OF SILUEIAN AND DEVONIAN. 51 



the ridges of the surface rtiu across the border, and the oval area is, therefore, not a sep- 

 arate plate (see fig. 2). Behind the oval sjaace the ridgelets are generally parallel to the 

 longer diameter of the sente, and their extremities near the triangiilar area are looped, or 

 retvxrn upon themselves, making an irregularly defined border across this part of the 

 dorsal scute. The parallel ridges on the main part of the dorsal scute frequently 

 anastomose. 



In all that part of the plate which lies behind the triangular and the oval areas, one 

 can distinguish two sets or ranks of the ridgelets, one finer than the other, there being 

 about four or five of the more minute ridges between the larger ones ; these larger ones 

 are not continuous, but result from a greater prominence being given to certain of the 

 minute ridges for a short distance. This two-ranked arrangement of the ridgelets does 

 not characterize those close to the outer edges of the scute, which approximate in appear- 

 ance to the ridgelets on the lateral i^lates, or cornua. 



From the present form of this scute it would appear that it has been flattened in the 

 shale, and that before compression it sloped downward in front from the oval area toward 

 the rostral plate. 



The little ridges which cover the surface of the ventral scute (fig. 1, d) are similar 

 in their general arrangement to those of the dorsal scute, but present differences in detail ; 

 on the border of the plate there is a much wider space covered by parallel ridges of uni- 

 form size than are found on the dorsal scute ; this band of uniform ridgelets is about half 

 of the width of the lateral plates, and there is at the front end of the ventral scute a sim- 

 ulation of the triangular area at the front of the dorsal scute, but the ridgelets upon it are 

 more irregular. 



There is a greater lack of symmetry in the ridges on the ventral scute than on those 

 of the dorsal, for the ridges on the right at the anterior end are not only looped, but some- 

 what coiled at their extremities, while those of the left side are looped and only slightly 

 curved. 



The fragments of i)lates e and e' (fig. 1) from their regular longitudinal ridges, 

 appear to be parts of lateral j)lates, whose place is probably just behind the anterior lateral 

 plates or cornua. There are some indications that these plates were connected with the 

 ventral rather than with the dorsal scute. '*' 



The fragment / (fig. 1) differs from all the others in having the striae concentric to 

 one of its sides ; it is possibly j)art of an ocular plate, but it is difficult to see where such 

 a plate could be inserted at the side of the rostrum, which seems closely connected with 

 the anterior lateral plates. 



The relation which the several plates are supposed to have had to each other is repre- 

 sented in Plate IV, fig. 3, where the upper set of plates 3 and 3' consists of the dorsal 

 scute and the series of x)lates that border it at the sides and in front ; the dotted areas show 

 the supposed extension of the deficient parts of these plates. The figure below is the 

 fragment of the ventral scute, with the deficient part supplied in outline. The extent to 

 which the plates are flattened is not known, but if they were strongly convex the two 

 main plates would be narrower than they are here represented. 



The fineness of the little ridges on the plates of the Acadian fish is quite as great as 

 that of the fish-plates described in Prof. Lankester's paper, there being from 150 to 200 

 of them in the width of an inch. There is a considerable difference in the number of 



