www.elsevier.comrlocaterapplanim
Behavioural response to humans and the
productivity of commercial dairy cows
K. Breuer
a,d, P.H. Hemsworth
a,b,), J.L. Barnett
a,
L.R. Matthews
c, G.J. Coleman
da
Animal Welfare Centre, Victorian Institute of Animal Science, Agriculture Victoria, Werribee, Vic. 3030, Australia
b
Institute of Land and Food Resources, Melbourne UniÕersity, ParkÕille, Vic. 3052, Australia
c
Animal BehaÕiour and Welfare Research Centre, AgResearch, Rurakura Agricultural Research Centre,
PriÕate Bag 3123, Hamilton, New Zealand
d
Department of Psychology, Monash UniÕersity, Caulfield, Vic. 3145, Australia
Accepted 24 September 1999
Abstract
This study examined the relationships between the attitude and the behaviour of the stockper-son towards cows and the behavioural response to humans and the milk production of cows at 31 commercial dairy farms over one lactation. The attitude of the stockperson was measured on the basis of the stockperson’s opinion of the cow’s behaviour and how the stockperson evaluated his own behaviour. The behaviour of the stockperson was measured by recording the nature and frequency of the tactile interactions and some visual and auditory interactions directed towards the cows. The behavioural response of cows to humans was assessed by observing their approach behaviour to an unfamiliar experimenter in a standard test and production records were collected for the entire lactation at each farm. Correlation and regression analyses using farm averages were used to examine relationships between human and cow variables. Several cow behaviour
vari-Ž . Ž .
ables, indicative of fear of humans, were moderately P-0.05 to highly P-0.01 correlated with milk yield and composition and regression analysis indicated that fear of humans accounted for 19% of the variation in milk yield between farms. The results suggest that at farms where milk yield was low, cows showed less approach to the experimenter in the standard fear test than at farms where milk yield was higher. A composite attitude score, based on the responses of stockpeople to questions about patting and talking to cows, ease of movement of cows and cows
)Corresponding author. Animal Welfare Centre, Victorian Institute of Animal Science, Agriculture Victoria, Werribee, Vic. 3030, Australia
0168-1591r00r$ - see front matterq2000 Elsevier Science B.V. All rights reserved.
Ž .
Ž . Ž .
recognising unfamiliar stockpeople, was moderately P-0.05 to highly P-0.01 correlated with the behaviour of the stockperson. While a number of stockperson behaviour variables were
Ž .
correlated P-0.05 with milk yield, the former variables were generally poorly correlated with cow behaviour. Therefore these preliminary findings provide evidence that, as seen in the pig industry, sequential relationships may exist between the attitude and behaviour of the stockperson and the behaviour and productivity of commercial dairy cows. Research is required to further examine these relationships because of the possible implications on cow productivity and welfare.
q2000 Elsevier Science B.V. All rights reserved.
Keywords: Attitude; Behaviour; Productivity; Fear; Stress; Cows; Stock person; Human–animal interaction
1. Introduction
In modern livestock industries, there is frequent and close contact between the stockperson and hisrher animals. The stockperson is largely responsible for the animal’s care and maintenance. There is a considerable body of literature, particularly in the pig industry, which suggests that human-animal interactions can have substantial effects on
Ž
the behaviour, physiology and productivity of commercial farm animals Hemsworth et
.
al., 1993 . For instance, there is evidence of significant interrelationships between the attitude and behaviour of the stockperson and the behaviour, welfare and productivity of
Ž .
farm animals Hemsworth et al., 1993 . Such evidence indicates the potential to improve animal productivity by improving these human–animal interactions.
In particular, fear of humans has been found to be negatively correlated with productivity: at farms where pigs were highly fearful of humans, reproductive
perfor-Ž .
mance was low Hemsworth et al., 1981b, 1989a . Handling studies on both young and adult pigs have also shown that high fear levels may depress the growth and
reproduc-Ž
tive performance of pigs Hemsworth and Barnett, 1991; Hemsworth et al., 1981a,
.
1986 . A chronic stress response appears to be implicated in this effect, since a number of studies have found that fearful pigs have a sustained elevation of plasma
cortico-Ž .
steroid concentrations Hemsworth and Barnett, 1991; Hemsworth et al., 1981a, 1986 . Fear of humans may also place an animal’s welfare at risk due to the actions of sustained elevations of corticosteroids on protein metabolism and the immune system
ŽHemsworth and Barnett, 1987 . Furthermore, research at a number of commercial pig.
farms has revealed that the nature of the behaviour that the stockperson directed towards hisrher pigs was associated with the pig’s fear of humans: the percentage of negative interactions used by the stockperson was predictive of the level of fear of humans by
Ž .
pigs at that farm Hemsworth et al., 1989a . By assessing the attitudes of commercial pig
Ž . Ž .
stockpeople, Coleman et al. 1998 and Hemsworth et al. 1989a have shown that the attitude of stockpeople towards pigs is predictive of their behaviour towards pigs. As a
Ž .
