of the Galapagos Archipelago. 243 



ferent. The relative abundance of Cyperaceas to Graminece is extremely fluc- 

 tuating, though in a measure amenable to very conflicting causes, which we 

 can only partially follow. Of the Galapageian Graminece the Pooe and Panicece 

 are nearly equal, which is very unusual within the tropics ; one of the species 

 belongs to Eutriana, a genus almost peculiar to America, and particularly 

 to the mountainous regions or cooler parts. Two new species belong to 

 Aristida, which is also represented by a species in the sterile island of 

 Ascension. 



The CoMPOsiTiE are in every respect the most remarkable family in the 

 Galapagos, both as regards number of new species and new genera, and 

 from their forming much of the wood of the islands. They also are the most 

 instructive, as the species are very clearly defined : the peculiar genera have 

 representatives in the different islets ; and whilst the new species are almost 

 wholly allied to plants from the Andes or extra-tropical parts of America, the 

 old are almost universally the weeds of the low coast of the same continent. 

 It is not therefore with this family as with some others, that the new species 

 are, though permanently, only partially distinct from the continental ones, and 

 possibly varieties due to climatic causes ; but they are the representatives of 

 species which are only found beyond the reach of direct migration, or are to 

 a great extent entirely new genera. 



In respect of the peculiarity of their Compositce, the Galapagos may be com- 

 pared with some other tropical islands, as the Sandwich group and St. Helena; 

 also with two extra-tropical islands, Juan Fernandez and New Zealand. All 

 of these have a larger amount of peculiarity in their floras than any other tracts 

 of land of the same size. It has been noticed that the four last-named islands 

 or groups are remarkable for possessing a great proportion of arborescent 

 Compositce, and in this too the Galapagos share, though the comparison can 

 be carried no further between any of them ; for whilst the order is here repre- 

 sented by Melampodineoe and Helianthece, in Juan Fernandez it is by Senecionece 

 and Cichoracece, in St. Helena chiefly by Asteroidece, in the Sandwich group by 

 Verhesineoe and Bidentineoe, and in New Zealand by Helichrysece and Asterece. 

 In all these cases, the further the islands are from the mainland, the less evidence 

 do the Compositce they contain afford of the botanical province to which each 

 may belong. Thus the Galapagos contain, in the peculiar plants of this order 



VOL. XX. 2 K 



