DISPERSION OF FRESH-WATER FISHES. 127 
sible, and people finding them in such localities 
often imagine that they have “rained down.” 
Once, near Indianapolis, after a heavy shower, I 
found in a furrow in a corn-field a small Pike,! 
some half a mile from the creek in which he should 
belong. The fish was swimming along in a tempo- 
rary brook, apparently wholly unconscious that he 
was not in his native stream. Migratory fishes, 
which ascend small streams to spawn, are espe- 
cially likely to be transferred in this way. By 
some such means any of the water-sheds in Ohio, 
Indiana, or Illinois may be passed. 
It is certain that the limits of Lake Erie and 
Lake Michigan were once more extended than 
now. It is reasonably probable that some of 
the territory now drained by the Wabash and the 
Illinois was once covered by the waters of Lake 
Michigan. The Cisco? of Lake Tippecanoe, Lake 
Geneva, and the lakes of the Oconomowoc chain, 
is evidently a modified descendant of the so-called 
Lake Herring. Its origin most likely dates from 
the time when these small deep lakes of Indiana 
and Wisconsin were connected with Lake Michigan. 
The changes in habits which the Cisco has under- 
gone are considerable. The changes in external 
characters are but trifling. The presence of the 
Cisco in these lakes and its periodical disappear- 
ance — that is, retreat into deep water when not in 
the breeding season — has given rise to much non- 
sensical discussion as to whether any or all of 
1 Zsox vermiculatus Le Sueur. 
2 Coregonus artedi sisco, Jordan. 
3 Coregonus artedi Le Sueur. 
