Vol. 23, No. 2 
Page 2 
made to evaluate site selection by squirrels within heterogeneous environments. 
Two study areas, Danville A and B, were grided into 76 and 66, 0.36-ha plots, 
respectively. One trap was placed randomly in each plot during 2, 10-day 
trapping sessions each year from 1971 to 1978. Plot use by squirrels was based 
on number of trap and nest-box captures and was related to 66 environmental 
variables measured for each plot (Table 1). Multiple regression would seem to 
be a logical means of evaluating the relationship between squirrel use and 
habitat characteristics. However, when dealing with a large number of independent 
variables, many of which are highly correlated, it is possible to lose important 
variables and to get misleading results. To alleviate this problem we used 
principal factoring analyses (PFA from SPSS statistical package). PFA transforms 
an original set of variables into uncorrelated component variables, each of which 
is composed of a linear arrangement of the original variables in decreasing 
order of importance. When run on Danville A and 8 data, 10 new variables were 
created. For brevity, these variables are described only for Danville A (Table 2). 
In next month's newsletter we wi11 describe the final step and results of the 
squirrel habitat analysis. 
Table 1. Original Variables 
Total density of dominant trees, total density, and basal area of the followinq 
trees (25" dbh). 
black oak 
red oak 
wh ite oak 
shagbark hickory 
tight-bark hickories (mockernut, pignut, and bitternut) 
beech 
black walnut 
basswood 
sugar maple 
sassafras 
crabapple 
ash 
elm 
tulip 
sycamore 
honey locust 
all trees combined 
Number of vines 
Number of tree species 
Mean dbh of dominant trees 
Mean dbh of suppressed trees 
Density of stems less than l' tall 
Density of stems greater than 1‘ tall and less than .5" dbh 
Density of stems greater than .5" and less than 5" dbh 
Number of species greater than 1' and less than .5" dbh 
Number of species greater than .5" and less than 5" dbh 
Distance to edge 