Considering the amount of contact that a stockperson has with hisrher animals in the dairy industry, it is surprising that only relatively limited research has been conducted on human–animal interactions in this livestock industry. Limited research suggests that fear
Ž .
of humans may affect productivity in the dairy industry. Seabrook 1972 reported some significant associations between the personality of the stockperson and milk production. In a study of 28 one-person herds, he found that stockpeople at the high producing farms tended to be introverted and confident. He also suggested that cows in the highest yielding herds were the most willing to return from pasture and enter the milking shed and were less restless in the presence of the stockperson. This finding suggests that productivity in the dairy industry may in part be related to the quality of human–animal interactions. Several experimental studies also suggest that handling may affect fear of humans, which in turn may limit the productivity of commercial dairy cows. For example, handling has been shown to affect the behavioural response of cows to humans
ŽBoissy and Bouissou, 1988; Boivin et al., 1992; Breuer et al., 1997 and Rushen et al.. Ž1999 reported that the presence at milking of an aversive handler, who had previously.
hit or occasionally used a battery-operated prodder on the cows over a 5-day period,
Ž .
increased the milk not collected at milking residual milk .
The existence of a fear–productivity relationship in commercial dairy cows may provide dairy farmers with the opportunity to reduce fear to improve cow productivity through improvements in their attitude and behaviour towards dairy cows. The objective of this preliminary study was to examine the relationships between stockperson attitude and behaviour, and cow behaviour and productivity at 31 commercial dairy farms over one lactation.
2. Material and methods
2.1. Subjects
Behaviour observations and productivity measurements were conducted at 31 com-mercial dairy farms in SE Victoria, Australia over a 2-year period. The study was
Ž .
conducted in two time replicates, with 15 farms studied in the first year replicate 1 and
Ž .
the other 16 farms in the second year replicate 2 . These farms had a winter calving with a peak occurring in mid to late winter and the first month of lactation at the farm was defined as the month in which most cows calved. The animals were grazed outdoors on pasture all year round and milking occurred twice a day, morning and afternoon. Farms were selected on the basis of availability of milk production records, herd size of 100–200 cows, predominantly Holstein–Friesian cows, herringbone milking-shed con-figuration and supplementary feeding during milking.
2.2. ObserÕationsrmeasurements
observers and a high degree of concordance was achieved between observers’ records
Žsee Coleman et al., 1998 for assessing interobserver reliability ..
2.3. Confidentiality of project objectiÕes
In order to reduce the influence of observers on the stockperson’s behaviour, the detailed objective of the study was not revealed to the participants until the study was completed. Instead, the reason given to the farmers for conducting observations was that the relationships between a range of farm factors, cow behaviour and cow production were being examined. Ethics approval was obtained for these procedures and stockpeo-ple were debriefed at the end of the study.
The following observations and measurements were taken at each farm.
2.3.1. BehaÕioural response of cows to humans
To assess the cows’ fear of humans, the approach behaviour of cows to a stationary
Ž
experimenter was measured in a standard approach test at each farm Hemsworth et al.,
.
1996 at months 3 and 5 of lactation. The test was conducted in a 6=6 m arena during one morning and one afternoon milking session at each of these two stages of lactation and a total of 35–50 cows were individually tested at each farm over these two test sessions.
Due to some difficulty in erecting the arena on sloping ground in the first replicate, the shape and construction of the arena were modified slightly in the second replicate. Although the floor areas of the two arenas were similar, the shape of the arena was
Ž .
changed from a square arena 6=6 m to an octagonal one with a maximum width of
Ž .
7.4 m. The height remained the same 2 m but the second arena was of a lighter construction to aid erection on difficult terrains. Both arenas were constructed of aluminium posts and rails and provided almost unrestricted visual access to events outside the arena. The approach behaviour of cows at the 15 farms in year 1 was assessed in the 6=6 m square arena, while that of the cows at the 16 farms in year 2 was assessed in the octagonal arena. At each farm the arena was set up before the afternoon milking, on pasture near the milking shed and adjacent to the exit lane in which the cows returned to pasture. The selected cows for testing were chosen at milking so that half came from each of the two milking sides in the herring-bone shed and these selected cows were well represented throughout the milking session and within each of the front, middle and rear positions in the shed. Individual cows were
Ž
only tested once over these four sessions two milkings at each of 3 and 5 mo. of
.
lactation and selected cows were held in groups of 5 in a yard within 50 m of the arena to ensure easy access for testing, while reducing visual contact with the cow that was under test.
used as the human stimulus in tests at all farms and this experimenter wore clean long-sleeved cloth overalls and clean boots. From these data, the following variables
Žname assigned to the variable is given in parentheses were calculated:.
