588 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 83 



zosterae, which tends to break up into distinct stocks as discussed 

 hereafter (p. 592), 



Of the two variants of regulus, the one with the nine trunk seg- 

 ments has a quadrangular first caudal, while the one with an incom- 

 pletely hexangular first caudal segment has 10 trunk segments. It 

 will be shown (p. 591) that in zosterae nine trunk segments are always 

 correlated with a hexangular first caudal segment. In regulus these 

 variations are not only infrequent but when they do occur they are 

 not correlated. Another point of considerable interest is that regulus, 

 in two important characters — number of trunk segments and number 

 of pectoral rays — approaches much more the Key West population 

 of zosterae than its Pensacola population (see table 5). 



There are legitimate grounds for difference of opinion in regard to 

 the taxonomic status of regulus, whether it is to be considered as a 

 full species or as a subspecies. According to the data presented, it 

 may be regarded, within reason, as a subspecies of zosterae. How- 

 ever, while the degree of intergradation in the characters investigated 

 is greater than usual between distinct species of fishes in general, it 

 is also of a lesser degree than the usual intergradation between sub- 

 species of fishes. Furthermore, speciation in the genus Hi'p'pocamjpus 

 is quite unlike that usual among fishes. A condition of very near 

 approach or even of overlapping is evidently normal in Hippocampus 

 (see, for instance, discussion of relationship of ingens, p. 536). A com- 

 parison between tables 1 and 5 shows that the divergence between 

 regulus and zosterae, in the number of dorsal rays and caudal segments, 

 is much more pronounced and of a much liigher degree than that 

 between the subspecies hudsonius and punctulatus, for instance. It 

 was also shown that regulus is nearer to the Key West population of 

 zosterae, whereas if regulus were a mere geographical subspecies of 

 zosterae, it would be reasonable to expect it to differ in a regular 

 latitudinal direction and to be nearer the Pensacola population of 

 zosterae. All available evidence considered, therefore, it seems best 

 to assign full specific rank to regulus, although this opinion may have 

 to be changed by a study of more material and specimens from 

 intermediate localities. 



As compared with all other American species of Hippocampus 

 except zosterae, regulus is readily distinguished by the number of 

 trunk and caudal segments, the number of fin rays, and its small size. 



Material studied and geographic distribution. — Cat Island, Miss., 

 collected by the author November 15, 1931. Harbor Island, Tex., 

 J. C. Pearson (92950, the type. May 1927; also in the Bureau of 

 Fisheries, collected on the following dates: 1 specimen with the type; 

 2 on April 3, 1927, 1 on October 20, 1926, 2 on October 25, 1926, 2 on 

 November 12, 1926). Hog Island, Tex. ; J. C. Pearson. Champoton, 



