1. An experimental study has been made of recurrent selection to an inbred tester. A suitable inbred line is used as a tester parent, and selection is made within a non-inbred population on the individuals crossing performance with the tester line.

It is concluded that there are two situations in which recurrent selection could be profitably applied. Firstly, recurrent selection should, theoretically, be successful when applied to characters closely related to fitness which have little additive genetic variance and secondly, in cases where a character has already been subjected to individual or family selection and has reached a plateau level in that population.

2. The two experiments—i.e. recurrent selection for large litter size in mice and for low bristle number in Drosophila melanogaster—reported here are respectively an example of each of the above situations. In each experiment selection was made between males within the closed non-inbred population on the basis of the performance of their testcross progeny resulting from matings with inbred line females.

3. Initial generation hybrid performance in both experiments was not intermediate between parental performance levels and the divergence from intermediacy was away from the direction of selection.

4. In both experiments there was no evidence to suspect the presence of over dominance.

5. Response to selection was obtained in each experiment but this was close to or less than the expected response calculated on the assumption that all the variance between sires in crossing performance was additive genetic variance.

6. From these experiments it is not possible to draw any firm conclusions about the effectiveness of recurrent selection for exploiting overdominance. It is, how ever, a very inefficient way of exploiting additive genetic variance. It is suggested that more success might be obtained by careful choice of base population material used in recurrent selection.

1. The effect of genetic sampling, when this sampling is without replacement, on variation in gene frequency is studied, and equations describing the genetic drift are derived. The effective size turns out to be about one greater than under sampling with replacement.

2. The relation between ‘spread of genes’ and genetic drift is worked out.

3. The University of Queensland control poultry flock is analysed by these methods.

4. The design of control populations is discussed with particular reference to the relative importance of genetic drift and phenotypic sampling.

1. Mass selection for both high- and low-ratio of wing to thorax length has been carried out on a population of Drosophila melanogaster. The response to selection was immediate and sustained. When the experiment was stopped after ten generations, the wing area in the two selected lines differed by about 30%. The heritability estimate worked out at 0·56 ± 0·08.

2. Thorax length remained comparatively unchanged during selection nor was there any change in wing shape. There was some evidence of assymetry of response since there was a relatively greater change in favour of smaller rather than larger size.

3. The tibia length of all pairs of legs showed correlated changes so that the lines with larger or smaller wing sizes had also larger and smaller legs.

4. The normal allometric relation between wing and thorax length, associated with variation in body-size, apparently also changed, so that for a given change in thorax length there was a greater or smaller proportional change in wing size in the high- or low-ratio lines.

5. The changes in relative wing size are due to changes in cell number.

6. It is suggested that the genetic changes due to selection act in the early pupal period when the imaginal discs are undergoing differentiation and proliferation to form imaginal hypoderm and appendages.

7. Tests of genetic behaviour failed to show any departure from additivity in crosses which involved the unselected population and the high-ratio line. But highly significant departures existed in the cross to the low-ratio line. Relatively smaller wing size behaves as largely recessive. Stability of the normal wing/thorax ratio involves dominance and probably also epistasis. The genetic properties of the relative size of the appendage are apparently similar to those which characterize body-size as a whole.

8. It is suggested that selection provides a valuable tool for studying the constancy or lability of the growth patterns which determine morphology.

Young birds homozygous for the recessive gene, sc, had a significantly lower requirement than their normally feathered siblings for the sulphur-containing amino-acids, methionine and cystine. When fed a synthetic diet lacking in methionine, feathered birds showed the expected abrupt decrease in weight gain while Scaleless birds grew at a rate equivalent to feathered birds receiving the full diet. Scaleless birds receiving the full diet exceeded feathered birds in weight gain during the first part of the feeding trial but they required more food per gram of gain. Between Scaleless groups those fed the deficient diet dropped behind those on full diet in gain after 3 weeks but even after 6 weeks these did not differ in other respects. On the other hand among normal birds those receiving the deficient diet were higher in heat output and had relatively larger thyroid and adrenal glands than those receiving the full diet.

