of the Fishery Board for Scotland. 



55 



the gill rakers number about 56 to 60 on the first arch, about 44 on the 

 second, 42 on the third, and 36 on the fourth. In the sprat there are 

 about 48 on the first arch, 44 on the second, 40 on the third, and 33 on 

 the fourth. So with the branchial filaments there is in the herring the 

 same preponderance of numbers over the sprat, and this applies also to 

 the pseudo-branchial filaments, which number 14 in the sprat and about 

 20 in the herring. The herring has one or two branchiostegal rays more 

 than the sprat, which has 7. 



The only notable point of difference in the alimentary system consists 

 in the number of pyloric caeca. In the sprat there are always 7 or 8 

 pyloric caeca ; the openings from these caeca into the duodenum form a 

 single row, the first 4 or 5, together with the very minute opening of the 

 bile duct, being arranged round the opening from the pylorus (stomach), 

 the remainder forming a straight row backwards. In the herring there 

 are 18 to 24 of these caeca; they vary slightly in position, but 14 or 15 

 of them generally form a double row round the pyloric valve, the 

 remainder passing backwards in a rather irregular double row. The bile 

 duct opens in the same position as in the sprat. The stomach (crop of 

 Huxley, Nature, Ap. 28, 1881) communicates posteriorly by a slender 

 duct, about *5 mm. in diameter, with the swim-bladder, and this duct 

 is generally slightly shorter and thicker in the sprat than it is in the 

 herring. 



In the herring the swim-bladder is known to give off two delicate branches 

 from its anterior end, which run forwards, one, at first, along each side 

 of the parasphenoid, from which they afterwards diverge, and soon after 

 enter a small oval or rather spindle-shaped capsule. From the anterior 

 end of this capsule the duct passes out and divides into two branches ; 

 one branch runs straight forward from the spindle-shaped vesicle, the 

 other passes outwards at nearly a right angle, and each terminates at the 

 ear in a spherical capsule. This description applies to all the young 

 herring which I examined. In them these ducts are very delicate tubes, 

 measuring in a herring of 120 mm. long *09 mm. in diameter, and are 

 surrounded by a cartilaginous sheath of '25 mm. external diameter. The 

 ducts meet posteriorly in the middle line, and open by a single aperture - 

 into the narrow anterior end of the swim-bladder. The spindle-shaped 

 capsule measures about 1 mm. long by *6 mm. broad, and the spherical 

 ones are about 1 -3 mm. in diameter, the anterior one being slightly the 

 larger. There is, however, a rather remarkable departure from this 

 arrangement in the sprat. The ducts are about the same size as, and 

 their form and direction is similar to, those in the young herring, although 

 from the point where they diverge outwards from the parasphenoid (about 

 5 mm. from the swim-bladder) they lie rather higher than in the herring, 

 and are more difficult to follow. But the duct of each side terminates in 

 a single capsule only, similar in all respects to that of the herring, and 

 there is neither formed on its course the spindle-shaped dilatation, nor 

 does it give off a branch to a second vesicle, as in the herring, none such 

 existing in the sprat. Thus, while in the herring there are three vesicles 

 on each side of the head, all containing air, in the sprat there is only one 

 (Plate III., figs. 3, 4, and 5). 



While these characters are so numerous, and so constant at all seasons, 

 that there can be no question as to the sprat being a distinct species from 

 the herring, a further proof is to be found in the fact that the former 

 is found with developed milt and roe. During the winter months the 

 reproductive organs of the sprat and young herring which I received 

 could not be distinguished from each other ; but about March it was 

 noticed that in a small proportion of the sprats the milt and roe were 

 increasing in size, the former showing the ova scattered more or less 



