INTRODUCTORY CHAPTER. lxxi 



and the warm regions of the Mediterranean region, and 

 reaching even the Atlantic Islands. 



b. Forms which are part of the general tropical African flora. 



c. Forms which are met with on the African continent in the 



mountainous region of Abyssinia, of east tropical Africa, and 

 of west ti-opical Africa, in south Africa, and also in Mada- 

 gascar. This is the element which is developed chiefly, 

 though not altogether, in the higher regions of the island. 



(B) In the Asian element we find 



a. Forms which belong to those types which people the plains of south- 



west Asia, extending as far as north-west India on the east. 



b. Forms which are part of the general tropical Asian flora. 



c. Forms which have relations in restricted districts in India or 



further east, but which have no connections in the inter- 

 mediate regions. 



7. There are some curious Mascarene connections. Thus, for instance, the 

 occurrence of Elceocarpus and of Cylista scariosa. 



8. There are some striking American affinities. Thamnosma, Dirachma, 

 Ccelocarpus illustrate these. 



How are these features to be explained 1 What is the origin of the 

 Socotran flora ? 



When we group together all the facts now known of the flora, we arrive 

 without hesitation at the conclusion that it has been isolated for a vast length 

 of time during which Socotra has been an island. The position which Socotra 

 occupies in the Indian Ocean naturally leads to the supposition that it has at 

 one time been a portion of the African continent, and that it has been broken 

 off from Cape Guardafui. The general evidence of the flora not only gives 

 certainty to the supposition, but it also shows that the separation from the 

 mainland is of great antiquity. 



It is by such a land-connection alone that we can account for the African 

 element in the flora. But, as I have shown, there are African elements of more 

 than one kind in the Socotran flora. Beside the general tropical African types, 

 we have those kinships with forms of sporadic African distribution of many of 

 the most peculiar plants of Socotra to which attention has been repeatedly 

 called, and some further explanation is needed to show what is the meaning 

 of this relationship, and how it is come about that we have this group of 

 isolated forms in the midst of others of more generally extended distribution. 



All the evidence tends to show that the present general tropical flora of 

 Africa is not the oldest African flora of which we have knowledge. The 