Differences between the two phenotypes are accounted for by (1) the difference in nutritive requirements determined by feather growth, and (2) the resultant differences in heat loss which accounted for the increase in heat output, packed cell volume and heart, thyroid and adrenal gland size of Scaleless birds.

The use of mutants such as Scaleless in studies concerned with requirements for specific physiological or growth purposes is suggested.

1. Techniques were developed to facilitate genetic analysis of a Myxomycete, Physarum polycephalum. The uninucleate cells that form part of the life-cycle (the ‘amoebae’), in two-membered culture with Pseudomonas fluorescens, were found to be suitable for handling by some of the standard techniques of microbial genetics, including plating.

2. As reported previously, a single-factor (+/−) mating-type system was discovered in P. polycephalum controlling the formation of plasmodia by the amoebae. Plasmodia are formed only when amoebae of different mating-type are mixed together. A single spore gives a culture of amoebae of one mating-type only, which may be assumed to be a clone.

3. A strain of amoebae resistant to emetine hydrochloride was isolated and was found to maintain its resistance through many subcultures in the absence of the drug.

4. A cross was made between the emetine-resistant strain (mating-type −) and an emetine-sensitive strain (mating-type +). Spores obtained from the hybrid plasmodium were plated and four sets of isolates of amoebae were made from these spores, and separately tested for mating-type and for emetine resistance.

5. The results of testing these isolates agree with the hypothesis that mating-type and emetine resistance are determined by pairs of alleles at two unlinked loci provided that factors disturbing the segregation for emetine resistance are assumed to be present.

6. Consideration of factors that could have disturbed the segregation for emetine resistance shows that selfing did not occur in the cross but that some misclassification for emetine resistance probably did occur.

Proteins of pollen of Oenothera organensis from heterozygous S2S3, S2S6, S3S6 and S4S6 and homozygous S6S6 genotypes produce characteristic precipitation lines in agar-gel plates against antisera prepared from the pollen. The maximum number of precipitation lines was three, and with some genotypes only one line was formed. The lines were characteristic of the different S alleles, but there was a strong cross-reaction between S4 pollen and S6 antiserum. The S protein diffused out from intact pollen grains as readily as from macerated pollen indicating that the normal site of action of the S protein, when the pollen tubes are growing down the style, is near the surface of the pollen tube. The self-compatible mutant S6′S6 in which the S6′ allele has lost the ability to produce an active pollen S protein, has a protein which cross-reacts strongly with S6 antiserum.

In tetrad data obtained from the Carbondale yeast stock, instances of greater than 50% recombination occur far more frequently than would be expected by chance. In the main, genes exhibiting this effect are located in the vicinity of the centromeres. Difficulties in assigning linear order also persist throughout the total data for certain gene combinations. The affinity hypothesis suffices to account for these two effects; the former resulting from ‘divergent’ combinations of affinity sites in the zygote, the latter, representing nonlinear quasi-linkages resulting from ‘convergent’ combinations.

Chromosome V (ur3-centromere-ch–hi1–is1–an) exhibits quasi-linkage with the gene th and reverse-linkage with the gene cu2, in Family 108. In each case ‘linkage’ is strongest with ch and becomes progressively weaker with the genes hi1, is1 and an, respectively, indicating that the preferentially segregating sites involved lie close to ch. It was impossible to determine whether the two sites were identical, as would be expected on an hypothesis of ‘polar’ rather than ‘mutual’ affinity.

Intra-ascal matings within tetrads NPD for cu2–ch, yielded quasi-linkage for these genes in F2, showing that the attraction is retained by the sites being segregated.

A second case of reverse linkage for cu1 and ar4 in Family 217, gave a similar effect—i.e. reverse linkage in F1, becoming quasi-linkage in the NPD—intra-ascal F2.

