A note on the relationship between the behavioural response of lactating sows to humans and the survival of their piglets

Abstract

The objective of this study was to examine the relationships, based on unit averages, between the behavioural responses of lactating sows to humans and the performance of sows in 25 farrowing units at a large commercial farm. The behavioural responses of 25 sows in each farrowing unit were observed at 2–4 and 16–18 days of lactation and a correlation analysis, using unit averages, was used to examine the behaviour–productivity relationships. Moderate and significant between-unit correlations were found between the behavioural response of lactating sows at days 16–18 to an approaching experimenter and the percentage of stillborn piglets. Units in which test sows were quicker to withdraw in the presence of the experimenter had a higher stillbirth rate than those units in which a reduced withdrawal response was displayed. The withdrawal response of sows in the close presence of the experimenter accounted for about 18% of the variance in percentage of stillborn piglets. While this study was a preliminary one examining the potential for the human–animal relationship to affect the performance of sows in the farrowing shed, the results indicate that high levels of fear of humans by sows may affect the survival of their piglets. The possible mechanisms whereby fear may affect piglet survival are considered. Clearly further research is warranted because of the important practical implications of this negative fear–productivity relationship on animal performance and welfare.

Introduction

The factors contributing to the death of neonatal piglets have been well understood for more than a decade (English and Morrison, 1984; English et al., 1984) and yet preweaning mortality is still a major cause of wastage in modern pig production (Cutler et al., 1989Cutler et al., 1992). Spicer et al. (1986)examined the causes of pre-weaning mortality in 238 litters at a large intensive piggery and found a mortality rate of 18.7%. The causes of these piglet deaths were stillborn (8.3% of total born), illness (4.3%), overlay (2.1%), anaemia (1.2%), savaging (1.1%), small and weak piglets (0.9%) and splay-legs (0.5%). Many farms are capable of weaning more than 9.5 pigs per litter with pre-weaning mortality of less than 10%, however national averages for many countries indicate numbers of pigs weaned per litter of less than 9.0 with preweaning mortality greater than 12%. Cutler et al. (1992), in reviewing piglet mortality rates in modern pig production, suggest that the solution to the pre-weaning mortality problem lies with the attitude, diligence and skill of the farrowing unit staff.

A series of studies examining the interactions between stockpeople and breeding pigs around the time of mating strongly suggests a causal link between the attitudes and behaviour of the stockperson on the one hand and the behavioural response to humans, fertility and fecundity of the pig on the other (Hemsworth and Barnett, 1987; Hemsworth et al., 1989, Hemsworth et al., 1993, Hemsworth et al., 1994). These attitudes and consequent behaviours by the stockperson predominantly affect the animal's fear of humans which, in turn, affects the animal's performance. The mechanism whereby fear affects performance appears to be through a chronic stress response (Hemsworth et al., 1981a, Hemsworth et al., 1986, Hemsworth et al., 1987). While these human factors may have direct effects on the behaviour and productivity of breeding pigs, they may also have their effects through their interrelationships with other important human factors such as the stockperson's work ethic (Hemsworth et al., 1993).

In contrast to our understanding of the human factors operating in the mating unit, there is little objective information on the importance of human factors operating in the farrowing unit which may affect sow and piglet performance. In a conference abstract, Seabrook (1991)reported significant between-unit relationships between the stockperson's personality, level of fear of humans by sows and piglet mortality: Stockpeople with a low “implied aggression-score” were associated with reduced withdrawal responses by sows in the presence of an experimenter and low mortality rates. While no information is presented in this report on the validity and content of the questionnaire, one interpretation of this report is that a low implied aggression score may be associated with positive attitudinal and behavioural profiles of stockpeople towards pigs, which in turn may have consequences for the behaviour and productivity of farrowing and lactating sows. Furthermore, significant relationships have been found in the pig industry between personality types of stockpeople, based on an evaluation using the Sixteen Personality Factor Questionnaire, and productivity in farrowing units of both small independent and larger integrated units (Ravel et al., 1996). At large integrated units, high preweaning performance (number of piglets weaned) was associated with stockpeople with traits of warmth, emotional stability and self-discipline, while low preweaning performance was associated with stockpeople being highly self-assured and sensitive. However, the strongest predictors of the stockperson's behaviour have been found to be the stockperson's attitudes (Hemsworth et al., 1993).

If fear of humans, as a consequence of human–animal interactions, limits the performance of sows in the farrowing unit, significant relationships should exist between the behavioural response of sows to humans and the performance of sows in the farrowing unit. Therefore, the present study examined the relationships, based on unit averages, between the behavioural response of lactating sows to an experimenter in a standard test and sow performance in the farrowing unit, which was measured in terms of litter size and piglet mortality. There is considerable controversy surrounding the concept of fear and the definition of fear (see Murphy, 1978; Jones, 1987). In studying the behavioural responses of animals to stimuli which may be painful or harmful to animals, it is useful to consider fear as an intervening variable, linked, on the one-hand, to a range of stimuli which may pose some risk or danger to the animal and, on the other hand, a series of responses, both behavioural and physiological, by the animal which enable the animal to respond appropriately to this source of danger. Therefore, since fear responses function to protect the animal from harmful stimuli (Toates, 1980; Jones, 1987), it has been proposed that the approach behaviour or conversely, the avoidance behaviour of animals toward a stationary or approaching experimenter in a standard test provides a useful measure of the animal's fear of humans (Hemsworth and Barnett, 1987; Jones, 1987; Hemsworth et al., 1993).