vious studies and in the present study. This 
result is surprising since they were some of 
the most common species in Vigo County 
(Whitaker & Wallace 1973). Other species 
collected in low numbers (30 individuals or 
less) were the common shiner (Luxilus chry- 
socephalus), steelcolor shiner (Cyprinella 
whipplei), Notropis stramineus, Phenacobius 
mirabilis, golden shiner (Notemigonus cryso- 
leucas), bullhead minnow {Pimephales vigi- 
lax), and western blacknose dace (Rhinichthys 
obtusus). 

Whitaker and Wallace (1973) captured the 
pirate perch (Aphredoderus sayanus). in Vigo 
County only in the Prairie Creek area. It was 
fairly common, 148 being taken by Whitaker 
& Wallace (1973) in Prairie Creek, and 38 in 
Oxendine Bayou. Three individuals were tak- 
en by Smith (1988) and none since, even 
though we seined at this site nearly every year 
from 1980 to 2000. From the results of the 
present study, it appears that the pirate perch 
no longer exists in the lower Prairie Creek 
area, although it is possible it has just been 
greatly reduced. Numerous factors could have 
contributed to decline of the pirate perch pop- 
ulation in the lower Prairie Creek area — such 
as the disturbance to the habitat, the fluctuat- 
ing nature of a bottomland, seasonal variation, 
or introduced and native species competing 
for resources. The pirate perch requires abun- 
dant cover, such as aquatic plants or aquatic 
debris. It is a solitary fish that hides under the 
vegetative growth during the daytime, and 
comes out to feed at night (Pflieger 1997). 
Since trees were removed along the creek at 
this site in 1997, it would not be surprising it 
the disappearance of this fish coincides with 
the removal of trees and the dredging within 
the creek. Jake Burskey. an Indiana State Uni- 
versity graduate student (pers. commun.). 
made the only recent collection of this species 
in Vigo Count)'. It was from northern Hone) 
Creek about 5km NE of Riley, Indiana on 10 
November 2006. while electrofishing for cray- 
fish. This locality is about 30 km north of 
Prairie Creek. 

In all three studies. Gerking (1945), Whi- 
taker & Wallace ( 1973). and the present study, 
the majority o\' percids taken were from the 
east-west portion of Prairie Creek, apparent 1\ 
because o( its habitat diversity: a combination 
of rock, gravel, bedrock, and sand, with man> 
areas of riffles, whereas the north-south por- 



194 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



tion tends to be sand and mud. The rock and 
gravel, rather than sand bottom is the pre- 
ferred habitat of Etheostoma caeruleum, 
which was more abundant than E. spectabile. 
Low numbers of percids in the north-south 
portion may be due to low sampling effort in 
this area. 

Most of the suckers prefer either a constant 
current or permanent and stable pools (Pflie- 
ger 1997). It is not surprising that so few 
suckers were present since Prairie Creek is 
continually subject to flooding and drought. 

As indicated above, there have been some 
changes in the fish fauna of the Prairie Creek 
area. Of the 65 species known from Prairie 
Creek, 14 were not taken prior to 1987, and 
10 were not taken after that year, whereas the 
other 41 were taken in both the earlier and 
later periods. Some of these differences ap- 
pear to be real, whereas some are probably 
due to chance — i.e., populations vary greatly 
from year to year and these variations can af- 
fect our results (Whitaker 1976; Grossman et 
al. 1982), i.e., part of the variation was prob- 
ably due to stochasticity. 

Real changes seem to be the addition of the 
bighead carp, the grass carp, and the mosqui- 
tofish, but those species were introduced. It is 
likely that the remainder of the species not 
collected prior to 1987 were present in small 
numbers, but were missed during the later col- 
lection. It is also likely that all ten species 
taken prior to 1987, but not since, probably 
occurred in small numbers and would have 
been taken if more seining were carried out. 
Total biodiversity as indicated by this study 
was about the same, with 5 1 species taken ear- 
lier, and 55 (including 3 recent introductions) 
later. Our data seem to indicate few major dif- 
ferences in the fish fauna over this period oth- 
er than the three introductions, and the major 
decrease in pirate perch. 

In contrast, Retzer (2005) found a loss of 
an average of 8.4 species per basin among 
seven river basins in Illinois in the past 100 
years. All seven drainages he studied lost four 
or more species of fishes. The greatest losses 
were in the highly disturbed Des Plaines River 
in the Chicago area where the greatest number 
of species was lost. On the other extreme was 
the Vermillion River in a highly agricultural 
area, with a loss of four species. The Vermil- 
lion River basin is about 85 km NW of the 
Prairie Creek area. On the other hand, species 



richness has been again increasing in some of 
the Illinois basins over the past 25 years. 

Patton et al. (1997) assessed changes in the 
fish fauna in ten drainages in Wyoming be- 
tween the 1960s and 1990s. They found that 
12 of 31 native species were collected in few- 
er locations during the 1990s survey even 
with more efficient gear. 

Koel & Peterka (1948) examined water 
quality and fishes in the Red River Basin of 
Minnesota, North Dakota, and South Dakota. 
That area also has become a major agricultural 
area. Although it is much larger than the Prai- 
rie Creek area, it contains only 85 species of 
fish, which is interesting since it is so large as 
compared to the Prairie Creek area with its 75. 
The Red River area has lost only one species 
over time. 

The Prairie Creek area resembled more the 
Red River area than the Illinois or Wyoming 
areas in probably losing fewer fish over time. 
We think the relatively small loss in Prairie 
Creek is because there have not been major 
habitat changes in this bottomland area 
through the period of study. Surely, pollutants 
enter the area from farming operations in the 
area; but otherwise this is a remote area that 
is not much affected by nearby industrial or 
domestic pollution, nor does much develop- 
ment occur there. The habitat is affected most- 
ly by flooding from the Wabash River. The 
flooding allows fish to move overland, but due 
to rains that caused the flooding, pollutants in 
the Wabash itself would be at their lowest lev- 
els. 

