975 



the Idiiuitudinal dinuictei' of t lie eye; li:is jjrovctl to \:\vy 

 lictwcrii .")1 .-mil -'ti '~ of the letigth of the liead. The 

 hrMiichiostepil iiiciuhrane does not form so distinct ;in 

 anale "itii the inferior margin of the svd)o]>er('idnm; 

 this angle is sometimes almost imperceptiMe, the jioste- 

 rior (npiier) iiranrhiostegiil rays extending further hack 

 than the intei'o|iereulum, and the last of them lieing 

 obii(iuei\- truncate at the extremity or produced upwai'ds 

 to a hlunt point, which is applied to the lower margin 

 of the suhoperculum. The number of gill-rakers on the 

 first branchial arch we have found to vary between 4fi 

 and 51, thus fewer than in the Herring. The dentition 

 resembles that of the Herring, but is feebler, the teeth 

 being often imperceptible, and usually wanting on the 

 palatines and vomer, but often present on the ento- 

 l)ter\goid and ectoj)terygoid bones. The most importa.nt 

 difference from the head of the Herring consists, how- 

 ever, in the coinparati\ely short lower jaw, ^vhich is 

 alwa^•s shorti'r than the base of the anal fin, and the 

 length of whicli varies between about 47 and 52' 2 % 

 of that of the head. The length of the maxillaries too 

 varies in the Sprat between 89 and 84 ?», in the Her- 

 ring between 75 % (in exceptional cases 73 %) and 79 % 

 (in exceptional cases 81 %), of that of the lower jaw, 

 which is besides perceptibly shorter in the Sprat than 

 the distance between its hind extremity and the upper 

 angle of the base of the pectoral fin, in the Herring 

 equal to this distance or a little longer. 



The dorsal tin is of the same form and occupies 

 the same position as in the most developed Herrings. 

 At a length of body averaging 88 mm. the average 

 distance between the dorsal fin and the tip of the snout 

 proved to be 49'9 % of the length of the body, and in 

 specimens averaging 120 ram. in length the correspond- 

 ing percentage was 50'7. The distance between the anal 

 fin and the tip of tlie snout is somewhat less than in 

 the Herring, and varies lietween aljout 65 and 71 % 

 of the length of the body; but the base of the tin is 

 longer, its length varying between 13 and 15 % of that 

 of the body. The caudal fin is usually not quite so 

 deeply forked as in the Herring. The length of its 



middle r;iys is about ti' ., % of that of the body and 

 rather more than ' ,> — '/r, of that of the longest rays 

 in the inferior caudal lobe. 



The pectoral fins resemble those of the Herring in 

 their length as well as in other respects. The \entral 

 lins are in general shorter and have fewer rays than in 

 the Herring. Their length varies between S'/j and 7V2 % 

 of that of the bodv. Their position affords one of the 

 most imjiortaut characters of the Sprat, the distance 

 from the tij) of the snout to tlie anterior (upper) angle 

 of the base of these tins extremely seldom — and then 

 only slightly — exceeding ' / the length of the body, 

 and being less on an a\erage than that between the tip 

 of the snout and the beginning of the dorsal fin, rarely, 

 and then but slightly, greater llian the latter distance. 

 It is consequenth- the rule that the ventral tins are in- 

 serted verticallv below the beginning of the dorsal tin 

 or further forward. Tlie preabdominal length varies 

 between 28 and SO',,, %, the postabdominal between 

 about 17 and 22 %, of the length of the body. 



The scales are exactly similar to those of the Her- 

 rino-, and the most essential difference in the abdominal 

 plates has been remai'ked above. 



The coloration too is so like tliat of the Herring 

 that no constant difference can be adduced. The line 

 of lustrous green between the dorsal and ventral colours 

 is sometimes ribbon-shaped, rather sharply defined. Ac- 

 cording to Valenciennes'' this line is sometimes a gold- 

 en band; and Collett mentions' "a half-grown spe- 

 cimen" that had lost its scales, but was marked along 

 this part of the sides with two, almost black stripes. 



The geographical range of the Sprat does not ex- 

 tend so far north as that of the Herring, but else co- 

 incides therewith in the East Atlantic. On the Ame- 

 rican side the Sprat has never been found; and Faber's 

 statement'' that it occurs off the coast of Iceland is 

 not convincing, for he describes its operculum as stri- 

 ated, in distinction from the smooth operculum of the 

 Herring. But on Guntheu's authority' the Sprat has 

 been included, together with several other European 

 fishes-', in the fauna of Tasmania. 



" 483 — 50'4 °o, according to our measurements. 

 ' 1. c. p. 288. 

 ' 1879, 1. c. 



'' Fisclie hlawls, p. 178. 

 ' Proc. Zool. Soc. Lond. 1871, p. 072. 



■'■ 6V/<EH'i aquila, Zeus faber, Oaranv traclmyus, Stolephoms encrasicltolus, Oonjer vulgaris. Orthcgoviscus mola, Rhina squalina, Galeus 

 iiis, 6'quati(s (^Acanthias) Blainvillei, etc. 



1-23 



