Genetic Divergence under Uniform Selection. II. Different Responses to Selection for Knockdown Resistance to Ethanol among DROSOPHILA MELANOGASTER Populations and Their Replicate Lines

We have tested the hypothesis that genetic differences among conspecific populations may result in diverse responses to selection, using natural populations of Drosophila melanogaster. Selection for ethanol tolerance in a tube measuring knockdown resistance was imposed on five West Coast populations. In 24 generations the selected lines increased their mean knockdown times, on average, by a factor of 2.7. An initially weak latitudinal cline was steepened by selection. The two southernmost populations showed the same increases in the selected character, but differed consistently in their correlated responses in characters related to ethanol tolerance. This result indicates that the populations responded to selection by different genetic changes. Selection decreased female body weight and increased resistance to acetone, suggesting components of the response unrelated to ethanol metabolism. The Adhs allele was favored by selection in all populations at the onset, but increased in frequency only in the selected lines of the southernmost population. There was a correlation between latitude and Adh frequency changes, suggesting that fitnesses of the Adh alleles were dependent on the genetic background. Genetic background also had a large effect on the loss of fitness due to selection. Genetic drift between replicate lines caused more variation in selection response than initial genetic differences between populations. This result demonstrates the importance of genetic drift in divergence among natural populations undergoing uniform selection, since the effective population sizes approached those of small natural populations. Drift caused greater divergence between selected replicates than control replicates. Implications of this result for the genetic model of selection response are discussed.     

The perturbation ofDrosophila melanogaster esterase-6 allozyme frequencies by selection for malathion resistance

The effects of selection for resistance to the organophosphate insecticide malathion on esterase-6 polymorphism was studied in laboratory populations ofD. melanogaster. A genetically well-mixed population was constructed from 40 locally-caught, iso-female lines, divided into control and malathion-selected lines and the frequency of the majorEst-6 alleles was followed for more than 100 generations. The main findings were: (1) The allele frequency in control replicates remained stable for over one hundred generations; (2) the allele frequency in the populations exposed to malathion changed dramatically and in the opposite direction between replicates during the early generations of the selection experiment; (3) all selected populations eventually returned to the controlEst-6 allele frequencies. This return was more rapid in populations exposed to lower selection intensity compared to those exposed to higher selection intensity; (4) the convergence ofEst-6 allele frequencies to control values was also observed in three populations obtained by mixing flies of appropriate genotypes from the control population to give different initial frequencies. These results have been interpreted to mean that esterase-6 does not have a direct role in malathion resistance, and that theEst-6 polymorphism in our experimental population was maintained by balancing selection.     

Simulation of marker-assisted selection utilizing linkage disequilibrium: the effects of several additional factors

Use was made of our published model and methods to investigate the effects of several additional factors on marker-assisted selection (MAS) utilizing linkage disequilibrium. The additional factors were: size of the sample used to estimate the marker quantitative trait locus (MQTL) association effects, the method used to estimate the MQTL effects, use of the average of the top MQTL estimates in selection rather than individual estimates, size of the selection population, and the crossing of duplicate selection lines to generate further linkage disequilibrium and further selection response. The average map distance between the quantitative trait loci (QTLs) and their nearest marker was 0.15 Morgans. Use of estimates of MQTL effects derived by least squares yielded smaller selection responses than estimates derived by mixed-model methods. Selection responses were also reduced by using a smaller sample for estimating the associations because MQTL effects were less well estimated. This applied to selection on the MQTL effects themselves and to selection combining the MQTL with phenotypic information. Thus, poorly estimated MQTL effects added noise to the system and reduced selection response in combined selection. Using the average of the top MQTL estimates, rather than individual estimates, also reduced selection response. New linkage disequilibrium, generated by crossing two lines selected from the same population, did not lead to additional selection response in the cross line. These results show limitations to MAS using linkage disequilibrium until close linkages of markers and QTLs are available.     

