144 PROCEEDINGS OF THE AMERICAN ACADEMY. 



over, to distinguish a northern and a southern form in the North Sea. 

 Kyle attributes the differences to the dissimilar temperatures and salt 

 content of the water. The brackish-water forms from the Baltic and the 

 Zuider Zee have a smaller number of vertebrje than that from the North 

 Sea. An increase in temperature brings about a decrease in the number 

 of vertebrae and fin-rays. Jordan (1893) had previously called attention 

 to the law that in fish in general the number of vertebrce tend to decrease 

 toward the tropics. In our pectens the most southern form unit has 

 fewest average rays; but this is not universal for pectens. 



The mackerel {Scomber scomber) of the British Isles has also been 

 studied quantitatively. Garstang (1898), working on a rather too small 

 number of fish from a variety of British localities, distinguished an 

 " Irish " race and a second from the North Sea and Enijlish Channel. 

 More extensive is the work of Williamson (1900) on mackerel from the 

 east and west coasts of Scotland. His fish came from three localities 

 forming the angles of a nearly isosceles triangle : Aberdeen on the north- 

 east, Barra and Stornoway on the northwest, and Clyde at the south. 

 While the fin-rays, finlets, and vertebrae do not difl'er enough to form 

 racial distinctions (on Heincke's method), "the differences in the length 

 of the head, skull, and pectoral fin may with some probability be granted 

 racial distinction." 



Among fresh-water fishes also, a comparison of lots from different 

 localities shows clearly marked racial differences. Eigenmaun (1895) 

 has shown in a roughly quantitative fashion that in Leuciscus halcatus 

 from near the mouth of the Frazer River of the Pacific coast of North 

 America the empirical mode of anal rays is 15; at Sicamous, altitude 

 1300, it is 17 ; and at Griffin Lake in the same basin, but at 1900 feet, 

 there are 19 rays. Yet the anal rays of fishes have usually so constant 

 a number that they are used as specific or even generic criteria. Even 

 in adjacent lakes two lots of a species have been found by Moenkhaus 

 (1895) to have an average difference of one in the number of dorsal rays. 

 Thus Etheostoma caprodes has in Turkey Lake, Indiana, an average of 

 14. 80 ±.02 dorsal rays; in Tippecanoe Lake 15.81 ±.03 dorsal rays. 

 These two lakes are only six miles apart, but the former drains into the 

 Great Lakes and the latter into the Wabash and Mississippi rivers. 

 The lakes are essentially similar in size and origin. The difference in 

 the number of fin-rays of their fish is probably due to the preservation of 

 mutations by isolation. 



In the marine gastropod MoUusca similar geographic races have been 

 studied quantitatively. Thus Bumpus (1898) has shown that the shore 



