488 DR. W. G. RIDEWOOD ON THE CRANIAL [Dec. 13, 



continuous with the pair of great cartilages which curve upward 

 from the suturally united posterior ends of the fourth a,nd fifth 

 ceratobranchials. The upper ends of these great cartilages meet 

 in the middle line above, but do not fuse ; from the anterior edge 

 of each there stretches forward an extensive tract of thin, but 

 tough membrane, which fills in the angle between the body of the 

 fourth epibi-anchial and the remarkably long process of the same 

 which slopes upward and backwaixl from near its anterior end. 



The first three epibi-anchials are of approximately the same 

 length, but the fourth is longer and wider. The first pharyngo - 

 branchial is a forwardly dii'ected cartilage of conical shape, and 

 there is an upright spicular bone rising from the junction of this 

 with the fi-ont of the first epibi'anchial. There are no teeth on 

 any part of the hyobi'anchial skeleton. 



Summary, 



On comparing the foregoing accounts of the cranial osteology 

 of Ghirocentrus, Gliqjea, Pellona, Pellonula, Pristigaster, Hyper- 

 lophus, Chatoessus, Dussimiieria, Engo'coulis, Goilia, and Chanos, 

 the most important features of resemblance and difterence appeal- 

 to be as follows. 



The parietal bones are rather small in size*, and are separated 

 the one from the othei- by the supraoccipital. In Ooilia, however, 

 they nearly meet in front of the supraoccipital, and in Chanos, 

 owing to the fusion of the commissuiul sensory-canal bones with 

 the pai'ietal bones, the latter appear to meet over the top of the 

 supraoccipital f. There is a fontanelle between the anterior ends 

 of the two frontal bones in Chirocentras, Clap)ea, Pellonula, 

 Pellona (a mere cleft in this genus), and Chatoessus, but not in the 

 othei- six genera. In Coilia and Engraidis the mesethmoid pro- 

 jects considerably in advance of the vomer, whicli is not the case 

 in the other genera. 



Of the genera under consideration, the only one in which the 

 postei'ior temporal groove is roofed ovei' is Chanos. The temporal 

 foramen appears to be a special feature of the Olupeoid skull. It 

 is an aperture, usually oval in shape, with the long axis horizontal, 

 liounded by the frontal and parietal bones. In no instance have I 

 found the postfrontal bone fonning part of the boundary, although 

 Boulenger regards this as the normal condition J. The only 

 departure fi'om the general rule that has come within my know- 

 ledge is in the case of Engi^aulis, in which a small poi-tion of the 

 squamosal may come between the parietal and the frontal in the 



* Small is, of course, but a relative term, aud opinions may differ as to the emplo}'- 

 ment of the word. " Tres petits," the expression used by Boulenger (Poissons du 

 Bassin du Congo, 1901, p. 123), appears to me to convey an exaggerated idea of the 

 smallness of the parietal bones. I leave the figures that illustrate this paper to speak 

 tor themselves. 



t Cope, it is worth noting, included the Lutodirida3 (■/. e. Chanidse) under the 

 heading "Parietals united " (Trans. Amer. Phil. Soc. n. s. xiv. 1871, p. 455). 



X Poiss. Bass. Congo, 1901, p. 123. " Un grand trou de chaque cote du crane 

 ])orde par Ic frontal, le postfrontal, et le parietal." 