-the time taken for the cow to approach within 1, 2 and 3 m of the experimenter
ŽTIMETO1, TIMETO2 and TIMETO3 ;.
Ž .
-the cumulative time spent within these areas TIMEIN1, TIMEIN2 and TIMEIN3 ;
Ž
-the time taken to physically interact with the experimenter defined as within 5 cm of
. Ž .
the experimenter TIMETOI ;
Ž .
-the number of times the cow physically interacted with the experimenter INTER . At the end of testing, each animal was returned to the herd on pasture. The average farm data for each of these variables were used in the analyses.
2.3.2. Cow behaÕiour at milking
The behaviour of all lactating cows in the milking shed was observed at each farm during each observation session. These observations at each farm were conducted by three of four trained observers, one observer per session and the observations were made at three morning milkings during the first, second and third or fourth months of lactation
Ž .
for the herd. The number of flinch and step responses FS and flinch, step and kick
Ž .
responses FSK by cows were recorded when it was estimated that the stockperson was approximately 0.5 m from the cow. A FS was defined as a step or lifting of the foot in which the hoof was not raised to the height of the udder and a FSK was defined as a step or kick in which the hoof was raised at least to the height of the udder. These observations were designed to give a measure of cow restlessness when in close contact with humans in the shed and were conducted during the following discrete activities, forcing the cows into position for milking in the shed and attaching and removing cups. It was therefore possible to obtain the sum of these behaviours at each farm for each of
Ž
the activities i.e., FSforce and FSKforce, FSon and FSKon, and FStotal and FSKtotal
Ž‘‘total’’ refers to the total for all activities . Furthermore it was possible to combine the.. Ž
averages of behaviours for each activity i.e., FSqFSKforce, FSqFSKon, and FSq .
FSKtotal .
2.3.3. Human behaÕiour obserÕations
At each farm the behaviour of the stockpeople was observed during three afternoon milkings made during the first, second and third or fourth months of lactation by three of four trained observers, one observer per session. Human behaviour was recorded during the following discrete activities: moving cows into the shed from the holding yard; forcing the cows into position for milking in the shed; attaching and removing cups; and moving cows out of the shed after milking.
Human tactile interactions with the cows were classified as either positive or negative
Ž .
animals such as slaps, pushes or hits with the hand or an object such as a plastic pipe. There were two types of negative tactile interactions recorded; those that were
moder-Ž .
ately aversive and these included moderate slaps, pushes and hits N1 , and those that
Ž .
were more aversive and these included forceful slaps, pushes, hits and tail-twists N2 . The moderate negative interactions differed from the positive ones in that for the former interaction some perceptible noise was associated with contact and the intention was generally to move the animals. From these observations an estimate was made of the mean number of bouts of these behaviours that cows at the farm received from the
Ž .
stockpeople during each milking session i.e., P1, N1, N2 . A combined variable of the
Ž . Ž . Ž .
number of positive P1 and negative N1 and N2 interactions I was used in the
Ž .
analysis. In addition, the number of arm waves variable assigned W was recorded and
Ž .
vocalisations e.g., talking, shouting, whistling and claping were recorded as quiet or
Ž . Ž
soft vocalisations variable assigned V1 or loud or harsh vocalisations variable
.
assigned V2 . A bout criterion interval of 5 s was chosen to separate one bout from another.
Ž .
In the analysis, the two types of negative tactile interactions N1 and N2 were
Ž .
combined to measure overall frequency of negative interactions NEG and to measure the percentage of these negative tactile interactions, these three variables were expressed
Ž .
as a ratio to the total number of tactile interactions N1%, N2% and NEG% . In addition to the above collation of the mean number of bouts of these tactile, visual and auditory interactions that cows at each farm received, each of these interactions were also collated for each of the four activities: moving cows into the shed from the holding yard; forcing cows into position for milking in the shed; attaching and removing cups; and moving cows out of the shed after milking. Thus, for example, the mean number of N1 interactions received per cow at each farm when being forced into position in the milking shed was also collated.
An additional two visits, one in summer and one in winter, were made by one of two trained observers to each farm to assess the speed at which the cows were being moved by the stockpeople from pasture to the milking shed. Speed was estimated over the last
Ž .
50 m along a race or laneway into the holding yard SPEED50 adjacent to the milking shed at an afternoon milking. By timing the entry and exit of the stockperson in this 50 m race and marking out the last 50 m into 10 equal length segments, the observer was able to record the average speed of the herd and the average distance between the last
Ž .
cow in the herd and the stockperson andror dog depending on which was closer at
Ž .
each 10 m interval DISTANCE . It was common for both motor bikes and dogs to be used by stockpeople in moving cows from pasture to the holding yards for milking.
2.4. Attitude questionnaire
a five-point scale, defined by the labels: disagree strongly; disagree; neither disagree nor agree; agree; agree strongly.
