Last update: April 14, 2011 09:41:00 AM E-mail Print


The effect of selection for fibre production on the fitness of woolled sheep from a breeding perspective


S.W.P. Cloete

Elsenburg Agricultural Development Institute, Private Bag, Elsenburg, 7607


J.J. Olivier

Grootfontein Agricultural Development Institute, Middelburg CP, 5900





Large parts of the RSA are suitable for extensive stock farming mainly as a result of low and unreliable rainfall. It is therefore essential that farmers use well-adapted breeds and genotypes in those parts. The adaptability or fitness of a genotype is defined as its ability to flourish and produce or reproduce optimally in a specific environment Reproduction rate (the percentage of lambs weaned per mating, or weaning percentage) is generally regarded as the best indicator of fitness. It is also of special importance to South African woolled-sheep farmers as the largest percentage of their income is provided by mutton. The other important source of income for these farmers is the wool produced.


It was recently established that the energy requirements for fibre production in Angora goats are approximately four times those for body growth (Herselman, 1990). The relatively high hair production potential of Angora goats is also associated with physiohormonal mechanisms related to poor adaptability which in turn is displayed by the typical problems of Angora goat farming such as low resistance to cold, abortions, poor growth and a poor survival rate (Wentzel & Herselman, 1991). The high price previously obtained for mohair justified the judicious supplementation of feed which solved most of these problems. At present, however, the income from mohair does not always justify the cost of supplementary feeding, stressing the need for an animal which is well adapted to sub optimal environments.


Obviously the question arises whether the same principle applies to woolled sheep, as selection for increased wool production has been practised for many years. This issue is discussed in the following review on the basis of research carried out at Grootfontein in the Karoo and in the Western Cape.


The genetic relation between wool


The genetic relation between wool production and weaning percentage Five genotypes, differing considerably in respect of wool production potential, were developed by crossbreeding Merino ewes with Merino, Dohne Merino, SA Mutton Merino, Afrino and Ronderib Afrikaner rams (Herselman, Olivier & Wentzel, 1993). Wool production, body mass at 18 months of age and the wool production potential of the ewe types are summarised in Table 1. There were only slight differences in wool production between the first four genotypes, whereas the Merino x Ronderib Afrikaner produced considerably less wool. Body mass, on the other hand, was the lowest in purebred Merinos. Wool production potential of the five genotypes decreased progressively from that of the pure Merinos being the highest, followed by Dohne Merino, SA Mutton Merino, Afrino and Ronderib Afrikaner crossbreds.


Fl ewes were mated to FI rams of the same genotype and the reproductive performance was subsequently recorded. From these results it is evident that weaning percentage, weaning mass and ewe productivity (mass lamb weaned/ewe) declined sharply with an increase in wool production potential. The relation between weaning percentage and wool production potential is presented graphically in Fig. 1. From this figure it is apparent that genotypes with a relatively higher wool production potential suffered the most in the sub optimal environment where they had been kept, resulting in major decreases in reproduction rate.





Genetic correlations in the literature, summarised in Table 2, may shed more light on the genetic relations between wool production and weaning percentage within breeds. Genetic correlations with a reproduction trait such as weaning percentage are often unstable because the trait has relatively low genetic variance components. It is, however, clear that the correlations are generally negative and in some cases relatively high. These results therefore support the hypothesis that generally there is a negative genetic correlation between the two traits. Clean-fleece mass in Merino sheep is influenced by fibre diameter, pleat score, body mass, density and staple length. Genetic correlations between these components of clean-wool mass and weaning percentage are summarised in Table 3. Although considerable variation exists, it is clear that both positive and negative correlations between weaning percentage and the components of clean-fleece mass occurred. Negative correlations were recorded throughout between weaning percentage and fold score which is in accordance with overseas results (Atkins, 1980). In respect of body mass, the correlations are predominantly positive, with a few relatively high values. It therefore appears that the latter two components of clean-wool mass have the greatest possibility of affecting weaning percentage through a correlated response.




In order to explain the biological background for the genetic relations obtained between fleece mass and weaning percentage, the interrelationship between clean-fleece mass, body mass, fold score and wool production potential is shown in Fig. 2. Clean-fleece mass is positively correlated with body mass. It seems possible therefore that selection for wool mass can have a positive effect on weaning percentage (through an increase in body mass) or a negative effect (through an increase in fold score). To our knowledge no genetic parameters for wool production potential are as yet available in the literature.



