MATTIIi:\\ , VLJMATE AND EVOLUTION 20.") 



Cuba, while near in actual distance to the North American continent, 

 has been comparatively inaccessible to sporadic colonization from that 

 source, on account of the direction of the ocean currents; but coloniza- 

 tions from South (or possibly Central) America have reached it. New 

 Zealand is more remote and inaccessible, and, during the whole Mesozoic 

 and Cenozoic eras, we have evidence of but two colonizations by land 

 vertebrates, neither implying any necessary continental connection. The 

 rock-lizard (Sphenodon) may, for aught we know to the contrary, be 

 derived from a marine form ; all its early Mesozoic relatives were aquatic, 

 some apparently marine. The few other reptilia may be best accounted 

 for by sporadic colonizations of later date. The moas are probably de- 

 rivatives from flying birds. 



When we come to the smaller oceanic islands, their poverty of fauna 

 is still more conspicuous. If their fauna is due to sporadic colonization, 

 this should be expected, as the chances are reduced directly in proportion 

 to the smaller length of coastline on which an immigrant might land, as 

 well as by their effective distance from the mainland. The colonization 

 of a group of islands one from another may be due to former land con- 

 nection and subsequent isolation, or to the same method of accidental 

 transport, subject to the same laws of chance. 



It is quite possible that in certain instances the small size and unfa- 

 vorable environment of islands formerly connected with the continent 

 may account for non-survival of the continental fauna. The Falkland 

 Islands are a case in point; but even here, we find the survivors closely 

 allied to the continental fauna and including types which afford the con- 

 clusive proof of continental connection which is uniformly lacking in 

 oceanic islands. ^^ 



The characteristics of continental and oceanic island faunae have been 

 very fully and ably elucidated by Wallace (Island Life), and it is in- 

 tended here merely to assert that the progressive increase of our knowl- 

 edge of the past life of the world tends only to emphasize the distinctions 

 in the source of their faunae which he has so clearly demonstrated and, 

 so far as my acquaintance with the subject goes, to reduce still further 

 the number of continental connections which he regarded as permissible. 



To the argument so often advanced that the transportation of a species 

 across a wide stretch of sea and its survival and success in colonizing a 

 new country in this way is an exceedingly improbable accident, it may 

 be answered that, if we multiply the almost infinitesimal chance of this 



=^ Introduction of Canis antarcticus by human agency in prehistoric times is, however, 

 a possible explanation of its occurrence. It is the only alternate to a Pleistocene land 

 connection. 



