﻿48 • 10 



outside as dermal parts in many other bony fishes, so that tlieir surfaces seem to 

 be exoskeletal for some distance, is not unknown; as examples I need only men- 

 tion Gastrosteiis, Trigla. All the other elements of the dorsal armour in Amphisile 

 are purely exoskeletal. This view, which is also quite natural, is further supported 

 by the fact that the lateral line courses through these bones. 



The lateral line canal passes out from the pterotic (sq Tab. I, fig. 1) into 

 the anterior, lower corner of dorsal plate 1, pierces this in an arch to the poste- 

 rior, lower corner and then further through the succeeding dorsal plates, forming 

 a flattened arch on each from the anterior to the posterior, lower corner, and 

 finally running in a straight line through the 5th dorsal plate and from the poste- 

 rior point of this into the ventral margin of the dorsal spine. Here it ends with 

 a pore; in A. scutata almost in the centre of the ventral aspect; in a specimen of 

 strigata 134 mm. long, I find the end pore S mm. from the articulation with the 

 movable spine; in punctulata the canal only extends a short distance on the dorsal 

 spine. On each of the dorsal plates (except the first) 1 find a single, ventral pore 

 close behind the highest point of the arch. The position of the lateral line is thus 

 remarkably near to the dorsal edge of the body and it has become connected in 

 an extremely characteristic manner with structures (sc. interspinous bones, T) which 

 do not usually take part in the lateral line system, perhaps because these struc- 

 tures are here in continuation of the main portion of the body. 



The peculiar features of the sutures in the dorsal armour must also be men- 

 tioned. All the connections between the plates on the same side of the body are 

 provided with long teeth fitting into one another; on the cross sutures these are 

 generally vertical to the line of articulation, but in the longitudinal sutures they 

 are obliquely placed. The lateral plates alternate somewhat with the dorsal plates 

 in such a way that a short, anterior portion of the upper suture of each lateral 

 plate meets one dorsal plate, whilst a longer, posterior portion is joined to the 

 next; on all the short portions the teeth go one way obliquely from in front back- 

 wards, on all the longer portions in the opposite direction, i. e. obliquely from 

 behind forwards. The only simple sutures are those in the middle line of the 

 dorsum, the connection between lateral plate I and the supraclavicle and further 

 the longitudinal suture on the dorsal spine in the two species where a movable 

 spinous ray is found. 



All the plates of the dorsal cuirass as well as the true skeletal parts united 

 to them show a characteristic sculpture. On most we find rounded longitudinal 

 striæ, which are again connected by numerous, short, irregular cross-lines, forming 

 together rounded pits; wherever the margins form jagged sutures, groups of stronger 

 striæ radiate out in the direction of the teeth from a spot in the upper half of 

 the plate, but almost at an equal distance from the posterior and anterior margins. 

 On some plates, e. g. the dermal part of the clavicle, the pitted sculpture is predo- 

 minant, and the longer striæ scarcely occur; on the other hand, the longitudinal 

 striæ are predominant on the dorsal spine. The lateral plates become very thin 



