ARGENTINA SILUS ON THE SHORE OF MORAY FIRTH. 335 



commences about 1J inch from the head ; ventrals nearly opposite 

 where the dorsal ends ; pectorals close to the gill-covers. The 

 fin-rays, so far as I could make them out, are as follows : — D. 9, 

 P. 13, V. 14, A. 11 to 12, C. 24. 



I may mention, however, that the fins were somewhat hardened 

 and contracted, the fish having been dead some days before it 

 reached me. 



The most important and peculiar characteristic of the fish is 

 the shape of its body. In this respect it is decidedly in part 

 hexagonal and in another heptangular. Thus in a great measure 

 its small contracted mouth resembles the SyngnathidaB, or pipe- 

 fishes ; but at the same time it differs widely in every other 

 respect. In the adipose appendage and the disposition of its fins 

 it approaches the Salmonida? ; but here again it differs as mate- 

 rially, if not more so, from this tribe in the form of its body. 



Another special and interesting peculiarity, which is not found 

 in either of the families alluded to, although met with in the keel 

 of the Clupeidae, is that each division of the body is distinctly 

 separated by a visible ridge, which, in most cases, are minutely 

 but very sharply serrated. These serrations may be felt by pass- 

 ing the finger downwards from the head. The sensation and 

 peculiar irritation, though not an unpleasant one, resembles that 

 given to the hand when rubbed gently along a very finely cut 

 saw. This is particularly the case with the parts at the centre 

 of the belly, which latter, like the back, is distinctly divided by 

 a raised ridge or keel ; the serrations here seem to be stronger 

 and larger than on the other ridges. 



The back is nearly f of an inch broad, the first compartment 

 about ^ inch, the other a little more and bounded by the lateral 

 line, which is straight. The stripe immediately below it is the 

 broadest of all, being l^inch, the others get narrower as they 

 descend. Belly not quite so broad as the back. 



As the number of divisions so are the number of colours, each 

 being mostly of a different hue. That on the back would seem 

 to have been of a deep amber tint, the next, sidewise, being 

 greenish blue, the other silvery grey, the one below the lateral 

 line being of a pure silvery white, with a bright metallic lustre, 

 the others only a little fainter, but very similar to those above; 

 belly greyish white, with a tinge of greenish blue and purple. 

 The head on the top had been nearly the same as the back, with 



27* 