Energy cost of wheel running in house mice: implications for coadaptation of locomotion and energy budgets

Laboratory house mice (Mus domesticus) that had experienced 10 generations of artificial selection for high levels of voluntary wheel running ran about 70% more total revolutions per day than did mice from random-bred control lines. The difference resulted primarily from increased average velocities rather than from increased time spent running. Within all eight lines (four selected, four control), females ran more than males. Average daily running distances ranged from 4.4 km in control males to 11.6 km in selected females. Whole-animal food consumption was statistically indistinguishable in the selected and control lines. However, mice from selected lines averaged approximately 10% smaller in body mass, and mass-adjusted food consumption was 4% higher in selected lines than in controls. The incremental cost of locomotion (grams food/revolution), computed as the partial regression slope of food consumption on revolutions run per day, did not differ between selected and control mice. On a 24-h basis, the total incremental cost of running (covering a distance) amounted to only 4.4% of food consumption in the control lines and 7.5% in the selected ones. However, the daily incremental cost of time active is higher (15.4% and 13.1% of total food consumption in selected and control lines, respectively). If wheel running in the selected lines continues to increase mainly by increases in velocity, then constraints related to energy acquisition are unlikely to be an important factor limiting further selective gain. More generally, our results suggest that, in small mammals, a substantial evolutionary increase in daily movement distances can be achieved by increasing running speed, without remarkable increases in total energy expenditure.     

Selection for increased length of reproductive life in mice

Summary An experiment was conducted in mice to examine whether selection can increase reproductive life and lifetime production of progeny. Mice in two lines with litter size standardized at birth and in two lines without standardization were pair-mated at 7 weeks of age and maintained as long as they produced litters up to 382 days. Progeny from the sixth litters were used to maintain the four selected lines, while progeny from the first litters were bred to maintain unselected control lines. Selected and control lines were compared at five and six generations of the selected lines. Contemporary comparisons revealed that the length of reproductive life and most lifetime production traits were significantly greater in the selected than in control lines. Realized heritability of the length of reproductive life ranged from 0.08 to 0.13. It was concluded that the length of reproductive life and lifetime production in mice can be increased by selection.Animal Research Centre Publication No. 1618     

Direct Response to Selection for Postweaning Gain in the Rat

The effectiveness of selection for 3–9-week gain was examined in a population of rats with a history of past selection for high 3–9-week gain. Lines were selected for high (U line) and low (D line) 3–9-week gain with two replicates of each line. Two randomly selected lines were also kept, one originating from the same base population as the two selected lines (R line) and the other originating from a population that had been randomly mated for the previous 27 generations (C line). Two replicates of each of these lines were kept. After seven generations of selection, a randomly selected line (relaxed line) was formed from each of the two upward- and each of the two downward-selected lines. Results have been presented for 13 generations of selection. The environmental trend for 3–9-week gain, as indicated by the randomly selected R and C lines, was consistently negative in all four lines. Realized heritabilities calculated by deviating the response to selection from the trend in the R or C lines resulted in non-significantly higher values in the D lines than the U lines. Six generations of relaxation of selection indicated no effect of natural selection in the U lines or the D lines. The relative magnitude of the drift, error and common environmental variances were estimated by the methods given by Hill (1971). The estimates of these parameters then led to calculation of the degree of bias in the sampling variances of the realized heritability estimates. As was predicted by Hill (1971), estimates of the variance of realized heritabilities obtained by using standard regression techniques were less than those obtained using Hill''s formulae. The results are discussed in relation to other similar studies with rats and mice.     

Acclimation and Selection for Increased Resistance to Thermal Stress in Drosophila Buzzatii

Direct selection for increased resistance to a heat shock (41.9° for 90 min) was carried out using two replicate lines of Drosophila buzzatii that were derived from a large base population. Selected individuals were first acclimated to high temperature before selection, while control individuals were acclimated but not selected, and selection was performed every second generation. Resistance to heat shock with acclimation increased in selected lines. Without acclimation, a correlated smaller increase in heat-shock resistance was suggested. Survival of males was higher than that of females in all lines when tested with acclimation, but with direct exposure to high temperatures, survival of females was greater than that of males both in selection and control lines but not in the base population. From analysis of reciprocal cross progeny between lines, one selection line was found to possess a dominant autosomal factor that significantly increased resistance of males much more than resistance of females. Also suggestive was recessive traits on the X chromosome in both selection lines that increased thermotolerance. No cytoplasmic effects were found. After accounting for other effects, survival of F(1) flies was intermediate, suggesting that additive variation is present for one or more of the autosomes.     

