Studies on the behavioural and genetic aspects of ewe rearing ability and lamb survival in South African sheep flocks

English: The dissertation is based on a number of separate trials conducted since 1989. The central theme is ewe rearing ability and lamb survival, and the study includes 15 papers prepared over an 11-year period from 1992 to 2002. These papers are all linked by their emphasis on sheep production as influenced by lambing and neonatal behaviour and its interface with breeding and management. The papers demonstrate the various phases the study went through, and how it eventually transpired into an account of genetic change in lamb output achievable through rigorous selection. The individual papers already contain abstracts. This summary is intended to provide readers with a broad overview of the most important outcomes of the study. The study was organised in three parts, as follows: Part 1:Background Initially, two papers provided the background for the investigations into lamb survival and ewe rearing ability. The first paper dealt with the average reproductive performance of sheep in the Southern Cape, and with levels of reproductive wastage in a number of flocks. It demonstrated that approximately one in four breeding ewes did not care for at least one lamb at lamb marking. Losses were more or less evenly distributed between barrenness and rearing failure. It was impossible to identify managerial practices conclusively associated with a good reproductive performance. Evidence was provided that an improvement in the reproduction of the current flock can be achieved by a simple method of selection that can be practiced by commercial farmers with minimal record keeping. The second paper considered rearing ability in four experimental flocks of the Merino, SA Mutton Merino and Dormer breeds. A minority of ewes was shown to contribute markedly to the observed levels of rearing failure. Animals classified as being "good" or "poor" mothers on their lamb rearing history were shown to differ in one or more pelvic dimensions in two Merino flocks and the SA Mutton Merino flocks. Ewes classified as "good" had larger dimensions. Rearing ability was also demonstrated to be repeatable in the Tygerhoek Merino flock, where sufficient data were available. Part 2: The dynamics of behaviour in lambing flocks This part of the dissertation included four papers. The first paper detailed aspects of parturition and separation from one or more lamb in Dormer and SA Mutton Merino ewes. Length of parturition was found to be dependent on breed, with Dormer ewes having shorter parturitions than SA Mutton Merinos. They also were less likely to be assisted and Dormer lambs were less likely to die during or shortly after parturition. Across years, length of parturition was also repeatable in ewes. Early movement (within two hours) of ewes from the lambing site and high concentrations of lambed ewes in the same paddock were found to result in higher levels of separation from one or more lambs, resulting in lamb mortality. The second paper detailed neonatal progress in lambs and related it to birth weight and lamb survival. Survival was affected by neonatal progress as well as by live weight gain from birth to 3 days in lambs. No breed differences were found for the interval from birth to standing or from standing to apparently suckling in Dormer and SA Mutton Merino lambs. It was, however, found that the traits were moderately heritable, using paternal halfsib methods on a relatively small data set. The mechanisms of difficult births were subsequently investigated in Dormer and SA Mutton Merino ewes, since the two breeds were markedly different in this respect. The etiology of difficult births in the breeds was found to differ vastly. Dystocia and feto-pelvic disproportions commonly led to prolonged parturitions and assistance in SA Mutton Merinos. Uterine inertia were more likely to be observed in Dormers. The ability of factors associated with birth difficulties (pelvic dimensions, litter size and weight, ewe live weight and conformation) to predict length of parturition within breeds was, however, limited. In the final paper of this part of the dissertation, it was possible to partition the genetic variances for behaviour traits observed during lambing in Dormer and SA Mutton Merino ewes. Maternal additive variance ratios for length of parturition were moderate and significant in both breeds. Neonatal progress (time intervals from birth to standing and from standing to apparently suckling) was lowly heritable in both breeds. The interval from standing to suckling was also influenced by the maternal permanent environmental variance in SA Mutton Merinos. Direct and maternal breeding values for behavioural traits differed significantly between lambs that survived and those that died in most cases. This result suggested a genetic association of lamb mortality with lambing and neonatal behaviour. Part 3: Responses to selection for ewe multiple rearing ability This part of the dissertation was based on observed responses in two Merino lines that were divergently selected for ewe multiple rearing ability since 1986. The line selected in an upward direction for multiple rearing ability is referred to as the High (H) line. The line selected downwards is referred to as the Low (L) line. This part of the study consists of nine papers. In the first paper, separation of ewes from one or more lambs was studied in the Merino lines referred to above. Ewes in the H line were more likely to be separated from one or more lambs because of interference, and they tended to be more interested in other parturient ewes. Ewes in the L line tended to be more likely to desert their lamb(s). Ewes that left their birth sites within two hours of birth were more likely to be separated from one or more offspring. The ability of ewes and lambs from the Hand L lines to recognize each other were investigated next. Ewes in the H line were able to find their lambs sooner after the lamb being tethered 20 m away, compared to their L line contemporaries. They were also able to establish contact with all litter members in multiples sooner after finding a tethered lamb. Lambs in the H line tended to be more likely to bleat and tug on the tethering rope when tethered. They also followed their dams more closely when chased away with their dams at three days of age. The paper allowed the estimation of preliminary genetic parameters for mutual recognition of one another by ewes and lambs, although it was not the primary objective. Significant maternal genetic variances were found for most of the traits. The following paper demonstrated a better lamb survival in H line lambs, despite a higher multiple birth rate. Line differences were reported, mostly involving behavioural adaptations conducive to lamb survival in the H line. Ewes in the