The Lower Prairie Creek area is a biologi- 
cally diverse aquatic and terrestrial landscape. 
Of the 21 1 species of fish known from Indiana 
(Simon et al. 2002; Simon pers. commun.), 
1 08 were known from Vigo County (Whitaker 
& Wallace 1973). Three species taken during 
the present study (the mosquitofish, grass 
carp, and bighead carp) were not reported by 
Whitaker & Wallace (1973), but they now 
bring the species reported from Vigo County 
to 1 1 1, or 52.6% of the 211 known fish spe- 
cies of Indiana. The number is this large be- 
cause of the diverse habitats provided by the 
Wabash River flowing through the county. In 
addition, part of the Southeastern Lowlands 
Natural Region extends northward into south- 
ern Vigo County. An additional reason for the 
high biodiversity is the lack of development 
in the bottomlands due to flooding. Until re- 



WHITAKER ET AL.— FISHES OF PRAIRIE CREEK 



195 



cently, the area has remained in an unfrag- 
mented state. Major threats to the lower Prai- 
rie Creek area are those common to 
bottomland areas: agricultural runoff, water 
drainage for agricultural fields, and erosion/ 
siltation runoff from logging. 

ACKNOWLEDGMENTS 

We thank Robert Jenkins and Thomas Si- 
mon for reading and improving the manu- 
script, Laura Bakken for typing it, and Linda 
Castor for the map. 

LITERATURE CITED 

Blatchley, W.S. 1938. The Fishes of Indiana. Na- 
ture Publishing Co., Indianapolis, Indiana. 121 
pp. 

Clem, P.D. & J.O. Whitaker, Jr. 1996. Distribution 
of the mosquitofish, Gambusia affinis (Baird and 
Girard), in Indiana, with comments on resource 
competition. Proceedings of the Indiana Acade- 
my of Science 104:249-258. 

Evermann, B.W. & O.R Jenkins. 1888. Notes on 
Indiana fishes. Proceedings of the United States 
National Museum 11:43-57. 

Gerking, S.D. 1945. The distribution of the fishes 
of Indiana. Investigations of Indiana Lakes and 
Streams 3:1-137. 

Grossman, G.D., RB. Moyle & J.O. Whitaker, Jr. 
1982. Stochasticity in structural and functional 
characteristics of an Indiana stream fish assem- 
blage: A test of community theory. The Ameri- 
can Naturalist 120:423-454. 

Hubbs, C.L. & M.B. Trautman. 1937. A revision 
of the lamprey genus Ichthyomyzon. Miscella- 
neous Publications of the Museum of Zoology, 
University of Michigan, No. 35. 109 pp. 

Jenkins, O.P. 1887. List of fishes collected in Vigo 
County in 1885 and 1886. The Hoosier Naturalist 
2:93-96. 

Jordan, D. S. 1877. On the fishes of northern In- 
diana. Proceedings of the National Academy of 
Sciences of Philadelphia 29:42-82. 

Jordan, D.S. 1890. Report of explorations made 
during the summer and autumn of 1888, in the 
Allegheny region of Virginia, North Carolina, 
and Tennessee, and in western Indiana, with an 
account of the fishes found in each of the river 
basins of those regions. Bulletin of the United 
States Fish Commission 1888 8:97-173. 

Jordan, D.S. & B.W. Evermann. 1902. American 



Food and Game Fishes. A Popular Account of 
All the Species Found in America North of the 
Equator, with Keys for Ready Identification. Life 
Histories, and Methods of Capture. Doubleday, 
Doran and Co., Garden City, New York. 573 pp. 

Koel, T.M. & J.J. Peterka. 1948. Stream fishes of 
the Red River of the North Basin: United States: 
A comprehensive review. Canadian Field-Natu- 
ralist 12:631-646. 

Patton, T.M., F.J. Rahel & W.A. Hubert. 1998. Us- 
ing historical data to assess changes in Wyoming 
fish fauna. Conservation Biology 12:1 120-1 128. 

Pflieger, WL. 1997. The Fishes of Missouri. Mis- 
souri Department of Conservation. Jefferson 
City, Missouri. 372 pp. 

Retzer, M.E. 2005. Changes in the diversity of na- 
tive fishes in seven basins in Illinois, USA. The 
American Midland Naturalist 153:121-134. 

Simon, T.P. 2006. Biodiversity of fishes in the Wa- 
bash River: Status, indicators, and threats. Pro- 
ceedings of the Indiana Academy of Science 
115:136-148. 

Simon, T.P., R.L. Dufour & B.E. Fisher. 2005. 
Changes in the biological integrity of fish com- 
munities in the Patoka River drainage as a result 
of anthropogenic disturbance from 1888 to 2001. 
Pp. 383-398, In Historical Changes in Large 
River Fish Assemblages of the Americas. (J.N. 
Rinne, R.M. Hughes & B. Calamusso. eds.). 
American Fisheries Society Symposium 45. 

Simon, T.P., R. Robertson & C.C. Morris. 2004. 
Distribution of fish assemblages in the Valparai- 
so Chain of Lakes, Porter County, Indiana, with 
emphasis on lake condition assessment. Proceed- 
ings of the Indiana Academy of Science 1 13:33- 
41. 

Simon, T.P, J.O. Whitaker, Jr.. J.S. Castrale & S.A. 
Minton. 2002. Revised checklist of the verte- 
brates of Indiana. Proceedings of the Indiana 
Academy of Science 111:1 82-2 14. 

Smith, L.M. 1988. Stream fish of Vigo County, 
Indiana. Unpublished thesis. Indiana State Uni- 
versity, Terre Haute, Indiana. 32 pp. 

Whitaker, J.O., Jr. 1976. Fish community changes 
at one Vigo County. Indiana localit} oxer a 
twelve year period. Proceedings of the Indiana 
Academy of Science 85:191-20". 

Whitaker, J.O.. Jr. & D.C. Wallace. L973. Fishes 
of Vigo County, Indiana. Proceedings of the In- 
diana Academy of Science 51:450 — 164. 

Manuscript received 22 January 2007, revised 10 
October 2007. 



2007. Proceedings of the Indiana Academy of Science ( 1 16): 196—199 



FOOD HABITS OF MAMMALS DURING AN 

EMERGENCE OF 17-YEAR CICADAS 

(HEMIPTERA: CICADIDAE: MAGICICADA SPP.) 

Jonathan J. Storm 1 and John O. Whitaker, Jr.: Department of Ecology and 
Organismal Biology, Indiana State University, Terre Haute 47809 USA 

ABSTRACT. We examined the diet of nine species of mammals in Vigo and Clay Co. Indiana during 
an emergence of 17-year cicadas (Cicadidae; Magicicada spp.) in May and June of 2004. Small mammals 
were snap-trapped, and larger mammals were collected dead along roadways. During the emergence, 
cicadas constituted over 51% (by volume) of the diet of short-tailed shrews {Blarina brevicauda) and 
raccoons (Procyon lotor), while the opossum (Didelphis virginiana), white-footed mouse (Peromyscus 
leucopus) and eastern chipmunk {Tamias striatus) consumed periodical cicadas in lesser amounts. 

