172 BULLETIN 50, UNITED STATES NATIONAL MUSEUM. 



not 1)(' (lillicult to divide the (^roup into .several shiirply eircuiiiseri))ed 

 and casil}' ehanieterized lesser groups; this has been attempted for 

 the family as a whole, and the "subfamilies"" Cassicina% Ao-ela?ina3, 

 Sturnellina', Icterinse, and Quisealinre have long been adopted, even by 

 the latest authority.^ A careful comparison of all the forms, however, 

 will soon convince one that such a subdivision can not be maintained, 

 the ditferent supposed groups running into one another so gradually 

 that any sul)division of the family is most likely to be more or less 

 arbiti'ary. So far as 1 am able to discover, only two genera stand out 

 prominentl}' from all the rest, these lieing Sfurnella and DoUclionyx; 

 3^et each of these is obviously related to other genera — Sturiudhi to 

 Trupialis, and this, through Leistes^ obviousl}' leading to the Agelaiine 

 type, the latter through Icterus to Cacicus^ and this in turn to the 

 opposite extreme from StiirnelJa^ the larger caciques or oropendolas 

 {(ji/inmJNtinojjs, Zarhynclius^ etc). The two "' lateral oH'shoots" repre- 

 sented by DoUehonyx and the typical Quiscali (genera Quiscalus, 

 Ilohxjawcalas^ and 2Iegaqii'iScalu.-<) are both obviously connected with 

 the more generalized t\'pes, though Doliclionyx evidently shows points 

 of relationship to the Sturnelline end of the series. 



The following arrangement of the genera is far from satisf actor}' to 

 the author, but is the best he has been a))le to make without a far 

 greater amount of time tban is at his disposal, together with a careful 

 study of internal characters, not now practicable on account of aljseftce 

 of the necessary material. While not entirely natural, the key has 

 been prepared to show as nearly as can be in a linear arrangement, 

 what appears to be the most natural sequence of the genera, without 

 obscuring the characters which serve to most readily identity them. 



The Icterida^ comprise l)irds of most various habits. Some are 

 strictly arboreal, and if placed upon the ground are almost incapable of 

 progression; others are terrestrial (though more or less frequently 

 alighting on trees and sometimes nesting there), and walk upon the 

 ground with the grace and dignity of a crow or starling; many inhabit 

 reedy marshes, and these usually nest i n large colonies. The oropendolas 

 (genera Ocycdus, Clyjncterus^ Zarhynchus^ Gymnosthiops^ and Osti- 

 ifiop8)^\\\ cacicjues (genera Cacicuswc\^ Cas.nc(i/u.s")iiUo nest in colonies, 

 but instead of building an open cup-shaped nest attached to the upright 

 stems of a(iuatic plants, attach their long pensile nests to the extremi- 

 ties of ])i'anches of tall forest trees. The " American Orioles"' (genus 

 Icterus) also build pensile nests, but, usualh' at least, are not gregarious. 

 Many species are remarkable either for the fullness and richness or 

 other remarkable character of their notes, some of them being song- 

 sters of high merit, Avhile others utter only the most harsh and dis- 

 cordant sounds. Some genern {3I(>I('f/i/'i(s, Callothnis, and Caxsldix) 



' 1'. L. Sclater in Cat. Birds Brit. I\Ius., xi, 1886, pp. 308-405. 

 ^The nesting haliits of A)n}>hjcA!rcn>< are ap])arently nnknown. 



