ASSORTATIVE MATING AND DIRECTIONAL OR STABILIZING SELECTION FOR A NONLINEAR FUNCTION OF TRAITS IN TRIBOLIUM

Assortative or random mating following selection in either direction on a non-linear index (experiment 1) or stabilizing selection for pupal length (experiment 2) were carried out for five generations in two lines of Tribolium castaneum (A and R, respectively), with three replicates each. The selected proportion was 25 % in all lines. In experiment 1, the selection criterion was designed to increase the aggregate value of adult weight and the first- and second-order powers of pupal length. The A and R lines gave significant responses for the aggregate value (184 +/- 6 and 161 +/- 14, respectively), pupal length (0.74 +/-0.02 and 0.64 +/- 0.05, respectively), and adult weight (0.79 +/- 0.03 and 0.78 +/- 0.12, respectively). Although the A line was not significantly better than the R line, there was a consistent advantage for assortative mating over random mating, the mean response for aggregate value and pupal length being a proximately 1.15 times greater for the A line. In experiment 2 the selection criterion was the square of the deviation from the mean pupal length (stabilizing selection); both lines did not show any change for pupal length. The phenotypic variance showed a significant decrease in the A and R lines, due to a decrease in between-family variance. The assortatively and randomly mated lines were similar for these changes in phenotypic variation.