The second section of the questionnaire contained 57 items that were designed to assess the stockperson’s attitude to working with cows. In each attitude category, the questions were asked about cows at different ages, cows or heifers. For example, in this section questions such as ‘‘To what extent do your cows flinch when you put the cups on?’’ and ‘‘How easy are your dairy cows to move into the milking shed at the ages listed below?’’ were asked. The stockperson was asked to answer these questions using
Ž
a seven point scale for example in the first question above, the scale ranged from ‘‘ a lot’’ to ‘‘ very little’’, while for the second question, it ranged from ‘‘ very easy’’ to
.
‘‘ very difficult’’ .
To reduce the number of variables in the analysis, a composite variable was created.
Ž .
This composite variable included a number of questions that were moderately P-0.05 correlated with the overall frequency of negative tactile interactions and these questions sought information on beliefs about patting and talking to cows, ease of movement of cows and cows recognising unfamiliar stockpeople. In collating the results of the responses to these questions, scores for each question were adjusted so that a high score for each response reflected a positive attitude. The responses to these questions were
Ž .
then grouped and summed to give this composite score ATTITUDE SCORE , with a high score indicating a positive attitude.
2.5. Incidence of lameness
The incidence of lameness was also monitored during two morning milking visits to each farm, one in summer and one in winter. A trained observer assessed the lameness of all cows as they exited the milking shed. Each cow was given a score of 0 to 3, where
Ž .
0 was defined as not lame, 1 was defined as mildly lame slight limp — no head bob , 2
Ž . Ž
was lame obviously lame — head bob and 3 was very lame head bob and held leg up
.
for a period of seconds .
2.6. Production records
Ž . Ž . Ž .
Records on total milk yield l, YIELD , protein kg, PROTEIN and fat kg, FAT over the lactation were collected for each farm from either the milk factories or a herd improvement monitoring scheme. The records from the herd improvement monitoring scheme were based on monthly testing of each herd and were provided on a cow average per farm per annum basis, while the factory records were calculated from each milk delivery from the farm to the factory and were provided as yearly totals for the farm. These latter records were converted to a per cow per year value by dividing the totals by the maximum number of cows observed during the study.
2.7. Statistical analyses
Ž
A correlation analysis and a step-wise regression analysis Genstat 5, Lawes
Agricul-.
relation-ships, based on farm averages, between the human and animal variables at the farms. Complete data were only available on 29 farms since the owners of two farms withdrew their permission to conduct the observations on the approach behaviour of cows to the experimenter in the standard test. These two farmers were concerned that their animals may injure themselves in attempting to escape from the novel arena.
The size and shape of the arena differed slightly in the two years and an analysis of
Ž .
variance indicated a significant P-0.05 difference between the two years in the
Ž
approach behaviour of cows to the experimenter e.g., TIMEIN3, means of 48.1 and
Ž . .
32.4 s, LSD Ps0.05 s12.8 but not in any other behavioural or production variables. To correct for any effect that arena design may have had, the farm averages for each of the four variables recorded in this test were standardised by subtracting the values for
Ž .
the farm from the mean for the replicate year and dividing by the standard deviation for the replicate. Correlation and regression analyses were performed on these standard-ised values.
Table 1
Mean and ranges of some of the main cow and stockperson variables studied
Variables Mean Range
Cow productiÕity
Ž .
Milk yield lrcowryear, ‘‘YIELD’’ 5632 4323–6662
Ž .
Milk protein kgrcowryear, ‘‘PROTEIN’’ 185 142–215
Ž .
Milk fat kgrcowryear, ‘‘FAT’’ 235 191–276
Cow BehaÕiour
aŽ .
Time spent within 3m of experimenter s, ‘‘TIMEIN3’’ 40.0 13.0–85.4 Number of FS and FSK responses per cow per milking during:
Ž .
Cups attached ‘‘FSqFSKon’’ 0.97 0.21–2.30
Ž .
Cups attached ‘‘FSKon’’ 0.10 0.01–0.35
Ž .
Overall ‘‘FSqFSKtotal’’ 2.21 0.56–5.38
Ž .
Overall ‘‘FSKtotal’’ 0.21 0.02–0.54
Stockperson behaÕiour
Ž .
Number of positive tactile interactions P1rcowrmilking 0.11 0.02–0.40
Ž .
Number of negative tactile interactions N1rcowrmilking 0.32 0.09–0.84
Ž .
Number of highly negative tactile interactions N2rcowrmilking 0.05 0–0.13
Ž .
Percentage of negative tactile interactions %, NEG 74.0 30.3–98.2
Ž .
Percentage of highly negative interactions %, N2% 9.5 0–29.5
Ž .
Number of soft, quite vocalizations V1rcowrmilking 0.31 0.10–0.63
Ž .
Number of loud, harsh vocalizations V2rcowrmilking 0.05 0–0.12
Ž .
Number of waves Wrcowrmilking 0.03 0–0.12
Ž .
