838 MESSES. A. AND E. NEWTON ON THE OSTEOLOGY OF THE SOLITAIRE. 
None of the other fragments (Plate XVIII. fig. 75), though undoubtedly belonging to 
larger (male'?) birds, could have come from individuals equalling in size that whose 
sternum is now at Paris, and believed to have been hitherto unique. Even making 
every allowance for the stalagmitic incrustation with which the specimen is encumbered, 
this bird could not have been less bulky than the largest Didus whose sternum has been 
preserved. 
The very near affinity of Pezophaps to Didus is nowhere better shown than on com- 
parison of the sterna of the two forms ; and it is self-evident on the most cursory 
inspection of them. So far as can be determined from the imperfect state of all the 
specimens of each, hitherto recovered, there is no distinction of any very great importance 
to be detected. The general configuration of the two — their extraordinary helmet-like 
form — is precisely similar and unlike that of any other bird known to us. It may be 
that the keel in Pezophaps is relatively somewhat deeper than that of Didus , but Didus 
varies not inconsiderably in this respect, and it would be very unsafe, from what we see 
of other parts of the skeleton, to assert that Pezophaps did not vary also. 
Yet a closer examination will reveal several characters more or less diagnostic. Pirst 
it may be remarked that the coracoidal grooves are shallower and narrower in Didus than 
in Pezophaps, and in all the specimens, save one of the last, there is a deeper depression, 
almost a little pit, at the inner extremity of the groove., in a line and, occasionally, to 
some degree confluent with it. Above this pit there is in one specimen a second, smaller 
and subtriangular shallow depression. This same specimen is the only one that exhibits 
any part of the costal processes, and even here their edges have been broken off. It has 
the same distal enlargement and concave extremity as Didus ; but the concavity is semi- 
divided by a well-defined ridge stretching halfway across the cup. A more remarkable 
difference is presented by the costal border in this specimen, which shows articular 
surfaces for four sternal ribs only instead of five, which is the normal number in Didus*, 
and, so far as can be determined from the broken state of the remaining specimens, there 
is nothing to induce the belief that they possessed more than four such surfaces. A third 
character which seems to prevail in Pezophaps (being found in all our six examples, and 
also remarkably well shown in the specimen of the Museum at Paris) is by no means 
constant in Didus , such indication of it as exists at all occurring there, so far as our 
experience goes, only in the smaller examples. This is the presence in Pezophaps of a 
very deep median hollow near the anterior end of the internal surface of the sternum — 
a hollow so deep that in five (including the Paris example) out of seven instances the 
bone is actually perforated, and in one the perforation (the external orifice of which is 
situated below the coracoid grooves) is of very considerable size. In most specimens of 
Didus where anything like this depression exists it seems to be connected with the 
* Too much importance must not be placed on this character. In one of the sterna of Didus in our posses- 
sion, five articular surfaces are exhibited on the left side and only four on the right. In another specimen all 
the five surfaces on the left side are bilobed, while on the right (as seems to be usually the case) this is so with 
four only, and the anterior surface is simple. 
