REVIEW OF HIPPOCAMPUS — GINSBURG 501 



aided materially in bringing this study to a point more nearly ap- 

 proaching completeness. For material lent by other institutions 

 acknowledgment is due the following: Dr. Carl L. Hubbs, who 

 lent two lots of seahorses from the Zoology Museum of the University 

 of Michigan, one of these lots forming the basis of a new subspecies, 

 multiannularis ; Alfred C. Weed, who lent the desirable collection of 

 Hippocampus in the Field Museum of Natural History, which helped 

 me considerably in distinguishing kincaidi and in confirming my 

 conclusions in regard to some other species and subspecies; John T. 

 Nichols, who lent seven specimens of H. hippocampus from the 

 collection of the American Museum of Natural History, which were 

 of considerable help in distinguishing that species from closely related 

 ones. 



Particular mention is made of the work of Miss Louella E. Cable, 

 who executed with skill and painstaldng care the drawings for the 

 figures, wliich should prove invaluable in the identification of speci- 

 mens. Whatever merit this report may prove to have, it will be 

 greatly enhanced by these beautiful and accurate illustrations. 



DEFINITIONS AND METHODS OF STUDY » 



The proper differentiation of the species of Hippocampus is difficult 

 at best. This difficulty is greatly increased by the lack of uniformity 

 in descriptions of the species by various authors. It becomes neces- 

 sary, therefore, to explain the methods of study and define the terms 

 used. 



How to determine accurately the number oj trunk segments. — The 

 first important point to decide is a uniform method in the determina- 

 tion of the boundary line between the trunk and the tail, since the 

 number of segments in the trunk and to a lesser extent that of the tail 

 form specific characters of primary importance. The distinction 

 between trunk and tail is readily made after the integument is removed 

 and the exoskeleton uncovered (fig. 54). This is, of course, imprac- 

 ticable when identifying specimens. Externally the last trimk seg- 

 ment is readily determined by the fact that the ventrolateral ridge of 

 the trunk extends only to that segment. The last spur on that 

 ridge, or, where the spur is missing, the last intersection of the longi- 

 tudinal with the transverse ridge on the side, unmistakably marks the 

 last trunk segment. From that intersection a wingfike extension 

 converges with its fellow from the other side to the base of the anal 

 for the support of that fin, but this extension is usually covered by 

 thick integument and not visible externally. In practice the best 

 M^ay to count the trunk segments with absolute accuracy is to trace 

 the transverse ridge on the last segment from its lower point, as de- 



' The reader will find it advantageous to study figure 54 in connection witli the discussion here of the 

 structure of some parts of the exoslieleton, which are of importance in classification. 