Speed in moving cows from pasture over last 50 m SPEED, mrs 0.60 0.07–2.08
Stockperson attitude
Subscale which included questionnaire items of petting and talking 29.6 14.0–45.5 to cows, ease of movement of cows and cows recognising unfamiliar
b
Ž .
stockpersons ‘‘ATTITUDE SCORE’’ a
Analyses conducted on the standardised values of TIMEIN3. b
3. Results
The mean and ranges of the main human and cow variables studied are presented in Table 1.
Some of the cow behaviour variables were moderately to highly correlated with cow
Ž .
productivity Table 2 . The average time that a cow spent within 3 m of the
experi-Ž . Ž .
menter in the standard test TIMEIN3 was positively correlated P-0.01 with milk
Ž . Ž . Ž .
yield YIELD and milk fat FAT . TIMEIN3 was also positively correlated P-0.05
Ž . Ž .
with milk protein PROTEIN . There were also negative correlations P-0.05
be-tween the cow variables, the number of flinch, step and kick responses when cups were
Ž .
attached FSKon , average sum of the number of flinch and step responses and flinch,
Ž .
step and kick responses per cow when cups were attached FSqFSKon and the
production variables, YIELD and PROTEIN. There were no significant correlations
ŽPs0.05 between TIMEIN3 and FS. qFSKon rŽ s y0.04 and FS. qFSKtotal rŽ s .
y0.22 .
A regression analysis, conducted to determine the contribution of cow behaviour to the prediction of cow productivity, with YIELD as the dependent variable, revealed that
Ž 2
TIMEIN3 accounted for considerable variation in productivity R s0.19, F1,28s7.39,
Table 2
Correlation coefficients between dependent variables, cow productivity, and cow behaviour and stockperson attitude and behaviour
Critical values for r,Õs27, 0.367, P-0.05; 0.470, P-0.01. Significant correlations atUP-0.05 andUUP-0.01.
a, high score represents a positive attitude.
Independent variables Cow productivity
.
P-0.01 . The prediction of YIELD significantly increased with only the inclusion of
Ž 2 .
FSqFSKon R s0.30, F2, 26s6.93, P-0.01 .
A number of the stockperson behaviour variables and the stockperson attitude
Ž .
variable were correlated with cow productivity Table 2 . There were significant
Ž .
negative correlations P-0.05 between the stockperson behaviour variable, the
per-Ž .
centage of highly negative tactile interactions used by the stockperson N2% and the production variables, YIELD, PROTEIN and FAT. A number of these behaviour variables during some of the discrete milking activities were also correlated with production. N2% when forcing cows into position in the milking shed and N2 when
Ž
moving cows out of the shed were significantly correlated with YIELD rs y0.40 and
.
y0.39, respectively, P-0.05 . There were also negative correlations between the
Ž .
stockperson behaviour variable, the number of loud or harsh vocalisations V2 used by
Ž . Ž . Ž .
the stockperson and YIELD P-0.05 , PROTEIN P-0.05 and FAT P-0.01 .
Ž .
The use of highly negative tactile interactions N2 , while not significantly correlated
Ž .
with YIELD rs y0.32, P)0.05 , was significantly correlated with PROTEIN and
Ž .
FAT Table 2 . Speed of movement of the stockperson when moving the cows from
Ž .
pasture to the milking shed over the last 50 m SPEED50 was negatively correlated
ŽP-0.05 with YIELD. The attitude composite score ATTITUDE SCORE , which. Ž .
included questionnaire items of petting and talking to cows, ease of movement of cows and the ability of cows to recognise unfamiliar stockpeople, was positively correlated
ŽP-0.01 with YIELD and PROTEIN Table 2 .. Ž .
Ž .
The incidence of lameness was not significantly associated Ps0.05 with any of the cow or human behaviour variables studied. Average herd size was not significantly
Ž .
associated Ps0.05 with cow behaviour or cow productivity variables.
Table 3
Correlation coefficients between dependent variables, cow behaviour, and stockperson attitude and behaviour
Independent variables Cow behaviour
ATTITUDE SCOREa 0.30 y0.37
Critical values for r,Õs27, 0.367, P-0.05; 0.470, P-0.01. Significant correlations atUP-0.05 andUUP-0.01.
Table 4
Correlation coefficients between the dependent variable, stockperson attitude, and stockperson behaviour Independent variable Stockperson behaÕiour
SPEED50 P1 N1 N2 N2% V1 V2 W
Stockperson attitude
UU U
ATTITUDE SCOREa y0.23 0.17 0.13 y0.36 y0.50 0.45 y0.25 y0.28 Critical values for r,Õs27, 0.367, P-0.05; 0.470, P-0.01.
Significant correlations atUP-0.05 andUUP-0.01.
a, high score represents a positive attitude.