Gene number estimation when multiplicative genetic effects are assumed — growth in flour beetles and mice

Summary Estimation of the number of segregating genes affecting a quantitative trait in populations initiated from a cross of two homozygous lines is considered. Experimental data, for the trait in question, is assumed available on total response to recurrent selection initiated in the F₂ or F₃ generation, the initial additive genetic variance and the heterosis exhibited in the F₁ generation. Appropriate procedures when multiplicative genetic effects are assumed are developed and reasons for assuming multiplicative rather than additive effects are indicated. These procedures were employed to estimate the number of genes affecting pupa weight in a population of flour beetles and growth in a population of mice. Estimates were 50–60 percent smaller than those obtained using familiar estimation procedures appropriate when no epistasis is assumed. However, the estimated numbers (about 200 and 100 for pupa weight and mouse growth, respectively) were still rather large.Scientific Journal Series, Minnesota Agriculture Experimental Station. Work supported by the National Science Foundation Grant No. G-1238, the Minnesota Agricultural Experiment Station and by Public Health Service Grant No. 16074     

Confirmation of QTL controlling seed yield in spring canola (Brassica napus L.) hybrids

Allelic effects observed in QTL discovery experiments must be confirmed to be useful in subsequent breeding efforts. Two QTL affecting seed yield of spring hybrid canola (Brassica napus L.) were previously identified in two populations of inbred backcross lines (IBLs) containing germplasm introgressed from a winter cultivar. The effects of favorable alleles at these QTL were retested by crossing two selected IBLs (M5 and M31) to three spring canola lines having different genetic backgrounds. Doubled haploid (DH) lines derived from each F₁ were genotyped with RFLP markers flanking the QTL and grouped into the four possible QTL genotypes. For the first field experiment, DH lines derived by crossing the M5 line to one spring line were crossed to two female testers and evaluated as individual testcross progenies in one environment. QTL genotypes had large variances and were not significantly different. A second field experiment was conducted using the DH lines from the first experiment and two other sets of DH lines derived from the M31 line crossed to two different spring canola lines. Individual lines within each QTL genotype of each set were bulked and crossed to the same testers used in Experiment 1. Bulked hybrid seeds of each QTL genotype were planted in a split-split plot randomized block design and 12 replicates. QTL genotypes had smaller variances in this experiment, and the effects of one QTL were confirmed in some genetic backgrounds. These results suggest that bulking of QTL genotypes and use of an appropriate experimental design with many replicates are needed to detect small differences between QTL genotypes.     

EXPERIMENTAL EVOLUTION OF RESISTANCE IN BRASSICA RAPA: CORRELATED RESPONSE OF TOLERANCE IN LINES SELECTED FOR GLUCOSINOLATE CONTENT

The evolutionary response of plant populations to selection for increased defense may be constrained by costs of defense. The purpose of this study was to investigate such constraints on the evolution of defense due to a cost of defense manifested as a trade-off between defense and tolerance. Variation in the response to artificial damage (tolerance) among lines of Brassica rapa that had been artificially selected for foliar glucosinolate content (defense) was examined. Leaf area was removed from replicates of three selection lines (high glucosinolates, control, and low glucosinolates) at three damage levels (0%, 20%, and 60% damage). An external cost of defense would result in a statistically significant selection line by damage treatment interaction, with those selected for high defense expressing less tolerance than those selected for low defense. Damage treatment had a significant overall effect on estimated total fitness, with fitness declining with increasing damage level. Further, selection line also had a significant overall effect on estimated total fitness, with low-defense selection lines having higher fitness compared to both control and high-defense selection lines. More importantly, a cost of defense in terms of tolerance was demonstrated by a significant selection line-by-damage treatment interaction. This interaction was in the direction to demonstrate a genetic trade-off between defense and tolerance, with low-defense selection lines decreasing estimated total fitness in response to damage less than both control and high-defense selection lines. Variation in tolerance among selection lines was due to the greater ability of low-defense lines to maintain fruit and seed production despite the presence of damage. In terms of tolerance, this cost of glucosinolate production in B. rapa could constrain the evolution of increased defense and, in so doing, maintain individuals within the population that are poorly defended yet tolerant.     

