Evidence of three growth stanzas in rainbow trout ( Oncorhynchus mykiss) across life stages and adaptation of the thermal-unit growth coefficient

Abstract

Current mathematical growth models describe the growth of finfish with few considerations of the changes in growth pattern occurring across life stages. This study analysed the growth pattern of rainbow trout and tried to improve the goodness of fit of an empirical growth function. Growth data were obtained from 21 separate lots of rainbow trout (Ontario ARST strain) fed to satiation and reared at constant water temperature (8.5 °C) at the Alma Aquaculture Research Station, University of Guelph between 1997 and 2005. Growth rates (207 observations) were calculated using the thermal-unit growth coefficient (TGC). Calculated growth rates were regressed against live body weight (BW). Piecewise linear analysis was used to determine changes in the growth pattern. This analysis revealed the existence of three growth stanzas: from first-feeding (0.2 g) to 20 g (Stanza 1); from 20 to 500 g (Stanza 2); and > 500 g (Stanza 3). The least squares method was used to optimize the weight exponent within each growth stanza and to improve the goodness of fit of the TGC model. Results indicated that weight exponents other than 1 − b = 1/3 currently used in the TGC model should be used for Stanzas 1 and 3. Only the weight exponent for Stanza 2 was not significantly different from the conventional TGC model where 1 − b = 1/3 ( P ≥ 0.05). The weight exponent values that gave the best fit within Stanzas 1 and 3 were 0.209 and 0.967, respectively. The predicted values for BW were overestimated for small fish (< 20 g) and underestimated for large fish (> 500 g) when using the exponent 1/3. The similarity between predicted and observed BW for fish weighing between 20–500 g meant that the cube root of BW is suitable for predicting BW in Stanza 2. These results provide a more realistic growth function that better fits the growth pattern observed across the life stages of rainbow trout.