564 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 83 



spines, against a darker background; full-grown specimens typically 

 mth narrow lines partly or wholly replacing the blotches, transverse 

 on trunk, lengthwise on head and anterior part of the trunk, the con- 

 trasting directions of the lines usually striking along the boundary 

 where they meet; white lines sometimes alternating with the brown 

 lines on the opercle; bluish or whitish dots quite profuse, except on the 

 side of the trunk, radiating rows of such dots or radiating white lines 

 often present around eye; dorsal with a submarginal dark band. (See 

 tables 1 and 3 for counts and measurements and table 4 for averages.) 



The variability and development of the filaments, spines, and the 

 color pattern are quite similar to the subspecies hudsonius. In gen- 

 eral, the spines are usually somewhat shorter than in hudsonius when 

 specimens of approximately the same size and the same sex are com- 

 pared. As in hudsonius, specimens sometimes have the brown lines 

 on the trunk and head broken up into series of spots. These spots 

 sometimes lose their rowed arrangement and such specimens approach 

 individuals of reidi in color. 



Four specimens were examined from Cuba. Two large males have 

 the spines on trunk and coronet very low, almost obliterated in the 

 largest male, 107 mm long (fig. 63), being nearly Hke specimens of reidi 

 or hippocampus in this respect; but the tubercles on the tail are con- 

 spicuously better developed than in those two species. A young 

 specimen 23 mm long also has the spines notably short for its size, 

 strikingly shorter than in a specimen of similar size from Key West. 

 The fourth specimen, a female 56 mm long, has the tubercles nearly 

 as well developed as specimens of similar size from Florida. From 

 these four specimens, therefore, it seems that the Cuban population 

 has, on the average, the tubercles not so well developed as the Florida 

 population. However, in the counts and measurements these four 

 agree well with those from Florida, and the difference between the 

 two populations apparently is of no more than racial magnitude. 



Distinctive characters and relationships. — The relation of this sub- 

 species to hudsonius has already been discussed (p. 557). Typical 

 full-grown specimens have a strikingly different appearance from 

 hudsonius on account of their deeper body, longer snout, and somewhat 

 lower tubercles and coronet. It also has a lower average caudal- 

 segment count and higher fin-ray count. The bluish or whitish dots 

 are generally more profuse and more prominent, the brown lines on 

 the head and trunk are oftener better defined, and the opercle some- 

 times has white lines alternating with the brown; but there is con- 

 siderable intergradation between the two subspecies, as noted. The 

 differences between this subspecies and reidi are discussed under the 

 account of reidi (p. 575). 