Population Size and Selection Intensity Effects on Long-Term Selection Response in Mice

Long-term response to within full-sib family selection for increased postweaning gain was evaluated in lines having different effective population sizes (N_e) and selection intensities (i). Line designations were I4(4), I8(2), I16(2), M4(4), M8(2) and M16(2), where I and M indicate selection of the top 50% and 25%, respectively; 4, 8 and 16 represent the number of parental pairs per replicate and number of replicates is given in parentheses. Realized within full-sib family heritabilities (h_R²) in the first phase of selection (0-14 generations) were larger in 16-pair lines than in 4- and 8-pair lines. In the second phase of selection (>14 generations), h_R² declined significantly (P<.01) in all lines, and only the I16 and M16 lines had h_R² values significantly (P<.01) greater than zero. Realized genetic correlations involving number born, 12-day litter weight, weaning weight and six-week weight tended to decline in the second phase of selection. The I16, M16 and control (C16) replicates were crossed in all combinations at generation 14. Crosses were then selected within litters for high postweaning gain. The h_R² values in the crossbred lines were all larger than those in the second selection phase for M16-1, M16-2 and I16-1, but not for I16-2. Within each N_e level, total response was significantly (P<.01) less for I lines compared with M lines. Total response increased as N_e increased, within each level of i. Relatively small differences in realized i values among N_e lines could not account for this result. The difference in total response among the N_e lines at a given selection intensity may be due to inbreeding depression and a combination of interactions involving "drift" and selection. By crossing replicates of the M lines with the C16 control, the effects of inbreeding depression were removed. Inbreeding depression and genetic drift, as defined herein, were equally important in accounting for differences among N_e lines in total response.     

Geographic variation and the evolution of reproductive allocation in the pitcher-plant mosquito, Wyeomyia smithii

We measured the egg size of six geographic populations of the pitcher-plant mosquito, Wyeomyia smithii, from Florida (30 degrees N) to Ontario (49 degrees N). Populations from northern latitudes produced larger eggs than populations from southern latitudes. Egg size increased with increasing latitude more rapidly when larvae were reared under low rather than high density. One southern (30 degrees N) and one northern (49 degrees N) population of W. smithii that persisted through 10 generations of selection for increased persistence under conditions of chronic thermal- and nutrient-limiting stress (conditions similar to southern rather than northern habitats) produced smaller eggs more rapidly than unselected control lines. However, there were no differences in lifetime fecundity or fertility between control and selected lines. Thus, laboratory evolution in an environment representative of extreme southern latitudes caused evolutionary changes consistent with geographic patterns of egg size. These results implicate temperature as a selective factor influencing the geographic variation of egg size in W. smithii, and demonstrate a novel trade-off in reproductive allocation between egg size and egg maturation time.     

Environment-dependent trade-offs between ectoparasite resistance and larval competitive ability in the Drosophila-Macrocheles system

Costs of resistance are expected to contribute to the maintenance of genetic variation for resistance in natural host populations. In the present study, we experimentally test for genetic trade-offs between parasite resistance and larval competitive ability expressed under varying levels of crowding and temperature. Artificial selection for increased behavioral resistance was applied against an ectoparasitic mite (Macrocheles subbadius) in replicate lines of the fruit fly Drosophila nigrospiracula. We then measured correlated responses to selection in larval competitive ability by contrasting replicate selected and control (unselected) lines in the absence of parasitism. Experiments were conducted under variable environmental conditions: two temperatures and three levels of larval density. Our results reveal a negative genetic correlation between resistance and larval-adult survival under conditions of moderate and severe intra-specific competition. At both low and high temperature, percent emergence was significantly higher among control lines than selected lines. This divergence in larval competitive ability was magnified under high levels of competition, but only at low temperature. Hence, the interaction between selection treatment and larval density was modified by temperature. As predicted, larvae experiencing medium and high levels of competition exhibited an overall reduction in female body size compared to larvae at low levels of competition. Female flies emerging from selected lines were significantly smaller than those females from control lines, but this effect was only significant under conditions of moderate to severe competition. These results provide evidence of environment-dependent trade-offs between ectoparasite resistance and larval competitive ability, a potential mechanism maintaining genetic polymorphism for resistance.     

Rflps for Somatotropic Genes Identify Quantitative Trait Loci for Growth in Mice

Restriction fragment length polymorphisms for somatotropic genes were tested for associations with body weight and postweaning growth rate in mice. Polymorphisms for growth hormone (GH) and insulin-like growth factor 2 (IGF-2) genes were identified in stock population lines which had been subjected to long-term selection for high 42-day body weight (H lines) or randomly mated (FP and C lines). Two F2 populations of mice (5F2 and MF2) were generated from crosses between a single H line of mice and two unselected control lines and subsequently, two divergently weight selected sublines were generated from each F2 population. The GHh allele which had originally been fixed in three of four H lines and absent from all FP and C lines was found to have a significant (P less than 0.01) effect on 42-day weight and postweaning growth rate in the F2 populations. However, GHh was associated with lower 42-day weight in the F2 populations, suggesting that the positive association between GHh and weight in the stock population was unique to the high weight selected genetic background of those lines. In agreement with this, the frequency of GHh increased in sublines selected for high 42-day weight and decreased in sublines selected for low 42-day weight. The IGF-2H5 allele was associated with higher weights in a sex-dependent manner in 5F2. In the high selected subline derived from 5F2, a significant increase in the frequency of IGF-2H5 was observed. Therefore this allele, in contrast to GHh, appears to be a positive indicator of growth irrespective of genetic background.     

