Holt and Caldeewood — Report on the Rarer Fishes. 399 



constant specific value. It is only in old Homelyns that more than two may occur 

 in this region, and they seem to be always very small, in fact not much larger 

 than the surrounding spinules. They belong to the secondary series, and may 

 occur also in old or half-grown Blondes. In Homelyns exceeding 12 inches in 

 transverse measurement it is unusual to find shoulder spines; in the Blonde these 

 structures would seem to disappear, as a rule, when the disk is at a width of about 

 14 inches ; but we have seen a specimen, 38 inches long, and therefore, probably, 

 about 26 inches wide, in which they were still retained. The smaller spines and 

 larger spinules have distinctly radiate bases in young examples of either species. 



In the Blonde the upper surface becomes practically covered with spinules 

 when the disk is a width of 16 inches, the smooth condition of certain areas in 

 the adult male being only attained, or rather re-attained, at a size which the 

 Homelyn never reaches. The spinules are never very clearly set, except in 

 certain regions mentioned in the diagnosis, and thus differ from those covering 

 the general surface in R. clavata, R. microcellata, and some other species. 



On the under surface, the border of very closely-set spinules along the anterior 

 mar gin J is an infallible character, since such is never present in the Homelyn at 

 all.* In Blondes of a little under 9 inches it is perfectly distinct, though narrow ; 

 it extends laterally to the level of the gills. In specimens of about 18 inches it is 

 about /g- inch wide, except at the lateral extremes, which are narrower, and ter- 

 minate at points where a straight line drawn through the nostrils would cut 

 the anterior margin. In full-grown Blondes the extremities of the border are 

 approximately on a level with the corners of the mouth. The narrow border of 

 closely set spinules along the edges of the caudal undersurface is another useful 

 character. It is present in all Blondes of 10 inches, and upwards, and entirely 

 absent from any stage of the Homelyn. 



In specimens of equal size, the teeth are always smaller and more numerous in 

 the Blonde than in the Homelyn ; but in both species they seem to vary a good 

 deal both in number and size, apart fx-om the regular developmental changes in 

 these characters. There remains, however, a distinction quite sufficient for sys- 

 tematic purposes, so long as the size of the individual is considered. Thus, a 

 Blonde 9 inches wide has 66 rows ; a Homelyn 9^ inches wide, only 49 rows ; 

 Blondes 10|^ and 11 inches wide have 62 and 66 rows; Homelyns of 11^ and 

 11|^ inches, 51 and 45 rows; a Blonde 16^ inches, 74 rows; a Homelyn, 

 16f inches, 51 rows ; and another, 17|- inches, only 50 rows. A Blonde, 21^ inches, 

 74 rows ; a Homelyn, 19^ inches, the largest we could procure, 66 rows. The 

 difference here is not great, but we imagine, from comparison with larger and 

 smaller specimens, that the Blonde in question exhibits about the maximum 



* It occurs, however, in R. asterias (Miill. and Henle), and in B. punctata (Risso), and in other spotted 

 ravs from the Mediterranean. 



