The genes coding for the 18s and 28s ribosomal RNA (rRNA) are present on both the X and Y chromosomes of D. melanogaster at a site known as the bobbed locus. Exchange was observed in males between a normally orientated X and Y chromosome (Dp(1; 1) scv1 and BSY y31d) with a frequency of 0·079%. One-quarter (7 in 27) of these exchange products between two + chromosomes which both carried sufficient rRNA genes for a bb+ phenotype exhibited a bb phenotype. Evidence is presented that one-half, and possibly all, of the exchanges involved the repetitive bb genes. These results together with those reported by Palumbo, Caizzi & Ritossa (1973) imply that the repeated bb genes of either (or both) the X or Y chromosome are not arranged with uniform polarity and, further, that spermatogonial exchange between the X and Y chromosomes may be restricted to the bb loci.

Mice selected for large size show increases in both food intake and efficiency, and small mice show decreases in both. This is true whether the comparisons are made at the same age or at the same weight. Food intake and efficiency contributed more or less equally to the responses to selection for growth. Mice seem to regulate their food intake to a certain level of energy. On suspension of a period of food restriction, mice ate the same amount as others of the same strain that had not been restricted, and which were bigger. At the same time, they converted it more efficiently than the mice which had been full-fed throughout, because of a linear negative association between efficiency and body weight. Thus, following restriction, mice eat as much as bigger mice of the same age, and convert it as efficiently as younger mice of the same weight. The product of these two effects gives rise to rapid (compensatory) growth.

Aggregation chimaeras were made from embryos of strains of mice selected for large and small body size and of unselected controls. The strains were combined in pairs marked by albino coat colour and by allo-zyme variants at the Gpi-1 locus. The proportion of cells derived from each component was scored visually in the coat melanocytes and by electrophoresis in ten other organs or tissues (blood, liver, lung, spleen, spinal cord, brain, pituitary, kidney, adrenal and testis). The object was to find out how body weight is related to cell proportions in the body as a whole and in the separate organs. Individuals varied widely in their mean cell proportions but there were significant differences between organs within individuals. Body weight was linearly related to the mean cell proportions which accounted for most, or possibly all, of the chimaeric variance of body weight. No one of the organs studied could be identified as being solely responsible for growth control, or as having a predominant influence on growth. The weights of some organs were probably influenced to a small extent by their own cell proportions independently of the individual's mean, but the differences of body weight were too great to be accounted for by the summation of localized effects on organs. The mean cell proportion, averaged over individuals, was close to 50%, proving that there was no tendency for cells from the larger component to outgrow those from the smaller. It is concluded that growth control must be systemic, but it was not possible to decide whether the systemic effect comes from some particular organ not studied, or is in some undefined way the consequence of the cell proportions in the body as a whole. There was some evidence, though it was inconclusive, that chimaeras show ‘heterosis’ for body weight.

An electrophoretically detectable variant of acid α-glucosidase has been found in SM/J mice. This variant is attributable to excess sialylation of the enzyme and is determined by a gene, alpha-glucosidase processing, Aglp, on chromosome 17. In addition, as also reported by Potier, Lu Shun Yan & Womack (1979), SM/J mice are relatively deficient in neuraminidase and it appears that the low level of this enzyme in SM/J is determined by an autosomal codominant gene, neuraminidase-1, Neu-1. Preliminary data indicate that Neu-1 is also on chromosome 17. It seems probable that the several processing genes Apl, Aglp and Map-2 which are all closely linked on chromosome 17 are one and the same, a gene Neu-1 coding for neuraminidase.

Previous studies have shown that the mouse X chromosomal locus, Xce, which causes non-random X chromosome inactivation, is closely linked to the Is(X; 7)Ct X-autosome translocation. This has placed it either near Ta on one side of the breakpoint or near jp on the other. Linkage tests with Movbr and Ta now demonstrate that the locus in fact lies close to Ta. The data also provide genetic evidence which establish that the C3H/HeH X chromosome carries the Xcea allele of this gene and the JU/FaCt and C57BL/GoH X chromosomes carry the Xceb allele, and further suggest that the X-linked modification of the heterozygous phenotypes of X-linked genes observed by various other investigators are all attributable to differences at the Xce locus. Evidence of a maternal influence upon Movbr phenotypes is also presented. This appears to operate independently of the X-inactivation process, probably through an effect of differing levels of copper in the milk in early life upon the mutant coat colour in the young.

The relationship between locus variability and genetic identity is analysed in pooled data from a collection of Drosophila and vertebrate allozyme surveys. For vertebrates the analysis provides evidence that loci with high expected heterozygosity diverge at a faster rate than loci with low heterozygosity in comparisons involving pairs of populations and pairs of species. This relationship is not observed in comparisons of Drosophila species pairs.

Thirty-one inbred strains were tested for their reaction to drinking water which contained a low concentration (10−4 M) of sucrose octaacetate (SOA). One strain, SWR, showed a strong aversion to drinking the SOA solution. The other thirty strains, and two samples of wild-derived mice, tended to prefer the SOA solution to untreated drinking water. The pheno-typic difference between SWR and the other strains was shown to be determined by an autosomal gene. The allele present in SWR is dominant. The gene is not closely linked to jerker (je), pearl (pe) or waved-2 (wa-2).

I develop a mathematical model which can account for a distribution of the number of repeated genes per chromosome under the joint effects of sister chromatid exchange (SCE), inter-chromosomal crossing-over (ICC), and selection. The model can be applied not only to the cases of small gene clusters but also to multigene families. Based on this model, an appropriate mathematical formula is derived and used to obtain the equilibrium distribution. Assuming stabilizing selection and two simple schemes concerning SCE and ICC, I numerically calculate the equilibrium distribution and compare the result with observations on frequencies of single and triple α-haemoglobin genes in primates. It is also shown that if SCE and ICC occur according to the same probabilistic law, the distinction between them does not make much sense in the equilibrium distribution.