The osteology of Cromeria nilotica and Galaxias attenuatus. 69 
supraoccipital. It is, however, quite conceivable that any or all of 
these features might appear in a derivative of the Galaxiad type. 
In spite of this extreme specialization, however, Cromeria exhibits 
features which show that it must have been derived directly from a 
more primitive stock than is represented by Galaxias. This is most 
strikingly shown by the branchial apparatus and primary pectoral 
girdle. It has already been pointed out that the fourth and fifth 
basibranchials are better developed than in Salmo, Alepocephalus, or 
even Amia: for, in the first of these, the representative of the fifth 
element extends but little further back than the ventral extremities 
of the last ceratobranchials; and in the last two, though the fifth 
basibranchial is longer than in Salmo, it, and also the fourth, may 
be entirely cartilaginous. On the other hand in Cromeria (Fig. J) 
these elements not only show no trace of diminution in size, but are 
both ossified. It is of course not impossible that the ossification of 
these elements is due to the same causes as the very complete 
ossification of the cranium; if this be the case, then ought also the 
hypobranchials of the fourth arch (br. h. 4), the first basibranchial, 
and the basihyal to be ossified — but they are not. Again, it may 
be wrong to assume that a series of basibranchials all ossified, is 
more primitive than one almost equally developed but not equally 
osseous. This, however, is a question for future investigation 
to decide, and does not affect the main point under consideration; 
for the mere presence of a fifth basibranchial in Cromeria, and its 
absence in Galaxias, points to a more lowly ancestral stock for the 
former than the latter represents. 
This is further emphasized by a consideration of the primary 
pectoral girdle, for the development of a mesocoracoid in Cromeria and 
its suppression in Galaxias leads to the same conclusion. True this bone 
is slender, and might almost be described as being on the verge of 
disappearance; but even then the general shape of the rest of the 
girdle approaches much more nearly to the primitive Salmonoid type 
than to that of Galaxias. Evidently therefore Cromeria was not 
derived from a Galaxiad stock. 
Other peculiarities ot the branchial skeleton suggest that Cromeria 
is a derivative of the Malacopterygian or Isospondyloid stock, but of 
a different offshoot from that to which the other types above mentioned 
belong. Thus, in the latter the basibranchial skeleton seems to be 
degenerating postero-anteriorly, that is to say the anterior elements 
are strong and well ossified and the posterior are much reduced and 