Genetic correlations found when 1 500 Tygerhoek control group progeny were analysed to obtain heritability estimates for the BLUP analyses discussed under the next heading, indicate a negative genetic relationship of -0,21 (SE--o,21) between wool production potential and body mass. Genetic correlations between wool production potential and clean-fleece mass as well as fold score, were, however, positive (0,17 :t 0,10 and 0,28 :t 0,20 respectively). Although not all the estimates are significant, they serve as an indication of the direction of expected genetic change. The relation between body mass and fold score is such that it can be expected that sheep with a high wool production potential will be small and that they will have a high fold score. Both these factors will therefore influence weaning percentage negatively. It can therefore be speculated that the relation between wool production potential and weaning percentage will be negative.


A working hypothesis is that the effect of selection for fleece mass on weaning percentage depends to a certain extent on the emphasis put on either fold score or body mass, while the two effects may partially neutralise one another. If body mass is increased along with fleece mass and fold score reduced, according to this hypothesis, weaning percentage ought not to be affected negatively. This principle is contained in the selection index for Merino sheep which is currently being used (Poggenpoel, 1990). An increase in body mass may coincide with an increase in maintenance requirements and may not necessarily result in a higher biological efficiency, although production per animal will benefit. This aspect, however, falls beyond the scope of this review.


Results of selection trials at Tygerhoek


The two lines at Tygerhoek are selected either for or against maternal ranking values for weaning percentage (the + and - lines respectively). In short, replacement animals of dams that have weaned more lambs than they have had lambing opportunities (i.e. that have weaned twins on at least one occasion) are selected for the + line. On the other hand, mothers of replacement animals in the - line have weaned fewer lambs than they have had lambing opportunities (i.e. they have skipped at least once, or lambed and lost some of the progeny born). Wool and body traits recorded for the different lines were subjected to BLUP analysis to obtain average predicted breeding values, reflecting the genetic change in these lines (S. W.P. Cloete, unpublished). Although the experiment has only been running since 1986 and no final conclusions have been reached, it appears that average breeding values for body mass in the + line are increasing as compared to the - line (Fig. 3). In respect of fold score, there was a corresponding tendency for + line progeny to have less folds. Although it is too soon to quantify the tendencies accurately, the increase in body mass of the + line as compared to the - line amounted to approximately 0,8 % per year, and the decrease in fold score to approximately 1,8 % per year. These tendencies correspond with the genetic correlations in Table 3. With regard to fleece mass the average breeding values for wool production of the + line individuals were generally lower than those of the - line individuals (Fig. 3), with no apparent tendency. The average breeding values for wool production potential, however, showed a declining tendency as discussed in Fig. 2. The decline in the average breeding values of the + line progeny as compared to those of the - line amounted to approximately 0,7 % per year. Selection for weaning percentage has up to now not affected fleece mass negatively, possibly because it was accompanied by an increase in body mass.




In the Tygerhoek clean-fleece mass selection and control lines, selection for clean-fleece mass has been the practice since 1969, while selection for replacement ewes in the control line has taken place on a random basis. Genetic changes in the line selected for clean-fleece mass, body mass and fibre diameter were described by Cloete, Delport, Erasmus, Olivier, Heydenrych & Du Toit (1992). The number of lambs weaned between 1982 and 1987 from ewes that have had five lambing opportunities was determined. In the clean-fleece mass line 147 ewes weaned on average 3,85 lambs during their five lambing opportunities. The performance of 190 ewes in the control line was almost the same, averaging 3,84 lambs over this period. It would therefore appear that the weaning percentage has up to now not been negatively affected by selection for fleece mass. This can possibly be ascribed to the fact that the average body mass of animals in the clean-fleece mass line has increased by 0,86 % per year during the experiment (Cloete et al., 1992). Fold score, on the other hand, has not been changed by selection as suggested by the average performance of the two lines for the 1991 and 1992 progeny groups (Table 4). The increase in clean-fleece mass in the selection line was also accompanied by a relatively smaller change in wool production potential, being 10 % in favour of the clean-fleece mass line. The working hypothesis partially fits the present result, although one would have expected the increase in body mass in the selection line to have been accompanied by a higher weaning percentage. The change in wool production potential could possibly be partially responsible for the fact that the expected improvement in weaning percentage did not realise with the increase in body mass.


Possible genotype x environmental Interactions


Another important aspect is the environment in which selection takes place. According to MacLeod, Howe, Lewer & Woolaston (1990) it appears that the ranking order of rams changes when they are transferred from a stud environment to a commercial environment. In this regard Staken, Soskin, Minina & Levinia (1974) also found that rams with the highest ranking when their progeny were evaluated under optimum feeding conditions, did not have the same ranking when evaluated under sub optimal feeding. A recent investigation in the beef industry indicated that the genetic correlation between the growth of bulls under feedlot conditions and the growth of their halfsisters under veld conditions was almost nil (Theron, Scholtz, Roux & Odendaal, 1992).









