HISTORICAL ZOOGEOGRAPHY 107 



and large size, are not entirely due to a common inheritance, but to 

 convergent development conditioned by similar habits. This evidence 

 for an antarctic center of dispersal is accordingly valueless. 24 



It was formerly believed that the shrew-like Centetidae of Mada- 

 gascar were closely related to the Solenodon of Cuba and Santo Do- 

 mingo, and their wide separation in space was accordingly highly 

 remarkable. More exact investigation has shown that there are numer- 

 ous differences between the two groups, and they are now regarded 

 as the "relicts" of a formerly widely distributed primitive type of 

 insectivore. 25 The frogs formerly grouped together as the family Den- 

 drobatidae were separated from the true frogs on account of their lack 

 of teeth; they inhabit Madagascar and tropical America. The absence 

 of teeth is now regarded as due to convergence, and the neotropical 

 Dendrobates is supposed to be derived from a neotropical genus such 

 as Prostherapis, the Madagascan Mantella from an African ranid 

 ancestor. 20, 27 



The different groups of animals are also unequal in zoogeographic 

 importance, especially with reference to their evidence on land con- 

 nections. Animals which are dispersed through the air, whether by 

 active flight like birds, bats, and insects, or passively like spiders and 

 the small invertebrates with resting stages, are relatively unimportant 

 in this respect. Relations between animals which are likely to survive 

 an ocean journey in driftwood, such as snails and the pupae of longi- 

 corns and snout beetles, must be interpreted with discretion in zoogeog- 

 raphy. The best evidence for former land connections is afforded by 

 groups such as the Amphibia and most earthworms, to which salt 

 water is fatal, and by non-flying mammals, for the transportation, 

 even by large rafts, of such large animals as ungulates is impossible, 

 and even small forms such as mice, shrews, and squirrels would starve 

 on any extended journey by this means. 



Mammals and birds have the great advantage to zoogeographic 

 studies of relatively rapid evolution. Although the crustacean genus 

 Apus, which is still living in fresh waters, was already in existence in 

 the Triassic, and most living genera of fresh-water mollusks and in- 

 sects are represented in the Eocene, and even the reptiles have numer- 

 ous Eocene genera which still persist,* none of the living genera of 

 mammals are present in the Eocene and very few indeed in the Mio- 

 cene. The Pliocene genera of mammals are largely the same as the 

 Recent, but nearly all the species are different, though 80 to 95% of 



* The lizards Chamaeleo, Agama, Iguana, and Lacerta, and the turtles Emys, 

 Testudo, and Chelonia. 