Keywords: diet, foraging, Indiana, Peromyscus, resource pulse 



Resource pulses {i.e., episodic periods of 
superabundance), such as the emergence of 
17-year cicadas (hereafter termed cicada), 
have recently gained attention as factors that 
can have widespread effects on ecological 
communities (Ostfeld & Keesing 2000; Yang 
2004). For instance, the emergence of cicadas 
can influence the population dynamics of avi- 
an (Anderson 1977; Strehl & White 1986; 
Koenig & Liebhold 2005) and mammalian 
species (Krohne et al. 1991). During spring of 
2004, Brood X of the 17-year cicada (Hemip- 
tera: Magicicada cassini, M. septendecim and 
M. septendecula) emerged across the middle- 
latitudes of the eastern United States. Cicadas 
are characterized by large size, slow flight, a 
general lack of anti-predator behavior (Karban 
1982; Steward et al. 1988) and high density 
(up to 3 million/ha; Dybas & Davis 1962). 
These characteristics should make cicadas 
easily obtainable prey and have led to the hy- 
pothesis that the synchronized emergence of 
cicadas evolved as a means of predator 
swamping (Karban 1982; Williams et al. 
1993). Cicadas are abundant along woodland 
edges and should be an easily obtainable food 
source for many terrestrial mammals. Given 
that cicadas are common both in trees and 



1 Send all correspondence to: Jonathan J. Storm, 
Department of Ecology & Organismal Biology, In- 
diana State University, Terre Haute, IN 47809 USA, 
Telephone: (812) 237-4424, Fax: (812) 237-2526, 
E-mail: jstorml @mymail. indstate.edu. 



along the ground, we expected a variety of 
insectivorous and omnivorous mammals to 
feed on cicadas. In addition, the high density 
of cicadas may allow foraging rodents to min- 
imize time spent foraging and therefore re- 
duce their risk of predation (Lima & Dill 
1990). 

A wide variety of birds feed on cicadas 
(Howard 1937; Leonard 1964; Karban 1982; 
Strehl & White 1986; Stephen et al. 1990), but 
there has been little investigation of mam- 
malian foraging on cicadas. Although there 
have been several studies on the diet of mam- 
mals in Indiana during non-cicada years (e.g. 
Whitaker 1966; Whitaker & Mumford 1972a, 
1972b), few studies have assessed the impact 
of cicada emergence on the diet of mammals. 
The few previous studies have found that ci- 
cadas are consumed by the white-footed 
mouse (Peromyscus leucopus), short-tailed 
shrew {Blarina brevicauda) and prairie vole 
(Microtus ochrogaster; Hahus & Smith 1990; 
Krohne et al. 1991). The objective of our 
study was to assess the diet of mammals in 
west-central Indiana during a cicada emer- 
gence. 

METHODS 

Small mammals were collected in a grassy 
field along the edge of a mature 16.2 ha wood- 
lot approximately 2 km NW of Brazil, Clay 
County, Indiana, from 21 May to 10 June 
2004. The woodlot is predominantly com- 
posed of tulip tree (Liriodendron tulipifera) 



196 



STORM & WHITAKER— MAMMALS AND PERIODICAL CICADAS 



197 



and a variety of oak {Quercus spp.) and hick- 
ory (Carya spp.) trees. During the emergence, 
there were numerous cicadas in trees, shrubs, 
and on the ground both within and along the 
woodlot. Small mammals were captured using 
snap-traps baited with peanut butter and rolled 
oats. Larger mammals (squirrels, opossum, 
raccoons, and eastern moles) were collected 
dead along roadways in Clay and Vigo coun- 
ties, Indiana, in areas where cicadas were 
abundant. 

Stomachs were removed and placed in a Pe- 
tri dish filled with ethanol. Food items were 
identified using a reference collection of ci- 
cada remains and keys in Whitaker (1988) and 
Arnett (2000). The percent volume of each 
prey type was visually estimated and the total 
percent volume for each prey type was cal- 
culated using the methods in Whitaker (1988). 

RESULTS 

Although all samples were small, six of the 
nine mammal species we collected had fed 
upon cicadas (Table 1). Cicadas constituted 
over 51% (by volume) of the diet of short- 
tailed shrews and raccoons {Procyon lotor) 
and were present in 66.7% and 77.3% of in- 
dividuals, respectively. Cicadas were less 
common in the diet of the opossum (Didelphis 
virginiana), white-footed mouse, eastern mole 
(Scalopus aquaticus) and eastern chipmunk 
{Tamias striatus). The prairie vole, woodland 
vole (Microtus pinetorum), and fox squirrel 
(Sciurus niger) did not contain any cicada re- 
mains in their stomachs. 

DISCUSSION 

Several species of mammals consumed ci- 
cadas, but none fed on cicadas to the exclu- 
sion of other prey. Many of the cicadas were 
unavailable to foraging non-arboreal mam- 
mals since cicada nymphs crawl toward the 
nearest tree trunk upon emergence (Dybas & 
Lloyd 1974) and adult cicadas spend most of 
their life in the tree canopy (Williams & Si- 
mon 1995). However, during the time of our 
trapping, many cicadas were readily available 
on the ground both within and along the edge 
of the woodlot. The lack of complete special- 
ization on cicadas may relate to the availabil- 
ity of non-cicada prey or the importance of 
complementary resources in the diet. 

The northern short-tailed shrew feeds pri- 
marily on invertebrates, especially earth- 



worms and gastropods (Whitaker & Mumford 
1972a; Mumford & Whitaker 1982). Despite 
the fact that northern short-tailed shrews often 
forage below ground (George et al. 1986). ci- 
cadas were a common component of their 
diet. Earthworms were also common in the 
diet during the emergence, but cicadas were 
the most prevalent food. Similar to our results, 
Hahus & Smith (1990) and Krohne et al. 
(1991) found that northern short-tailed shrews 
feed on cicadas. 

