It has been clearly demonstrated that rough handling of animals by people leads them to become fearful of people and avoid them. But can the animals’ tendencies to approach or avoid people reflect only their degree of fear of them or can it reflect the influence of other motivations? In many cases, the extent that animals avoid people will likely reflect a mix of motivations including fear and curiosity (Murphey et al., 1980, 1981). Similarly, an animal that is used to being fed by people may approach people closer than animals that have not been fed by people. However, this may not so much indicate the degree of fear as the extent that the animals expect to receive food. Jago et al. (1999) found that calves that had been hand fed by people approached people closer than calves that had been gently patted and stroked. Hand feeding of calves also increases the chance that they will suck or butt at the people (Jago et al., 1999; Krohn et al., 2001, 2003) again suggesting that feeding motivation may be playing a major role in the extent that the animals approach people. That an animal’s response to people may reflect feeding motivation, rather than fear, is particularly problematic for tests that examine animals’ response to people in the presence of food. For example, Lensink et al. (2000) tested calves for the tendency to withdraw from feed when a person approached the feed trough. This was used to assess the degree of fear the calf felt for the person, but it may also reflect their level of feeding motivation. This is best demonstrated in the study of Lankin (1997) on sheep. In this study, sheep’s responses to people were measured by having the person stand near the feed trough and recording how much the sheep were willing to approach the person in order to eat. The assumption was that fearful sheep would stay away from the person and so eat less. However, Lankin (1997) found that this measure was also (not surprisingly) affected by the sheep’s level of hunger. Food deprived sheep approached the person closer than satiated sheep, and sheep approached the person closer when the test was done a long time since the last meal compared to when the test was done soon after the meal. As well as feeding motivation, it seems probable that an animals responses to people may also be affected by their degree of curiosity. Thus it seems likely that an animal’s response to people in such tests will reflect a mix of factors depending on the particular context of the test, and we need to be very careful in assuming that the response simply measures the degree of fear that the animals show. One reason why distance measures are influenced by many factors is that there is often no cost to the animal to avoid people. We have recently examined the possibility of using measures of vigilance to evaluate the quality of the man–animal relationship (Welp et al., 2004). Vigilance behaviour is costly to the animals because they often must stop eating in order to be vigilant. By increasing the cost to the animal of expressing its fear of people in this way, we may more easily interpret changes in vigilance as reflecting changes in animals’ fearfulness.
Effect of location
The context or location in which the test is applied may play a role. Some studies test animals’ reactions in their home area, while others move the animals to another, often unfamiliar area. Some studies do find that the location of the test has no or only slight effects on the tests. For example, the effects of increased early handling on calves’ approaches to people were apparent in both the calves’ home pens and in an unfamiliar pen (Krohn et al., 2001, 2003). Other studies, however, show a strong effect of location of the test. Calves will approach a person who is gentle to them, and avoid another person who has handled them roughly when tested in the pen where the handling occurred, but may react quite differently if tested in another place (de Passille´ et al., 1996). When cows were tested in their own stall and in an unfamiliar stall, the correlation between the distance kept from a person in the two stalls was low and non-significant, even though the stalls were very similar (Munksgaard et al., 1997; Rushen et al., 1998). Similarly, calves that were hand fed interacted more with an unfamiliar person than calves that were not hand fed when they were tested in their home pen, but no differences were found when tested in an unfamiliar pen (Jago et al., 1999). When animals’ fear of people is tested in a novel area, their responses will reflect not only the response to the people but also their response to the novel area. Generally, farm animals isolated in novel areas show responses suggesting increased arousal and a great deal of exploratory behaviour. In some cases, the animals are placed in the arena before the person enters (e.g. Boivin et al., 1998), which may help reduce the effects of novelty of the arena. However, we must be aware that an animal’s responses to humans when tested in novel/unfamiliar areas may be unrelated to those in its home environment, because the reactions to a person in a novel area will be a measure of the combination of exploration of the area and fear of the person.
Identity of the ‘‘test’’ person
In a welfare assessment, the interest is in how animals are handled by their regular caretaker. However, in many cases it may be necessary to measure animals’ responses to another person, with the hope that the animals’ responses to the stranger will reflect how they will respond to the caretaker. Many recent studies show that aversive or gentle handling affects cattle’s responses to familiar and unfamiliar people in a similar way (Krohn et al., 2001; Breuer et al., 2003). Lensink et al. (2001) found moderate positive correlations between veal calves’ readiness to approach the usual stockperson and an unfamiliar person. Rousing and Waiblinger (2004) found that flight distances of dairy cattle to familiar and unfamiliar people were correlated. However, a number of studies have shown that cattle and pigs do respond differently in these sorts of tests according to whether the person is familiar or not (reviewed in Rushen et al., 1999b). Certainly, cattle will react very differently to people who treat them differently (Munksgaard et al., 1997, 1999; Rushen et al., 1999a) and evidence is accumulating that some cattle can use very fine features, such as facial features, to recognise individual people (Rybarczyk et al., 2001). Even a change of clothing can have quite large effects on how cattle react to people (Munksgaard et al., 1999). Thus, it is hazardous to assume that the identity of the person will have no effect. At this stage we do not fully understand why farm animals sometimes treat all people similarly, while at other times they respond quite differently to different people. Perhaps the number of different people that the animals normally encounter is a factor. But until we understand the factors that do cause animals to react differently to different people, we cannot simply assume that their responses to an unfamiliar person will reflect their responses to their usual caretakers.
Early rearing conditions
The degree of contact with conspecifics during early rearing can influence how animals respond to people, independently of how they are handled. For example, Krohn et al. (2003) showed that calves tend to approach people less if they have been reared with other calves, and that rearing them with their dams may limit the effects of extra gentle handling on increasing their approaches to people. Again, these results show that animals’ reactions may not reflect the way that the animals have been handled by their caretakers. Thus, many studies have shown that animals’ responses can be (but are not always) affected by the contextual factors that should not strongly influence the animals’ fear of people. This throws some doubt on the validity of the measures used. The fact that some contextual factors (such as the identity of the test person) do not always have an effect does not help us, unless we can predict with some certainty whether or not they will be operating in any given test.