Steimie Population, growth, and production o^ Echinarachnius parma 



181 



Figure 2 



Ech inarachnius parma size-frequen- 

 cy histograms per collection at station 

 11. Middle Atlantic Bight. 



40 

 20 



60- 

 20^ 

 40 

 20 



40 

 20 



40 

 20 



40 

 20 



40 

 20 



40 

 20 



.^^S^ 



Lw 



A^ 





-Jh^ 



20 30 



SIZE(mm) 



JUL 1977 40 

 N=137 



20 



N = 228 



APR 1978 *° 

 N=467 20 



SEP 1978 40 



20 



APR 1979 40] 

 N=212 20 



JUL 1979 '*0 

 N--115 20 



SEP 1979 40 

 N=293 20- 



DEC 1979 ""O 

 N=259 20 



JUL 1980 

 N: 147 



DEC 1980 

 N = 186 



AUG 1981 

 N = 48 



. r-^lTyg rv n— -^ 



FEB 1982 

 N=125 



JUL 1982 

 N:363 



JUL 1983 

 N:89 



NOV 1983 

 N=122 



n 



AUG 1984 

 N:5 



^v.^^-^^^^. 



JUL 1985 

 N:213 



20 30 



SIZE(mm) 



Figure 3 



EchiNiirm-hniu.'i parma size-frequency histograms per 

 collection at station 17. Middle Atlantic Bight. 



the samples (winter-spring) and thus identified by the 

 pair of years covering the probable peak spawning 

 period. The recruitment season was estimated, in some 

 cases, from back-calculations based on growth curves. 

 Ages assigned to each cohort are based on inspecting 

 the length-frequency histograms and noting when 

 recruitment, i.e., >2 mm size mode, was apparent. 

 These can be considered relative ages, although it is 

 highly probable they represent absolute age ( + 3 

 months) if it is assumed that a delayed metamorphosis 

 of the larvae is not normally a major problem; there 

 is no evidence to suggest otherwise. 



The reliance on size-frequency analysis for growth 

 and production estimates assumes (1) frequency modes 

 represent cohorts, a reasonable assumption when re- 

 cruitment period can be determined and modal shifts 

 can be traced over a period of time, and (2) that the 

 size distribution within each cohort was normal or 

 representative of the true population and unbiased. A 

 departure from normality may not be critical to the ac- 

 curacy of mean size estimates (McNew and Summerfelt 

 1978). The sample sizes were often below the minimum 

 (400) suggested by Cohen (1966) to separate accurate- 

 ly two cohorts in a single size-frequency analysis, with 

 a larger sample needed for separating more than two 



