Offspring size effects mediate competitive interactions in a colonial marine invertebrate

Over the past 30 years, numerous attempts to understand the relationship between offspring size and fitness have been made, and it has become clear that this critical relationship is strongly affected by environmental heterogeneity. For marine invertebrates, there has been a long-standing interest in the evolution of offspring size, but there have been very few empirical and theoretical examinations of post-metamorphic offspring size effects, and almost none have considered the effect of environmental heterogeneity on the offspring size/fitness relationship. We investigated the post-metamorphic effects of offspring size in the field for the colonial marine invertebrate Botrylloides violaceus. We also examined how the relationship between offspring size and performance was affected by three different types of intraspecific competition. We found strong and persistent effects of offspring size on survival and growth, but these effects depended on the level and type of intraspecific competition.. Generally, competition strengthened the advantages of increasing maternal investment. Interestingly, we found that offspring size determined the outcome of competitive interaction: juveniles that had more maternal investment were more likely to encroach on another juvenile's territory. This suggests that mothers have the previously unrecognized potential to influence the outcome of competitive interactions in benthic marine invertebrates. We created a simple optimality model, which utilized the data generated from our field experiments, and found that increasing intraspecific competition resulted in an increase,in predicted optimal size. Our results suggest that the relationship between offspring size and fitness is highly variable in the marine environment and strongly dependent on the density of conspecifics.

Designs for generalized linear models with several variables and model uncertainty

Standard factorial designs sometimes may be inadequate for experiments that aim to estimate a generalized linear model, for example, for describing a binary response in terms of several variables. A method is proposed for finding exact designs for such experiments that uses a criterion allowing for uncertainty in the link function, the linear predictor, or the model parameters, together with a design search. Designs are assessed and compared by simulation of the distribution of efficiencies relative to locally optimal designs over a space of possible models. Exact designs are investigated for two applications, and their advantages over factorial and central composite designs are demonstrated.