. The Biological bulletin. Biology; Zoology; Biology; Marine Biology. 408 J. P. A. GARDNER ET AL Table I The sorting procedure used to obtain 12 groups of 25 mussels each (where possible) for analysis, as exemplified by the Whitsand low shore (WLSl sample Initial WLS sample (n = 2166). (450) (25) (107) (25) (804) (25) (129) (25) (265) (25) (110) (25) (81) (25) (42) (25) (IDS) (25) (80) (25) (51) (25) (39) (25) E = edu/is mussels; I = intermediate mussels; G = galloprovincialis mussels; S = small; L = large. the data pair was discarded (there was no evidence of a nonrandom factor explaining whi
. The Biological bulletin. Biology; Zoology; Biology; Marine Biology. 408 J. P. A. GARDNER ET AL Table I The sorting procedure used to obtain 12 groups of 25 mussels each (where possible) for analysis, as exemplified by the Whitsand low shore (WLSl sample Initial WLS sample (n = 2166). (450) (25) (107) (25) (804) (25) (129) (25) (265) (25) (110) (25) (81) (25) (42) (25) (IDS) (25) (80) (25) (51) (25) (39) (25) E = edu/is mussels; I = intermediate mussels; G = galloprovincialis mussels; S = small; L = large. the data pair was discarded (there was no evidence of a nonrandom factor explaining which data pairs could or could not be accurately scored). Nine of 65 mussels were incorrectly aged, eight by ± 1 year, and one by 2 years. Regression analysis of the number of annual growth rings from the acetate peel against the number estimated from the valve exterior was carried out, the two being highly correlated (R: = ; n = 65; P < ). The ages of all remaining nonsectioned mussels were determined by examination of external growth marks. Each mussel provided length-at-age data for a minimum of 4 years of age up to a maximum of 13 years of age, with an average of about 6-7 years. Data were analyzed by univariate ANOVA of each length-at-age category. This analysis is based on minimal assumptions, the results are easily interpretable and biologically meaningful, and maximum information can be extracted from the data set because a large proportion of the values can be used. This is a conservative approach, so that if significant length-at-age differences are observed it can be assumed that they represent genuine differences. A slight loss of statistical power is associated with this approach, but this is offset by the advantages described above. To control for the multiple testing involved with this approach, we used the Bonferroni test, which compensates for the in- creased probability of Type I error rate (the sequential Bonferroni test, which is less conservativ
Size: 3061px × 817px
Photo credit: © Library Book Collection / Alamy / Afripics
License: Licensed
Model Released: No
Keywords: ., bookauthorlilliefrankrat, booksubjectbiology, booksubjectzoology
