150 “TERRA NOVA” EXPEDITION. 
anal. It seems possible, therefore, that the backward position of the dorsal fin of 
larval Clupeoids may be an ancestral feature. 
In Notolepis (p. 125, Pl. I, figs. 4, 5), Paralepis (cf. Prymnothonus, p. 137, Pl. VII, 
figs. 1-3), etc., the larva has a short gut with the anus not far behind the head ; as 
the fish grows the gut becomes relatively longer, and the anus travels backward until 
it reaches its final position a short distance in front of the anal fin. In a paper on 
the classification of the Iniomi (Ann. Mag. N. H. (8) VII. 1911, pp. 120-133) I have 
pointed out that the long-bodied and long-snouted Paralepis is closely related to, 
but more specialized than, Chlorophthalmus, a short-snouted fish of normal piscine 
form, but with the anus nearer to the head than to the anal fin. It is not impossible 
that the anterior position of the anus in the larval Paralepis may be due to the 
evolution of this genus from a Chlorophthalmus. 
3. DEVELOPMENT OF THE FINS. 
The pectoral fins, which are principally concerned with balancing, are -usually 
present in the youngest larvae, and the permanent fin-rays appear at a very early 
stage. As a rule the caudal, used for propulsion, is the next fin to develop. The 
hypurals and fin-rays make their appearance below the notochord, and then by flexion 
of the latter are brought into a terminal position. Afterwards the dorsal and anal - 
rays develop in the embryonic fin-fold and the pelvic fins grow out. There are many 
deviations from and exceptions to the general course of development of the fins 
outlined above. 
The dorsal and anal fins usually originate in their final position, but in the 
Clupeidae, and the related Albulidae and Elopidae, they develop more posteriorly and 
the dorsal especially migrates forwards for a considerable distance after it has been 
formed (ef. Pl. V, figs. 2-4). In these families the dorsal fin develops earlier than the 
anal, but in the Iniomi the reverse is the case (e.g., Synodus synodus, p. 137, Pl. VII, 
fig. 4; Prymnothonus, p. 137, Pl. VII, figs. 1-3). Sometimes the formation of the 
dorsal fin, or the anterior part of it, is so delayed that the embryonic fin-fold disappears 
first ; this may often happen with the spinous dorsal fin of acanthopterous fishes (e.g., 
Oda balteatus, p. 143, Pl. VIII, fig. 4; Tripterygium varium, p. 145, Pl. IX, figs. 1, 2). 
Notable exceptions to the general rule that the anterior or spinous portion of the 
dorsal fin develops later than the posterior or soft-rayed portion are the Trichiuroids 
and some Paralichthmae. In the former (e.g., Thyrsites atun, p. 144, Pl. VIII, figs. 1-3) 
the early formation of the spinous dorsal may be for purposes of defence, and one may 
well believe that it would be effective in warding off the attacks of the predaceous 
young of such fishes as Pomatomus saltator (cf. Agassiz and Whitman, Mem. Mus. 
Comp. Zool. XIV, No. 1, 1885, p. 16). 
In some Flat-fishes of the sub-family Paralichthinae (e.g., Ancylopsetta, p. 146, 
Pl. IX, figs. 3, 4) the anterior dorsal rays are formed at a very early stage and grow 