H line generally experienced shorter births, and H line lambs were less likely to succumb during or soon after parturition. Ewes in the H line showed better cooperation with the first suckling attempts of their lambs, and H line lambs apparently suckled sooner after standing than L line contemporaries. This line difference remained after correction for the better cooperation of H line ewes. A marked advantage in terms of weight of lamb weaned per breeding ewe in the H line was demonstrated for the five years of this study, when compared to the L line. It was found that a substantial proportion of births will not be supervised if the supervision of Merino ewes in South Africa was confined to daylight hours in intensive systems. A larger than expected frequency of very short intervals between subsequent onsets of parturition was observed. If the contention that births would trigger each other could be substantiated, it could be of benefit to quantify the mechanism involved. Knowledge of this would facilitate the synchronisation of natural births in pasture lambing flocks, where it is impractical to alter the distribution of births by nutrition. Relatively few line differences were found in this paper. The distribution of birth sites within paddocks could not be related to selection line. Ewes in the H line ewes were more likely to groom their lambs shortly after birth, while L line ewes commenced grazing sooner after birth. These behavioural adaptations in the H line would generally facilitate lamb survival. Behavioural data from the H and L lines were recorded over a lO-year period and subjected to a genetic analysis. The line differences obtained earlier could be confirmed in this study. Genetic parameters estimated from the data indicated that behaviour traits in lambs and ewes were lowly to moderately heritable. In lambs, a significant heritability estimate was derived for the interval from birth to standing. Maternal heritability estimates were significant for length of parturition and for the interval from birth to standing. Maternal permanent environmental variances were significant for maternal cooperation and the interval from standing to apparently suckling. When behavioural traits were assessed in ewes, length of parturition, maternal behaviour score and the period that ewes remained on or near the birth site were found to be moderately heritable. Genetic divergence was found for the H line compared to the L line. The derived genetic trends indicated changes conducive to lamb survival in the H line. Direct additive, maternal additive and maternal permanent environmental variance ratios were subsequently derived for lamb birth weight, birth coat score and lamb weaning weight. Birth coat score was highly heritable and not influenced maternally. Genetic and phenotypic trends indicated divergence between the H and L lines for weaning weight over the period of assessment. Genetic divergence was particularly strong in the case of direct breeding values for weaning weight. Birth weight and birth coat score were independent of selection line. Genetic parameters were estimated for annual reproduction traits, ewe greasy fleece weight and ewe body weight at joining, using a repeatability model. Heritability estimates for reproduction traits were low, consistently below 10 % of the overall phenotypic variance. Ewe permanent environmental variances were higher, between 7 and 12 %. Genetic correlations of reproduction traits with ewe joining weight were favourable and very high in the case of weight of lamb weaned per ewe. Corresponding genetic correlations with ewe greasy fleece weight were low and variable. Ewe permanent environmental correlations of reproduction traits with ewe joining weight and greasy fleece weight were unfavourable and high in some instances. Phenotypic trends indicated divergence in the expected direction between the Hand L lines in the reproduction traits considered. These tendencies were confirmed by genetic trends based on averaged predicted breeding values within birth years. The results confirmed that genetic progress at a rate of 1-2 % per year was attainable in reproduction traits, despite low additive variance ratios. Genetic parameters were estimated for hogget wool and live weight traits, as well as for testis measurements. Adequate genetic variation for exploitation in a well-structured breeding program was estimated in all traits. Genetic trends indicated change towards heavier and plainer sheep in the H line, while L line contemporaries became smaller and more developed. Genetic trends for testis circumference suggested divergence in the expected direction between lines. This result was, however, associated with the genetic change in live weight, since no line difference was found in testis circumference corrected for live weight. Genetic trends for clean fleece weight and fibre diameter indicated no divergence between lines. Age influences on reproduction, wool traits and live weight were assessed in H and L line ewes subjected to at least 10 years of divergent selection. Genetic parameters derived from this data set were generally consistent with previous estimates, despite the usage of a much smaller data set. The Hand L lines differed markedly for reproduction, but no differences in the shape of age trends were found between lines. Compared to average L line performance, the superiority of H line ewes amounted to 56 % in the case of weight of lamb weaned per ewe. Ewes in the H line were heavier than L line contemporaries at two-tooth age, but the line difference disappeared at later ages. Wool production of L line ewes were heavier than those of their H line contemporaries, particularly in the middle age groups (3 to 6 years), where the highest fleece weights are expected. Fibre diameter increased with ewe age, with no apparent line difference. It was contended that the stress of increased reproduction resulted in the different age trends for the respective selection lines for live weight and greasy fleece weight. In the final instance, this dissertation provides a framework for the genetic improvement of lamb output in pasture-fed sheep by selection for reproductivity. Unwanted correlated genetic changes were minimal, although the stress associated with an increased reproduction impaired live weight and wool traits in mature animals to an extent. The interface between behaviour, lamb survival and ewe rearing ability were clarified for a number of South African sheep breeds, leading to a better understanding of the dynamics involved. These results contributed to the formulation of adapted husbandry and breeding practices, in the context of overall flock productivity. Results pertaining to behaviour appeared to be fairly robust across studies involving different resource populations, and should be appropriate for extension to the broader industry.