MESSES. A. AND E. NEWTON ON THE OSTEOLOGY OE THE SOLITAIEE. 349 
be more elongated and transverse, and the process to which the squamosal is articulated 
does not project backward so far as in JDidus. There is still less resemblance in the 
structure of this part to the same in Didunculus, where the inferior articular surface 
deviates, as has been already stated by Dr. Melville (op. cit. pp. 99, 100), most greatly 
from the normal Columbine form. In point of bulk, the quadrate of Pezophaps would 
seem to be nearly absolutely as large as that of Pidus. The intero-posterior cavities 
present very considerable individual variation both in shape and extent. 
The remaining parts of the skull are unfortunately wanting. 
§ 10. Comparison of the Skeleton. 
As the whole of the preceding description has been in a manner comparative, there 
remains little to be said that can be placed under this heading. Nevertheless, to pre- 
serve the uniformity of arrangement with Professor Owen’s treatise on the Dodo, we 
here include such general remarks on the subject as we have to make. 
The most important distinctions between Pezophaps and Pidus , as may be gathered 
from what we have already said, are on the whole such as not to remove the former 
further from the normal Columbce than the latter. In several respects (of which the 
enormous development of the cancellous boss on the top of the head, the general form 
of the lower mandible, and, above all, the extreme attenuation of the coracoid in Pidus 
may be mentioned as chief) Pezophaps holds a middle place between that form and 
ordinary Pigeons. On the other hand the armature of the wing is a character equally 
remote from either*. We had expected to have found that Pezophaps would prove to be 
a link between Pidus and Pidunculus ; but this is not the case : the latter seems to be 
a perfectly distinct form, and, in our opinion, sufficiently isolated from other Columbce to 
be considered the type of a separate family, in which light it has before been regarded ; 
but at the same time this family Pidunculidce must not be considered to be so aberrant 
as the Pididce, which we look upon as the most remotely connected of the Order 
Columile, consisting of three well-marked families of equal rank, (1) Didida:, (2) Colum- 
BiDiE, and (3) Didunculida], the first and last of which are only related to each other 
by means of the second, and have no direct affinity, however much superficial resem- 
blance there may be between the beak of Pidunculus and that of Pidus. 
The grounds on which Strickland (‘The Dodo’ &c. p. 54) established the genus Pezo- 
phaps, by separating it from Pidus, were merely “ the greater length of the legs and the 
less development of the beak.” The result of our investigations prove that he was amply 
justified in coming to this conclusion. Still we would desire to see his genus more 
firmly established, and accordingly we venture here to propound its characters in set 
phrase as follows : — 
Pezophaps, Strickland, 1848. Genus Columbarum Pidinum. 
Char, gener. Rostrum mediocre, curvatum, processu nasali et ramis maxillaribus 
antice divergentibus. Frons plana, porca osseo-cancellata circumdata. Ossa coracoidea 
* In Goura, as Professor Huxley lias lately most kindly pointed out to us, the wing is armed with a knob, 
but it is, even comparatively, much smaller than in Pezophaps. 
