Sizes of steelhead families raised in hatchery tanks vs in artificial streams

Salmonid fish raised in hatcheries often have lower fitness (number of returning adult offspring) than wild fish when both spawn in the wild. Body size at release from hatcheries is positively correlated with survival at sea. So one explanation for reduced fitness is that hatcheries inadvertently select for trait values that enhance growth rate under the unnatural environment of a hatchery, but that are maladaptive in the wild environment. A simple prediction of this hypothesis is that juveniles of hatchery origin should grow more quickly than fish of wild origin under hatchery conditions, but should have lower survival under wild conditions. We tested that hypothesis using multiple full sibling families of steelhead (Oncorhynchus mykiss) that were spawned using either two wild parents (WxW) or two first-generation hatchery (HxH) parents. Offspring from all the families were grown together under hatchery conditions and under semi-natural conditions in artificial streams. HxH families grew significantly faster in the hatchery, but had significantly lower survival in the streams. That we see this tradeoff after only a single generation of selection suggests that the traits involved are under very strong selection. We also considered one possible alteration to the hatchery environment that might reduce the intensity of selection among families in size at release. Here we tested whether reducing the fat content of hatchery feed would reduce the variance among families in body size. Although fish raised under a low-fat diet were slightly smaller, the variation among families in final size was unchanged. Thus, there is no evidence that reducing the fat content of hatchery feed would reduce the opportunity for selection among families on size at release.

Methods
Full sibling families of fry were pooled together and raised in 400 L tanks, or in artificial streams at a research hatchery.  These datasets provide family average growth to compare with family survival, and the individual growth data.