Density-dependent mortality in young life stages should strongly limit the potential for additive effects caused by stocking of fish sizes that are smaller than size at recruitment into the fishery. Indeed, stocking models have suggested that stocking of fry should not elevate year class strength in self-sustaining stocks. However, limited data based on replicated and controlled experiments are available to support this prediction. We performed a pond experiment (N = 4 per treatment) to compare the stock enhancing outcome of stocking hatchery-reared northern pike (Esox lucius) fry and the natural production of young in self-recruiting pike populations. We also added a treatment where pike fry were stocked into ponds that otherwise did not have pike to mimic the absence of natural recruitment. Fry stocking into self-reproducing stocks did not elevate year class strength over unstocked controls. However, in the absence of competition, year class strength of juveniles in late summer emerging from fry stocking was similar to the production of natural recruits. Overall, we demonstrated the competitive disadvantage of hatchery-reared fry when released into waters already containing natural recruits, the partial replacement of natural recruits by hatchery-reared fry, and the lack of additive effects of stock enhancement in naturally reproducing stocks. A stock-enhancing effect of pike fry stocking may only be expected in the absence of natural recruitment.

In this study, population genetic and demographic parameters were inferred using sequence data from 151 individuals of Anguilla mossambica originating from continental south and south-east Africa and Madagascar. The analyses were based on a 532 bp segment of the mitochondrial cytochrome b gene. The absence of genetic structuring was observed thereby supporting the hypothesis of panmixia for the endemic A. mossambica. The overall nucleotide diversity π = 0.002 and the haplotype diversity reached h = 0.691. Significant negative values from several tests of neutrality and mismatch analysis pointed to fluctuating historical population sizes. Bayesian averaging resulted in higher support for population growth models vs. a constant population-size model. Population decline and subsequent growth most likely predated the last glacial and were probably related to extended periods of extreme drought followed by wetter and more stable hydroclimate between 150 and 75,000 years before present (kBP). According to this scenario the female effective population size has increased since 110 kBP by c. two orders of magnitude to a recent level of about 650,000 (219,317-2,292,000). K E Y W O R D S