Although stockperson behaviour was in general poorly correlated with cow behaviour
ŽTable 3 , there were some consistent moderate correlations. For example, the number of.
Ž .
waves displayed by the stockperson in moving cows WAVE was negatively correlated
ŽP-0.01 with TIMEIN3. N2, N2% and the number of soft or quiet vocalisations V1 ,. Ž .
Ž .
although not significantly Ps0.05 correlated, were moderately associated with
Ž . Ž .
TIMEIN3. The number of vocalisations, quiet or soft V1 and loud or harsh V2 and
Ž .
SPEED50 were significantly P-0.05, P-0.01 and P-0.05, respectively correlated
Ž .
with the cow behaviour variable FSqFSKon Table 3 . N2 when moving cows out of
Ž .
the shed was significantly correlated with FSqFSKon rs y0.39, P-0.05 . In
Ž
addition, the composite score used to assess attitude of the stockperson ATTITUDE
. Ž . Ž .
SCORE was negatively correlated P-0.05 with FSqFSKon Table 3 .
Ž .
ATTITUDE SCORE was negatively correlated P-0.01 with the stockperson
Ž .
behaviour variable, N2%, and positively correlated with V1 Table 4 . ATTITUDE SCORE was negatively correlated with N2% when moving cows out of the shed
Žrs y0.55, P-0.01 . Regression analysis with N2% as the dependent variable.
Ž 2
showed that ATTITUDE SCORE accounted for significant variance R s0.20, F1,27s
.
8.19, P-0.01 . The variables V2 and ATTITUDE SCORE were significant predictors
Ž 2 .
of the variable FSqFSKon R s0.25, F2,26s5.68, P-0.01 .
Ž .
There was a significant negative correlation rs y0.37, P-0.05 between the
Ž .
stockperson behavioural variable, the number of positive tactile interactions P1 , and
Ž .
incidence of lameness. V1 was positively correlated with P1 rs0.43, P-0.05 , and V2 was positively correlated with N2% and N2 when moving cows out of the shed
Žrs0.37 and 0.37, P-0.05 ..
4. Discussion
fear-productivity relationships have been found in other industries. For example, fear of humans was found to account for about 23% of the variation in reproductive
perfor-Ž .
mance among 19 commercial pig farms Hemsworth et al., 1989a , about 28% of the
Ž
variance in feed efficiency among 22 commercial meat chicken farms Hemsworth et al.,
.
1994b and about 28% of the variance in egg production among 16 commercial laying
Ž .
hen farms Barnett et al., 1992 . The fact that the dairy farms studied here were likely to have differed in terms of many of the farm inputs that affect productivity, such as nutrition, genotype and health, and the fact that similar results have been found in the pig and poultry industries, suggest that this fear–productivity relationship in the dairy industry may be a robust one.
While it is possible that there is a causal basis for this fear–productivity relationship in the dairy industry, other factors, such as the stockperson’s technical knowledge, presence of stray voltage, nutritional or mineral deficiencies and genotype, may directly affect cow productivity. Therefore, these non-behavioural factors, by affecting cow fear, may be responsible for the observed fear–productivity relationship. However, there is no unequivocal evidence in the literature that these factors affect both fear responses and productivity of cows.
Research on the pig that demonstrates a causal relationship between fear of humans and productivity can be utilised to provide some perspective to the findings of the present study on dairy cows. Observations in the pig industry have revealed negative
Ž .
fear–productivity relationships Hemsworth et al., 1981b, 1989a and handling studies on both young and adult pigs have also shown that high fear levels, through a chronic
Ž
stress response, may depress the growth and reproductive performance of pigs
Hems-.
worth and Barnett, 1991; Hemsworth et al., 1981a, 1986 . Two recent studies suggest
Ž .
that aversive handling may depress the milk yield of cows. Rushen et al. 1999 found that the presence of an aversive handler increased residual milk and tended to decrease milk yield. Moderate or forceful slaps imposed briefly before or after milking when animals failed to avoid humans, increased flight distance and tended to reduced milk
Ž .
yield in heifers Breuer et al., 1997 . Further studies under controlled experimental conditions are required to establish whether fear of humans in cows affects productivity and if so, whether the mechanism is via a stress response.
Studies on both experimental and commercial pigs have demonstrated that the behaviour of the stockperson is an important determinant of the pigs’ fear of humans. The nature of the behaviour that the stockperson directed towards hisrher pigs was
Ž
associated with the pig’s fear of humans Coleman et al., 1998; Hemsworth et al.,
.
1989a . The present study revealed only moderate correlations between stockperson behaviour and fear of humans in cows. For example, while the frequency of waving by the stockperson was significantly and negatively correlated with approach behaviour of
Ž .
human used in the standard approach test in the present study did not reflect the response shown to the familiar stockperson. In a series of experiments, de Passille et al.
