786 EEPOBT— 1893. 



researches of Mr. Scott Wilson, and the explorations of the Honourable W. Roth- 

 schild's collectors, it was not known that almost every island of the group possessed 

 one or more representatives of each of these peculiar genera. Thus, every island 

 which has been thoroughly explored, and in which any extent of the primeval 

 forest remains, possesses, or has possessed, its own peculiar species of Hemignathus, 

 Mimatione, Phceornis, Acrulocercics, Lo.rops, Drepanis, as well as of the massive- 

 beaked finches, which emulate the Geospiza of the Galapagos. Professor Newton 

 has shown that while the greater number of these are, probably, of American 

 origin, yet the South Pacific has contributed its quota to this museum of ornitho- 

 logical rarities, which Mr. Clarke very justly proposes to make a distinct biological 

 sub-region. 



That each of the islands of this group, however small, should possess a flora 

 specifically distinct suggests thoughts of the vast periods occupied in their diSeren- 

 tiation. 



In the Canary Islands, either because they are geologically more recent, or 

 because of their proximity to the African coast, which has facilitated frequent 

 immigrations from the continent, the process of differentiation is only partially 

 accomplished. Yet there is scarcely a resident species which is not more or less 

 modified, and this modification is yet further advanced in the westernmost islands 

 than in those nearest to Africa. In Fuertaventura and Lanzarote, waterless and 

 treeless, there is little change, and the fauna is almost identical with that of the 

 neighbouring Sahara. There is a whin-chat, Pratincola dacotia, discovered by my 

 companion, Mr. Meade- Waldo, peculiar to Fuertaventura, which may possibly 

 be found on the opposite coast, though it has not yet been met with by any collectors 

 there. Now, our whin-chat is a common winter visitant all down the West African 

 coast, and it seems probable that isolation has produced the very marked characters 

 of the Canarian form, while the continental individuals have been restrained from 

 variation by their frequent association with their migratory relations. A similar 

 cause may explain why the blackbird, an extremely common resident in all the 

 Canary Islands, has not been modified in the least, since many migratory indi- 

 viduals of the same species sojourn every winter in the islands. Or take the blue 

 titmouse. Our familiar resident is replaced along the coast of North Africa by a 

 representative species, Paiics ultramajinus, differentiated chiefly by a black instead 

 of a blue cap, and a slate-coloured instead of a green back. The titmouse of Lan- 

 zarote and Fuertaventura is barely separable from that of Algeria, but is much 

 smaller and paler, probably owing to scarcity of food and a dry desert climate 

 Passing, 100 miles further to sea, to Grand Canary, we find in the woods and forests 

 a bird in all respects similar to the Algerian in colour and dimensions, with one ex- 

 ception — the greater wing coverts of the Algerian are tipped with white, forming a 

 broad bar when the wing is closed. This, present in the Fuertaventura form, is re- 

 presented in the Canarian by the faintest white tips, and in the birds from the next 

 islands, Tenerife and Gomera, this is altogether absent. This form has been recog- 

 nised as Parus tenerifce. Proceeding to the north-west outermost island, Palma, 

 we find a very distinct species, with different proportions, a longer tail, and white 

 abdomen instead of yellow. In the Ultima Thule, Hierro, we find a second very 

 distinct species, resembling that of Tenerife in the absence of the wing bar and in 

 all other respects, except that the back is green like the European, instead of slate 

 as in all the other species. Thus we find in this group a uniform graduation of 

 variation as we proceed further from the cradle of the race. 



A similar series of modifications may be traced in the chaSinch {Fringilla), 

 which has been in like manner derived from the North African F. spodiogena, and 

 in which the extreme variation is to be found in the westernmost islands of Palma 

 and Ilierro. The willow wren {Phylloscopus trochilns), extremely numerous and 

 j-esident, has entirely changed its habits, though not its plumage, and I have felt 

 justified in distinguishing it as Ph. fortunntm. In note and habits it is entirely 

 different from our bird, and though it builds a domed nest it is always near the top 

 of lofty trees, most frequently in palm-trees. The only external difference from 

 our bird consists in its paler tarsi and more rounded wing, so that its power of 

 flight is weaker, but, were it not for the marked difference in its habits and voice, 