Although eastern moles feed primarih on 
invertebrates (Whitaker & Schmeltz 1974: 
Mumford & Whitaker 1982), cicadas were not 
common in their diet, likely due to the fos- 
sorial habits of moles. Although we found no 
evidence of cicada nymphs or adults in the 
stomachs of moles, it is possible that moles 
fed on cicada nymphs prior to the nymphs 
emergence from the ground. Eastern moles 
feed on beetle larvae (Hartman et al. 2000) 
and it has been hypothesized that subterranean 
foraging by moles may influence cicada pop- 
ulations (Lloyd & Dybas 1966). Similar to 
Hahus & Smith (1990). we found that white- 
footed mice fed upon cicadas. Given the om- 
nivorous foraging habits of white-footed mice 
and their preference for forested habitats (Barry 
& Francq 1980), the presence of cicadas in 
the diet was expected. 

Raccoons are opportunistic foragers. 
(Mumford & Whitaker 1982) so the large 
number of cicadas in their diet is not surpris- 
ing. Eastern chipmunks are predominantly 
herbivorous and granivorous (Mumford ^ 
Whitaker 1982); however, we expected chip- 
munks to forage on cicadas. The scarcity of 
cicada remains in our study likely reflects the 
small number of individuals examined, rather 
than an avoidance of cicadas by foraging chip- 
munks. Fox squirrels generally feed on mast 
(Mumford & Whitaker 1982). and none of the 
individuals we examined had eaten cicadas: 
however, both gray and fox squirrels were ob- 
served eating cicadas (J. Whitaker. pers. obs. I. 
Prairie and woodland voles are primarih her- 
bivorous (Mumford & Whitaker 1 L )S2> and 
did not feed on cicadas during this study. Ha- 
hus and Smith (1990) reported that prairie 
voles fed upon cicadas, but the\ constituted 
less than 29c of the diet in their stud) . 

ACKNOWLEDGMENTS 
We thank J. Boyles. V. Brack. D. Sparks 
and M. Storm for comments on the manu- 



198 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



0) T3 

a o 

J2 >^ 

g u 

C 

i & 

i § 

.2 O 

■9 o 

o S 



I -s 



»I 



6<a B<a 








S -2 


sS ^ 


£ S3 


&5 


3 k 


V ^ 


a oc 




o e 


CO 


B9 S2 






ta 


*3 Q 


? s 




s 


§ v. 


& S 




O 


k "~~ 


0, 


&5 






£ &< 


§ 8 


P s 


b.s 


ft, 




»5 S 


3 2S 


sa 


^j ^j 


S -5 


v. 


*» £ 


S 65 


g | 


£ 5 


*-5 





6, Q 




-S ^ 








<5j -a 


^3 04 


~ k 


Q "2 






2 1 




•5 a 


C f o 


-3 -2 


0Q £> 


■Q 


u 


a 


>^ 


H 


>> 


<u 




CU 



000 


vO O O 


'oqioqqqrnqoo 


00 p p p 00 


odd 


—'do 


^d^ddd-drn 


06 © © in en © © 

OO ^O (N 



0000 
dodo 



000 
odd 



om h —00 
d d \6 r-J d 



OOOOOOOOO 

ddddddddd 



'— > ooppprnopp 



0000000 



o o o in in o o 

O CS f» 



O O O O O O l> 

»-5 © d d d i-h 



ON © O l> 


CN 00 © 


© © in ^o 

00 —1 


O^ O O 

in 



^ppppppr^r-; (Npppr^pin 

00©©©©©©^— On © © © ^O © <N 

in m rH rt 



© © © © 


© © © 


©ppppppp© 


© r-_ © © © ro © 


>n © © © 

(N 


in © d 

(N 


CN (N 


in vd in d © en © 
r-- cn cn (N 



o © © © 

© © © © 



© © © © 
© © © © 



© © © © 
© in © © 



© © © 
© © © 



© © © 
© © © 



©©o©©©©©© 
ddddddddd 



©©©©©©©©© 
ddddddddd 



© © © o © © © 
d> d> d> d> d> d> d> 

© © 



©©©©©©© 



©©©©©©© 
© © 



©©© ©©©©©©©©© 



© © 



©©©©©©© 
^d><Dd>d>d>d>d>ui © © © © © © © 



©©©©©©© 
© d © © © © © 



© © © © 


© © © 


©ppppppp© 


© d -3 © 
"t 


on d © 
in 


o\dd(N^dddrn 
in m 



a. 
S 















CJ 






















<D 






















<*> 






















C 




































g 




^ 


~4 S 




ntified 

idae 

ptera 




-a 


> 

H 




Ed 






C/2 


O C C/2 




r3 


'0 


~^ 


O "f! 


"3 
E 

'8 


■p 


_3 




rthrop 
Arach 
Crayfi 


2 



u 


Unide 
Carab 
Coleo 
Diptei 


1 

Eh 

O 


O 
U 




< 








< 


c 













pq b 



t 

p 03 3 « 
S S 00 



STORM & WHITAKER— MAMMALS AND PERIODICAL CICADAS 



199 



script. This project was approved by the In- 
diana State University Institutional Animal 
Care and Use Committee. 

LITERATURE CITED 

Anderson, T.R. 1977. Reproductive responses of 
sparrows to a superabundant food supply. Con- 
dor 79:205-208. 

Arnett, R.H., Jr. 2000. American Insects: A Hand- 
book of the Insects of America North of Mexico. 
2 nd ed. CRC Press, New York. 

Barry, R.E. & E.N. Francq. 1980. Orientation to 
landmarks within the preferred habitat by Pero- 
myscus leucopus. Journal of Mammalogy 61: 
292-303. 

Dybas, H.S. & D.D. Davis. 1962. A population 
census of seventeen-year periodical cicadas (Ho- 
moptera: Cicadidae: Magicicada). Ecology 43: 
432-444. 

Dybas, H.S. & M. Lloyd. 1974. The habitats of 
17-year periodical cicadas (Homoptera: Cicadi- 
dae: Magicicada spp). Ecological Monographs 
44:279-324. 

George, S.B., J.R. Choate & H.H. Genoways. 1986. 
Blarina brevicauda. Mammalian Species 261: 
1-9. 

Hahus, S.C. & K.G. Smith. 1990. Food habits of 
Blarina, Peromyscus, and Microtus in relation to 
an emergence of periodical cicadas {Magicica- 
da). Journal of Mammalogy 71:249-252. 

Hartman, G.D., J.O. Whitaker & J.R. Munsee. 
2000. Diet of the mole Scalopus aquaticus from 
the Coastal Plain Region of South Carolina. 
American Midland Naturalist 144:342-351. 

