﻿66 



28 



process mentioned above this epibranchial seems to be bifurcated. This is even 

 more the case with e[)ibranchial IV, which is much more slender than the fore- 

 going epibranchiais; its long process is directed oblicpiely forwards to connect with 

 the process on epibr. ///, whilst the distal (inner and longer) head-part is connected 

 by a cartilaginous end with the corresponding pharyngeal and likewise touches the 

 upper cartilaginous part of the pharyngeal in front. Of the 4 pharyngeals (p/i'—'^', 

 figs. 17 and 18) the one belonging to arch I is short, wedge-shaped and without 

 teeth; in A.strigata I find that it is not bone but cartilage; the next 3 form together 

 tlie epipharyngeals (upper "throat-bones"), and are beset on the surface towards 

 the throat with a number of pointed teeth; the pharyngeals // and /// are elon- 

 gated, narrow, the grou[) of teeth oval; pharyngeal IV is short, rounded, its tooth 

 plate of similar form. 



Of the gill -rakers it is mainly the outer (front) row which is well-developed; 

 the irmer (posterior) is quite wanting on arch /, is only indicated by some scattered 

 parts on arch //, more numerous and more distinct on /// and is most developed 

 on IV (it is hardly necessary to say that it is absent on V). The outer or front 

 row is specially long on arches / — ///, much shorter on IV; they are flat, pointedly 

 triangular, rod-shaped wlien seen in profile; their inner skeletal axes are not ossi- 

 fied on the anterior arches or but little ossified at tlie base; on IV tlae axes are 

 ossified to a greater extent; the ossifications arc not fused to the skeleton of the 

 arch; on V a small row of 4 — 5 short gill-rakers is present distally in front of 

 the teeth. Tlie inner row of gill-rakers is, as above mentioned, weakly developed 

 everywhere, and the rakers short and fine (easily overlooked between tlie numerous 

 papillæ on the throat), but usually with a small bony axis, especially on IV, where 

 however owing to the sliortness of llie gill-deft they do not occur on the epi- 

 branchiais*. 



The most important features of the brancliial a])paratus can be represented 

 in tabular form as follows: 



The shoulder girdle (PI. II, fig. 1 and text-fig. 19) has been very carefully 

 and correctly described by Starks (30, pp. 033—34) in A. strigata. On one point only, 

 but that a fairly important point, Starks has misunderstood the conditions; he 

 states namely (p. 634): "The hypercoracoid (scapula here, sc in flg. 1, PI. II and 



