164 A NATURALIST ON DESERT ISLANDS. 



which have governed the distribution of some of these 

 species. Food, of course, is one of the most obvious 

 causes which determine the presence or not of a species 

 in any particular locahty ; and in this connection we have 

 wondered if the presence of numerous grackles on the 

 Testigos and the Hermanos Islands has not been deter- 

 mined by the fact that so many hundreds of gannets are 

 accustomed to nest on them. It may be that the grackles 

 have been attracted to the islands for the purpose of 

 picking up a living, derived from the " crumbs " left 

 about the nests during the process of feeding the young 

 gannets ; and that they have consequently elected to take 

 up their abode on the islands, eking out the rest of the year 

 on an insect diet and on the proceeds derived from other 

 sea-birds, which also aest there in their proper season. 

 At any rate, it is a curious fact that no grackles are found 

 on Blanquilla, the Cura9ao group, or any of the other 

 smaller islands off the coast ; and on none of these, with 

 exception of Isla de Aves, do gannets breed. 



The absence of honey-creepers from Blanquilla seemed 

 to us even more strange than in the case of some of the 

 other birds mentioned ; for apart from the fact that these 

 pretty little birds are found on almost every scrap of land 

 surrounding the Caribbean Sea, there are large numbers 

 of low bushy trees of the mimosa tribe, together with 

 logwood and guaiacum trees, which these birds constantly 

 frequent elsewhere. Another curious fact, is the apparently 

 haphazard way in which the species of this genus have 

 spread themselves over the islands of the Leeward group 

 in general. Their distribution seems almost as fortuitous 

 as the distribution of thistle seeds driven by the wind. 

 For example, we have in the Testigos a black form {G. 

 laurce), obviously derived from the Southern Antilles. 

 In Margarita and Tortuga, the species found are normally 

 coloured species (C. luteola and C. ferryi), derived from the 

 mainland. Further westward still, after skipping Mar- 

 garita and Tortuga, we again come across another black 

 form (C. lowii), on the Los Roques islands ; and again 