Howard, W.J. 1937. Bird behavior as a result of 
emergence of seventeen-year locusts. Wilson 
Bulletin 49:43-44. 

Karban, R. 1982. Increased reproductive success at 
high densities and predator satiation for period- 
ical cicadas. Ecology 63:321-328. 

Koenig, WD. & A.M. Liebhold. 2005. Effects of 
periodical cicada emergences on abundance and 
synchrony of avian populations. Ecology 86: 
1873-1882. 

Krohne, D.T., T.J. Couillard & J.C. Riddle. 1991. 
Population responses of Peromyscus leucopus 
and Blarina brevicauda to emergence of peri- 
odical cicada. American Midland Naturalist 126: 
317-321. 

Leonard, D.E. 1964. Biology and ecology of Mag- 
icicada septendecim (L.) (Hemiptera: Cicadidae). 
Journal of the New York Entomological Society 
72:19-23. 

Lima, S.L. & L.M. Dill. 1990. Behavioral deci- 
sions made under the risk of predation: a review 



and prospectus. Canadian Journal of Zoology 68: 
619-640. 

Lloyd M. & H.S. Dybas. 1966. The periodical ci- 
cada problem. I. Population Ecology. Evolution 
20:133-149. 

Mumford R.E. & J.O. Whitaker, Jr. 1982. Mam- 
mals of Indiana. Indiana University Press. Bloo- 
mington. 

Ostfeld, R.S. & F. Keesing. 2000. Pulsed resources 
and community dynamics of consumers in ter- 
restrial ecosystems. Trends in Ecology and Evo- 
lution 15:232-237. 

Stephen, F.M., G.W. Wallis & K.G. Smith. 1990. 
Bird predation on periodical cicadas in Ozark 
forests: Ecological release for other canop} ar- 
thropods? Studies in Avian Biology 13:369-374. 

Steward, V.B., K.G. Smith & EM. Stephen. 1988. 
Red-winged blackbird predation on periodical ci- 
cadas (Cicadidae, Magicicada spp.) — Bird be- 
havior and cicada responses. Oecologia 76:348- 
352. 

Strehl, C.E. & J.A. White. 1986. Effects of super- 
abundant food on breeding success and beha\ ior 
of the red-winged blackbird. Oecologia 70:178- 
186. 

Whitaker, J.O. 1966. Food of Mus musculus, Per- 
omyscus maniculatus bairdi and Peromyscus leu- 
copus in Vigo County, Indiana. Journal of Mam- 
malogy 47:473-486. 

Whitaker, J.O. 1988. Food habits analysis of in- 
sectivorous bats. Pp. 171-189. In Ecological and 
Behavioral Methods for the Study of Bats (T. 
Kunz, ed.). Smithsonian Institution Press. Wash- 
ington, D.C. 

Whitaker, J.O. & R.E. Mumford. 1972a. Food and 
ectoparasites of Indiana shrews. Journal of Mam- 
malogy 53:329-335. 

Whitaker, J.O. & R.E. Mumford. 1972b. Ecologi- 
cal studies on Reithrodontomys megalotis in In- 
diana. Journal of Mammalogy 53:850-860. 

Whitaker, J.O. & L.L. Schmeltz. 1974. Food and 
external parasites of the eastern mole. Scalopus 
aquaticus, from Indiana. Proceedings of the In- 
diana Academy of Science 83:478 — 181. 

Williams, K.S.. K.G. Smith & FM. Stephen. 1993. 
Emergence of 13-year periodical cicadas (Cicad- 
idae: Magicicada): Phenology, mortality, and 
predator satiation. Ecology 74:1 143-1 152. 

Williams, K.S. & C. Simon. 1995. The ecology, 
evolution, and behavior of periodical cicadas. 
Annual Review of Entomolog\ 40:269-295. 

Yang, L.H. 2004. cicadas as resource pulses in 
North American forests. Science 306: 1565-156". 

Manuscript received 14 June 2007, revised 10 Au- 
gust 2007. 



2007. Proceedings of the Indiana Academy of Science (1 16):200-201 



SHORT COMMUNICATION 

AEDES AEGYPTI (DIPTERA: CULICIDAE) IN 
ST. JOSEPH COUNTY, INDIANA 

Catherine L.E. Young: Center for Global Health and Infectious Disease, Department 
of Biological Sciences, University of Notre Dame, Notre Dame, Indiana 46556 USA 

Robert E. Sheffer: St. Joseph County Health Department, South Bend, Indiana 46601 
USA 

Frank H. Collins: Center for Global Health and Infectious Disease, Department of 
Biological Sciences, University of Notre Dame, Notre Dame, Indiana 46556 USA 

ABSTRACT. Two adult female specimens of Aedes aegypti, the Yellow Fever mosquito, were collected 
on separate occasions in September of 2006 at a collection site in St. Joseph County, Indiana. The eggs 
of this species are not believed to be capable of surviving the winter in this area and were most likely 
introduced with a shipment of used automobile tires to a tire recycling facility located near the collection 
site. This finding emphasizes the need for caution and the potential for transport of exotic species in cargo 
shipments of this kind, and also represents the northernmost report of this species in the midwestern 
United States to date. 

Keywords: Mosquito, Indiana, Aedes aegypti, introduced species 



The mosquito Aedes (Stegomyia) aegypti 
(Linnaeus) (Diptera: Culicidae) may be found 
throughout the tropics and subtropics and is a 
well-known pest and the primary vector of 
several serious arboviruses, including Yellow 
fever, Dengue, and Chikungunya (Womack 
1993). Wild specimens have also been found 
infected with the West Nile virus in the United 
States (CDC 2005). The usual range of this 
species extends only as far northward as the 
southern border of Tennessee (Darsie & Ward 
2004), as they are unable to survive the pro- 
longed freezing temperatures of more northern 
winters (Hawley et al. 1989; Womack 1993). 
However, scattered reports of A. aegypti from 
several counties in southern Indiana (Chris- 
tensen & Harmston 1944, Hart 1944; Siverly 
1972) indicate that accidental introductions of 
or invasions by this species are not as uncom- 
mon as might be expected, while the collec- 
tion of larvae from standing water in Clark 
County (Christensen & Harmston 1944) sug- 
gests that imported specimens may be capable 
of temporary breeding that could produce sig- 
nificant numbers of offspring over the course 
of a summer. More disturbingly, Christensen 
& Harmston (1944) also report the collection 



of several blood-seeking adult A. aegypti col- 
lected inside residences in Terre Haute, Indi- 
ana in January of 1943. Larvae and pupae 
were discovered in a bowl of water containing 
an aquatic houseplant, and more continued to 
appear and emerge over the following weeks. 
It is not, therefore, impossible that this species 
could survive even the coldest winter if suit- 
able shelter is available. 

