INSULAR FLORAS. 299 



cal data strongly favour its correctness, and the soundings 

 given in the map accompanying Agassiz's report (2,7) of 

 the Albatross expedition show a relatively shallow area 

 in which the Galapagos Islands are situated, and which 

 extends eastward to the mainland of Veraguas. Probably 

 the separation would be greatly anterior to the segregation 

 of the West Indian Islands. 



In the Botany of the Challenger expedition (38) I 

 attempted a rough classification of islands in relation to the 

 composition of their floras. These are defined as follows : 

 1, Vegetation comprising a large endemic element including 

 distinct generic types ; 2, vegetation comprising a small, 

 chiefly endemic element, the derivation of which is easily 

 traced ; and 3, vegetation containing no endemic element. 

 Without due consideration the Galapagos were referred to 

 the first category. Sir Joseph Hooker (39) fully realised 

 the absolute American affinities of the flora ; but he analysed 

 and discussed it as a derived one rather than as a remnant. 

 Darwin, through some misinterpretation of the statistics sup- 

 plied to him, fell into a singular error respecting the generic 

 endemic element in the Galapagos (40). Referring to 

 the Compositse, he says : " There are twenty-one species, 

 of which twenty are peculiar to this archipelago ; these 

 belong to twelve genera, and of these genera no less than 

 ten are confined to the archipelago ! " How this error arose 

 it is impossible to say, but as a matter of fact the statement 

 quoted is wrong (and was wrong at the time it was written) 

 in all its details. With regard to assumed endemic genera 

 of Compositae, five were founded on galapageian plants, 

 namely, Microcoecia and Desmocephalum, since reduced to 

 Elvira ; Macrcea to Lipochczta ; and Scalesia and Lecocar- 

 pus are so near to Mirasolia and Melampodinm respectively 

 that the late Mr. Bentham gave it as his opinion that they 

 might well be reduced. Two genera from these islands 

 belonging to other orders have also been reduced. These 

 are Galapogoa = Coldtnia (Boraginacae), and Dictyocalyx 

 = Cacabus (Solanacese) ; and Pleuropetalum (Amarantaceas) 

 has since been found in several localities in Western 

 America. Taking this view of their affinities, there is not 



