EclüiKKlernia II: Crinoidea. 309 



This indieates clearly an uninterrupted land connection at some one 

 or more epochs in the past about tliis entire portion of the Pacific, though 

 this continuity is now brokcn south of Tierra del Fuego, in the western 

 end of the Ah^utian chain of islands, and again in the deep Channels of 

 the Kurüs. 



In the Atlantic no such thing occurs; the Antarctic fauna does not 

 reach South Africa, nor does it pass the Magellanic region. It is thus 

 clear tliat froni the evidence afforded by the recent crinoids there can 

 never have been any land connection between South Africa and Antarctica. 



3. From the southeastern Indian Ocean i'Madagascan region). This path 

 of migration extended from Madagascar and the adjacent islands and 

 portion of Africa in a general northwesterly direction to the Antillean region. 



Among the most characteristic shallow water genera of the Caribbean 

 Sea and adjacent regions, connecting the Caribbean area with southeastern 

 Africa and Madagascar, but not occurring on the west coast of Africa, are 

 the foUowing; each of these is but a slightly differentiated derivative from 

 the widely ranging Indo-Pacific genus listed in the second column. 



Caribbean Genus. East Indian Genus. 



Nemaster Capillaster 



Lepionemaster Comissia 



( Cominia 

 Comactinia 



Comatula 

 Analcidometra Oligometrides 



In addition to these the genus Tropiometra is represented at Mada- 

 gascar and in the adjacent territory by one species (carinata)^ while an- 

 other (picta) occurs at 8t. Helena and along the American coast from St. 

 Lucia in the West Indies to Santa Catharina Island in southern Brazil 

 (28 S. lat.). 



The Indo-Pacific genus Comatdla, occurring at Madagascar, is represent- 

 ed in the Atlantic by Neocomatella which occurs throughout the Caribbean, 

 at St. PauFs Rocks in the mid-Atlantic, aad oft' southwestern Europe and 

 northwestern Africa as far south as the Canary Islands. This genus, how- 

 ever, has a somewhat deeper habitat than any of the preceding, and is not 

 therefore strictly comparable with them. 



A very curious thing about the section of the Indo-Pacific fauna 

 from which this fauna has been derived is that the progressive decrease 

 in the number of species inhabiting progressively increasing depths is not 

 as would naturally be supposod uniform, but is represented by a curve 



21 Michaelsen, Westafrika. 



