32 
Psyche 
[March 
gone a marked change in the past thirty years. Although 
the long winged females may have been present as a small 
minority in the population before 1920, they are now clearly 
the majority in the same population. The reverse applies, 
of course, to the short winged females. It is conceivable 
that the long winged females, which are able to fly short 
distances and which might, therefore, more readily find 
males, have a definite selective advantage over the flightless 
females. This could account for such a change taking place 
in the course of a maximum of 30 generations. 
The distribution of carolinensis presents some interest- 
ing features. Its most northern record is now Linville 
Falls, North Carolina, and its most southern near Unicoi 
Gap, Georgia. The species seems to be confined to the nar- 
row strip of mountainous country, about 120 miles long, 
between these two places. However, even in this area, the 
species appears to be local except for the Black Mountains, 
where it occurs more generally. With the possible excep- 
tion of one individual (the data on which are questionable), 
all specimens of the species so far known to me have been 
collected at elevations between 3000 and 6700 feet. At the 
lower elevations, as at Carolina Hemlock Camp (3000'), 
both male and female adults occur as early as May 15, but 
virtually all disappear by June 4. At higher elevations 
(such as 6000'), the adults do not appear until early July; 
both sexes have been taken as late as July 18 on the top of 
Mt. Mitchell (6700'). 
The species occurs only in shady woods, where the soil 
is moist, though without any definite plant associations. 
Galax, Rhododendron, Jack-in-the-Pulpit, jewel weed, asters 
and May apple occur commonly where carolinensis is found, 
but these plants are characteristic of nearly all moist woods 
in the southern Appalachian region. Some of them, how- 
ever, may be necessary as food for the adults of carolinen- 
sis. Our first efforts to keep the adults alive in cages failed 
completely because we did not find anything on which they 
would feed. Bits of animal flesh or of dead insects, which 
Panorpa and Neopanorpa readily consume, were not eaten 
by carolinensis. Subsequent observation showed that they 
fed on the epidermis of soft leaves, such as those of aster 
