||<p/>To evaluate the effect of selection for parameters of a growth curve, four selection lines and a control line were started from one base population. In the selection lines is selected for a large and a small relative growth rate between 21 and 29 days (RGH and RGL) and for a large and small bodyweight at 56 days (W56H and W56L). Besides the direct results of selection, attention was paid to the consequences for some fertility traits, the growth curve, the efficiency and composition of the growth.<p/>Selection in RGH and RGL lines was within litter to exclude maternal effects, which are of importance at 21 and 29 days. One male and one female were chosen per litter. In the C, RGH and RGL lines 16 pairs per generation were mated, according to a system of maximum avoidance of inbreeding. In W56H and W56L mass selection was applied. Per generation 32 females and 16 males were selected and randomly mated. The material consisted of data from the first 14 generations of these lines.<p/>Inbreeding coefficient in generation 14 was 10.5 % in C, RGH and RGL lines and 18.4 % and 18.9 % in W56H and W56L. Relative growth rate (RG) decreased in C during generations. There was no significant change of weight at 56 days (W56D) in C.<p/>Realised heritability of selection for RG was 0.06 ± 0.03 in RGH and 0.10 ± 0.02 in RGL. A combined estimation from divergent lines gave 0.08 ± 0.01. It followed from further analysis that the effectivity of selection was low in first generations. Cumulative selection differential in RG was small, mainly as a consequence of within family selection. There were large differences between W56H and W56L. In the last generations, mice in W56H were over twice as heavy as in W56L at 56 days. Realised heritability for W56D was 0.28 ± 0.01 in W56H and 0.35 ± 0.02 in W56L. A combined estimation from divergent lines gave 0.31 ± 0.01.<p/>There were only small differences in tail length between C, RGH and RGL. The same applied to conception rate, timespan between mating and littering, litter size and survival rate between 0 and 56 days. Tail length increased in W56H and decreased in W56L. These changes in tail length indicate differences in size between the lines. Days between mating and littering increased in W56H and W56L; number of litters decreased substantially in both lines. Litter size increased in W56H and decreased in W56L. Survival rates were similar in the lines. As a consequence the number of mice at 56 days decreased during selection in these lines. This decrease was in W56L very strong and selection differential in females became almost zero in some generations.<p/>Rate of death was increased from generations 5 to 8 of all lines.<p/>Growth curves of the lines were described by the logistic function.<p/><em>W <sub><font size="-1">t</font></sub></em> = <em>A</em> /(1 + <em>be <sup><font size="-1">-kt</font></SUP></em> )<br/><em>W <sub><font size="-1">t</font></sub></em> = weight at age <em>t</em><br/><em>A</em> = asymptotic weight<br/><em>b</em> = integration constant<br/><em>k</em> = parameter, that determines the spread of the curve along the time axis<p/>The tentative choice of this function followed from literature, while a definite choice was based on a test of this function in the base population. In this base population weights were estimated daily between 18 and 61 days. Correlations between observed and calculated weights were of the level (r = 0.97) as reported in literature. However there were some systematic deviations. In both sexes weights were overestimated in the first part of the growth curve and they were underestimated in the later part. Correlations between calculated and observed weights did not decrease as the interval between observations increased from 1-7 days.<p/>Growth curves were described in generations 6-14 of the 5 lines. Weights were observed weekly between 1 and 8 weeks. Residual variances in these lines, except W56L, were smaller than in the base population, probably because variation in the observed weights in the base population was smaller. This effect of the variation in weight might also explain the higher residual variance in females rather than in males and the low values in W56H, as follows from Table 4.2. Residual variance increased significantly during generations in W56L.<p/>The judgement of the differences in growth curve between lines was based on the comparison of parameters <em>k</em> and <em>A</em> and calculated age at inflexion point (t). Changes in mean parameter values were used as indications for changes in growth curves, while line differences and sex differences were estimated in generations 13 and 14. Selection for large RG decreased <em>k</em> and increased <em>A</em> and <em>t <sub><font size="-1">i</font></sub></em> . Selection for small RG had an opposite effect. Differences between RGH and RGL were significant (p≤0.001). In W56H <em>k</em> decreased, while A and <em>t <sub><font size="-1">i</font></sub></em> increased. W56L showed opposite effects. Differences between W56H and W56L were large and significant (p≤0.001).<p/>Males had a smaller <em>k</em> and a larger <em>A</em> and <em>t <sub><font size="-1">i</font></sub></em> than females. Sex differences were with some exceptions significant. Within-litter correlations between RG and parameters of the logistic function were estimated (Table 4.6). Partial correlations were calculated too while <em>A</em> was kept constant (Table 4.7). It followed from these calculations that the direct relation between RG and <em>k</em> was positive, while the indirect relation via the relation of RG and <em>k</em> with mature weight was negative. The differences between the selection lines showed, that the indirect relation was stronger than the direct one, because the realised correlation between RG and <em>k</em> was negative.