448 MEMOIRS OF THE CARNEGIE MUSEUM 
diogenes-group). If this is the case it very likely extended in the Tertiary at least 
as far north as at present, but the advancing ice of the Glacial Period must have 
driven it south again, and it must have survived in the mountains of Virginia, 
West Virginia, and states further south. Possibly southern Pennsylvania formed 
part of its preserve in Glacial times, for the peculiar preference of this species for 
cold water admits this assumption. Be this as it may, it is certain that after the 
retreat of the ice this species advanced, occupying or reoccupying the whole state 
of Pennsylvania, and keeping on in its northward migration, until finally reaching 
its present range. 
This advance in a northeastern direction clearly agrees with Adams’ third high- 
way of dispersal (1902, p. 123) along the Appalachian chain, and C. bartoni also 
belongs to the northeastern biota of the second Postglacial wave (1905, p. 58). What 
is interesting in this case is that an aquatic creature follows here the main strike of 
the mountains, independent of the drainage systems. I have previously called 
attention to this fact (1905), p. 129), and have pointed out that this is rather the 
rule with the subgenus Bartonius (this has been observed already by Faxon, 1885a, 
p- 179). There is not the slightest doubt that this peculiarity is connected with the 
ecological laws governing this species. It lives generally in the region of the head- 
waters of the streams, where the dendritic conformation of the drainage systems and 
their mutual interlocking favors frequent shifting of the divides in consequence of 
stream-piracy. 
Moreover, C. bartoni is a form which habitually leaves the water. It is found 
not only in small streams, but also in springs, often at places where there is a very 
scanty supply of water, and this has forced it to often assume burrowing habits. 
Like C. obscwrus it is able to survive exposure to the open air for a considerable time, 
provided the temperature is not too high, and thus it is easy to imagine that it may 
cross over divides during rainy or cloudy weather, wandering from spring to spring 
in the mountains.. On the other hand, we see that C. bartoni is not entirely absent 
from larger streams, and if once established in a small part of the drainage of a cer- 
tain river it may easily be distributed over the rest of it by simply following the 
course of the stream. ‘Thus it is not strange that this species has occupied the whole 
of the state, and this uniformity of distribution is chiefly due to the fact that the 
whole of Pennsylvania is hilly or mountainous, offering everywhere favorable condi- 
tions for this species. The general dispersal is due to two causes: first the ability of 
this species to cross watersheds by active migration ; and second to stream-piracy, 
which has apparently played a considerable part in its dispersal. 
It should be noted, as we have seen above, that the size of this species decreases 
