106 CEYLON MARINE BIOLOGICAL REPORTS. 



i.e. , about half the number of the sura total of the Ceylon C aider pa-s})ecies. These are entirely lacking along 

 the southern coasts of South America, which, of course, faU in part outside the tropical zone, and in general 

 do not harbour any Caulerpa at all. The flora of the Cape is also very poor in Caulerpas ; only five are 

 known, and among the species common to the Atlantic and Pacific only two, davifera and Chemnitzia, are 

 included in this number and they are, moreover, exclusively found in Natal and on the east side of South 

 Africa, which is washed by the warm Mozambique stream, but not on the west side, which has a much 

 less tropical character owing to cold antarctic currents. It is, therefore, a remarkable fact that the 

 centres of distribution which the species have in common, the Atlantic Ocean on the one hand, and the 

 Indian-Pacific Ocean on the other, are at the present time separated from each other, so that there 

 is no communication between the two different districts for the different species. Murray, who has 

 closely studied the problem of the distribution of the tropical algae, has pointed this out with reference 

 to all the marine flora that the Atlantic and the Indian Oceans have much in common. To explain this fact 

 Murray has suggested that the two rather similar tropical floras in the Atlantic and the Indian Oceans , 

 which now communicate only via the Cape, must probably " have been periodically mingled at the 

 epochs of warmer climate at the Cape." 



This is perhaps not beyond the bounds of possibility, but it seems to me not improbable that the 

 resemblance between the flora districts of the tropical Atlantic and of the Pacific -Indian Ocean may 

 be explained naturally in another way, namely, that it (the resemblance) must be regarded in combination 

 with the geological data which argue in favour of a pre-historic direct communication between the Pacific 

 and the Atlantic, either across the Isthmus of Panama or further south. That such a communication 

 existed even as late as in the Tertiary Age is beyond aU doubt (Compare P. M. Duncan ''On the Fossil 

 Corals of the West Indian Islands" ; R.T.Hill, " The Geological History of the Isthmus of Panama and 

 portions of Costa Rica " ; A.E.Ortm.vnn, " Tertiary Invertebrates "), though opinions differ when the ques- 

 tion arises where the communication took place, some geologists assuming a more southerly connection 

 than across the Isthmus of Panama. 



It is very remarkable that the tropical algal district in the Atlantis is almost confined to the 

 West Indies. This probably depends on the eastern coast of South America, just as the western coast 

 of Africa — as Murray points out, not offering suitable habitats for algal growth. But then one can 

 scarcely assume that, even if warmer water washed the south coasts of South America and especially of 

 Africa, a more luxuriant algal vegetation should have been harboured then than is the case to-day, seeing 

 how little suited they are said to be for algal growths of any kind. I therefore think that the communi- 

 cation and the relationship between the floral districts of tlie Indian-Pacific Ocean and the West Indies 

 can be more naturally explained in another way, i. e., that these districts once had direct communication 

 over the districts where now the Central or South American continent separates the two great oceans. 



If we examine the geographical distribution of other marine organisms, we cannot fail often to 

 notice a greater resemblance between the West Indies and the Indian-Pacific Ocean than between the 

 West Indies and the rest of the Atlantic. Let us, for instance, examine the marine phanerogamic plants, 

 to the geographical distribution of which Ascherson ("Die geographische Verbreitung der Seegraser") 

 has devoted careful study. 



Of the genus Thalassia belonging to the family of the Hydrocharitacece there are only two most 

 closely alUed species, viz., Th. Hemprichii with uniform distribution from the Red Sea, the Northern 

 Indian Ocean (but not the east coast of Africa, nor the Cape!) to the eastern islands of the Pacific, 

 and Th. testudinum, confined to the West Indies alone. 



A quite analogous distribution is true of the two closely alhed Cymodocea species, C. isoetifolia 

 and C. manatorum, which together form the group Phycoschoenus, Aschers., which is sharply distinct 

 from the other CiymofZocca-species. The distribution of G. isoetifolia almost perfectly coincides with 

 that of Th. Hemprichii, that is to say, the Red Sea, the North Indian Ocean, to the eastern islands of the 