An intra-ascal mating between members of the single PD tetrad for cu1 and ar4 in Family 217 indicated that this tetrad did not result from ‘chance’ failure of otherwise active sites to preferentially segregate.

1. Replacement of the dominant genes M1 and M2 in Paramecium aurelia, stock 540 (syngen/variety 1), results in loss of ability to maintain mu particles and manifestation of mate-killing after a delay of eight to fifteen fissions in most cells. The change, when it does occur, is relatively abrupt, extending over less than the space of one inter-fission period.

2. The delay between change of genotype and loss of mu particles is interpreted as being due to presence in the initial cytoplasm of some thousand ‘metagons’, which are non-replicating gene derivatives having the physiological activity of the corresponding genes. During successive fissions of paramecia deprived of M1 and M2 the metagons are passively distributed amongst the progeny, until virtually all animals lack them.

3. On reaching a stage at which some individuals of genotype m1m1m2m2 contain only a single metagon, the paramecia still contain large numbers of mu particles and are mate-killers. Fission of such animals gives rise to one daughter again with mu particles, and another in which the latter are destroyed during the next inter-fission period.

4. By induced cytoplasmic exchange between conjugants, metagons can be transferred from one animal to another via the cytoplasm. Where such transference is into an animal not originally containing mu particles, that animal is converted into a condition in which it favours the maintenance of mu particles and transmits the latter to one or more of its offspring.

5. Distribution of metagons amongst progeny of dividing paramecia is not random, due possibly to clumping of the metagons. Induced cytoplasmic exchange seems to break up the clumps.

6. Reintroduction of a dominant gene (M2) into a cell recently deprived of the same gene, succeeds—even after fifteen fissions—in re-establishing the ability to support growth of mu particles, provided that the recipient cell contains at least one metagon and one or more mu particles. There is a regular lag of only one fission between introduction of such a dominant gene and its phenotypic manifestation.

7. Mathematical formulae are developed for calculating the expected initial number of metagons, the proportions of animals lacking mu particles at each fission following loss of the dominant genes, and the proportions of cells containing 0, 1, 2 …, etc. metagons per cell at any stage. The consequences of one of the possible types of irregular distribution of metagons in dividing paramecia are also considered mathematically.

The length of time of development, from oviposition to emergence in Drosophila melanogaster was subjected to stabilizing selection. In each generation only the individuals emerging close to the mean development time were used as parents of the next generation. This line was designated the ‘S’ line. In a parallel line disruptive selection was practised; where in each generation the earliest flies to emerge were mated to the flies last to emerge; those emerging at intermediate times were discarded. This line was designated the ‘D’ line. Two control lines were also carried, where the flies were mated at random with respect to time of emergence. The experiment extended for 40 generations and produced the following results:

(1) The variance of development time decreased in the S line and increased in the D line, relative to the control lines.

(2) The mean development time decreased in the S line and increased in the D line.

(3) The coefficients of variation decreased in the S line and increased in the D line.

(4) The viability, measured as per cent flies emerging, decreased in the D line.

Toward the end of the experiment the amount of additive genetic variance in the selected lines and in the control lines was estimated from the response to directional selection. The estimates showed that (i) the loss of total variance in the S line can be accounted for completely by a loss in additive genetic variance, and (ii) the increase in the total variance of the D line cannot be ascribed to an increase in the additive genetic variance. It was probably due to an increase in the environmental component of variance, i.e. to a loss of ‘buffering capacity’.

Shrivelled, a new mutant in the house mouse, is characterized by degeneration of the fibres of the ocular lens, followed by the development of cataracts. The condition is caused by a mutant autosomal gene showing intermediate dominance.

The efficiency of food utilization, the digestibility of foodstuffs, energy metabolism, and body activity have been studied in three lines of mice, one selected for large, another for small body size, and a third, control, line.

