subsample was spread evenly in a tray approx- 
imately 5 by 30 inches. The tray had a series of 
25 notches on each side that served as stops 
when the tray was passed under a binocular 
microscope. One eyepiece of the scope had a 
crosshair. When the tray and material was 
passed under the scope, the species nearest the 
crosshair was considered a hit. Fifty hits were 
recorded from each subsample. Preliminary 
samples of known composition indicated that a 
total sample of 400 hits was necessary to ob- 
tain sample means for the major species 
within 10 percent of the population mean 95 
percent of the time. 
Other work with the microscope point tech- 
nique for analysis of esophageal and ruminal 
fistula samples from cattle and sheep has been 
reviewed by Van Dyne and Torell (1964). Per- 
centage point data for species and species 
groups were converted to percentage weight by 
use of equations developed by Heady and Van 
Dyne (1964). Regressions based on a similar 
procedure were used by Galt et al. (1966). 
The adaptation of the point-analysis method 
to several combinations of artificially con- 
structed populations of plant fragments and on 
actual rumen contents of white-tailed deer 
(Odocoileus virginianus) was reported by 
Chamrad and Box (1964). Their sample device 
consisted of a frame containing five hatpins 
placed at a 45-degree angle through a wooden 
bar. The 45-degree angle of the pins facilitated 
microscopic observation of pin hits. Each pin 
was dropped into the artificial population or 
rumen material spread evenly over the bottom 
of an enameled lab tray, and the first hit of 
each pin on a plant fragment was recorded. 
One hundred pin drops per sample were used. 
There were no significant differences between 
the abilities of two investigators to estimate 
the volumetric composition of the samples, nor 
was there a difference in the ability of the indi- 
vidual investigator to repeat his estimates. The 
point analyzer gave reliable results if (1) the 
rumen sample was adequately mixed and (2) 
there were no large items with unusual surface 
texture in the mixture. 
Harker et al. (1964) evaluated the micro- 
scopic point method. Samples of Brachiaria de- 
cumbens (a coarse perennial common in East 
Africa) and sweet potato vice (Ipomoea bata- 
tas) were obtained from six esophageal-fistu- 
lated Zebu cattle. The forage was rinsed with 2 
percent acetic acid to remove saliva, and then 
mixed by wet weight into desired portions. The 
method gave a satisfactory estimate of species 
composition on a percent dry-weight basis. 
Four hundred points estimated percent dry 
matter at the 90-percent confidence limit to 
within 20 percent of the mean if the weight is 
16 to 30 percent, to within 10 percent of the 
mean if the weight is between 30 and 50 per- 
cent and to within 5 percent of the mean if the 
weight is 50 to 95 percent. 
Van Dyne and Heady (1965) found variabil- 
ity among individual animals in the composi- 
tion of diet. From their results based on the 
microscopic point technique, they calculated 
sample size required to estimate dietary compo- 
sition within 10 percent of the mean with 90 
percent confidence. In general, many animals 
would be required to sample the diet for most 
botanical constituents. More sheep than cattle 
usually would be required for a given constitu- 
ent, and more animals of either class generally 
would be required in early summer, when there 
is a high availability of herbage, than in mid- 
dle or late summer, when less herbage is avail- 
able. Fewer animals are needed with more in- 
clusive plant groups. Even for such broad 
plant categories as grasses, forbs, and shrubs, 
at least 10 animals would be required for ei- 
ther cattle or sheep in early summer. 
Adams (1957) and Adams et al. (1962) pro- 
posed a method of estimating the weight of food 
eaten by snowshoe hares; using this method, 
the number of recognition items in the feces on 
an area is estimated. This number is then con- 
verted to weight eaten by calibration obtained 
by feeding penned hares known weights of 
foods and then counting the recognition items. 
The fecal pellets were teased apart in water 
and detergent, suspended in water over a grid, 
and scanned under a dissecting microscope at 
7X magnification. 
Low-power magnification has also been used 
in fecal examination to determine the presence 
of tracers for food items. Grain and fresh veg- 
etable baits have been coated with aluminum 
tracer and placed in the field. Fecal pellets 
have been examined for the nondigestible alu- 
minum flake to determine the distance traveled 
from bait sites, and the percent of the rodent 
population using bait. 
Identification Under High-Power Magnification 
Stomach analysis of small rodents that grind 
their food into very fine particles made it man- 
datory for microtechniques and _ histology 
methods to be used. These same techniques, 
with some modifications, have been used in 
food-habits studies of big game and domestic 
livestock. With big game animals, rumen con- 
tents are usually taken from dead animals. 
Samples from domestic livestock have been col- 
lected from either esophageal- (Torell 1954) or 
rumen-fistulated (Shumway et al. 1963; Male- 
chek 1966) animals. 
Botanical composition of stomach contents 
of pocket gophers was determined by washing 
stomach contents in cold water, screening on 
151 
