oye ORTMAMN — AFFINITIES OF CAMBARUS. 129 
species, we do not know much about their habits, but a few remark- 
able cases may be mentioned. (1) C. wirzlis and C. immunts, 
although sharply separated, are rather closely allied, and occupy 
large identical tracts of the central states. We know that C. wirilts 
prefers running water with stony bottom, while C. zmmunis is a 
pond and ditch form (see above, p. 117). (2) C. monongalensis 
inhabits, in western Pennsylvania, almost the same territory that 
is occupied by C. diogenes. ‘The first, however, belongs to the 
hills, the second to the lowlands (see Ortmann, Anz. Carnegte 
Mus. N23, Pp. 400). 
The various drainage systems have a different effect upon the 
species of the different subgenera, which ts apparently due to funda- 
mental differences tn their habits. (1) Bartontus is preéminently a 
mountain-stream group. It goes up into the smallest streams, up 
to their very sources. In this region, changes of drainage, due to 
piracy, are common, and rather the rule than the ‘exception, and 
thus the species quite generally occupy the headwaters of streams 
running in different directions from the divides. This is exampled 
by the distribution of the following species: extraneus, bartoni, 
longulus, latimanus, carolinus, and probably also by dogenes. (See 
Adams, ‘‘ Migration of Divides,’’ in Americ. Matural., 35, 1901, 
p- 844). (2) The dlandingt-section belongs originally to the low- 
lands of ths Gulf and Atlantic plain. Here removal of barriers 
largely has taken place, and thus the species of this group belong 
to the drainages of different coast rivers, for instance: J/econtet 
blandingt, clarki, troglodytes, allent. (See Adams, zézd., p. 842: 
‘*In a country approaching base-level a wide distribution of the 
fauna will be facilitated.’’) (3) The subgenus /axonius belongs 
to the great rivers of the interior basin, and does not ascend far into 
the headwaters, at least in the mountainous regions, and also does 
not descend far toward the coastal plain. Consequently, the drain- 
age systems being more permanent, the distribution of these species 
is more closely connected with the latter. We may, perhaps, com- 
pare this —in a very general way — with the period of maximum 
roughness of Adams (/. c.), although this does not hold good for 
all of this immense region. Indeed, there are important excep- 
tions, and the subgenus has crossed over into the lake-drainage 
(C. propinguus, obscurus, rusticus, virilis, immunis), and even into 
the Hudson Bay drainage (C. wirilts). This has been brought 
