ERIK A : SON STENSIO 
104 
situated some distance ventrally of the canal itself. The tubes are wide and unbranched. 
There seem to be at least two pores in the splenial. 
According to the definition of the infraorbital canal above, the cranial section of 
the main lateral line of the body extends in the Coelacanthids up to the suture between 
the parieto-intertemporal and the supratemporo-extrascapular. From this suture, where 
it forms a continuous junction with the infraorbital canal, the cranial section of the 
main lateral line first passes through the supratemporo-extrascapular, which it leaves on 
the medial edge near to the posterior end. Afterwards, as far as one can see, it 
traversed a small membranous interspace. Here it has a supra-temporal branch, which 
continues through the extrascapulars towards the median line and there presumably joins 
that of the opposite side, forming a continuous supratemporal commissure (text figs. 43, 44. 
.v. com, Pis. 11— 13 ; PL 14, fig. 1). Immediately behind this commissure the section in 
question of the lateral line has turned in a more purely caudal direction, being in its 
course just for a short distance situated close to the medial edge of the supratemporo- 
extrascapular. Near the posterior end of this bone it leaves the membranous interspace 
and enters' a small rounded scale-like plate probably representing the suprascapular 
(.Sscap , text figs. 43, 44; Pis. 11, 12, i 3 ; PI. 15, fig. 1). — Nothing is known with regard 
to pores in this part of the lateral line. 
Neither on the cranial roof nor on the cfieek have I been able to observe «pit- 
lines» or anything - corresponding to these. 
If we now examine the Rliipidistia as regards the sensory canals, we shall find 
in them, at least to judge from Osteolepis (Goodrich 1919), Eustlienopteron (text fig. 57) 
and Dictyonosteus (text fig. 58), a very great resemblance to A. robusta, W. sinuosa and 
other Coelacanthids. A comparison between the same forms with regard to the relation 
between the sensory canals and the dermal bones, on the other hand, also clearly shows 
that several of the elements that are found separate in the Rhipidistids must in the Coela¬ 
canthids, on the contrary, be united to one another in about the way I have tried to 
indicate by the nomenclature used in this work. From the relations of the infraorbital canal 
it is thus easily understood that the homologues of the parietal and the intertemporal of 
the Rhipidistids have coalesced to form the parieto-intertemporal of the Coelacanthids, and 
that in the same way the homologues of the dermosphenotic and frontal of the Rhipidistids 
together have formed the fronto-dermosphenotic of the Coelacanthids. The conditions of 
the supraorbital canal also show clearly that the homologue of the nasal of the Rhipidistids 
and that of their antorbital may be united to the nasalo-antorbital of the Coelacanthids, 
and finally the cranial section of the lateral line together with its supratemporal commissure 
shows that in the Coelacanthids the homologue of the supratemporal of the Rhipidistids 
has incorporated with itself certain parts of the lateral-extrascapular («tabular»), so 
that it may quite correctly be called the supratemporo-extrascapular. 
The course of the mandibular canal indicates that the row of independent elements 
that, at least in a number of Rhipidistids, usually follow in front of the angular 
(«infradentals», of Traquair’s descriptions) is divided in Coelacanthids among the splenial 
and supraangulo-angular. It is, however, impossible to establish more closely which of 
these in this case forms part of the one or the other bone mentioned. On account of 
its considerable extension dorsally it is probable, as I have already pointed out, that 
