306 ANATOMY AND AFFINITIES OF COEREBIDjE— LUCAS, vol.xvii. 



, . ^ 



mens have been examined. The tongue shown in fig. 5, page 163, 

 "Review of North American Birds," is unfortunately not the tongue of 

 Dendroica tigrina. There has evidently been a transposition of speci- 

 mens, and fig. 4, which is said to be that of Dacnis, is j^robably that 

 of D. tigrina. As the shape of the tougae was the principal character 

 of the genus Perisoglossa, the genus would for this reason, if for no 

 other, be untenable; but even had the tongue been as figured, it would 

 hardly seem a character of sufficient importance for the establishment 

 of a genus. 



The tongue of the Tanagridiij may be slightly bifid as in Pyranga, 

 Tanagra, and Rhamphocoelas, or thick, fleshy, and fringed, as in Salta- 

 tor ((triceps, but so far I have found no species in which the tongue 

 bore any resemblance to that of Gwreha. 



Among the Dreiianididiie, Himatione^Hemignathus, and Vestiaria have 

 very perfect tubular tongues, the upturned edges meeting or even 

 hipping over one another slightly, being so firmly apposed that it is 

 often a difficult matter to force them apart. A few filaments at the 

 end, and here and there along the edge, constitutes the entire feather- 

 ing of the tongue. 



Oreomyza has the commencement of a tubular tongue, but, owing to 

 its shortness, the tubular structure is not carried out. None of these 

 tongues are deeply cleft or widely feathered at the tip, as in the Ccpre- 

 bida?, and none approach the peculiar condition found in Gerthiola, 

 which has a two-branched tongue, with a twisted brush on either 

 branch, and a shallow groove down the center of middle third of the 

 tongue. 



The general pattern of this tongue is very much like that of the 

 Australian Meliornis while the nearest approach to such a tongue as 

 that of Coereha coerulea is found in the Australian 

 Acanthorhynchns tenuirostris, and in this bird the 

 coerebine pattern is carried to the extreme, the 

 tongue being extremely long, slender, bifid, feath- 

 ered at the tip, and tubular for a part of its 

 length. 



The alimentary canal of the Mniotiltidte is, as a 

 rule, comparatively simple, but in Dendroica coro- 

 nata the convolutions of the intestine are almost ex- 

 actly the same as in Coereha. The stomachs of all 

 Fig. 7.-inte8tinai Mniotiltidai examined contained insects. There is 

 convolutions of Tan- jjq (j^.Qp jj^ ^jjj^ group and the stomach is large and 

 somewhat pyriiorm in shape. 

 The tanagers are fruit-eaters, are devoid of a crop, and have 

 the largest intestine and simplest convolutions of any birds exam- 

 ined. 



In the complexity of the alimentary canal there is a parallel 