These results imply that the same trait in different environments is not necessarily determined by the same genes. In the woolled-sheep industry stud animals are generally kept under more favourable environmental conditions compared to their progeny.


From the study by Olivier (1989) it is clear that the genetic merit for clean-fleece mass in the Grootfontein Merino stud, as reflected by average predicted breeding values, had gradually increased since 1969 (Fig. 4). In the control line at Tygerhoek there were only slight indications of a corresponding increase (Cloete et al., 1992), as reflected by breeding values shown on the same scale in Fig. 4. The overall long-term average (1966-1986) for clean-fleece mass in the Grootfontein stud was 6,7 kg for rams and 4,5 kg for ewes (Olivier, 1989). The average for animals of both sexes in the control line was only 3,9 kg for the period 1971 to 1989 (Cloete et al., 1992). The wool production potential of the Grootfontein Merino stud was about 37 % higher than that of the control line (0,29 vs 0,21 kg/kg WO.75). It can therefore be assumed that the genetic merit for clean-fleece mass and wool production potential in the Grootfontein Merino stud will be considerably higher than that of the Tygerhoek control line.


Recently a control test, as described by Poggenpoel & Vander Merwe (1987), was carried out using rams of the Grootfontein Merino stud and the Tygerhoek control line on commercial Merino ewes at Groo1fontein (J.J. Olivier, unpublished). The results of the test are given in Table 5. The progeny were raised under veld conditions until they reached the age of 18 months. Contrary to expectation the control progeny were heavier and also produced more wool than the stud progeny. The wool of the stud progeny was stronger with shorter staples, a lower clean yield and lower crimp frequency, but with a lower fold score than that of the control progeny.





A possible explanation for this phenomenon is to be found in the fact that the progeny had been raised under Karoo veld conditions. It is possible that the progeny of the control rams, with their lower wool production potential, are better adapted to the limited nutritional regime of Karoo veld when compared to the progeny of the stud rams. This probably enabled them to perform similarly in terms of body mass and wool production, in spite of their lower potential. Similar results are presented in Table 1, showing that the wool production per head of a genotype with a lower wool production potential (Dohne Merino x Merino) was considerably higher in absolute terms than that recorded for purebred Merinos. The progeny of the control rams apparently performed better than their potential, in spite of being more developed. This result is contrary to the working hypothesis formulated previously. It probably suggests that the genetic interrelationships between the various production traits are much more complicated than were realised previously.


The same progeny of these two ram groups (Grootfontein and Tygerhoek) were kept on a high level of nutrition for another year and the results are presented in Table 5. A clear change in ranking order occurred in respect of body mass, greasy-fleece mass, clean-fleece mass, staple length and clean yield. These results are indicative of a typical genotype x environment interaction. Animals with a high wool production potential were apparently unable to realise their potential under a regime of limited feeding. Although these results do not prove beyond a doubt the existence of such an interaction, the practice of selecting woolled sheep under favourable conditions for use under veld conditions should be seriously questioned. Re-evaluation of present practices, as well as more research on the physio-endocrine mechanisms underlying these complex interrelationships, seems to be essential if genetic progress, aimed at improved efficiency, is to be realised.




Estimates of genetic parameters in the literature indicate a negative genetic relation between clean-wool mass per animal and number of lambs weaned per mating (weaning percentage). A working hypothesis was formulated that small, pleated sheep have a high wool production potential. As body mass is generally positively, and pleat score negatively correlated with weaning percentage (an indication of fitness), it was argued that fitness could be negatively affected in sheep with a high fibre production potential. This hypothesis fits tendencies in selection lines for and against weaning percentage at Tygerhoek well. Results of a control test at Grootfontein, where the progeny of control rams under veld conditions performed better than the progeny of plainer Grootfontein stud rams, however, did not support this hypothesis. The fact that the ranking order of the respective progeny groups reversed under favourable conditions, in accordance with the previously determined potential for wool production of the two sets of rams, suggests a possible genotype x environment interaction. The practice of selecting rams under favourable conditions for breeding purposes in commercial flocks must therefore be seriously questioned.  




ATKINS, K.D. 1980. Selection for skin folds and fertility. Proceedings of the Australian Society of Animal Production, 13:174.


CLOETE, S. W.P. 1986. A study of phenotypic and genetic aspects of reproduction rate in the Tygerhoek Merino flock. M.Sc. (Agric.) treatise, University of Stellenbosch.