The gross efficiency of food utilization was highest in the large line, intermediate in the control line and lowest in the small line between 21 and approximately 35 days of age. During this period, gross efficiency declined in the large and control lines with increasing size and decreasing growth-rate, presumably due to an increase in maintenance costs in comparison with the weight gained. In the small line, the efficiency of food utilization increased up to 35 days of age but declined thereafter. The energetic efficiency (measured in Calories) was higher in the large than in the small line up to 4 weeks of age, i.e. when the growth-rate was high, and after 6 weeks of age, when fat was being deposited at an increased rate.

The increased efficiency of large mice was not entirely associated with a greater proportion of the ingested food being absorbed from the gut. Large mice absorbed a greater proportion of protein, though the difference was not sufficient to account for the large weight difference between the large and small lines.

The energy expenditure of mice of the large line was greater than that of the small line at all ages and similar for the same body weights. The reduced growth-rate of small mice was not due to abnormally high or low energy costs. There was no evidence that body activity determined or restricted the rate of growth in either line.

Mice selected for small size were phenotypically unlike pituitary dwarf mice, although the low nitrogen retention during the growing-period indicated a deficiency of some growth stimulus.

An attempt was made to modify a near-limit of selection response in the egg production of a closed poultry flock which had been under selection for 31 years; to this end the laying hens were subjected to weekly periods of 24-hour feed deprivation, and selected for egg number under these conditions.

In two replicate populations genetic gains of about 20 eggs were made under shock treatment to 40 weeks of age, over 5 and 6 generations of selection. These gains were also manifest under normal management.

Selection, under 24-hour starvation per week, also improved the ability of the selected populations to lay under weekly starvation periods of 33 and 48 hours respectively. Differences between selected and unselected lines increased with the more severe stress treatments. Also it was found that the difference in average egg production between 33-hour and 48-hour treatments was larger in unselected lines than in the selected ones, indicating the existence of genotype-environment interactions.

Correlated responses of reduced egg-weight and body-weight, and somewhat earlier maturity of selected lines correspond with what might be expected under normally effective selection for egg number. These adaptations however varied considerably between replicate lines.

The linear relationship between fourteen allelic p-aminobenzoic-acid-requiring mutants of Aspergillus nidulans was investigated. The mutants were assigned to twelve different mutational sites within one functional region. No case of intra-cistron complementation was found. The order of sites as determined by the distribution of the outside markers in different crosses between paba alleles is consistent and the recombination fractions between the sites are approximately additive. Crossing-over between alleles is associated with strong localized negative interference. The topography of the pabal region is characterized by a marked concentration of the mutational sites in a small part of the cistron.

A method for the experimental estimation of generation interval is presented together with results obtained at 25° C. for D. melanogaster Oregon-R-C maintained in population cages and in population bottles, and for D. simulans v in population bottles. Although there is significant variation between the replicate estimates obtained in population bottles, and although a number of potential sources of error have been discussed, it is suggested that this method provides useful operational estimates of the parameter, which may be taken as 23 days for D. melanogaster Oregon-R-C in population cages, 14 days for the same stock in population bottles, and 16 to 18 days for D. simulans v in population bottles.

The ancestory of Koller's subline of the C57BL strain is not known. A study of the minor skeletal variants in this subHne showed that it differs from the other ten sublines of C57BL more than they difier from each other. It appears to have separated from. the rest before the genotypes of the sublines had been fixed by inbreeding. There is enough evidence to show that it is indeed a branch of C57BL strain.

The rate of divergence of the sublines of C57BL is surprisingly high. It is reconsidered whether the whole of this divergence can be explained by recent mutations or whether residual heterozygosis has also played a part in it.

1. The problem of improving rate of larval development of Drosophila by selecting for this ‘character’ on deficient diets is examined by culturing larvae axenically on low-casein and low-pyridoxine media. Under these conditions it is possible to develop strains which grow faster than the parent population.

