26 BIOLOGICAL SOCIETY OF WASHINGTON. 



in a great degree, to the same types ; and although the reahiis thus 

 associated are characterized by a number of forms pecuHar severally 

 to each of them, they form a natural whole in contradistinction 

 to the others. Very closely connected with this division is the 

 East Indian, and this forms with them an integral portion of a 

 great super-realm or hemisphere. 



In striking contrast with the association of forms characteristic 

 of the several divisions alluded to are the fresh-water types of Aus- 

 tralia. Most closely related to Australia in this respect is South 

 America, and in the fresh waters of that contii^ent are to be found 

 several types which are common to rhe two continents. The only 

 remaining continent — Africa — although presenting some forms that 

 are common to it and India, on the whole furnishes us with an as- 

 sociation of fresh-water forms which recalls the South American 

 realm more than it does any other. Several families of fishes and 

 a number of types of other animals are common to the two and 

 are found nowhere else. 



Combining these facts into a systematic whole, it has been pro- 

 posed to segregate the several realms in the manner hinted at, and 

 to combine under the name Eogaa (i) the Australian, (2) South 

 American, and (3) African realms ; and under the name Cenogaa 

 (4) the North American, (5) the European, and (6) the East Indian 

 realms. JSoga;a gives to us a number of forms which remind us of 

 the ancient inhabitants of the northern hemisphere, and hence the 

 name, while Cenogcea has lost most of the forms that were character- 

 istic of the past, and presents the newest aspect of the earth-faunas 

 in contradistinction with the other. 



An explanation of these relations may be found in one or other 

 of two hypotheses, (i) The like forms may have originated where 

 they are now found, and have been ever confined within pro.xi- 

 mately their present limits; or (2) they may be the survivors of 

 anciently widely-distributed races. Neither hypothesis of itself is 

 a sufficient explanation of all the associations in question, but each 

 is applicable to different cases 