METHODS 

Both specimens were captured in dry ice- 
baited CDC-style light traps (American Bio- 
physics Corp.) used for mosquito and West 
Nile virus surveillance by the St. Joseph 
County Health Department and the University 
of Notre Dame. One gravid trap (John W 
Hock Co.), baited with an infusion of grass 
clippings, and one dry ice-baited CDC-style 
light trap were set at each of nine sites two 
nights per week beginning on 3 May 2006. 
Three additional trapping sites were subse- 
quently added in response to collections of 
dead birds infected with the West Nile virus, 
for a total of 12 collection sites distributed 
throughout St. Joseph County. Specimens 



200 



YOUNG ET AL.—AEDES AEGYPTI IN INDIANA 



201 



were killed by freezing and identified by mor- 
phological characters (Siverly 1972). 

The site from which A. aegypti was sub- 
sequently collected, located on private prop- 
erty in North Liberty, Indiana, was one of 
those added to the surveillance program in 
mid-season. A dry ice-baited light trap was set 
at this site two nights per week beginning on 
17 August 2006 and ending on 26 October 
2006. No gravid trap was available for this 
site. Single adult female specimens of A. ae- 
gypti were collected from the North Liberty 
site on 15 September and 22 September 2006. 

DISCUSSION 

While the lack of freeze tolerance in A. ae- 
gypti makes permanent breeding in northern 
Indiana unlikely, this event does highlight the 
need for caution regarding the transportation 
of tires and other containers in which eggs 
may be laid. It should also serve as a reminder 
to scientists and public health workers in 
northern states of the need for awareness of 



this and other exotic species which may oc- 
casionally form a significant part of our local 
fauna, however temporarily. 

LITERATURE CITED 

CDC. 2005. Mosquito species producing WNV 
positives by year. Centers for Disease Control 
and Prevention. 

Christensen, G.R. & EC. Harmston. 1944. A pre- 
liminary list of the mosquitoes of Indiana. Jour- 
nal of Economic Entomology 37:1 10-111. 

Darsie, R.F. & R.A. Ward. 2004. Identification and 
Geographical Distribution of the Mosquitoes of 
North America, North of Mexico. Universit) 
Press of Florida. Gainesville, Florida. 

Hart, J.W. 1944. A preliminary list of the mosqui- 
toes of Indiana. American Midland Naturalist 3 1 : 
414-416. 

Siverly, R.E. 1972. Mosquitoes of Indiana. Indiana 
State Board of Health, Indianapolis. Indiana. 

Womack, M. 1993. The yellow fever mosquito. 
Aedes aegypti. Wing Beats 5:4. 

Manuscript received 1 February 2007, revised 23 
May 2007. 