2. Selection for fast- and slow-growing larvae on a low-pyridoxine diet proceeds with a realized heritability of about 20%, but progress ceases after eight to nine generations. The selected larvae show no alteration of pyridoxine requirements up to the tenth generation, but the lines develop at different rates under optimal conditions. This difference is exaggerated when the diets are low in pyridoxine. By the fourteenth generation, requirements of the two lines for optimal growth have become distinct, the fast line requiring less pyridoxine than the control. Casein requirements show about the same optimum for the two lines but this is lower than that of the control, foundation population.

3. Selection for fast- and slow-development lines on a low-casein diet continues to be effective throughout the fifteen generations of the test. The realized heritability in this case is about 10%. The optimal requirements of the two lines are the same, and there is little difference in their development rates when reared on this optimal diet. The response is found only under sub-optimal conditions, both of deficiency and of excess casein. Pyridoxine requirements do not seem to be altered in the two lines.

4. Crosses between the selected lines show that each genotype has its own optimal environment, as judged by pyridoxine and casein requirements. Crosses among the lines after fourteen generations show that all the hybrids are superior to the mid-parent and three grew faster than the better parent. Other environments would have given different results.

The time of action and morphological pattern of damage of eight recessive autosomal lethals in Coelopa frigida (F.) is described. Seven embryonic lethals are phase-specific in their time of action but lethal glassy may cause death at any time from the beginning of the third instar onwards.

A method for estimating the preadult viability of heterozygotes for lethals is described. Heterozygotes for lethal E/L3 show some impairment of viability, whereas, under conditions of larval competition heterozygotes for lethal E/L4 are superior in viability to their wild type sibs.

A sex-linked recessive anaemia (symbol sla) in the mouse is responsible for a rather uncharacteristic blood picture which persists virtually unchanged throughout life. The red cell count is reduced to about three-quarters of the normal value, the cells produced differ from normal ones by a slight reduction in the mean corpuscular Hb concentration and a slight reduction in the mean cell volume. As the mean cell diameter is more strongly reduced, it is concluded that the cells must be thicker than normal. There is a reduction of haemopoietic tissue both in the liver and in the bone marrow.

The gene ‘sex-linked anaemia’ (sla) appeared first in the progeny of a daughter of a male that had been irradiated with 500 r. of X-rays. The daughter was born one month after the irradiation of her father. Mated to an unrelated wild-type male she produced sixteen sons of which six were anaemic. Her fertility was normal. Experience in subsequent matings showed the anaemia to be rather difficult to classify. In litters examined on the day of birth the anaemic young were usually clearly distinguishable by their pale colour. But in older litters, and even in a few at birth, the classification could not be made with confidence. The haematology of the anaemics is described by Grewal (1962).

The competitive-index method of measurement of over all fitness in Drosophila has been used to measure the effect of inbreeding and of artificial selection for metric characters in a large population of Drosophila melanogaster. The technique itself was examined in detail with particular reference to its repeatability and to the effect on it of the modification of various environmental variables.

With continued full-sib mating the decline in the competitive index was very rapid (it was reduced to a half by a single generation of full-sib mating) and there were no indications that interactions between deleterious genes at different loci were important in determining the rate of decline of fitness as inbreeding increased. Other unselected lines with ten pairs of parents in each generation were carried to serve as a control for the lines under artificial selection. At the same theoretical degree of inbreeding the control lines had a much higher average fitness than the lines produced by continued full-sib mating.

From the base population lines were selected in both directions for abdominal bristles, sternopleural bristles and for wing length, there being two replicates in all cases. Four control lines were kept with the same number of parents as the selected lines. In all cases the selected lines declined in fitness below the value for the base population. However, in three of the lines the fitness was not significantly below the value for the control lines. The effect of artificial selection on fitness was asymmetrical, the decline being greater with down selection for all characters.

The relevance of these results to various theoretical models is discussed. If the variation in these characters is actively maintained in the base population by the selection of heterozygotes then the results are consistent with an average selection disadvantage of homozygotes relative to heterozygotes of about 0·5%.