CLOETE, S.W.P., DELPORT, G.J., ERASMUS, G.J., OLIVIER, J.J. HEYDENRYCH, H.J. & DU TOIT, ELIZABETH. 1992. Environmental and genetic trends in clean fleece mass, live mass and fibre diameter in selection and control flocks involving a selection experiment for increased clean-fleece mass in South African Merino Sheep. South African Journal of Animal Science, 22:50


DELPORT, G.J. 1989. A mixed model approach for selecting Merino ewes. Ph.D. (Agric.) dissertation, University of the Orange Free State.


EIKJE, E.D. 1975. Studies on sheep production records. VII. Genetic, phenotypic and environmental parameters for productivity traits of ewes. Acta Agric. Scand. 25:242.


ERASMUS, G.J. 1988. A mixed model analysis of a selection experiment with Merino sheep in an arid environment. Ph.D. (Agric.) dissertation, University of Orange Free State.


GJEDREM, T. 1966. Selection index for ewes. Acta Agric. Scand. 16:21.


GREGORY, I.P. 1982. Genetic studies of South Australian merino sheep. II. Correlated responses to selection. Australian Journal of AgriculturaIResearch,26:973.


HERSELMAN, M.J. 1990. Die energiebehoeftes van angorabokke. M.Sc. (Agric. )-verhandeling, Universiteit van Stellenbosch.


HERS ELMAN, M.J., OLIVIER, J.J. & WENTZEL, D. 1993. Varying fibre production potentials under field conditions. Karoo Agric. 5(1 ):8.


HEYDENR Y CH, H.J. 1975. ' n Studie van kuddestatistieke, nie-genetiese faktore, genetiese parameters, en seleksievordering met betrekking tot die Tygerhoek Merino-kudde. Ph.D.(Agric.)-proefskrif, Universiteit van Stellenbosch.


KENNEDY, J.P. 1967. Genetic and phenotypic relationships between fertility and wool production in 2-year-old Merino sheep. Australian Journal of Agricultural Research, 18:515.


LEWER, R.P. 1990. Aspects of Merino breeding in South Africa. Report presented to the South African Department of Agricultural Development. R.P. Lewer, Research Fellow, August-November 1990.


LEWER, R.P., RAE, AL. & WICKHAM, G.A 1983. Analysis of records of a Perendale flock. IV. Estimates of genetic and phenotypic parameters for mature ewes. New Zealand Journal of Agricultural Research, 26:309.


MACLEOD, I.M., HOWE, R.R., LEWER, R.P. & WOOLASTON, R.R. 1990. Genotype by environment interactions among Merino sheep in Western Australia. Proceedings of the Australian Association of Animal Breeding Genetics, 8:303-306.


MORE O'FERRALL, G.J. 1976. Phenotypic and genetic parameters of productivity in Galway ewes. Animal Production, 23:295.


OLIVIER, J.J. 1989. Genetic and environmental trends in the Grootfontein Merino stud. Ph.D. (Agric.) dissertation, University of the Orange Free State.


OLIVIER, J.J. 1992. Die belang van subjektiewe eienskappe by wolskape. In: Lesings by 'n seleksiesimposium aangebied deur die Komitee vir KIeinveevoorligting in the Winterreengebied en die Wes-Kaap tak van die SAVDP, 17Maart 1992,31.


POGGENPOEL, D.G. 1990. Systems of breeding wool sheep in South Africa. Proceedings of the World Merino Conference, 2-4 May 1990, Pretoria, South Africa, 1:4.1.


POGGENPOEL, D.G. & VAN DER MERWE, C.A 1987. Selection response with index selection in three commercial Merino flocks. South African Journal of Animal Science, 17:70.


SHELTON, M. & MENZIES, J.W. 1968. Genetic parameters of some performance characteristics of range finewool ewes. Journal of Animal Science, 27:1219.


ST AKAN, G.A, SOSKIN, AA, MININA, E.K. & LEVINIA, I. T. 1994. Genotype-environment interaction and its significance in sheep breeding (abstract). Animal Breeding Abstracts, 42:4109.


THERON, HELENA E., SCHOLTZ, M.M., ROUX, C.Z. & ODENDAAL, MARGA 1992. Die verwantskap tussen produksie-eienskapppe op verskillende voedings- en bestuursregimes. Handelinge van die 31ste Kongres van die SAVDP, 13-16 April 1992, Zithabiseni, B14.


WENTZEL, D. & HERSELMAN, M.J. 1991. Physio-endocrine responses to genetic change in the Angora goat. Proceedings of the 30th Congress of the SASAP, 26-28 March 1991, Port Elizabeth.


YOUNG, S.S. Y., TURNER, HELEN NEWTON & DOLLING, C.H.S. 1963. Selection for fertility in Australian Merino sheep. Australian Journal of Agricultural Research, 14:460.






Grootfontein Agric Vol. 1, No 1, 1995 (1-6)