Correlated responses to selection for large body size in oMt1a-oGH transgenic mice: reproductive traits

Correlated responses in female reproductive performance were evaluated following short-term selection within full-sib families for increased 8-week body weight in two replicates of four lines of mice: two ovine metallothionein-ovine growth hormone (oMt1a-oGH) transgene-carrier lines, one from a high-growth background (TM) and one from a control background (TC), and two non-transgenic lines, one from each of these genetic backgrounds (NM and NC, respectively). A fifth line (CC), not containing the transgene, served as a randomly selected control. The initial frequency of the oMt1a-oGH transgene construct in the TM and TC lines was 0.5. The frequency of transgenic females sampled at generations 7 and 8 of selection was 84.0% and 6.1% in the TC and TM lines, respectively. No significant female infertility differences were detected between transgene-carrier and non-transgenic lines or between transgenic and non-transgenic mice within carrier lines, whereas high-growth background lines had a higher infertility than control background lines (P < 0.05). Correlated responses in the TC transgene-carrier line were suggestive of reduced reproductive performance as indicated by increased post-implantation mortality (P < 0.05), number of dead fetuses plus implants (P < 0.05), and loss of fetuses from day 16 to parturition (P < 0.001). For the first two traits, the negative correlated responses were accounted for by the reduced performance of transgenic compared with non-transgenic females. Embryos carrying the transgene may also have a lower viability. In contrast, the NC non-transgenic line did not exhibit reduced reproductive performance for these traits. The low frequency of the transgene in the high-growth background TM line was associated with reduced fitness and a lower additive effect for 8-week body weight compared with the control background TC line.     

Food consumption and body composition in mice selected for high wheel-running activity

The effects of genetic selection for high wheel running (13th generation) and prolonged access (8 weeks) to running wheels on food consumption and body composition were studied in house mice (Mus domesticus). Mice from four replicate lines selected for high wheel-running activity ran over twice as many revolutions per day on activity wheels as did mice from four replicate control lines. At approximately 49 days of age, all mice were placed individually in cages with access to wheels and monitored for 6 days, after which wheels were prevented from rotating for the "sedentary" individuals. During the experiment, five feeding trials were conducted and body mass was measured weekly. After 8 weeks, body composition was measured by hydrogen isotope dilution. Across the five feeding trials, mice in the "active" group (wheels free to rotate) consumed 22.4% more food than mice in the "sedentary" group (wheels locked); mice from the selected lines consumed 8.4% more food than mice from the control lines (average of all trials; body mass-corrected values). In females, but not males, we found a significant interaction between selection and wheel access treatments: within the "active" group the difference in food consumption between selected and control animals was greater than in the "sedentary" group. At the end of the study, mice from the "active" and "sedentary" groups did not differ significantly in body mass; however, mice from the selected lines were approximately 6% smaller in body mass. Estimated lean body mass did not differ significantly either between selected and control lines or between wheel-access groups (P>0.3). Mice from selected lines had lower total body fat compared to mice from control lines (P=0.05; 24.5% reduction; LSMEANS) as did mice from the "active" compared to "sedentary" group (P= 0.03; 29.2% reduction; LSMEANS). Under these conditions, a sufficient explanation for the difference in body mass between the selected and control lines was the difference in fat content.     

