3i8 



SNODGRASS AND HELLER 



leads up to the largest billed forms of Geospiza, viz., G. strenua and 

 G. magnirostris^ but which in plumage are more generalized than the 

 members of the Cactornis group. It was only for the conventionality 

 of placing all the varieties of a species in continuous succession that 

 we did not insert the G. fortis magnirostris series immediately after 

 G. fuliginosa -parvula. 



MEASUREMENTS OF ADULT SPECIMENS OF GcosftZa fuUginOSa 



difficilis. 



63. THE GEOSPIZA FORTIS SERIES. 



Under this series we include forms that have by Ridgway and 

 Rothschild been kept separate in two groups, G. fortis and G. 

 dubia. We have specimens from the southern end of Albemarle at 

 Iguana Cove that absolutely bridge over the differences between the 

 previously known forms of these two supposed species. These Iguana 

 Cove specimens we have described as a new subspecies, G. fortis 

 ■platyrhyncha. This grouping of all the forms heretofore separated 

 into the speciesyb^/zj and dubia leads us into the necessity of recog- 

 nizing two subspecies of the same species on the same island wherever 

 the two forms occur. There are four cases of this sort: G.f. fortis 

 and G.f. dubia occur together on both Chatham and Duncan; G.f. 

 fortis and G. f. bauri occur on James ; G. f. fortis and G. f. 

 simillima occur on Charles. G.f. simillima Rothschild and Hartert 

 is a variety of very doubtful standing, but the other three cases are 

 apparently well substantiated. On these islands when two subspecies 

 of G. fortis occur together in this manner, the two forms are always 

 entirely distinct, and intergrade as subspecies only through individuals 

 on some other island. On Albemarle where two subspecies occur 

 they are geographically separated and intergrade on the intermediate 

 territory just as do subspecies inhabiting contiguous areas on the 

 mainland. It may of course be questioned whether groups of indivi- 

 duals living on different islands and showing some average difference as 