2007. Proceedings of the Indiana Academy of Science (1 16):202-204 



INDEX 

to the 

Proceedings of the Indiana Academy of Science 

Volume 116 

2007 



Abandoned mine, bats 58 

Abrams combat tank 127 

Activity patterns, bats 66 

Aedes aegypti 200 

Africa 90 

Anabaena 173 

Anderson, Alan B 126, 139 

Anhydrite 1 

Artibeus jamaicensis 66 

Australia 90 

Aves (Estrildidae) 90 

Ayers, Paul D 126, 139 

Badger, Kemuel S 11 

Baltzer, Holly 108 

Bark-boring beetles 46 

Bats and insecticides 50 

Beech-maple forest 16 

Behavior, bats 66 

Big Blue Hole Quarry 2 

Bird's-eye 1 

Brookville, Indiana 13 

Brown County, Indiana 162 

Cadmium and soil 148 

Calcite and pyrite 4 

Camp Atterbury, Indiana 126, 139 

Carbonates 1 

Cass County Quarry 1 

Cattail marsh 17 

Ceramic value 117 

Chance, Janis L 11 

Cholinesterase inhibition 48 

Cicadas 1 96 

Circadian activity, bats 66 

Clay County, Indiana 196 

Click, Lindy 11 

Collins, Frank H 200 

Common pine shoot beetle 46 

Conner house 1 20 

Copepods 179 

Copperhead "Cave" 59 

County records, vascular plants 11 

Crane 84 

Crawford Upland 166 

Culicidae 200 

Cyanobacteria 173 

Darke County, Ohio 11 

Deam, Charles 14 

Diet, mammals 196 



Diptera 200 

Doffin, Jason 1 

Duchamp, Joseph 184 

Eagle Creek Reservoir 173 

Eastern pipistrelle 58 

Echolocation, bats 67 

Ecological classification 158 

Economic class 117 

Eidels, Ronny R 48 

Emerald ash borer 42 

Eptesicus fuscus 58 

Estrildid finches 90 

Exotic vegetation 18 

Ferns 37 

Field crop insects 43 

Financial statement 115 

Fishes 184 

Floodplain woodland 17 

Flora, Indiana 11 

Floristic Quality Index 11 

Food habits, mammals 196 

Foraging, mammals 196 

Forest ecosystems 158 

FQI 11 

Franklin County, Indiana 13 

Fruit insects 47 

Furbearer, Indiana 71 

Gibb, Timothy J 42 

GIS 158 

Glendora, Indiana 86 

Godeke site 120 

Grain insect pests 48 

Granulate ambrosia beetle 45 

Grazing preference 173 

Great Miami River 11 

Griswold, Matthew 184 

Guano, insecticides 48 

Gypsy moth 44 

Hayes Arboretum 11 

Hee, C.K 148 

Hemiptera 196 

Hess, Benjamin 108 

Hewitt, Thomas R 71 

Hibernation, bats 58 

Historic ceramic vessels 117 

Honey bees 43 

Hoosier National Forest 158 

Household insects 47 



202 



INDEX 



203 



Howard, Heidi 126 

Hudson, Cassie M 71 

Impact assessment 126, 139 

Indianapolis 173 

Indo-Pacific islands 90 

Insect activity 42 

Insecticide residues 50 

Intertidal limestone 1 

Introduced species 200 

Jacob, J.R 148 

Jamaican fruit bats 66 

Johnson, Scott 71, 84 

Kilibarda, Zoran 1 

Kokomo limestone 1 

LaFontaine site 120 

Land Type Association 158 

Land type mapping 158 

Landscape insects 48 

Landscape analysis 158 

Late Silurian limestone 1 

Lead and soil 148 

LeGrange County, Indiana 44 

Logansport, Indiana 2 

Lonchura 98 

Lontra canadensis 71 

Lower Prairie Creek 184 

LTA delineation 161 

M1009 utility cargo vehicle 129 

M1025 utility vehicle 141 

M35A3 cargo truck 129 

M548A cargo carrier 141 

M813 cargo truck 141 

M88 tank recovery vehicle 129 

M998 utility vehicle 141 

Magicicada 196 

Mammals, terrestrial 84 

Marion County, Indiana 117 

Martin County, Indiana 86 

Metal mobility, soil 150 

Miami River Basin 13 

Microcystis 173 

Mines, bats 58 

Minutes of the meetings 110 

Mixed hardwood forest 17 

Morgan Hill Bed 3 

Mosquito 200 

Mudcracks 1 

Muncie, Indiana 149 

Municipal solid waste 148 

Myotis lucifugus 58 

Myotis sodalis 52 

Myotis septentrional is 52 

Naval Support Activity 84 

Naval Surface Warfare Center 84 

Newlin, Kenneth D 126 

North Liberty, Indiana 201 

North American river otter 71 

NSA Crane 86 

Ochsner, William R 1 39 



Off-road impact 126, 139 

Ohio River 13 

Old- World bollworm 43 

Ordivician shale 14 

Organophosphate insecticide 48 

Ornamentals, insect pests 48 

Otter mortality 78 

Otters, distribution 71 

Oxendine Bayou 192 

Parker, George R 158 

Pascual, Denise Lani 173 

Passer domesticus 98 

Perimyotis subflavus 58 

Personnel carrier 127 

Physical inhibition 173 

Phytostabilization 148 

Phytotreatment of soil 148 

Pichtel, J 148 

Pierce, Christopher M.F. 42 

Plant communities 11 

Poephila guttata 98 

Point-centered-quarter analysis 108 

Pond, vegetation 17 

Prairie Creek 1 84 

Public health pests 48 

Randolph County, Indiana 13. 108 

Reddick site 1 17 

Rees, Douglas 184 

Reidy, Chris R 11 

Re-introduction, otters 71 

Resource pulse 196 

Re-vegetation 14S 

Richardville site 120 

Richmond, Indiana 11 

River otters 71 

Roost temperature, bats 62 

Rothrock, Paul E 11 

Rotifers 179 

Ruch, Donald G 11, 108 

Seeps vegetation 17 

Sewage sludge 148 

Sheffer, Robert E 200 

Silurian Kokomo limestone 1 

Small mammals 84 

Socioeconomic status 1 17 

Soil 140 

Soil metal fractionation 150 

South Logansport Quarry 1 

Southern Asia 90 

Sparks, Daniel 48 

St. Joseph County. Indiana 200 

Stanley W. Hayes 11 

Steep slope woods 1 6 

Stored food, insects 48 

Storm. Jonathan J 196 

Stryker armored combat vehicle 140 

Successional woods 11 

Sullivan County. Indiana 86 

Sullivan. Patricia 139 



204 



PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE 



Superfund soil 148 

Supratidal limestone 1 

Thompson, Jeff S 71 

Tipton Till Plain 13 

Torke, Byron G 1 1, 108 

Trierweiler, Annette 174 

Tung, Jennifer J 66 

Turfgrass, insects 48 

Upland forest 108 

Upland woods 11 

Urly, Eileen Grossman 11 

VanderVeen, James M 117 

Vascular flora 11 

Vegetables, insects 47 

Vegetation recovery 126 

Vegetation impacts, vehicle 126, 139 

Vegetational communities 11 

Vehicle impacts 126, 139 

Veilleux, Sherry L 184 

Vermillion County, Indiana 60 



Vidua fischeri 98 

Vigo County, Indiana 184, 196 

Vincent, Elizabeth A. 58 

Walker, Heather D 71 

Walters, Brianne L 84 

Waltz, Robert D 42 

Waltz, Rod 11 

Ware ratios 123 

Wax-billed finches 90 

Wayne County, Indiana 11 

Webster, J. Dan 90 

Whitaker, John 48, 58, 84, 184, 196 

Whitewater Valley Drainage 11 

Whitewater River Basin 13 

William Conner house 120 

Wood-boring beetles 46 

Young, Catherine L.E 200 

Yuhas Woods 108 

Zhalnin, Andriy V 158 

Zinc and soil 148 

Zooplankton 173 



INSTRUCTIONS TO AUTHORS 

(revised May 2007) 



General information. — Manuscripts and all correspon- 
dence should be sent to the Editor. To be eligible for pub- 
lication in the Proceedings, at least one author must be a 
member of the Academy. There are two categories of papers: 
regular feature articles (usually longer than three pages) and 
short communications (three pages or less). Submission of a 
manuscript to the Editor implies that it has not been sent to 
another publication, nor has it been published elsewhere. If it 
is accepted for publication in the Proceedings, it may not be 
republished without the written consent of the Academy. 

All manuscripts are evaluated by referees. Submit three 
copies of the manuscript. Figures in manuscripts for review 
may be submitted as photocopies, and they should be approxi- 
mately the size that they will be printed in the Proceedings. 

After the manuscript has been accepted, the author will 
be asked to submit the manuscript as a paper copy and on a 
computer disc in a widely-used word processing program. 
Indicate clearly on the computer disc the word processing 
program and the type of computer (Mac or PC). 

Galley proofs will be sent to the primary author for ap- 
proval and correction. 

Voucher specimens and permits: Voucher specimens of 
species used in scientific research should be deposited in a 
recognized scientific institution. Authors must indicate that 
the collection of specimens, especially of threatened or en- 
dangered species, was authorized by the appropriate govern- 
mental organizations. All type specimens must be deposited 
in a recognized museum collection. 

Style. — Do not use right or full justification. Also, use a 
non-proportional font (e.g, Courier, but not Times Roman). 

Title page. — The title page should include ( 1 ) the title in 
capital letters, (2) each author's name and address, (3) the 
running head (see below), and (4) the complete name, ad- 
dress, and telephone number of the author with whom proofs 
and correspondence should be exchanged, a FAX number 
and electronic mail address if available. 

