The Haldane-Muller principle of mutation load is generalized so as to be applicable to both cases of strong and very weak selection with any time variation. It is proved that in an infinite asexual haploid population, the average Malthusian parameter m¯ of a population, the evolution rate ν, and the total mutation rate μ satisfy the relation ∂m¯/∂/∂μ = ν/μ−1, so long as each Malthusian parameter is independent of μ. A similar result is also true in a diploid population under genie selection. It is discussed how the above relation gives a restriction on the possible range of values of relative evolution rate ν/μ.

Ontogenetic series of phenotypic, additive genetic, maternal and environmental correlation matrices are presented and interpreted in the light of recent models for the Ontogenetic origin and variation in correlation between traits. A total of 432 mice from 108 full-sib families raised in a cross-fostering design were used to estimate the various components of phenotypic correlation for five live-body traits at eight ages. The level of genetic and phenotypic correlation decreased with age, while levels of maternal and environmental correlation remained more or less constant. Genetic correlations probably decreased due to compensatory growth. Phenotypic correlations decreased primarily due to the relative decrease in importance of highly correlated maternal effects and consequent increase in poorly correlated environmental effects as portions of phenotypic variation. The effect of compensatory growth on genetic correlation was also responsible for a portion of the decline in phenotypic correlation. Phenotypic correlation patterns remained constant over the ages studied here. It also seems likely the genetic, maternal and environmental correlation patterns do not change with age for the characters analysed.

Aggregation chimaeras were made from embryos of C57BL/6 and BALB/c mice. Chimaeric and control females were mated with ICR males at 8 weeks of age and their litter sizes were evaluated over a 5-month period after the first mating. Progeny tests showed that 18 of 27 chimaeras produced oocytes of both genotypes. The mean litter sizes of C57BL/6, BALB/c and their F1 crosses (C57BL/6 × BALB/c and BALB/c × C57BL/6) were 8·14, 9·36, 13·38 and 13·40, respectively. The mean for chimaeras was 11·54 and chimaeric heterosis was evident, but it was not as much as heterosis in the F1 When the chimaeras were classified into the mixed and single-genotype progeny chimaeras, chimaeric heterosis was observed only in the mixed-progeny chimaeras. Quantitative GPI analyses in ten organs showed that the degree of chimaerism in the mixed-genotype progeny chimaeras was higher than that in most of the single-genotype progeny chimaeras and that the degree of chimaerism in the ovaries was positively correlated with litter size in the mixed-genotype progeny chimaeras. On the other hand, such correlation was not observed in the single-genotype progeny chimaeras.

The average effect and average excess of a gene substitution are formulated in terms of gene frequencies and inbreeding coefficient. This clarifies their meanings and shows how each is affected by non-random mating. The meanings of various definitions are examined, and one commonly used definition of average effect is found to be invalid with non-random mating. The concept of breeding value is shown to have no useful meaning when mating is not random.

The radiation-induced agouti allele a1 is recessive to the alleles a and a4H (which resembles ae). It is lethal when homozygous and also in combination with the dominant yellow allele Ay. The ethylnitrosourea induced allele a16H is also lethal when homozygous, and when heterozygous with a shows a phenotype like that of ax, with black back and lighter belly. Like ax it is not lethal with Ay, and it is also not lethal with a1, a1 is believed to be the first recessive allele which is lethal with Ay, and may be useful in elucidating the complexity of the agouti locus.

An experimental evaluation of Jódar & López-Fanjul's (1977) theoretical treatment of the optimum proportions to select when the numbers of males and females scored are unequal has been carried out for sternopleural bristle number in Drosophila melanogaster. Three different values of the sex-ratio (c) were considered (c = 1, 4 and 10) for the same total number of individuals scored per generation. For each c value two types of line were selected with proportions theoretically maximizing the response to be attained after 10 or 20 generations, respectively. Thus, there were six types of lines and each type was replicated sixfold. A good qualitative agreement was found between the observed and the expected rankings of the different types of selected lines at the designated generations.

Genetic polymorphisms revealed two distinct substrains of the inbred strain 101. One group included substrains 101*/Rl, 101/H and 101/HOxe; the other group comprised 101/El and 101/Sl. The two groups differed at 5 of the 8 genetic loci tested. The accompanying paper (Evans, Burtenshaw & Adler, 1985) shows that the two groups also differ for several chromosome polymorphisms. We suggest that genetic contamination occurred during the derivation of 101/El from 101/Rl and was already present in 101/El when 101/Sl was produced from this substrain. We further propose that these substrains be renamed 102/El and 102/Sl respectively.

A centromeric heterochromatin (Hc) comparison of two substrains of the mouse inbred strain 101 with another two, revealed differences in at least two and possibly in up to six chromosome pairs. The differences could have arisen by the inclusion of chromosomes from the inbred strain C3H/He.

DNA sequence comparison among homologous genes sampled at random from one or two populations allows one to estimate the ultimate amount of genetic variation maintained in a population and to construct the gene genealogy within and between populations. Moreover, if we use the finding of the molecular clock (Zuckerkandl & Pauling, 1965), it is also possible to estimate the divergence time of populations examined. Such an estimated divergence time is, however, intricately affected by samples and stochastic forces occurring in the course of evolution.