492 BULLETIN 50, UNITED STATES NATIONAL MUSEUM. 



ally, the same, the measurements must necessarily be the only factor 

 in determining the number, limits, and relationships of the sevei'al 

 forms. Looking at the matter from this standpoint, it seems to me 

 that anj^ grouping of the various forms into species and subspecies 

 must be purely arbitrary, and almost certain to differ with different 

 persons; and I am still of the opinion, expressed on page 468 of my 

 paper,' that "were 'lumping'^ once begun there could be no end to it, 

 unless purely arbitrary limits were given to the species recognized, 

 and if followed to a logical conclusion might easily end in the recog- 

 nition of a single variable species, equivalent in its limits to the genus." 

 This observation, according to mjr views, applies with exactly equal 

 force to any attempt to segregate the forms into specific groups. There- 

 fore, it seems to me that if there ever was a case which justified the 

 avoidance of trinomials the present genus is certainly of that kind.' 



* Birds of the Galapagos Archipelago, Proc. XJ. S. Nat. Mus., xix, 1897, pp. 459-670, 

 pis. 56, 57. 



^The term "lumping" was intended to apply to the reduction of any definable 

 forms to the rank of subspecies, although the inference is not clear in the original. 



'Also the genera Camarhynchus and Certhidea. 



In order to demonstrate the extreme difficulty of satisfactorily using trinomials in 

 the present genus, I give below the average measurements (of males only) of the 

 forma recognized conjointly by Messrs. Rothschild and Hartert and myself. It is 

 true that these measurements are not absolutely correct, relatively, for I have not 

 been able to measure the very large series contained in the Tring Museum, and have 

 in some cases been compelled to use measurements taken by Messrs. Eothschild and 

 Hartert, and it is a well-known fact that no two persons can get precisely the same 

 measurements for the same individual specimen. Furthermore, in the case of Messrs. 

 Eothschild's and Hartert's measurements only the mean and not the average can in 

 some cases be given. 



Species. 



Wing. 



Tail. 



Cul- 

 men. 



Gonya. 



Width 

 of man- 

 dible at 



base. 



Depth 

 of bill 

 at baae. 



Tarsus. 



Middle 

 toe. 



GEOSPIZiE. 



1. G, magnirostris : Measurements 



from Sharpe and Eothschild 

 and HarlLTt 



2. G.pachyrhyncha: Measurements 



from Rothschild and Hartert. . 



3. G. strmua: 36specimens, but tail, 



tarsus, and middle toe of 18 

 only 



4. G. darwini: 4 specimens, meas- 



ured by Eothscjiild and Hart- 

 ert (ayerage) 



5. G. eonirostris: 5 .specimens, meas- 



ured by Robert Eidgway 



6. G.propmgwa; 58pecimens, meas- 



urements of bill by Rothschild 

 and Hartert 



7. G.bauri: 1 specimen (R.R.) 



8. G. dubia: 10 specimens (R.R.) .. 



9. G.albemarlei: 2 specimens (R.R.) 



92.20 

 88.00 



83.31 



84.00 

 80.77 



77.47 

 SI. 28 

 72. 04 

 72.14 



50.80 



26.67 

 25.20 



48.26 

 60.80 

 43. 94 

 44. 45 



23.25 ( 13.21 

 22. 35 12. 19 



21.60 12.00 



20.32 , 



18.03 9.66 

 17.27 ' 9.14 



13.75 

 10.92 



10.92 

 10. 16 



23.62 

 22.20 



18.60 

 17.02 



15.00 

 17.27 

 14. 22 

 13. 72 



18.54 



23. 37 

 23. 62 



22. 61 

 22. 80 



18.03 



18.03 



16.00 

 16.61 



