308 



ANATOMY AND AFFINITIES OF C(EBEBIDM— LUCAS, vol.xvii. 



coil into a knot or short loop, whence it uncoils or unfolds in an increas- 

 ing spiral. In Loxioides, Psittacirostra, Himatione, Vestiaria, and Hemi- 

 gnathus the first turn of the intestine from the center is to the left, while 

 in the American species figured it is to the right. The point is one of 

 little or no value, but among the sjjecies figured the difference exists. 



My only specimen of Acanthorhynchus was doubly unfortunate; first 

 in being neatly shot through the palate, completely destroying that 

 region ; and, secondly, in having the intestine in so tender a state that it 

 was difficult to trace its convolutions. Hence I do not feel quite posi- 

 tive that the figure is entirely correct, although it is very nearly so, and 

 if there is any error it lies in the portion beyond the central knot and 

 consists in the omission of some convolutions. It is much simpler than 

 in Certhiola, but not unlike Ca^reba, while a little more complexity 

 beyond the central knot would make the general pattern of the intestine 

 very much like that of Glosaoptila. , 



To sum up: In the character of their palate the Cterebid;e differ 

 from the Mniotiltidie and resemble in some points the Drepanididae 

 and some of the Tanagrida', 

 The Drepanididae differ from all the above-mentioned groups except 



Certhiola in the production of the angle 

 of the jaw. 



In their tongue the Cterebida' are 

 markedly ditterent Irom the Mniotilti- 

 die, but it is largely a difference of 

 degree rather than of kind. They differ 

 in toto from the Tanagridae, are quite 

 distinct from the Drepanididw, and find 

 their nearest homologue in Acantho- 

 rhynchus. 



As regards the Drepanididae, it may 

 be thought that this distinction is very 

 much a matter of opinion, but to me 

 the two patterns of tongue seem quite difi'erent, though both derivable 

 from such a tongue as that of Dendroica. 



It would, perhaps, require less modification to derive the tongue of 

 the Drei^anididse from such an one as that of Icterus icterus^ as this is 

 considerably upcurved along the edges, is not greatly feathered, and is, 

 considering its size, less fleshy at the basal portion than that of Den- 

 droica. 



It must be borne in mind, too, that there are three distinct types of 

 tongue among the Ccerebida^ and that no comparison can be made with 

 them in this particular as a group. 



In comiilexity of alimentary canal they much exceed the Muiotiltidae 

 (except Gcereba cyanea, noted previously), bear no resemblance at all to 

 the Tanagridte, and are a])i)roaclied by the Drei)anidid<e. 



Fig. 9. — Lower mandible of (1) C'a'reba 

 cu-nilea; (2) Certhiola xiortoricengis; (3) 

 Oreomyza bairdii: all twice natural size- 



