REVIEW OF HIPPOCAMPUS — GINSBURG 575 



hippocampus may be a case of parallelism, and it is possible that reidi 

 is more nearly related to kincaidi and punctulatus. In any case, 

 whatever is the true relationship of reidi, for the practical purpose of 

 identification it is necessary to compare it with them, since its geo- 

 graphic range overlaps with that of kincaidi and possibly also with 

 that of punctulatus. 



Full-grown or nearly full-grown specimens of reidi may be sharply 

 distinguished from punctulatus by their markedly slender trunk (see 

 table 3) along with the difference in the color pattern, reidi being 

 profusely spotted with small spots, while large specimens of punctu- 

 latus are marked generally by narrow lines or sometimes by large 

 blotches. H. reidi also has the tubercles obsolescent, while in punc- 

 tulatus they are in most specimens fairly well developed, although 

 full-grown males sometimes closely approach reidi in that respect. 

 Small specimens are not readily distinguished by depth, but may be 

 separated on direct comparison by the difference in the structure of 

 the tubercles, in most, but not all cases, some small specimens oi punctu- 

 latus having the tubercles rather low. As further aids in separating 

 the two, reidi has a distinctly lower dorsal fin ray count and a longer 

 snout than punctulatus, but there is more or less intergradation in 

 those two characters (see tables 1 and 3). 



The present species differs from kincaidi in the same characters, 

 namely, in having a slenderer trunk, obsolescent tubercles, fewer 

 dorsal rays, a longer snout, and a different color pattern. It has been 

 noted that kincaidi has a slenderer trunk and generally lower tubercles 

 than punctulatus, and it consequently approaches nearer to reidi in 

 those two important characters. However, to offset this conver- 

 gence, kincaidi has a somewhat shorter snout than punctulatus, and 

 it consequently diverges more from reidi in this character. While 

 kincaidi converges toward reidi in the depth of the trunk, there was 

 no intergradation in the few specimens measured (see table 3). 



When all the characters are taken into consideration there should 

 be found no difficulty in most cases in distinguishing reidi from kin- 

 caidi, as well as from punctulatus. At least, I did not find it difficult. 

 It is reasonable to expect some difficulty, however, in referring occa- 

 sional extreme variants of kincaidi and reidi in places where both 

 occur, as in Bermuda. Out of seven specimens of seahorses from 

 Bermuda available, only one may be referred to reidi and six to kin- 

 caidi, and the latter is probably the commoner seahorse on the coast 

 of Bermuda. The single specimen of reidi from that coast is for- 

 tunately a nearly full-grown individual having the important char- 

 acters typical of its species, and there is no question as to where it 

 belongs. 



Material studied and geographic distribution. — Porto Bello, Panama; 

 Meek and Hildebrand (79685, March 19, 1912; Field Mus. Nat. Hist. 



