268 



Appendices to Fifth Annual Report 



include, of course, the first two, and also this modified terminal vertebra. 

 While calling it terminal, however, because it appears to be so to a super- 

 ficial examination of its condition in the adult herring, it is really not so. 

 This vertebra turns slightly, but distinctly, upwards — indeed, a slight 

 tendency upwards is sometimes noticeable in the previous vertebra — but if 

 the remnant of notochord is removed its central perforation is still perfectly 

 patent. Beyond it come two small nodules, which continue still further 

 the upward or dorsal direction of the vertebral column from the previous 

 general straight horizontal line. In some adult herrings these nodules are 

 scarcely to be distinguished as more than a small bony mass or rod ; in 

 others they are distinctly to be made out as separate bones ; and in 

 favourable examples they bear the appearance of much modified vertebrae, 

 contracted in size, but with distinct evidence of original separateness, and 

 possibly showing slight traces of the central notochordal perforation. 

 That these are really vertebrae is completely proved by an examination of 

 herring fry of 25 mm. to 35 mm. or more in length. 



In these cases there are found at least three more vertebrae beyond the 

 one described above as the last. This row of vertebrae is formed round 

 the turned-up portion of the notochord, which is usually described as 

 having a single urostylar bone developed around it. These vertebrae 

 are similar to the others, except that they are not provided with distinct 

 zygapophyses, and that the first two have their posterior diameters rather 

 less than their anterior. This in the last vertebra seems more pronounced, 

 owing to its rather greater length than that of the others. This increase 

 in length I believe, from the examination of several specimens, to probably 

 proceed from its representing two vertebrae. The narrow terminal end 

 of this vertebra very commonly becomes closed up, partly by dry cellular 

 matter, partly by encroaching cartilage. All these vertebrae in the herring 

 fry distinctly surround the notochord, and are so complete that with care 

 they can be readily separated from one another and strung on a hair. 

 These terminal narrow vertebrae which I have described appear to be 

 homologous with the more elongated ossification in many other fishes, 

 which forms the urostyle, and which is probably only a coalescence in 

 these more specialised forms of several vertebrae (PI. XVIII. fig. 28).* 



Taking in the adult herring the vertebra above described as the terminal 

 one, to be the 57th in position, we find that the 54th, as originally described, 

 is normal in all respects — i.e., it bears a neural and haemal arch, each 

 provided with a spine, free at its tip. The 55th vertebra, however, while 

 it is provided with the normal free-ending neural spine, has its haemal 

 spine thickened, especially towards its distal extremity, which is hollow, 

 or filled merely with dry cellular matter. This spine acts as a support to 

 certain accessory rays ventral to and at the root of the main part of the 

 caudal fin. The 56th vertebra, as has been said, may be very slightly 

 turned upwards — it bears a normal neural spine ; but the haemal spine, 

 while thickened like that of the 56th vertebra, in addition is expanded on 

 its ventral (anterior) face into a thin bony (laterally compressed) lamina, 

 which fills up most of the space intervening between the 56th haemal spine 

 and itself. These, of course, are still perforated, or form arches for the 

 blood-vessels. 



The so-called terminal vertebra (the first which distinctly inclines 

 upwards) is somewhat reduced in diameter at its posterior end — so much 



* This formation of the vertebral bodies round the upturned portion of the noto- 

 chord is more clearly seen in the adult trout, where the upward inclination is more 

 gradual, the vertebrae more persistent in their normal size and form than in the 

 herring, but where the general arrangement of the vertebrse and their neural and 

 haemal arches are similar in nearly all respects, except that there is a more pro- 

 nounced urostyle. 



