18 BULLETIlSr 50, UNITED STATES NATIONAL MUSEUM. 



hensive by the inclusion of forms so "abeiTant" as to complicate any 

 diagnosis of the group and obscure its true characters. 



Excepting the Hirundinidse and Alaudidse, both of which are sharply 

 cut off from all other Oscines by most obvious external characters, no 

 group of Oscines can be considered as very trenchant unless such rad- 

 ical treatment as is here given be applied. Allowed their commonly 

 accepted limits, their intelligible definition is in some cases (e. g., the 

 Fringillidse, Tanagridae, Coerebidae, Mnlotiltidse, and "Ampelidse" 

 (=Ampelidfe + Ptiliogonatidse + Dulidse) simply impossible; but by 

 reconstructing the limits of these groups a fairh^ rational diagnosis of 

 each may be accomplished. It is not unlikely that several genera may 

 yet have to be withdrawn from the families in which they are now 

 placed and raised to independent family rank,^ examples being Phai- 

 nojJtila (here, as heretofore, associated with Ptiliogonys and Phainope- 

 pla), Poliojptila (usually placed in the Sylviidse, where it certainly does 

 not belong, and here provisionally referred to the Mimidse), Cdbypto- 

 philms, and Rhodinocichla, the last being here provisionally referred 

 to the Mnlotiltidse. ^ 



KEY TO THE FAMILIES OF OSCINES. 



a. Tarsus sharply ridged posteriorly, the ridge coinciding with the posterior median 

 line, or else outside the latter; inner posterior edge of acrotarsium coinciding 

 with the lateral median line or anterior to it, and at least as far separated 

 from the posterior ridge of the planta tarsi as is the outer posterior edge of the 

 acrotarsium; planta tarsi usually undivided.' (Acutiplantar Oscines.*) 



^ It is of course to be understood that by family rant the Oscinine standard only is 

 meant. 



' Calyptophilus has been considered a member of the Tanagridse, and placed next to 

 Phsenicophilus, but being a ' ' ten-primaried ' ' bird it certainly does not belong there. 

 Rhodinocichla was first described as a member of the Clamatorial family Furnariidse, 

 but, after its Oscinine character had been demonstrated, was placed by some author- 

 ities among the Mimidfe, by others among the Troglodytidte; both these groups, how- 

 ever, belong to the "ten-primaried" section of the Oscines, while Rhodinocichla is 

 typically "nine-primaried," and therefore, being obviously out of place in either of 

 these groups, must be otherwise disposed of. 



'A notable exception to the usual undivided planta tarsi in this section is seen in 

 the genus Salpinctes (Troglodytidre), in which the planta tarsi are more or less dis- 

 tinctly divided into transverse segments. Something of the same sort is seen in 

 most Corvidee, in which also the lateral plates of the planta tarsi are usually more or 

 less distinctly separated along the posterior ridge. In all these exceptional cases, 

 however, the posterior ridge of the planta tarsi is well defined and the tarsal envel- 

 ope as a whole very distinct in its character from that of the group {Latiplantar 

 Oscines) containing the Alaudidfe. 



*The terms Laminiplantar and Scutelliplantar, commonly adopted from Sundevall, 

 are rejected by me because these terms are misleading, some " Laminiplantares" 

 (e. g., Salpinctes, Corvida;, part) having scutellate planta tarsi, while the Alaudidse 

 (forming part of Sundevall's SciileUiplaniares) sometimes (in very old birds) have 

 the planta tarsi entire or "booted;" furthermore, the latter group included, besides 

 the Alaudidse, the superfamily Clamatores. 



