No. i.] MORPHOLOGY OF THE PETROSAL BONE. 7 



ary separation of the bone into two parts by the optic fenestra. 

 The bone is, however, a variable and inconstant element of the 

 piscine skull. It may be wanting ; may be preformed in car- 

 tilage (Amia), or in membrane only (No. 32, p. 255) ; may be 

 fused with its fellow of the opposite side of the head (No. 27, 

 p. 68) ; and may be, as in Polypterus, fused also with both the 

 alisphenoids and the teleostean basisphenoid to form a single 

 impair bone, called by Traquair (No. 31, p. 170) the sphe- 

 noid bone, and by Pollard (No. 23), the orbitosphenoid. In 

 Traquair's figures (Figs. 2, 3) the bone seems even to extend 

 backward beyond and around the pituitary fossa. In Pollard's 

 figures (Fig. 12), taken perhaps from a younger specimen, it 

 does not extend so far. 



Where the orbitosphenoid, in fishes, is found impair, and 

 the cartilage of the interorbital part of the basis cranii has not 

 disappeared, as in Polypterus and certain of the Characinidae, 

 the median part of the bone lies on the dorsal surface of the 

 cartilage, and is not seen from the outer ventral surface of 

 the skull (No. 31, p. 170, and No. 27, Figures). 



These facts all seem to indicate that the orbitosphenoids and 

 so-called basisphenoids, once having appeared in any fish, as 

 ossifications preformed either in cartilage or in membrane, can 

 extend themselves in all directions, not only in the contiguous 

 or adjacent cartilage but also, and by preference, in the mem- 

 branes that line the optic and pituitary parts of the skull, and 

 fill its orbital and interorbital openings. The same would seem 

 to be true of other vertebrates also, for in man the orbitosphe- 

 noids of opposite sides of the head, in their later development, 

 send, according to Sutton (No. 29, p. 582), "a thin lamella 

 across that portion of the presphenoid which is anterior to the 

 optic groove, thus excluding it from the cranial cavity." As 

 the presphenoid first appears, according to Sutton, "on the 

 deep aspect of the perichondrium," these thin lamellae of the 

 orbitosphenoid must necessarily lie entirely superficial to 

 the subjacent cartilage. Moreover, the sphenoid bone in man 

 is subject to certain variations (No. 24, vol. ii, pt. i, p. 47) 

 which are further important indications in this connection. A 

 middle clinoid process is often found, sometimes connected 



</* **- 