Abstract. — All manuscripts have an abstract, which 
should summarize the significant facts in the manuscript. 
The "ABSTRACT" heading in capital letters should be 
placed at the beginning of the first paragraph set off by a 
period. Use complete sentences, and limit the abstract to one 
paragraph and 250 words. 

Keywords. — Give 3-5 appropriate keywords following 
the abstract. 

Text. — Double-space text, tables, legends, etc. through- 
out. Three categories of headings are used. The first category 
(METHODS, RESULTS, etc.) is typed in capitals, centered, 
and on a separate line. The second (lower) category of head- 
ing, in bold type, begins a paragraph with an indent and is 
separated from the text by a period and a dash. (This para- 
graph begins with an example of this heading.) The third 
heading category may or may not begin a paragraph, is itali- 
cized and followed by a colon. (The paragraph below is an 
example.) Use only the metric system unless quoting text or 
referencing collection data. 

Citation of references in the text: Cite only papers already 
published or in press. Include within parentheses the sur- 
name of the author followed by the date of publication. A 
comma separates multiple citations by the same author(s) 
and a semicolon separates citations by different authors, e.s*., 
(Smith 1990), (Jones 1988; Smith 1993), (Smith 1986, 1987; 
Smith & Jones 1989; Jones et al. 1990). 

Literature cited section. — Use the following style, and 
include the full unabbreviated journal title. Repeat the name 
for multiple references by the same author. Note that book 
titles have the first letter of each word capitalized. 



Walter, J. & B. Hallet. 1979. Geometry of former subgla- 
cial water channels and cavities. Journal of Glaciology 
23:335-346. 
Walter, J. 1992. The significance and complexity of commu- 
nication in moths. Pp. 25-66, In Insect Communications: 
Mechanisms and Ecological Significance. (P.N. Work & 
J.S. Rivers, eds.). Princeton University Press. Princeton, 
New Jersey. 

Footnotes. — Footnotes are permitted only on the first 
printed page to indicate current address or other informa- 
tion concerning the author. These are placed together on a 
separate page at the end of the manuscript. Tables and figures 
may not have footnotes. 

Running head. — The authors surname(s) and an abbre- 
viated title should be typed all in capital letters and must not 
exceed 60 characters and spaces. The running head should 
be placed near the top of the title page. 

Tables. — These should be typed double-spaced, one table 
to a page and numbered consecutively. Most tables contain 
only three horizontal lines (see recent issues for examples). 
Do not use vertical lines or shading. Include all pertinent 
information in the table legend (no footnotes). 

Illustrations. — All art work must be camera-reads 
(mounted and labeled) for reproduction. Figures should be 
arranged so that they fit (vertically and horizontal lv) the 
printed journal page, either one column or two columns. 
with a minimum of wasted space. When reductions are to 
be made by the printer, pay particular attention to width of 
lines and size of lettering in line drawings. Multiple photos 
assembled into a single plate should be mounted with onl\ 
a minimum of space separating them. In the case of mul- 
tiple illustrations mounted together, each illustration must be 
numbered Fig. 1, Fig. 2, etc.) rather than given letter (A.B. 
etc.) designations. The name(s) of author(s). and an indica- 
tion of top edge, and whether the illustration should be one 
or two columns wide, should be written on the back of the il- 
lustration. The overall dimensions should be no more than 1 1 
inches (28 cm) x 14 inches (36 cm). Larger drawings present 
greater difficulty in shipping and greater risk of damage for 
which the Proceedings assumes no responsibility. In manu- 
scripts for review, photocopies are acceptable, and should be 
reduced to the exact measurements that the author prefers m 
the final publication. Make notations in the text margins to 
indicate the preferred position of illustrations in the printed 
text. Color plates can be printed: but the author must assume 
the full cost, currently about $600 per color plate. 

Legends for illustrations should be placed together on the 
same page(s) and separate from the illustrations. Each plate 
must have only one legend as indicated below 

Figures 1—4. — Right chelicerae of species of Centruroides 
from Timbuktu. 1. Dorsal view; 2. Prolateral \ icw of move- 
able finger: 3. Centruroides holotype male: 4. Centruroides 
male. Scale = 1.0 mm. 

Page charges and reprints. There are no page charges. 
Corrections in proof pages must be restricted to printer's er- 
rors only; all other alterations ma\ be charged to the author 
(currently S3 per line). Reprints are available onl\ from Al- 
len Press and should be ordered (with payment) when the 
author receives the proof pages. 



CONTENTS 

Proceedings of the Indiana Academy of Science 
Volume 116 Number 2 2007 

Anthropology 

People, Pots, and Prosperity: The Ceramic Value Index and an 

Assumption of Economic Class by James M. VanderVeen 117 

Environment 

Vehicle Impacts on Vegetation Cover at Camp Atterbury, Indiana: Part 

1. Initial Impacts and Vegetation Recovery by Alan B. Anderson, 

Paul D. Ayers, Heidi Howard and Kenneth D. Newlin 126 

Vehicle Impacts on Vegetation Cover at Camp Atterbury, Indiana: Part 

2. Predicting Impacts of Untested Vehicles by Alan B. Anderson, 

Paul D. Ayers, Patricia Sullivan and William R. Ochsner 139 

Amendments for Field-Scale Phytotreatment of Pb, Cd and Zn from an 

Indiana Superfund Soil by J.R. Jacob, C.K. Hee and J. Pichtel 148 

Botany 

Land Type Association Delineation and Spatial Analysis for the Hoosier 
National Forest in Southern Indiana by Andriy V Zhalnin and 
George R. Parker 158 

Zoology 

Zooplankton Growth Responses to the Cyanobacteria Microcystis 
and Anabaena in Eagle Creek Reservoir in Indiana by Annette 
Trierweiler and Denise Lani Pascual 1 73 

Fishes of the Lower Prairie Creek Area, Vigo County, Indiana by John 
O. Whitaker, Jr., Sherry Veilleux, Joseph Duchamp, Matthew Griswold and 
Douglas Rees 184 

Food Habits of Mammals During an Emergence of 17-Year Cicadas 
(Hemiptera: Cicadidae: Magicicada spp.) by Jonathan J. Storm and 
John O. Whitaker, Jr. 196 

Short Communication 

Aedes aegypti (Diptera: Culicidae) in St. Joseph County, Indiana by 

Catherine L.E. Young, Robert E. Sheffer and Frank H. Collins 200 

Index 
Index to Volume 116 (2007) 201