lf)4 THE VOYAGE OF H.M.S. CHALLENGER. 



Phamorhina goliath, from what has been said above, therefore differs from Carpophaga latrans 

 in having the cones of the gizzard proportionally longer, at the same time that they are ossified 

 (which necessitates the presence of vessels in the ossification, which appear after death as the fibrous 

 cord) and oblique. There is, however, a great similarity between the two organs. 



I am informed by the Rev. S. J. Whitmee that Carpophaga pacifica in the Samoan Islands feeds 

 on nutmegs, from which it is highly probable that in that species the gizzard-epithelium is modified 

 in a manner similar to that of the Fiji or New Caledonian species. Specimens of Carpophaga pacifica 

 preserved would therefore be of special interest for the determination of this point. 



With reference to the other parts of Carpophaga latrans, the intestine is very capacious, only nine 

 inches long, and transversely sacculated from the contraction of its outer longitudinal muscular coat, 

 this producing the appearance of thirty bold transverse folds on the mucous surface. There are no 

 colic caeca ; and, as in the genus Carpophaga generally, 1 the gall bladder is well developed. The 

 liver-lobes are eipial in size. 



The syrinx (rig. 2, p. 153) is Columbine, with its lateral muscles attached inferiorly to the mem- 

 brane between the penultimate and antepenultimate tracheal rings. The trachea is composed of rings 

 which are very yielding in the middle line posteriorly. As can be seen in the figure, the musculi 

 sterno-tracheales, which are independent of the intrinsic muscles, are not quite symmetrically attached. 



The furcula, as is the rule in the subfamily, is very slender, but complete, wherein this species 

 differs from Phamorhina goliath, in which it is cartilaginous at its symphisial end, according to MM. 

 Verreaux and Des Murs. 2 



Myologically, the ambiens is to be found, not large; the femoro-caudal with its accessory head 

 are well developed ; the semiteudiuosus and its accessorius are the same. 



There are two carotids ; so that in this as well as all the other features above mentioned Carpo- 

 phaga latrans agrees with my definition of the division of the Columbidse into which it naturally falls, 

 its gizzard differing, however, from that of all but one of the species which have been examined. 



1 Vidt Pro,-. Zool. Soc. 1874, p. 258. - hoc. cit. } p HO. 



