ERIK A : SON STENSIO 
244 
trally by the mandibular canal and should therefore comprise not only the homologue 
of the dental of the Rhipidistids but also that of the splenial, and possibly also those 
of certain of the elements following between this and the angular. 
The medial side of the mandible is covered with a single large and long coronoid 
element, for which I have suggested the name mixicoronoid, as it presumably includes 
homologues of several of the elements that are found independent in Tetrapods and 
Rhipidistids. 
The hyomandibular is always large and well developed, but at least generally it 
is ossified mainly perichondrally. Like other cartilage bones of the hyoid arch it too 
has had cartilaginous epiphyses. As a rule it is somewhat angular, in such a way that 
its angle is directed backwards and enters more or less closely into contact with the 
operculum or the operculum and the suboperculum (Traquair, 1877 a). While the majo¬ 
rity of forms seem to show these primitive conditions, we find in Boreosomus and Acro- 
rhabdus a large, well developed processus opercularis issuing from the angle on the 
posterior margin, which must of course mean that the musculature of the gill-cover was 
developed in this two genera in about the same direction as in the higher ganoids and 
Teleosts. Neither in Boreosomus nor in Acrorhabdus have I found the hyomandibular 
pierced by any canal for the truncus hyoideomandibularis facialis. On the other hand 
I found a canal of this sort in Gyrolepis. The hyomandibular is in most cases situated 
very obliquely in such a way that the axis of length through it runs from above and 
from in front downwards and backwards; it is less frequently vertical or almost vertical. 
An ossified little symplectic is found in Acrorhabdus and in the so-called Oxygnathus 
browni (Broom, 1909 a). On the other hand no stylohyal is shown as yet With certainty. 
In the middle part of the ceratohyal there always occurs a large hour-glass-shaped 
perichondral diaphysis of bone. There are indications, however, that the cartilaginous 
epiphysis situated caudally of this had also begun to be ossified, i. e. we have here 
presumably the earliest known homologue to the bone in higher ganoids and Teleosts 
that has been incorrectly called the epihyal (cf. Gaupp, 1904, pp. 902—904). 
The preoperculum is generally a curved bone consisting of two shanks, a vertical 
narrow one, situated behind the maxillary, and a more or less horizontal and wider 
one lying dorsally of the maxillary and reaching far along towards the entrance of the 
orbit, thus covering a rather considerable portion of the postorbital part of the cheek. 
The bone under discussion may, however, be also developed in the form of a large, 
wide triangular plate (Acrorhabdus), by which, as we shall see, it bears a great resem¬ 
blance to that of the Catopterids. 1 ) 
In certain species, as, for instance, in Cheirolepis trailli Agassitz (Traquair, 1875 b, 
PI. 17, fig. 7, bone y), Elonichthys pectinatus Traquair (Traquair, 1901, pp. 83 , 84, fig. 1, 
bone x) and the so-called Oxygnatus brownii (Broom, 1909 a, p. 259), there occurs between 
the forward directed shank of the preoperculum and the operculum a long and narrow 
bone, element, which occupies such a position that it may perhaps have been pierced 
! ) In Rhadinichthys argentinicus Tornquist the reconstruction given by Tornquist (1904, pp. 846—351) 
seems to be incorrect, at least as regards the shape of the preoperculum. For as far as can be seen from.ToRN- 
quist’s own figure of the type specimen (PL 36), the preoperculum had two shanks, although the vertical one 
seems to be poorly preserved. 
