ON CERTAIN CROSSOPTERYGIANS. 1263 



The fenestration has thus approximately the same position as 

 tlie canalis transversus of the Elasmobranchii (canal for the 

 pituitary vein), and must certainly be considered as such a canal, 

 though it has becojiie very much widened and in addition 

 shortened in a transversal direction. From its posterior surf ace, 

 which is somewhat wider than the anterior one, a short fossa 

 penetrates in a backward direction between the base of the 

 primordial neurocranium and the parasphenoid. To define it 

 more closely, this fossa is so situated in the post-Devonian 

 Ccelacanthids, that it is bounded dorso-caudally and caudally by 

 the corpus of the basisphenoid and on either lateral side b}^ the 

 ventral process of the basisphenoid, while the ventral wall is 

 formed by the parasphenoid. In the Devonian form Diplocer- 

 cides hayseri (Stensio, 1922 «, pp, 178-179, 205-206) the fossa 

 is situated in quite the same place as in the post-Devonian 

 forms, but as there the basisphenoid is contained in and forms 

 a postei-o-ventral part of the lai-ge sphenoid, it is consequently 

 bounded there dorsnlly, caudally, and laterally by the pars basi- 

 sphenoidea sphenoidei." 



" In fishes in general, as we know, the recti muscles of the eye 

 have their place of origin postero-ventrally of the opticus and 

 oculomotorius foramina or else postero-ventrally of a line that 

 can be drawn through these two foramina. In the Ccelacanthids 

 the place where one must suppose the musculi recti to have 

 originated is occupied by the widened transversal canal, and we 

 can, therefore, scarcely help thinking that an expansion of this 

 canal has been caused by the recti muscles. The recti externi 

 have probably, as usual, invaded the transversal canal and then 

 pushed postero-ventrally into the posterior wall of this canal, 

 forming the short fossa described in this wall. The other three 

 musculi recti have perhaps partly had their origin on some 

 vertical membrane, which may have filled the anterior part of 

 the transversal canal ; most probably, however, they ought to 

 have originated for the most part from the dorsal surface of the 

 yjarasphenoid on the sides of or in the transversal canal itself. 

 Whatever the conditions may have been with regaixl to the 

 origin of certain of the recti muscles, it seems, however, 

 impossible not to suppose that in the expanded transversal canal 

 and the fossa opening into this from behind we are concerned 

 with a sort of myodome."' 



" "Watson has recently (1921, p. 335) doubted this explanation 

 of mine, which I put forward in an incomplete form in the 

 former part of this work. He then laid stress especially on the 

 fact that the space that would be regarded as a myodome could 

 not under any circumstances be homologous with the myodome of 

 the Palseoniscids, Avhich of course is situated dorsally of the 

 basipterygoid process and laterally of what he calls ' the body of 

 the basisphenoid.' It is of course true that the relation of the 

 myodome to the basipterygoid process is quite difi^erent in 

 Palseoniscids and Ccelacanthids, but, as I believe I can show, 



