376 BULLETIN 50, UNITED STATES NATIONAL MUSEUM. 



(2) " That the members of the Coerebidse do not form a homogeneous 

 group, but contains at least three well-marked types." 



Leaving out the genera Diglossa, Diglossopis, Dacnidea^ Ateleodac- 

 nis, Oreomanes,^ and Conirostrum, whose internal structure has not, to 

 my knowledge, been investigated, it is obvious- from Mr. Lucas's obser- 

 vations that the remaining Ccerebidse comprise three quite distinct 

 groups, whose distinctive characters are as follows: 



a. Crop small or absent; tongue bifid; dorsal pteryla broader; feathers shorter, more 

 compact. 

 b. Crop present but small; tongue fringed; intestinal convolutions simple. 



Chlorophanes, Cyanerpes, Bacnis, 

 bb. Crop absent; tongue brushy; intestinal convolutions extremely complicated. 



Coereba. 



aa. Crop well developed; tongue trifid; dorsal pteryla narrower; feathers longer, 



looser Gloasiptila. 



"In their tongue," says Mr. Lucas, "the Coerebidee are markedly 

 different from the Mniotiltidse, but it is largely a difference of degree 

 rather than of kind. They differ in toto from the Tanagridee, are 

 quite distinct from the Drepanidse, and find their nearest relatives in 

 Acanthorhynchris. ^ 



"The relationship with the tanagers is not very close, although such 

 short-billed forms as Ohlorophanes' and Dacnis, which unfortunately 

 were not available, might bring the two groups a little closer. 



"In size, form, ptei-ylosis, structure of tongue, and pattern of con- 

 volutions of alimentary canal, there is a strong resemblance between 

 Cce7-eba [i. e., Cyanerpet^ and Acanthorhynchus [Meliphagidse], and so 

 far the two forms exhibit a most interesting case of parallelism. The 

 palate, too, on superficial examination looks not unlike that of Glos- 

 siptila. . . . 



"Finally, it must be said that the members of the Coerebidse do not 

 form a homogeneous group, for the family contains at least three well- 

 marked types, Goe.r€ba\Oyanerpes\^ OeTtJiiola\^C(ie,reha\,&nA GlossiptUa, 

 and these types differ from one another in a very marked degree. 

 While Dacnis and Chlorophanes have not been examined by me, the 

 figures of skulls and tongues of these genera indicate that they belong 

 near Ccereha [Oyanerpes]. These genera form a well-marked gi'oup 

 containing those species nearest to the Mniotiltidse and characterized 

 by a long, cleft, feathered, but not suctorial tongue, small crop-like 

 dilatation of the oesophagus, and simply convoluted intestine. 



' ' OertMola [ Cosreid] has a bifid brushy tongue, no crop, extremely com- 

 plicated intestine, and produced angle to the mandible. The tongue 

 resembles that of some Meliphagidse; and the other characters are like 

 some found in the Drepanididse. Glossiptila, with its loose ptilosis, 



^ I am inclined to regard Oreomanes as a Mniotiltine type, related to Conirostrum. 

 The bill of this genus is strongly suggestive of that of Helinaia. 

 ^ A member of the Meliphagidse. 

 ' CMorophanes, however, is not a short-billed form. — R. R. 



