1. Several lines of Drosophila melanogaster have been selected for increase of sternopleural chaeta-number.

2. Three lines of the same origin, dp 1, dp 2 and dp 6, showed remarkably similar patterns of response involving an accelerated response from 24 to 28 chaetae and a plateau at about 30 chaetae. A line formed by crossing two of these was not responsive to selection, suggesting that the two parent lines were genetically similar.

3. A fourth line, vg 4, related to these others, showed a similar accelerated response but continued to respond beyond 30 chaetae, reaching a plateau at about 37 chaetae.

4. A further line, vg 6, set up by crossing two ‘improved’ lines, one with 30, the other with 22 chaetae, reached 46 chaetae and suggests there may sometimes be merit in selecting from the hybrids of ‘improved’ strains.

5. It is argued that the event making possible the accelerated response is probably a recombinational event.

6. The line vg 4 clearly demonstrates that some of the genes that may be exploited in a selection experiment have to be exploited in a particular order because of gene interactions.

A theory of mass selection in a small population was developed, and the mean change in gene frequencies, the variance of gene frequency changes and the expected gain in the mean phenotypic value of an offspring population were formulated in terms of a generalized selection differential and the additive and dominance effects of genes.

The magnitude of the variance of changes in gene frequency was compared with the magnitude of the variance expected from the genetic random drift in a population with the same gene frequency and of the same size in absence of selection. The former was found to be usually smaller than the latter when the gene frequency ranged from intermediate to high and when selection was directed for a high performance.

The usual prediction equation for gain from selection in an infinite population was compared with the expected gain formula derived for a small population. The size of the population did not cause a serious difference between the two expected gains when there was no dominance effect of genes. Dominance alone could cause the usual prediction to be slightly more biased. The joint effects of the finite size of population and dominance gene action could amount to a considerable bias in the usual prediction equation. Such a bias can be, in the main, accounted for by the inbreeding depression.

In a population under artificial selection, the effective population size may be less than the actual number of parents selected because there will be variation between families in the character under selection and consequently in the probability of selection. Expressions are developed for the magnitude of the effect, which will be greater the more intense the selection and the higher the heritability of the selected character. The inbreeding due to outstanding individuals may rise for several generations after their use.

The regression of progeny on mid-parent value is often used in progeny tests to estimate the heritability of a quantitative character. The statistical precision of such an estimate can be considerably increased without increasing the size of the test, by using assortative mating or selection of parents (or both together) so as to increase the mid-parent variance; but the danger arises that this may introduce bias into the estimate through correlation between non-additive gene effects.

It is shown by a mathematical argument that such bias will be negligible provided that all individual gene substitution effects are small compared with the phenotypic standard deviation of the character. Under this condition, deviations from additive effects either within or between loci will not appreciably affect the expected value of the regression on mid-parent, compared with its expected value in a test using random mating.

Correlation between the non-additive gene effects is likely to cause more serious bias to the correlation between sibs, when non-random mating is used.

1. Many of the Ustilaginales, or smut fungi, appear to have the qualities necessary for the application of modern techniques of microbial genetics. Ustilago maydis is considered the most suitable species.

2. Investigations of the mating system confirm reports that the production of diploid brandspores in the host is controlled by alleles at two loci.

3. Genetic markers were obtained by inducing mutations in a wild-type strain with ultra-violet light. Of 100 biochemical mutants which were isolated, the growth requirements of 94 were identified. Thirty of these were used in genetic tests.

4. The compact growth of colonies on artificial media allowed new techniques to be developed by means of which large samples of progeny could be isolated and identified easily. The analysis of brandspore colonies consisting of the products of single meiotic divisions is the quickest method for detecting linkage, but its accurate measurement appears to be achieved by examining the individual members of tetrads.

5. Linkage was detected relatively rarely, but eight markers, including the a mating-type locus, were assigned to one or other of two linkage groups. Although recombination values were not always determined accurately owing to irregular basidiospore germination, the auxotrophic markers in each group could be mapped in a linear order. Since no indication of other linkage groups was obtained, the genetic evidence is so far consistent with cytological reports that the basic haploid chromosome number is two in the smut fungi.

6. Three linked markers were used to investigate chromatid interference by tetrad analysis. None was detected in a total of eighteen double exchanges.

1. Two different methods of selection were used to obtain stable prototrophic solopathogenic strains of Ustilago maydis from haploids with different biochemical requirements. The strains were shown to be heterozygous diploids.

2. Two diploids which were examined in detail had four markers in common, but they were in different coupling and repulsion phases. Rare spontaneous segregation was detected in one of the diploids, but a high frequency of segregation was obtained in both after treatment with ultra-violet light. The proportion of segregants amongst the survivors increased with the dose. These auxotrophic segregants were detected by means of the replica plating technique.

3. The types of segregant which were obtained from both diploids were consistent with the view that they arose as a result of mitotic crossing-over. After low doses of radiation the reciprocal products of crossing-over were often detected. There was no evidence from the phenotypes of the segregants that haploidization was occurring. The diploidy of a sample of the segregants was confirmed by mating-type tests, and by the fact that they showed further segregation after another dose of radiation.

4. A slow-growing unstable segregant recovered after a high dose of radiation proved to be a monosomic strain which consistently reverted to a stable diploid homozygous for one chromosome. It was possible to use this auxotrophic diploid together with a haploid with different biochemical requirements, to synthesize a prototrophic triploid strain. The triploid was much less stable than the diploid strains.

5. By means of pathogenicity tests with certain segregants it was possible to distinguish the function of the two loci which control the mating system. The a locus is responsible for the fusion of haploid sporidia and has no effect on the pathogenicity of the heterokaryon which is under the control of the b locus.

