1985] 
Pettry & Tarter — Baetisca Carolina 
357 
each food category (filamentous algae, plant detritus, mineral detri- 
tus, diatoms) was determined by figuring the percentage of small 
grid squares within each field that contained each of the different 
food categories. 
Multivariate discriminant analysis, using the computer program 
BMDP7M (Dixon, 1981), was used to separate B. Carolina from the 
closely related B. berneri. Thirteen morphological characters (shield 
length and width, head width, length of caudal filament, body 
length, prothoracic leg and claw length, mesothoracic leg tibia and 
femur length, mesothoracic leg tarsus and claw length, width of 
abdominal segment 10, labrum width) were measured on all nymphs 
(15) of B. berneri (holotype locality, Laurel Fork, West Virginia) 
and from five different geographical populations (Georgia, North 
Carolina, Tennessee, Virginia, and West Virginia - Panther Creek) 
of B. Carolina nymphs (85). Two colormorphs from Panther Creek 
were included in the analysis. Males and females were grouped 
together in the analysis. Because of size variation within and among 
populations, regression analyses were applied to all measurements to 
remove the linearly related effects of size. Head width was used as 
the independent variable for regression of the other variables. The 
SAS General Linear Models procedure (Barr et al., 1976) produced 
residual values for each character; these values were then used as 
“size-free” variables. In the final analysis, the computer program 
(BMDP7M) generates canonical variates with maximum between 
group variance relative to their within group variance. The canoni- 
cal variate means are plotted on the first two canonical axes, and 
analysis of variance describes significant differences between groups 
(P < 0.01). Using canonical functions, the posterior probability of 
each nymph belonging to its respective population is computed and 
classified accordingly. 
In order to determine the fecundity of B. Carolina, direct egg 
counts were made on adults. The dorsal body cavity of ten females 
(four imagos and six subimagos) was opened longitudinally with 
microdissecting scissors. The eggs were carefully removed, placed in 
a Petri dish, and counted under a Bausch and Lomb dissecting 
microscope. The regression of fecundity on body length was calcu- 
lated and a coefficient of correlation was determined. The diameter 
of 50 eggs per female was measured with an ocular micrometer in a 
dissecting microscope to the nearest 0.01 mm. 
