MESSES. A. AND E. NEWTON ON THE OSTEOLOGY OF THE SOLITAIEE. 347 
forward on either side, a somewhat dish-like formation is given to the top of the 
head *. 
The frontals rise abruptly as in Didus, but, as will have been gathered from what has 
just been said, to a much less extent. The cranium also in this place is much com- 
pressed, the largest specimen measuring only T76 inch at the narrowest part between 
the orbits, and the interorbital septum is not so thick as in Didus , but composed of the 
same cancellous structure. None of the specimens are sufficiently perfect to show the 
entire extent of the posterior olfactory chambers, which are, as in Didus , partially 
divided. At the junction of the nasals with the nasal process of the premaxillary, 
the proximal portion of which is not preserved in any of the specimens, the surface 
is extremely rugose, indicating the position of the caruncular ridge, which, from 
Leguat’s figure and description, we already knew the bird to have possessed at the base 
of its bill. 
In the upper mandible (Plate XXII. figs. 151-158) we have presented one of the 
most remarkable differences between Didus and Pezophaps. Instead of that singular 
characteristic of the former (which has been duly dwelt on by previous investigators), 
the vertical enlargement and coalescence, before reaching the core of the dertrum, of the 
maxillary branches of the premaxillary, we have in the latter a structure much more 
in accordance with the normal Columbce , and presenting an appearance laterally though 
not from below much as in Treron. The maxillary branches of the premaxillary in 
Pezophaps approach each other gradually, but do not coalesce until about halfway 
between the commencement and extremity of the core of the dertrum, and thus they 
do not meet the nasal process of the premaxillary at an acute angle immediately behind 
the core of the dertrum, but the core of the dertrum intervenes vertically (Plate XXII. 
figs. 152, 155, 157). In other words, the axes of the nasal 'process and maxillary 
branches in Didus are convergent, in Pezophaps divergent. 
There is remarkable variation in the size of the upper mandible in different indivi- 
duals to the extent of very nearly one-half the linear dimensions between the largest 
and smallest specimens, of which the collection contains thirteen in all, some of which, 
however, are merely fragmentary, and the best exceedingly imperfect. No portion of 
the palatals is preserved. 
Of the lower mandible (Plate XXII. figs. 159-162) the collection contains thirty-four 
fragments, one-half belonging to the distal and the other half to the proximal end. It 
has not been possible to fit together with certainty any two of them so as to form a 
perfect ramus. To one of the dentary portions, however, part of the core of the sym- 
physis (Plate XXII. fig. 162) remains attached, and there is a second such core which 
is perfect though detached from both its rami. Still the exact outline of the lower as 
of the upper mandible must be left in some degree a matter of conjecture. Enough, 
however, is present to prove that it was sufficiently different in shape from that of Didus , 
* It is proper to observe that in tbe only portrait we possess of the bird, that of Leguat, there is no trace 
of this formation to be found. 
3 a 2 
