Mean spine diameter was related linearly with fish fork length (r2 = 0.92, p < 0.001) (b) and the regression equation was used to estimate fork lengths for Jamestown samples. Extant James River Atlantic sturgeon ages ranged between 1 and 19 years (mean 4 ± 3 s.d. years) compared with 5–42 years for the colonial samples. The coefficient of variation (CV) based on multiple reads of extant samples was 1.8 per cent (70% exact agreement) while the Jamestown CV was 4.0 per cent (46% agreement). Average disagreement was 1.3 years for modern and 2.2 years for Jamestown samples. The absence of fish less than five years of age in the colonial samples probably reflects selective fishing for large individuals. Therefore, all comparisons exclude modern fish less than five years of age. Mean age-at-capture was significantly younger for extant fish (8 ± 3 years) than for historic samples (19 ± 10 years; t91 = 7.49, p < 0.001, a). Mean length-at-capture (1048 ± 338 mm) for extant fish was significantly less than in historic samples (1511 ± 625 mm; t91 = 4.64, p < 0.001). Finally, the slope of growth curves from areas of sample overlap (present day n = 58, Jamestown n = 19), for extant fish was significantly higher (ca 25%) than comparably aged, ca 1610 samples (slopes: t73 = 3.80, p < 0.001, b).
(a) Photograph of a spine of an 11 year old Atlantic sturgeon from the well in colonial Jamestown, ca 1610. (b) Relationship of fork length to average spine diameter from extant James River sturgeon (y = 11.198x, r2 = 0.92, p < 0.001).
Figure 2. (a) Age-frequency distribution of James River Atlantic sturgeon from ca 1610 (black bars) compared to extant population (white bars). (b) Comparison of fork length-at-age for ca 1610 and extant James River Atlantic sturgeon population (ca 1610, y = 58.903 (more ...)
Growth curves comparing populations at various latitudes from South Carolina to Quebec indicate clinal growth (c), suggesting that colder water temperatures and/or a shorter growing season decreased annual growth. Analysis of covariance showed that the ca 1610 James River population was significantly different from all populations except for the St John River, New Brunswick, Canada (p = 0.232).