obtained from filter screens at electric power 

 plants and by otter trawl and cast net. Fish 

 were preserved in 10 percent formalin at the 

 time of collection, later leached in water, and 

 stored in 40 percent isopropyl alcohol. Local- 

 ity and date for each collection are shown in 

 appendix tables 1 through 4. 



Vertebral counts were made from X-ray 

 photographs of fish (fig. 1) larger than 30 mm. 

 fork length (Sutherland, 1958) and from stained 

 and cleared fish that were less than 30 mm. 

 (Clothier, 1950). Counts included the total 

 segments between, but not including, the oc- 

 cipital bone and the hypurals. Abnormal verte- 

 brae were counted when distinguished by a 

 suture between fused centra or the presence 

 of a neural or hemal spine. Fork length, in 

 millimeters, and sex were determined for 

 each fish. 



Vertebral counts were analyzed (Snedecor, 

 1956) for sample size requirements and vari- 

 ations by sex and length, in the frequency of 

 fused vertebrae, and within and between sam- 

 ples from different localities, geographic 

 groups, and year classes. 



RESULTS OF ANALYSES 

 Sample size 



Sample size was maintained at or near 100 

 specimens where possible as determined from 

 sample variance ( cr 2) in the formula 4o;2 _ 



Variance of test samples ranged between 0.17 

 and 0.35. An L-value of 0.10 vertebra, selected 

 as the allowable error with a 5-percent chance 

 that the error of the sample mean would be 

 exceeded, indicated that samples of 68- 140 fish 

 were required. 



Figure 1. --X-ray photograph of a juvenile Atlantic 

 menhaden. 



Vertebral abnormalties 



Abnormal vertebrae occurred in 1.4 percent 

 of specimens examined. The most frequent 

 type was the fusion of two centra (53.6 per- 

 cent). Others included fusion of three or more 

 centra, heavily ossified fused and nonfused 

 centra, and centra with more than one neural 

 or hemal spine. An estimated 0.5 to 1 percent 

 of all specimens examined had curved verte- 

 bral columns, but these were not considered 

 an abnormality unless accompanied by one 

 of the previously described types. These 

 abnormalities are similar to those found in 

 young Pacific herring, Clupea paZ/osu, (McHugh, 

 1942), although they occurred less frequently 

 in Atlantic menhaden than in herring. 



Vertebral abnormalities were not randomly 

 distributed. The percentages for all years, by 

 major coastal areas, were 2.8 percent in the 

 North Atlantic (Cape Cod, Mass., to Sandy Hook, 

 N.J.), 1.0 percent in the Middle Atlantic (Sandy 

 Hook, N.J.,to Chesapeake Bay), and 1.9 per- 

 cent in the South Atlantic (Chesapeake Bay to 

 Nassau River, Fla.). Chi~square tests applied 

 to the data for the 1956 and 1957 year classes 

 indicated that the differences were significant 

 (P < 0.05) between the North and Middle At- 

 lantic and between the South and Middle Atlan- 

 tic, but not between the North and South 

 Atlantic. This suggests that a common factor 

 may be responsible for the greater frequency 

 of vertebral abnormalities in the North and 

 South Atlantic areas. 



Variation between sexes 



Analysis of 1,778 juveniles (table 1) showed 

 that the mean vertebral numbers of males and 

 females differed significantly (P< 0.05) in only 

 one of 18 samples. Since the frequency of 

 occurrence was about the same as that ex- 

 pected in random sampling (1 in 20), sexual 

 differences were not considered an important 

 source of variation and were not included in 

 the analysis of the data. The unweighted means 

 of the test samples were 46.95 for males and 

 46.96 for females. 



Variation with length 



Regressions of vertebral number on fish 

 length were calculated for the same samples 