Ž1996 found that dairy calves exhibited clear avoidance of a handler that had previously.
handled them in a negative manner in comparison to handlers wearing different colour clothing who were either unfamiliar to the calves or had previously handled them in a positive manner. Initially there was a generalisation of the aversive handling, with calves showing increasing avoidance of all handlers, but with repeated treatment, calves discriminated between handlers, and in particular between the ‘‘negative’’ and ‘‘posi-tive’’ handlers. It is of interest that discrimination was greatest when tested in the area in which handling had previously occurred rather than in a novel location. These data on calves indicate that discrimination between people by farm animals will be easier if the animals have some distinct cues on which they can discriminate such as colour of clothing or location of handling. To test whether or not commercial cows showed similar approach behaviour to familiar and unfamiliar stockpeople in a standard approach test as used in the present experiment, 60 commercial dairy cows were studied by the authors
ŽHemsworth et al., 1995 . The approach behaviour of cows to the familiar stockperson.
was not significantly different to that shown by cows to the unfamiliar stockperson, indicating that the behavioural responses of commercial cows to unfamiliar humans in
Ž .
the standard test as used in the present study is likely to be similar to that shown to the familiar stockperson. Therefore, while the associations between stockperson behaviour and cow fear were only moderate in the present study, the associations between both stockperson behaviour and cow productivity and those between cow fear and cow productivity suggest that human–animal interactions may be implicated in fear–produc-tivity relationships. Further research is obviously warranted to examine the behaviour of the stockperson that may regulate the commercial cow’s fear of humans.
The attitude of the stockperson towards pigs has been found to be an important determinant of the behaviour that the stockperson directed towards hisrher pigs in the
Ž .
pig industry Hemsworth et al., 1989a . For example, beliefs that considerable force was required to move pigs and that pigs do not require petting were associated with the use of a high percentage of negative behaviours, both moderate and forceful, when handling
Ž .
The correlations between human behaviour and cow restlessness in the presence of the stockperson during milking in the present study suggest that a component of these flinch and step responses and flinch, step and kick responses at milking is a response to the stockperson. Use of negative tactile interactions when moving cows out of the shed, loud harsh vocalisation and speed of movement of the stockperson as he or she moves his or her cows, were positively correlated with restlessness in the presence of the stockperson. In contrast, the use of soft, quiet vocalisation were negatively correlated with restlessness in the presence of the stockperson. Several previous studies have shown that fear of humans may increase restlessness during milking. For instance,
Ž .
Hemsworth et al. 1987; 1989b found that heifers that had calved in the presence of a human, and were subsequently less fearful of humans, displayed less flinch, step and kick responses in the presence of a human during milking than heifers that calved without human presence. The authors suggested that the heifers that had calved in the presence of a human had formed a general social attachment to humans and were therefore less fearful of humans. However, there is no evidence from the present study that the flinch and step and flinch, step and kick responses are correlated with the behavioural responses of cows to an experimenter in the standard approach test. Restlessness of cows during milking may also be affected by a number of factors other than fear of humans. For example, restlessness at milking may be affected by stray voltage, mineral deficiency, social pressures from adjacent cows, lameness, presence of biting flies, etc. There was no evidence from the present study that herd size and lameness were associated with restlessness.
Observations from the present study indicate that where restlessness was high,
Ž .
productivity was low. According to Willis 1983 , restlessness, as measured by the number of flinch, step and kick responses, is indicative of a stressful situation. In contrast to the present results, there are reports of significant positive between-herd
Ž . Ž
associations Purcell et al., 1988 and within-herd associations Willis, 1983, which were
.
not confirmed by Purcell et al., 1988 between restlessness during milking and milk production. There are no obvious explanations for these contradictory findings between the present study and those in the literature. However, as previously mentioned, a number of factors such as stray voltage, mineral deficiency, social pressures from adjacent cows and lameness may affect restlessness. Thus, differences in the aetiology of restlessness may contribute to these contradictory results.
The present study provides evidence of relationships between a number of human and animal variables. Research is clearly required to further examine these relationships because of the possible implications on animal productivity and welfare. If these associations have a causal basis, there may be opportunity in the industry to manipulate the human factors regulating fear responses in order to improve productivity. For example, fear of humans by dairy cows may be reduced by identifying and targeting those human factors which regulate these fear responses through cognitive behavioural intervention. After identifying the main attitudinal and behavioural profiles of
stockpeo-Ž .
ple regulating the pig’s fear of humans, Coleman et al. 1998 and Hemsworth et al.
Ž1989a; 1994a designed a training programme aimed at improving the attitudinal and.
for improved reproduction in pigs. Research is underway in the dairy industry to fully identify the attitudinal and behavioural profiles of stockpeople which may affect the fear and productivity of cows. Such research may lead to the opportunity to improve the productivity of commercial cows by improving the attitude and behaviour profiles of stockpeople.