<p/>A comparison of maturity rates, defined according to TAYLOR and FITZHUGH 1971, showed that selection in RGL and W56L gave an increased maturity rate, while maturity rate in RGH and W56H was decreased. However changes in maturity rate were accompanied by opposite changes in mature weight. The lack of systematic crossing of growth curves might be explained by the strong positive effect of changes in mature weight on weights at young ages. These effects were probably stronger than the effect of a change in maturity rates. Plotting weights and ages on a relative scale showed that differences in mature weight and in time taken to mature, did not determine all differences in the growth curve. W56H and RGH had a smaller percentage of mature weight than W56L and RGL at the same percentage of time taken to mature. Females had a higher percentage of mature weight than males at the same percentage of time taken to mature (Figure 4.5).<p/>To examine whether there were differences in composition and efficiency of growth between the selection lines, growth, feed intake and body composition were observed in samples of mice. These samples represented mean genotypic value of generation 11 of each of the lines. Weekly weights and feed intake were determined between 3 and 15 weeks. Body composition was estimated in two replicates per sex per line of 6 mice each at 3, 4, 5, 6, 7, 8, 9, 11,13 and 15 weeks of age.<p/>Males had a larger growth rate than females. Differences between RGH and RGL were small; RGH had a higher growth rate until 10 weeks. Growth rate in W56H was much larger than in W56L during the whole observed timespan. Mice in W56L hardly increased in weight at all after 6 weeks of age. Feed intake was rather constant after 6 weeks. Before this age it was increased, probably by compensatory growth and extra activity during puberty. Differences between RGH and RGL were small. Feed intake in W56H was much higher than in W56L. Males ate more than females.<p/>Feed intake and bodyweights were strongly positively correlated. Feed efficiency decreased substantially in all lines until 6 weeks and remained rather constant afterwards. Males had a higher efficiency than females until 8 weeks. Differences between RGH and RGL were not systematic. Feed efficiency in W56H was much higher than in W56L.<p/>Ash percentage increased with aging. Differences between lines and between sexes were small and not systematic. Differences in fat percentage were much larger. It increased with aging, especially in W56H. Females had a higher fat percentage than males. Initially, RGH was fatter than RGL. Differences between W56H and W56L were very large. In W56H fat content increased up to 1/4 of total bodyweight. Protein content remained rather constant with aging. However there were two exceptions; protein percentage increased substantially until 9 weeks and decreased afterwards to the level of the other lines. This indicated that weight increase in this line consisted more of protein and less of fat until 9 weeks than in the other lines. W56L had a large protein percentage at young ages. Water percentage was low. Water percentage decreased with aging. These differences between lines and between sexes appeared also in the analysis of differences in body composition at 8 weeks (Table 5.4).<p/>Correlations, calculated within lines, sexes and ages between fat percentage and other components were all significantly negative. Correlation between water percentage and fat percentage was -0.91. Correlation between protein percentage and water percentage was 0.32. The other correlations between components were not significant.<p/>Efficiency of weight increase was calculated during intervals from 3 weeks onwards. Gross energy efficiency decreased with aging, except for females in W56L which had a very low efficiency from the beginning. Differences between RGH and RGL were small and mostly in favour of RGH. Efficiency in W56-H was much higher than in W56L. Males grew more efficiently than females. Fraction of energy intake deposited in protein decreased with aging. In males it was higher than in females. In the first weeks RGH was slightly higher than RGL, but afterwards this was reversed. Differences between W56H and W56L were large and in favour of W56H. However W56H was only slightly different from C. So the higher feed efficiency in W56H was to a large extent based on fat deposition. From the analysis of line differences and sex differences during the interval of 3 - 8 weeks it followed that RGH took in 9 % more gross energy, had 10 % higher gross energy efficiency; however there was 23 % less protein deposited per unit gross energy intake than RGL. W56H took in 40 % more gross energy, had 145 % higher gross energy efficiency and deposited 85 % more protein per unit gross energy intake than W56L. For sex differences these numbers were 20 %, 32 % and 40 %, in favour of males.<p/>If the amount of deposited protein per unit gross energy intake was used as a criterion for meat production ability of a line, it could be concluded that RGL and W56H were similar and the best. Thus the higher feed efficiency in W56H in comparison to RGL was determined by the larger fat deposition in W56H.<p/>It was concluded from the same criterion that males were much more efficient for meat production than females.