Table 2. — Comparison of mean vertebral counts of Atlantic menhaden of the 1955 and 1956 year 



classes at successive ages 



"P is less than 0.05. 



between ages 4 and 5, but in remaining com- 

 parisons the differences showed no consistent 

 trend. 



The significant decrease with age in mean 

 vertebral count of nonspawners of the 1955 

 year class in the Long Island samples may 

 be due to (1) differential mortality as the 

 fish became older, (2) a shift in the spawning 

 cycle of individual fish with advancing age, 

 or (3) chance. Because the difference in the 

 mean vertebral count between ages was sig- 

 nificant only in the 1955 year class in one 

 locality, it is unlikely that this source of 

 variation altered the results of subsequent 

 tests. 



Variation between Year Classes 



Samples from the different localities in- 

 cluded fish of more than one year class. 

 Therefore, if mean vertebral counts of the 

 various year classes represented in the 

 samples differed significantly, the nnean 

 counts of sanriples simply might reflect differ- 

 ences in age composition. 



In every instance except one (nonspawners. 

 Long Island) mean counts of both spawners 

 and nonspawners of the 1956 year class were 

 slightly higher than those of the 1955 and 

 1957 year classes (table 3). Differences as 

 great as 0.145 vertebra are shown for mean 

 counts between year classes (1956 and 1957 

 year classes in Cape Cod samples); however, 

 no difference was statistically significant. 



Table 3. — Camparison of mean vertebral counts of Atlantic menhaden 

 of the 1955-57 year classes 



Sutherland (1963) found the mean vertebral 

 counts of juveniles of the 1956 year class 

 significantly higher than those of the 1957, 

 1958, and 1959 year classes. Examination of 

 daily and nnean monthly surface water tem- 

 peratures and salinities (Bunnpus, 1957; Day, 

 1959a, 1959b, 1960) at lightship and shore 

 stations along the Atlantic coast showed no 

 striking anomalies during the period of early 

 development of eggs and larvae in 1956. 

 There was no pattern of correlation between 

 the vertebral counts and the hydrographic 

 data. An explanation of this phenomenon naay 

 involve one or nnore of the following points: 

 (1) selection by spawning fish of temperature 

 and/or salinity strata other than the surface 

 waters considered, (2) variations in the tinne 

 and place of spawning, (3) differential sur- 

 vival in the early life history stages, and 



