The Effects of Dietary Composition on Energy Deposition in Juvenile White Sturgeon

The objective of this experiment was to examine energy partitioning patterns in juvenile white sturgeon on high (HE) and low (LE) energy diets. The experiment was designed to allow the use of mathematical models to investigate various aspects of energy metabolism in the fish.;The experiment was conducted on a commercial sturgeon farm near Sacramento, California. Fish and feed samples were collected monthly for 13 months and analyzed for water, protein, and fat content. Five female white sturgeon were collected from each tank every month for a total of 520 observations. For proximate analysis, body tissue and ovaries were pooled each month by tank for a total of 104 observations.;Average daily metabolizable energy (ME) intake was 55.9 kJ/kg BW for the LE fish and 68.2 kJ/kg BW for the HE fish. Body weight (BW) was not significantly different between the two treatments. Normalized for body weight, mean body protein deposition was significantly higher in LE fish. Body fat deposition was significantly lower in LE fish (19.6 MJ vs. 23.7 MJ) in both absolute terms and as a percent of BW. However, ovarian protein and lipid deposition was not significantly different between the treatments.;Linear mixed models, multiple regression, and multivariate models were used to investigate different aspects of energy metabolism in the fish. However, due to the inconsistent composition of the LE diet during the first portion of the experiment, only data collected over the last 6 months of the experiment were used for these analyses.;Linear mixed models were used to examine differences in gross energetic efficiencies for protein and lipid deposition in the body and ovaries of the LE and HE fish. Efficiency was defined as the amount of energy deposited per unit of metabolizable energy intake (ME). Mixed effects were applied to account for random tank-to-tank variation. None of the efficiencies---protein and lipid deposition in the body and ovaries---were significantly different between the treatments. This, coupled with the observation that there were no significant differences in weight or growth rates between the treatments implies that the LE diet may be economically beneficial to use.;Multiple regression models were used to estimate maintenance requirements and the partial efficiencies of energy deposition for the fish on the HE and LE diets. The first equation estimated the partial efficiency of retaining total body energy. This was expanded to estimate the partial efficiencies of both protein and lipid deposition. This was further decomposed to estimate the partial efficiencies of protein and lipid deposition in both the body and ovaries. However, the models gave biologically unfeasible estimates, a well-documented problem with this approach (Koong, 1976; van Milgen and Noblet, 1999; Azevedo et al., 1999). Thus, biased regression techniques and fixing the maintenance requirement were introduced to try to improve the estimates. Regardless, the models did not produce biologically meaningful partial efficiencies.;Lastly, a multivariate approach was applied to estimate partial efficiencies. This method has been shown to make improved partial efficiency estimates over those from multiple regression (Koong, 1977; van Milgen and Noblet, 1999; Azevedo et al., 2005). A multivariate model similar to the ones described by Azevedo et al. (2005) and Strathe et al (2010) was used, however, the equations again failed to produce biologically reasonable estimates.;The comparable growth rates and body protein deposition patterns between the LE and HE fish suggest that the LE diet is sufficient for maintaining desired growth and development rates over the six month trial. Thus, the LE diet may be commercially beneficial to use. However, further experiments are necessary to ensure that growth and sexual development continues to progress normally in fish started on the LE diet at a young age and that the production cycle is not prolonged. The thesis concludes with suggestions for future experiments to increase the likelihood of generating data that will allow the use of these models successfully and improve our understanding of energy partitioning in white sturgeon.