REVIEW OF HIPPOCAMPUS — GINSBURG 591 



tendency for the last trunk segment to change to a caudal segment by 

 the loss of the last pouit of intersection on the lower lateral ridge. 

 As a result the number of trunk segments is reduced by one; the first 

 caudal, instead of the last trunk segment, now bears the extra plate 

 for the support of the dorsal, and the base of the dorsal is placed over 

 the last trunk and first caudal segments instead of over the last two 

 trunk segments. This important trend of evolution shown by a 

 comparatively large percentage of specimens evidently indicates a 

 more recent development. The frequent presence of a hexangular 

 caudal segment in this species may appear to show a more primitive 

 condition, since tliis occurs also in the subgenus Hippocampus. 

 However, in zosterae a hexangular caudal segment is correlated with 

 a hexangular last trunk segment, and the latter condition, in its turn, 

 is unique and apparently represents a more recent development. 

 Consequently, the hexangular first caudal segment in zosterae prob- 

 ably represents a pseudoreversion and not a primitive condition; that 

 is, it is caused by the last trunk segment changing to a caudal seg- 

 ment as a consequence of a shortening of the lower ridge on the trunk. 

 The evidence strongly favors the conclusion that zosterae is now under- 

 going a gradual change, which, if carried far enough, will result in the 

 formation of a distinct species, or even subgenus, having nine trunk 

 segments. The tempo of the change evidently differs with the 

 population (see p. 592). 



For convenience, the individual variability, besides the main 

 trends of variation, may be indicated as follows: Altogether 59 speci- 

 mens were examined, in which the number of trunk segments were: 

 19 with 9 complete segments; 34 with 10 complete segments; 4 with 

 10 incomplete segments; and 2 with 11 complete segments. Of those 

 having 9 segments 14 have an extra plate on the first caudal segment 

 only; three have an extra plate on the last trunk and first caudal 

 segments ; in one an extra plate is present only on one side of the first 

 caudal segment ; and in one an extra plate is present only on one side 

 of the last trunk segment and on both sides of the first caudal seg- 

 ment. Counting the variants showing asymmetry as though they 

 were bilaterally symmetrical, and combining the above figures, w^e 

 get 15 specimens having an extra plate on the first caudal segment 

 only and four having extra plates on the last trunk and first caudal 

 segments. These figures consequently show that nine trunk seg- 

 ments are always correlated with a hexangular first caudal segment 

 and decidedly correlated with a hexangular last trunk segment. Of 

 the 34 specimens having 10 trunk segments, 30 have an extra plate 

 on the last trunk segment only; one has an extra plate on one side of 

 the penultimate trunk segment on both sides of the last trunk seg- 

 ment and none on the first caudal ; one has an extra plate on one side 

 only of the last trunk segment and on both sides of the first caudal 