6. The effects of ultra-violet light on mitotic crossing-over do not seem to be easily compatible with a copy choice or similar mechanism of recombination.

An account is given of the habitat and ecology of a wrackbed population of Coelopa frigida (Fab.). The spatial distribution of the third-instar larvae was found to be patchy. This is due to a preference of adult flies for oviposition sites in zones of high temperature within the wrackbed. The relationship between larval density and temperature was found to account only in part for the patchy distribution, a second factor being the formation of family aggregates within each temperature zone. The effect of this type of spatial distribution on the interpretation of genetic sampling methods for the population is discussed.

Samples of third-instar larvae from the wrackbed population have been analysed for the presence of recessive embryonic lethals in three consecutive years from 1957 to 1959. From a total of 528 gametes tested, 44 recessive embryonic lethals were isolated. Cross tests for allelism showed that certain lethals were repeated within samples taken in the same year.

The gross frequency of embryonic lethals is shown to be independent of seasonal changes in population density. Estimates of gene frequency are made for certain individual lethals and these are shown to be greater than expected on the basis of the census size of the population. Possible causes of this discrepancy are examined.

Midget, a new dwarfing gene in the house mouse, was found in a line (NS) selected for small body size. Backcrosses to, and heterozygous matings of NS mice gave midget offspring in the expected 1:1 and 1:3 ratios. Backcrosses to two other lines of mice selected for small body size resulted in fewer midget offspring than expected and in a reduced phenotypic expression of the gene, though matings between these midget offspring gave all midget progeny.

A characteristic check in growth occurred between 15 and 21–24 days of age, and midgets rarely exceeded 10 g. at 6 weeks of age. They were 1·2 g. lighter than NS litter-mates at 3 weeks, and 3·6–4·4 g. lighter at 6 weeks. Fostering or transferring 3½-day-old midget embryos to the uteri of large mothers slightly increased the growth-rate, though the characteristic growth check still persisted. Transfer as embryos did not influence the midget phenotype.

Midget males were of near-normal fertility. The females were semi-sterile due to irregularity or absence of oestrous cycles, failure to ovulate, or the death of embryos before implantation. Each of these factors, which were also found in NS mice, could be remedied by the administration of exogenous hormones.

The genetic factors underlying the expression of midget are discussed, and the physiological traits associated with this and with other dwarfing genes are compared.

1. A new hereditary defect affecting the hair coat of Swiss albino mice is described. The defect was found to be inherited as a recessive character and is designated as bare (ba).

2. The gene affected the vibrissae as well as pelage hairs. The lack of vibrissae and other sensory hairs at birth, helped to differentiate them easily from the normal animals. The first hairs on the body appeared at 13–14 days of age. The hairs were thin and tiny and remained there till about the thirtieth day. The hairs in the second hair cycle appeared at about 45 days, and again disappeared within 10 days. The animals were entirely naked when they were 6 months old.

3. Gross as well as microscopic examination of hairs in the first hair cycle did not show the four different normal types of hairs. The hairs were comparatively very small, thin, and the internal structure did not show any regularity in the arrangement of the air cells.

4. No difference was found in the number of hair follicles in the bare and the normal Swiss mice.

5. Histology revealed the presence of keratinized globular masses instead of straight hair. This abnormality persisted in all the hair cycles.

6. The bare mice were compared with the other mutants and placed in the Alopecia-Naked group where the abnormality was in the keratinization of the hair.

Dappled, a new completely sex-linked semi-dominant lethal mutation, is described. It is probably a new allele at the Mottled locus and has been given the symbol Modp.

Heterozygous adult females have a variegated coat very similar to that of Brindled and Mottled females; they have curly vibrissae at birth and some also have clubbed feet. All the characteristics are variable.

Hemizygous males die at about 17 days' gestation; they show a characteristic bending and thickening of the ribs and other skeletal defects. The cause of death is unknown.

The experiment was designed to form a bridge between the results of specific-locus experiments, using only a few gene loci, and those using the whole genome of the mouse. Male mice were given 600 r acute X-rays and bred from in such a way that at successive stages mutation in spermatogonia to dominant visibles and lethals, dominant semisteriles, recessive visibles and recessive lethals could be measured. The data concerning dominant mutations were relatively few but confirmed previous results. No recessive visible mutations were found, and the upper fiducial limit to the induced mutation rate to recessive visibles was set at a value 4500 times the rate to viable specific-locus mutations. From the attempt to measure recessive lethal mutations two interesting points emerged. The first was that granddaughters of the irradiated males had fewer corpora lutea per pregnancy than granddaughters of the control males, and the second was that this difference in number of ova shed was not reflected in any difference in litter-size at birth. Since this suggests intra-uterine compensation, no attempt was made to calculate mutation rates to recessive lethal genes from these data. The implications of the results are discussed.

In independent investigations, reported almost simultaneously (Cohen, 1960; Shreffler, 1960), the authors have described a protein variant system in the serum of mice, involving an electrophoretic difference in the iron-binding beta-globulin component, transferrin, and have shown the difference to be genetically determined by a pair of co-dominant alleles at a single locus. Both reports introduced the same gene symbol, Trf, for this locus, but in one, letter designations were used (Cohen, 1960), making the allele symbols TrfA and TrfB, and the three phenotypes TrfA, TrfAB and TrfB, while in the other, number designations were used (Shreffler, 1960), with the corresponding symbols Trf1, Trf2, Trf-1, Trf-1·2 and Trf-2. To avoid confusion in future reference to this locus, due to the two differing systems of nomenclature, the authors have agreed upon the revised nomenclature proposed in this communication.