WEST INDIAN AND WEST AFRICAN FLORAS 95 



general verdict strongly favours the current hypothesis, there are 

 some important exceptions. With the mangroves and their asso- 

 ciated plants, as well as with the cosmopolitan beach plants, the 

 vote is practically unanimous. With the American beach plants 

 and with the riverside plants it is a majority vote, whilst with the 

 estuarine plants the evidence is conflicting and at times apparently 

 hostile (pp. 89-94). 



7. Yet the ultimate tendency of the qualifying cases may not 

 necessarily be hostile. The implication of the behaviour of such 

 peculiarly American plants as Sapindus saponaria, Hippomane 

 mancinella, Morinda royoc, Saco glottis amazonica, and Tournefortia 

 gnaphalodes, which possess fruits or seeds well suited for the direct 

 ocean traverse, may be that the currents have given them no oppor- 

 tunity of extending their range eastward beyond the New World, 

 since, as we have seen, the facilities offered by the currents to West 

 African littoral plants of reaching America are very much greater 

 than the opportunities supplied to American tropical shore plants of 

 extending their range to West Africa. But here another difficulty 

 arises in the case of certain of the littoral plants which border 

 the beaches on the Pacific coasts of tropical America. It is not 

 easy to understand why Conocarpus erectus and Hippomane mancinella 

 that are evidently well fitted for oceanic dispersal, have not reached 

 some of the easternmost groups of the tropical Pacific through the 

 agency of the equatorial currents (pp. 93, 94). 



8. In the case of plants like Andira inermis, Ecastaphyllum brownei, 

 and Symphonia globulifera, which, although existing on both sides 

 of the tropical Atlantic, have neither fruits nor seeds that could 

 cross the ocean in a sound condition even by the most direct route, 

 it is probable that other considerations arise which are concerned 

 with the original distribution of the genus in both worlds (pp. 92, etc.). 



9. Probably the most significant general conclusion in this sum- 

 mary is given in the third paragraph ; and it is in this connection 

 essential to remember that if buoyancy data had been available for 

 several plants not included in the list, the proportion responding to 

 the test among littoral plants confined to the New World would in 

 all likelihood have been as low as 15 per cent. 



