REVIEW OF HIPPOCAMPUS — GINSBURG 593 



nine segments on one side. Consequently, the decided or predominant 

 tendency shown by the more northern populations of having one seg- 

 ment less than the population from Key West is actuall}^ more 

 pronounced than indicated by the figures in table 5. Besides the 

 decided difference in the number of trunk segments, table 5 also 

 shows a less decided but apparently significant difference in the 

 frequency distributions of the number of pectoral rays. While the 

 number of specimens studied is too small for a thoroughgoing 

 racial analysis, it seems evident that zosterae tends to break up into 

 distinct stocks in spite of its comparatively restricted geographic 

 range. 



Distinctive characters and relationships. — H. zosterae may be dis- 

 tinguished easily from its congener occurring in its range, punctulatus, 

 by the smaller number of fin rays and trunk segments and its much 

 smaller size. The number of caudal segments is also generally less, 

 but there is a small degree of intergradation in this character. This 

 species is closely related to regulus and the difference between them 

 has been discussed (p. 585). 



Material examined and geographic distribution. — All localities on 

 the coast of Florida, as follows: Cape Florida (67658, three dried 

 specimens). Biscayne Bay at Bonefish Banks, November 27, 1906 

 (57236). Newfound Harbor Key, Pine and Bean, December 7, 1906 

 (57453). Key West (92717, April 15-27, 1884, Albatross, 1 specimen; 

 also 15 specimens collected on seven different dates by the staff of 

 the Bureau of Fisheries Biological Station). Boca Chica, April 11, 

 1922. Captiva Pass; O. P. Hay (Field Mus. Nat. Hist. no. 2131). 

 St. Martins; January 17, 1902; Fish Hawk (73242). Pepperfish 

 Key; November 21, 1901; Fish Hawk (73241). Apalachicola Bay; 

 S. Stearns; 1880 (26595, this specimen found inseparably mixed in 

 same bottle with 30753). Pensacola; S. Stearns (30753, mixed with 

 the preceding specimen as noted; also 31920). 



Total number of specimens examined, 59; 29, with a brood pouch 

 or at least a rudiment of one, 25 to 44 mm long; 30, with no trace of a 

 brood pouch, 24 to 44 mm long. Biscayne Bay to Pensacola, there- 

 fore, must be regarded now as representing the geographic range of 

 this species, and unquestioned records from other places that may 

 be referred to the present species are not known to me. The refer- 

 ence of rosamondae, from Cuba, to the synonymy of zosterae, as 

 noted in the next paragraph, must remain in doubt until the type is 

 reexamined and compared with authentic specimens of zosterae. 



Synonymy. — In the description of H. rosamondae, Borodin states 

 that it differs from zosterae << * * * ^y having longer dorsal, 

 longer snout and very scarce and small filaments on the head and by 

 the absence of body's spines." The dorsal in rosamondae (14 rays) 



73864—36 7 



