592 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 83 



segment; one has an extra plate on both sides of the last trunk and 

 first caudal segments; one lacks extra plates (this being the only one 

 of all the specimens examined, including all the species, which en- 

 tirely lacked extra plates for the support of the dorsal). Again com- 

 bining the specimens showing asymmetry with the others, as above, 

 omitting the specimens entirely lacking plates, and not taking 

 account of the extra plate on the penultimate trunk segment of one 

 specimen, we get 31 specimens having an extra plate on the last 

 trunk segment only and two having extra plates on the last trunk and 

 first caudal segments. Consequently, these figures show that 10 

 trunk segments are nearly always correlated with an octangular last 

 trunk segment and nearly always with a quadrangular first caudal 

 segment. 



The two specimens with 1 1 trunk segments have an extra plate on 

 the last trunk segment only, like the dominant condition in those 

 specimens having 10 trunk segments. 



Four specimens, two from Pensacola and two from Captiva Pass, 

 have 10 trunk segments mth the last one incomplete (see p. 504 for 

 explanation of an incomplete trunk segment). Each one of these 

 four has the extra plate on bo th sides of the tenth or last incomplete 

 segment, one also having an extra plate on one side of the penultimate 

 segment. If each side is considered separately in these four asymmetri- 

 cal specimens, one side will have nine trunk segments and the extra 

 plate on the first caudal segment, while the other side will be found to 

 have 10 trunk segments with the extra plate on the last trunk and none 

 on the first caudal. The two chief trends of variation in zosterae are 

 thus indicated on either side of each one of these four variants, the 

 last trunk segment having had the lower lateral ridge shortened on 

 one side only, the last trunk segment thus having changed to a caudal 

 segment on that side. 



Population divergence. — The relative frequency of occurrence of the 

 two chief variations as described in the preceding paragraphs differs 

 markedly with locality and may be used in racial or varietal distinc- 

 tion as follows (for the sake of brevity these differences may be indi- 

 cated by reference to the number of trunk segments, but the other 

 correlated differences also occur as described): 



By reference to table 5, it will be noted that nine trunk segments are 

 possibly the dominant condition at Pensacola, although the number of 

 specimens studied is not sufficient to be certain. Anyway, the per- 

 centage of such specimens must be high. In the Captiva Pass lot a 

 little less than a third of the specimens have nine trunk segments, 

 while in the Key West population a little less than a fifth have nine 

 trunk segments. Among the specimens enumerated as having 10 

 trunk segments in table 5, two from Pensacola and two from Captiva 

 Pass have the last segment incomplete and may be counted as having 



