4G8 BIRDS OF THE C.ALAPAOOS AECHIPELAGO—RIDGWAY. vol. xix. 



of chaos and enemy of order is the " lumper," and not his much ma- 

 ligned CO worker, the "■ liair-splitter." 



When it is remembered that tbe coloration is practically if not abso- 

 lutely the same in all of the twenty-odd forms of the genus Gcospiza, 

 it will be seen th.it if any segregation of species is made at all it must 

 be based upon measurements; and when it is further seen that there 

 is a gradual transition in size from the enormous beak of G. magnirostris 

 to the comparatively minute one of (/. parvula (see Plate LVll) and from 

 the excessively thick one of G. piuhyrhyncha (whose latenil outlines 

 approximate an equilateral triangle) to the slender and curved one of 

 (7. seandens or the acununate one of G. acutirostrifi; and that size of 

 beak is not necessarily correlated with length of wing, tarsus, etc., the 

 ditticulty of detining the species becomes obvious. In fact, the seg- 

 regation of dctinable forms would not be possible were there not a rea- 

 sonable uniformity of measurements among specimens from one locality, 

 it being usually the case that when a great difference in size between 

 specimens from anyone island is observed, the specimens can easily be 

 divideil into two or moi e (rarely as many as seven) sets, whose meas- 

 urements do not inosculate, the individuals whose measurements are 

 intermediate coming from some other island. Some islands, uufortu 

 nately, are so poorly represented by specimens that much doubt must 

 necessarily exist respecting the forms which are found upon them. 



Having been perplexed by these diihcuUies, I have carefully weighed 

 all doubtful cases, and whenever there seemed to be a well defined 

 average difference between specimens from different islands, I have not 

 hesitated to separate them as local forms. No other course, indeed, is 

 practicable; for 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 recogni- 

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



How many fairly good species there really are in the genus it is not 

 possible for me to conjecture from the insufficient material that I have 

 been able to examine. A considerable number of the forms recognized 

 in this work are undoubtedly mere local races. Insular forms, how 

 ever, can hardly be treated in the same manner as continental ones, 

 whose conditions of environment are so much more favorable to inter 

 gradation; hence 1 have treated alike as si)ecies all the forms that it 

 has seemed worth while to distinguish by a separate name. Regard 

 ing supposed excessive individual variation in the genus Gcospizo, I 

 am unable to agree entirely with Mr. Salvin,' who has, I think, made 

 the specific limits too wide, and thus brought together under one spe- 

 cific name forms from ditferent islands which are really more or less dis 

 tinct. Indeed I have failed to discover in the series of specimens from 

 any one island a greater range of variation in measurements than often 

 exists among an equal number of specimens of mainland' forms. (See 

 under genus Gcospiza, p. 5()S.) 



'Trans. Zool. Soc. Loudou, IX, Ft. ix, 1876, pp. 479-484. 



