260 



Appendices to Fifth Annual Report 



In the case of the articulation of the 2nd to the 3rd vertebra this also 

 holds good, but its articulation to the 1st, as also of the 1st to the head, 

 is different (PL XVI. figs. 2, 3). The anterior end of the 2nd vertebra is on 

 the dorsal side slightly hollowed on each side of the median line. Into 

 or rather over these hollows slight rounded projections on the dorsal edge 

 of the posterior end of the 1st vertebra project. Necessarily between 

 these two slight hollows on the 2nd vertebra there is a slight projection, 

 and this fits into a small hollow on the posterior upper edge of the first. 

 A similar but more pronounced arrangement occurs with the anterior end 

 of the 1st vertebra. In it there is a distinct dorsal median process, 

 strengthened by a couple of diverging bars or rather thickenings on its 

 upper side. This process projects forwards, aud fits into a corresponding 

 median hollow or pit in the posterior dorsal surface of the basi occipital — 

 the basioccipital overlapping the 1st vertebra on each side of this 

 process just as the 1st overlaps the 2nd. The ventral edge of the 1st 

 vertebra shows generally a very slight curve forwards, fitting into a 

 corresponding shallow concavity on the posterior (ventral) face of the 

 basioccipital. The basioccipital, viewed from above, appears to overlap 

 (except in the middle) the atlas, and the atlas overlaps the 2nd vertebra. 



The Haemal Arches. 



On the 2nd to the 24th or 25th vertebrae are found certain structures, 

 which may be called basal pieces, to distinguish them from bony processes 

 (haemapophyses),* forming more directly the haemal arches. These are some- 

 what triangular-shaped pieces of a spongy bony nature, and which lie in the 

 latero-ventral chambers of the vertebrae (PI. XVI. figs. 2, 4, 14, bp.). There 

 are, therefore, a pair to each vertebra, one on each side. They are more or 

 less firmly fixed in these positions, but there is no ossification connecting 

 them absolutely with the vertebra?., and they can be removed pretty 

 readily from their position ; they often, indeed, drop out when dried, in 

 both the adult and the very young (30 mm. long) herring. In the young 

 herring of this size the vertebrae have the thinnest possible wall of ossifi- 

 cation formed round the notochord, and the latter is not yet constricted, 

 the whole inside of the ossified vertebra being of almost the same 

 diameter throughout its length. Pound each end, however, ossification 

 has proceeded further, forming a slight bony riug, and the septa described 

 above as being present on the external parts of the vertebrae, are in such 

 a young fish also evident as narrow bony bars, which have already the 

 sharp terminal articular processes developed. on them. The two partitions, 

 however, which form the ventro-lateral chamber in which the basal pieces 

 lie in the adult, have in the young fish not separated throughout the 

 entire length of the vertebra. The posterior third of the vertebra shows 

 them only as a single bar, which, passing forwards, bifurcates, the two 

 bars curving round at the anterior end to meet again, and thus form an 

 oval-shaped shallow cavity in which sits the basal piece. The basal piece 

 at this stage then is only from half to two-thirds the length of the vertebra, 

 or smaller according to the size of the fry (PI. XVI. fig. 14, bp.), while 

 in the adult it occupies as far as possible the full length of the vertebra. 

 The basal piece in the young fish is tooth-shaped, having an oval-shaped 

 base, and stands nearly directly outwards and downwards, and curved 

 slightly backward terminates in a blunt-pointed tip. In the adult it lies 

 more outwards, showing a flat side on ventral view, and is more equila- 



* I have limited the term 'transverse process ' to the long appendages of the ribs 

 (posteriorly of the vertebrae). ' Haemaphphysis ' being applied to the processes 

 which form the haemal arch. 



