The occurrence of European types in South Africa, undifferentiated 

 or weakly differentiated, and of Japanese types not only in South Africa 

 but also in Australia, in similar stages of speciation, suggests that the 

 Pleistocene coohng of the oceans was world wide. Even the vast Indo- 

 Pacific area appears to have been cooled sufficiently to permit the passage 

 of a limited representation of the warm-temperate fauna, though insuffi- 

 cient to exterminate any considerable part of " the great mother fish 

 fauna of the world." 



Combining certain zoogeographical and oceanographic data, and 

 accepting the strong evidence that the temperature relations of organisms 

 are very conservative, we may infer the approximate amount to which 

 certain tropical waters were cooled during the late Pleistocene, to allow 

 the transgression of the Tropics by such fishes and other organisms. 

 The evidence is particularly strong and uncomplicated for the long and 

 nearly straight shore of the eastern Pacific, where, it is commonly assumed, 

 the cool-water faunal connections were especially numerous, x^long this 

 coast many of the now disrupted forms, such as the sardine and the 

 giant kelp, are separated by a tropic belt that extends from near Cape 

 San Lucas (the tip of Baja California) to the northern limit of the 

 Humboldt or Peru current, in Peru. The almost certainly late Pleistocene 

 faunal connection between these present limits would, we may assume, 

 have required the cooling of the ocean surface to approximately the 

 temperature now prevaihng in the region from Cape San Lazaro to 

 Cape San Lucas — that is, off Magdalena Bay, Baja California. The 

 surface temperature data averaged monthly for coastwise squares of 5° 

 latitude and longitude indicate for this region a fluctuation from about 

 18° c. in the spring to about 25° in the summer, and for the warmest area 

 intervening between this region and the sharp gradient in Peru, a change 

 from 26-5° in the winter to 28° in May. The difference between the 

 monthly a^-erages is about 8-5° for the coldest month and 3° for the 

 warmest. Since, toward the warm end of their ranges, most cool- water 

 forms breed in the coolest season, it is probable that during winter in the 

 last Ice Age the surface of the shore waters of the eastern Pacific was 

 cooled in the order of 8°, though not necessarily more than about 3° 

 during the warmest month. That there was some cooling during the 

 summer is suggested not only by the general data, but also by such facts 

 as this : that the tuna clippers are unable to keep sardines alive at 

 temperatures much above 26°. The decrease of about 8° in winter would 

 account for the presumed annihilation during the late Pleistocene of much 

 of the fish life of the Panamanian region, but not of all the fauna, for there 

 is a large tropical representation about Magdalena Bay, where winter 

 temperatures run low. 



Similar, though less definite, data for the western Pacific indicate that 

 the equatorial waters there were cooled during the Pleistocene, but 

 probably not more than 2° to 4° during any month. There is httle 

 evidence of extensive cooling of the western Atlantic Tropics during the 

 Pleistocene, but in all probability the convergence of the isotherms and 

 the equatorial coohng was very pronounced along the eastern shores of 

 the Atlantic. 



328 



