BIRDS OF NORTH AND MIDDLE AMERICA. 27 



manner, but the g-enera themselves are often poorly defined (e. g.,Aimo- 

 pkila)^ or when they seem clearly natural it is found on close compar- 

 ison of the component species that they present such great variations 

 in structural details (e. g., Cyanospiza, Cyanocomjpsa^ etc.) that the fram- 

 ing of a satisfactory diagnosis is by no means an easy matter. Certain 

 members of this extensive series of genera present a close superficial 

 resemblance to the Coccothrauste^ in fact, they are "grosbeaks" so 

 far as the large size of the bill is concerned, though not otherwise, for 

 there is very great difierence in the form of the beak between such 

 genera as Pheuctlcus^ Zamelodia^ Geosplza^ Oryzohorun^ Cwrdinalis^ 

 etc., and that of liesperiphona^ Coccothraustes. etc., not to mention rad- 

 ical differences in other respects. 



Notwithstanding the immense difference in appearance, structural 

 details, and habits between different minor groups of this assemblage 

 of American types, I have failed to discover characters wherebj^ very 

 trenchant groups ma}' be defined. Four genera {Calamospiza^ Spiza, 

 Chondestes^ and Pooecetes^ all Nearctic) do not tit into any of the 

 groups that seem susceptible of more or less exact definition, nor do 

 they constitute a group by themselves. Leaving them out of account, 

 the remaining genera may be rather roughly and arbitrarily separ- 

 ated into two series; one composed of the smaller billed and more 

 plainly colored (usuall}' conspicuously streaked ^) species, and repre- 

 sented by the genei'a Passerculus^ Centronyx, Coturniculus^ Ammo- 

 dramus^ Plaglosplza^ Aimophila, Amphisjyiza, Junco^ Spizella., 

 Zonotrichia, Brachysj)lza, Melospiza^ Passerella^ (Jreosplza., Plpllo^ 

 Melozone^ ArremonojM^ Arrenwn, Lysurus, Atlap>etes, Buarremon^ 

 PseUiophorus^ and Pezopetrx; the other comprising the larger billed 

 or more brightly colored forms, or those with more uniform colors, 

 the genera being Platysplza., Camarhynchus^ Geosj)iza, Cocornls^ 

 Acanthidops^ IIaploi<plza^ Yolatinia, Euetheia, Melanospiza^Loxipasser^ 

 Pyrrlmlagra,, Melopyrrha, Sporophila^ Amaurospiza, Slcalls^ Cyano- 

 spiza, Cyanocompsa, Oryzohorus^ Guiraca., Zamelodia, Pheucticus^ Py^- 

 rhuloxia^ Cardhrdlx, Pttylut^^ Caryothi'auste^s, Phodofhranpis^ and 

 Saltator'^. While the characters given above as distinguishing these 

 two groups are artificial, even trivial, I feel convinced that when the 

 internal structure of all the genera becomes known the line of first 

 division will be drawn somewhere near that here indicated. 



Although an effort has been made in the following analytical ke}^ to 

 keep the component parts of the different groups together, it has been 

 found impracticable in some cases to arrange the groups in what seems 

 to be their most natural sequence; in fact, to do this in a linear arrange- 



^The young always(?) streaked, even if the adults are plain colored. 

 ^ The genera peculiar to South America are not enumerated. 