Designing artificial selection experiments for specific objectives

The observed genetic gain (ΔP) from selection in a finite population is the possible expected genetic gain E(Δ G) minus the difference in inbreeding depression effects in the selected and control lines. The inbreeding depression can be avoided by crossing the control and selected ♂ and ♀ parents to unrelated mates and summing the observed gains. The possible expected gain will be reduced by an amount D from the predicted gain because of the effects of the genetic limit and random genetic drift, the magnitude of which is a function of effective population size, N. The expected value of D is zero in unselected control populations and in the first generation for selected populations. Therefore, this source of bias can be reduced by increasing N in the selected populations and can be avoided by selecting for a single generation. To obtain observed responses which are unbiased estimates of the predicted response from which to estimate the realized heritability (or regression) in the zero generation, or to test genetic theory based on infinite population size, single-generation selection with many replications would be most efficient. To measure the "total" effect or genetic efficiency of a selection criterion or method, including the effect of different selection intensities, effective population sizes, and space requirements, more than one generation of selection is required to estimate the expected response in breeding values. The efficiency, in the sense of minimum variance, of estimating the expected breeding values at any generation t will decline as the number of generations t increases. The variance of either the estimated mean gain or the regression of gain on selection differential can be reduced more by increasing the number of replicates K than by increasing the number of generations t. Also the general pattern of the response over t can be estimated if the N's are known. Therefore, two- or not more than three-generation selection experiments with many replications would be most efficient.     

Indirect selection for increased susceptibility to permethrin in diamondback moth (Lepidoptera: Plutellidae)

We have been exploring the behavioral response ot insect pests to heterogeneous distribution of toxins (low dose with refugia), and its genetic correlation with physiological tolerance to these toxins. A field-collected population of diamondback moth, Plutella xylostellu (L.) (Lepidoptera Plutellidae), from Celeryville, OH, was selected with permethrin to determine whether low heterogeneous doses could lead to increased susceptibility to permethrin by selecting indirectly on behavior. Two replicates of each of three selection regimes: uniform high concentration hypothesized to result in increased physiological tolerance, heterogeneous low concentration hypothesized to result in increased susceptibility through indirect selection on behavior, and a control with no exposure to permethrin, were maintained in 1-in3 cages in a greenhouse, for 33 generations. All life stages of the diamondback moth were exposed to the selection regimes, and new generations were started with a random selection of pupae from the previous generation. Lines selected with uniform high concentrations developed 76-fold levels of resistance to permethrin by the 17th generation, with little changes thereafter. For generations 1-20, lines selected with heterogeneous low concentrations remained slightly lower in LC50 but not significantly different from the unselected control lines. Based on confidence intervals from probit analyses, the LC50 of the lines selected with heterogeneous low concentration, however, were significantly lower than those of the control lines in generations 21-33. Our results demonstrate that selection on behavioral responses can result in greater susceptibility than no selection at all, despite exposure to the toxin and ample genetic variation and potential for increased physiological tolerance. The implications of our findings, which are based on selection scenarios that could take place in field situations, are that behavioral responses can prevent and even decrease the levels of resistance in insect populations, an important result with respect to resistance and resistance management.     

Mutational variance for pupa weight in Tribolium castaneum

Summary Directional selection for heavier pupa weight in Tribolium castaneum was practiced for 18 generations in two replicates of an inbred line, each separately maintained in small population cages for more than 90 generations. Mutational variance was estimated in two ways, based on Hill's (1982a) prediction equation for response to directional selection where an equilibrium state between effective population size and variation created by new mutation is assumed. Estimates of mutational variance based on response to selection in a selected population and from a sire-offspring regression analysis in an unselected control population were in strong agreement within each replicate population. Significant differences between the two replicates were observed. Estimates of the ratio of mutational variance to environmental variance ranged from 0.0002 to 0.0012, depending upon the assumptions made about effective population sizes maintained in the two replicate lines. Estimates of realized heritability from the 18 generations of selection were 0.23±0.02 and 0.12±0.02 in the two replicates. The results support the hypothesis that mutation may have played a significant role in supplying useful genetic variation for long-continuing response to selection for this trait in experiments reported earlier.     

Effects of Population Size and Selection Intensity on Short-Term Response to Selection for Postweaning Gain in Mice

The effects of population size and selection intensity on the mean response was examined after 14 generations of within full-sib family selection for postweaning gain in mice. Population sizes of 1, 2, 4, 8 and 16 pair matings were each evaluated at selection intensities of 100% (control), 50% and 25% in a replicated experiment. Selection response per generation increased as selection intensity increased. Selection response and realized heritability tended to increase with increasing population size. Replicate variability in realized heritability was large at population sizes of 1, 2 and 4 pairs. Genetic drift was implicated as the primary factor causing the reduced response and lowered repeatability at the smaller population sizes. Lines with intended effective population sizes of 62 yielded larger selection responses per unit selection differential than lines with effective population sizes of 30 or less.