(54 Merriaw.— Oeographic Didribution of Life. 



iiicnting or l)oing the compleiaent of the other — for water ])emg 

 the medium in which the species live, the hodies of water with 

 their i)rolongatioiis and extensions, as bays, rivers, and lakes, 

 must l)e studied as entities, just as we study a continent with its 

 iKMiinsulas and outlying islands — the means of access to a given 

 liody of water being the principal factor in determining the 

 water-area to Avhich its aquatic life lielongs. And it should l)e 

 remarked that a(|uatie mainmals, as seals and cetaceans, and 

 a(iuatic birds, as ducks and gulls, conform in the main to the laws 

 and areas of aquatic distribution and should not 1)e taken into 

 account in studying the distribution of terrestrial forms of life. 



Gill has said with much truth : " There appears to he a total 

 want of correlation between the inland and marine faunas, and 

 a positive incongruity, and even contrast, between tlic two.'' 

 (Proc. Biol. Soc. Wash.. II, 1S84, 82.) 



Principles ox which Bio-Geographic Regions should re 



established. 



Wallace, in writing of the principles on which Zoological 

 regions should be formed, ex})resses the opinion tliat "' conveni- 

 ence, intelligibility, and custom, should largely guide us." But 

 I quite agree with America's most distinguished and philosophic 

 writer on distribution. Dr. J. A. Allen, that in marking off the 

 life regions and sulirisgions of the earth, truth should not be 

 sacrificed to convenience ; and I see no reason why a homo- 

 geneous circumpolar fauna of great geographic extent should 1 )e 

 split up into primary regions possessing comparatively few 

 peculiar types simply because a water separation ha])pens to 

 exist in the present geologic period ; nor is it evident Avhy one 

 of the resulting feel)le divisions should be granted higher rank 

 than a region of mucli less geograpliic extent comprising several 

 times as many peculiar types. Hence the divisions here recog- 

 nized, and the rank assigned them, are based as far as possible 

 ui)on the relative numbers of distinctive types of mammals, 

 birds, reptiles, and plants they contain, Avith.due reference to the 

 steady multijilication of species, genera, and higher groups from 

 the poles toward the tropics. IMammals have lieen chiefly used 

 as illustrations l)ecause they answer the purpose ])etter than any 

 other single group, and liecaiiso it is clearly impossible in a brief 

 essay of this character to cnumei-atc such a multitude of forms 

 as would he necessary were ('i|U:il consideration accorded to each 

 class. 



