The selective pressures involved in the evolution of semelparity and its associated life-history traits are largely unknown. We used species-level analyses, independent contrasts, and reconstruction of ancestral states to study the evolution of body length, fecundity, egg weight, gonadosomatic index, and parity (semelparity vs. degree of iteroparity) in females of 12 species of salmonid fishes. According to both species-level analysis and independent contrasts analysis, body length was positively correlated with fecundity, egg weight, and gonadosomatic index, and semelparous species exhibited a significantly steeper slope for the regression of egg weight on body length than did iteroparous species. Percent repeat breeding (degree of iteroparity) was negatively correlated with gonadosomatic index using independent contrasts analysis. Semelparous species had significantly larger eggs by species-level analysis, and the egg weight contrast for the branch on which semelparity was inferred to have originated was significantly larger than the other egg weight contrasts, corresponding to a remarkable increase in egg weight. Reconstruction of ancestral states showed that egg weight and body length apparently increased with the origin of semelparity, but fecundity and gonadosomatic index remained more or less constant or decreased. Thus, the strong evolutionary linkages between body size, fecundity, and gonadosomatic index were broken during the transition from iteroparity to semelparity. These findings suggest that long-distance migrations, which increase adult mortality between breeding episodes, may have been necessary for the origin of semelparity in Pacific salmon, but that increased egg weight, leading to increased juvenile survivorship, was crucial in driving the transition. Our analyses support the life-history hypotheses that a lower degree of repeat breeding is linked to higher reproductive investment per breeding episode, and that semelparity evolves under a combination of relatively high juvenile survivorship and relatively low adult survivorship.
Corresponding Editor: G. Wallis