SUMMARYWithin-line family selection was carried out at an optimum (33°C) and two stress temperatures (28°C and 38°C) for increasing egg number laid by virgin females of Tribolium castaneum from the 7th to the 11th day after adult emergence. A control was maintained throughout the experiment. Two replicated lines were selected at each temperature and all selected lines were tested in each of the three environments.Direct and correlated responses to selection in different environments have been analysed in order to study implications of genotype-environment interaction on the selection outcome. Although selection at the optimum environment has been the most effective it did not confer the specialisation needed for performance in stress environments. However, selecting in adverse environments led to a broader range of performance over environments, which in one case (38°C) included the optimum one.The degree of adaptation to adverse environments was mainly determined by the magnitude of the genetic correlation between performances in the adverse and the optimum environments. The evolution of such correlations through selection has also been investigated.

As a potential biological model of non-additive gene effects, the polygenic trait egg lay was studied in the unselected heterogenous Consejo Base Population of Tribolium castaneunl in three environments: optimum (33"C), mild stress (3S°C), and severe stress (28°C). The measured variable was number of eggs laid by virgin females during the peak period of 7-11 days after adult emergence.Egg lay was highest in 33" (18.5 f 0.1), followed in order by 38" (14.8 * 0.1) and 28" (12.9 k 0.1). Similar phenotypic variances in the optimum and mild stress environments were twice those observed in severe stress.Estimates for additive gene effects were larger than those for non-additive effects, but their relative amounts varied with environments. Heritability of egg lay was highest in the optimum environment (0.36 2 0.03), followed by mild stress (0.30 i 0.03) and severe stress (0.25 f 0.03). The reverse was true for estimates of non-additive gene effects where 28" > 38" > 33" with values of 0.15, 0.09 and 0.04, respectively. All genetic correlations for egg lay in different environments were positive and large (0.69 k 0.04 to 0.86 2 0.02).The results suggest that egg lay in Tribolium, when observed in both optimum and stress environments, could provide an appropriate biological model for evaluating mating or selection schemes whose theoretical efficiencies vary with the relative importance of non-additive gene effects.

1. The genetics of plumage colour of a melanotic columbian-restricted male called 'Melanotic Prat' is described. It was found among the F 2 progeny of crosses between the Castellana and Prat breeds, and is very similar to the 'quail' phenotype. 2. Results of crosses involving this male indicate that its plumage colour pattern is the result of the Ml/Ml genotype adding black pigment to the Columbian plumage pattern produced by the e Wh and Co genes. 3. The melanotic gene (Ml) is recessive in the presence of the columbian-restricted genotype. It has no effect on the chick down colour although it could be a modifier to produce greyish or black back colourations when Co is present in the genotype. 4. It is suggested that the quail plumage colour is due to the same melanotic factor interacting with the primary pattern of a Columbianrestricted genotype.

The genetic basis for plumage color in the Prat fowl, a spanish breed with columbian restricted pattern, was studied, and results of crosses between the Prat and a recessive wheaten tester line (ey/ey co+/co+) or the Castellana breed (E/E), showed that it carries a wheaten allele at the E locus. Crosses between the Prat and the Light Brown Leghorn (e+/e+ co+/co+) or the Brown (eb/eb co+/co+) tester lines showed that the wheaten allele is the dominant one (ewh). These crosses suggested that the brown allele (eb) is also present at low frequency. The dominant columbian restriction gene (Co) seems to be involved in the Prat breed, interacting with the dominant wheaten allele to produce the characteristic plumage of the breed. Results of crosses between the Prat and the Vasca, another Spanish breed with gold columbian restricted pattern and ewh/ewh Co/Co genotype showed that they both differ by one major gene affecting phaeomelanin intensity. The Prat breed is well characterized by the orange buff down and by the ginger red adult plumage.

SUMMARYAn experiment was carried out to test the effect of varying selection intensity on the response to individual selection with a fixed number of individuals scored per generation. The selected trait was egg laying of virgin females of Tribolium castaneum scored from the 7th to the 1 lth day after adult emergence. Five different selected proportions of females were considered (5, 10, 20, 33 and 50%) and each treatment was represented by two replicate lines. Control lines were maintained throughout the experiment. The lines selected at the lowest proportions (5 and 10%) led to the largest initial gains, but the largest final gains were achieved, by the lines where the proportions selected were 10 and 20%, in agreement with theory. Lines where the proportion selected was 50% gave the lowest rate of response over the period considered (32 generations). A good agreement was found between predicted and realized short-term responses to selection. Prediction at later stages of selection deteriorated in the most strongly selected lines mainly due to the levels of inbreeding attained.