Acknowledgements
The work was supported by a grant from the Australian Dairy Research and Development. The technical support of S. Borg, E. Newman, and B. Schirmer and the participation of the study farmers are gratefully acknowledged.
References
Barnett, J.L., Hemsworth, P.H., Newman, E.A., 1992. Fear of humans and its relationships with productivity in laying hens at commercial farms. Br. Poult. Sci. 33, 699–710.
Boissy, A., Bouissou, M.F., 1988. Effects of early handling on heifers’ subsequent reactivity to humans and to unfamiliar situations. Appl. Anim. Behav. Sci. 20, 259–273.
Boivin, X., Le Neindre, P., Chupin, J.M., 1992. Establishment of cattle–human relationships. Appl. Anim. Behav. Sci. 32, 325–335.
Breuer, K., Hemsworth, P.H., Coleman, G.J., 1997. The influence of handling on the behaviour and productivity of lactating heifers, Proc. 31st Intern.Congress Intern. Soc. Appl. Ethology, Research Institute of Animal Production, Prague, Czech Republic and the Institute of Animal Biochemistry and Genetics, Slovak Academy of Sciences, Slovakia.
Coleman, G.C., Hemsworth, P.H., Hay, M., Cox, M., 1998. Predicting stockperson behaviour towards pigs from attitudinal and job-related variables and empathy. Appl. Anim. Behav. Sci. 58, 63–75.
Ž .
Hemsworth, P.H., Barnett, J.L., 1987. Human–animal interactions. In: Price, E.O. Ed. , The Veterinary Clinics of North America, Food Animal Practice, W.B. Saunders Co., Philadelphia, 3, 339–356. Hemsworth, P.H., Barnett, J.L., 1991. The effects of aversively handling pigs either individually or in groups
on their behaviour, growth and corticosteroids. Appl. Anim. Behav. Sci. 30, 61–72.
Hemsworth, P.H., Barnett, J.L., Coleman, G.J., 1993. The human–animal relationship in agriculture and its consequences for the animal. Anim. Welfare 2, 33–51.
Hemsworth, P.H., Barnett, J.L., Hansen, C., 1981a. The influence of handling by humans on the behaviour, growth and corticosteroids in the juvenile female pig. Horm. Behav. 15, 396–403.
Hemsworth, P.H., Brand, A., Willems, P.J., 1981b. The behavioural response of sows to the presence of human beings and their productivity. Livest. Prod. Sci. 8, 67–74.
Hemsworth, P.H., Barnett, J.L., Hansen, C., 1986. The influence of handling by humans on the behaviour, reproduction and corticosteroids of male and female pigs. Appl. Anim. Behav. Sci. 15, 303–314. Hemsworth, P.H., Barnett, J.L., Breuer, K., Coleman, G.C., Matthews, L.R., 1995. An investigation of the
relationships between handling and human contact and the milking behaviour, productivity and welfare of commercial dairy cows. Research Report on Dairy Research and Development Council Project, Attwood, Australia.
Hemsworth, P.H., Barnett, J.L., Coleman, G.J., Hansen, C., 1989a. A study of the relationships between the attitudinal and behavioural profiles of stockpersons and the level of fear of humans and reproductive performance of commercial pigs. Appl. Anim. Behav. Sci. 23, 301–314.
Hemsworth, P.H., Coleman, G.J., Barnett, J.L., 1994a. Improving the attitude and behaviour of stockpersons towards pigs and the consequences on the behaviour and reproductive performance of commercial pigs. Appl. Anim. Behav. Sci. 39, 349–362.
Hemsworth, P.H., Coleman, G.J., Barnett, J.L., Jones, R.B., 1994b. Fear of humans and the productivity of commercial broiler chickens. Appl. Anim. Behav. Sci. 41, 101–114.
Hemsworth, P.H., Hansen, C., Barnett, J.L., 1987. The effects of human presence at the time of calving of primiparous cows on their subsequent behavioural response to milking. Appl. Anim. Behav. Sci. 18, 247–255.
Hemsworth, P.H., Price, E.O., Borgwardt, R., 1996. Behavioural responses of domestic pigs and cattle to humans and novel stimuli. Appl. Anim. Behav. Sci. 50, 43–56.
de Passille, A.M.B., Rushen, J., Ladewig, J., Petherick, C., 1996. Dairy calves’ discrimination of people based on previous handling. J. Anim. Sci. 74, 969–974.
Purcell, D., Arave, C.W., Walters, J.L., 1988. Relationship of three measures of behaviour to milk production. Appl. Anim. Behav. Sci. 21, 307–313.
Rushen, J., de Passille, A.M.B., Munksgaard, L., 1999. Fear of people by cows and effects on milk yield, behaviour and heart rate at milking. J. Dairy Sci. 82, 720–727.
Seabrook, M.F., 1972. A study to determine the influence of the herdsman’s personality on milk yield. J. Agric. Lab. Sci. 1, 45–59.
