1999 J Dairy Sci 82:720–727
Received February 19, 1998.
Accepted December 21, 1998.
1Lennoxville Contribution number 600.
2Author to whom all correspondence should be addressed: Dairy
and Swine Research and Development Centre, Agriculture and
Agri-Food Canada, PO Box 90, 2000 Road 108E, Lennoxville, QC,
Canada J 1M 1Z3.
Fear of People by Cows and Effects on Milk Yield,
Behavior, and Heart Rate at Milking1
J. RUSHEN,*,2A.M.B. DE PASSILLE´,* and L. MUNKSGAARD†
*Dairy and Swine Research and Development Centre,
Agriculture and Agri-Food Canada,
Lennoxville, QC, Canada J1M 1Z3
†Department of Animal Health and Welfare,
Danish Institute of Agricultural Science,
Research Centre Foulum, 8830 Tjele, Denmark
To examine the ability of cows to recognize people
and the effects of the fear of people by cows at milk-
ing, cows (n = 14) were handled by two people; one
handled the cows gently, and the other handled them
aversively. The handlers wore clothes of different
color. After handling, the cows stood further from the
aversive handler than from the gentle handler. When
the handlers changed the color of their clothing, the
cows did not discriminate between them. The gentle
handler stood close to the cows for one milking, and
the aversive handler stood close to the cows for
another milking. For two control milkings, neither
handler was present. Measurements included milking
duration, milk yield, residual milk, heart rates, inci-
dence of movement, and kicking behavior of the cows.
Compared with control milkings, the presence of the
gentle handler did not change milk yield or residual
milk. The presence of the aversive handler increased
residual milk by 70%. Kicking behavior of cows dur-
ing milking was reduced with either handler present,
and kicking during udder preparation was reduced
with the aversive handler present. For cows that best
discriminated between the handlers, the presence of
the aversive handler increased movement and heart
rate during milking. For cows that did not dis-
criminate well between the handlers, the presence of
either handler increased heart rate and decreased
dividual people, and the fear of people who are
present during milking may reduce milk yield.
( K ey words: stress, fear, milk yield, handling)
Rough or aversive handling of cows can reduce
milk yield and animal welfare by causing stress (1,
10). Cows handled aversively yield less milk, are
more difficult to handle, and are more fearful of peo-
ple (24); gentle handling causes dairy cows to be less
fearful and easier to manage (3, 11, 18) but may not
cause an increase in milk yield (12, 18). Higher
yielding cows approach people more readily (1, 13,
24), suggesting a lesser degree of fear, although some
studies have not found such an effect (8, 21). Cows
that are less fearful are less likely to kick the milker
(12). The degree of fearfulness that cows demon-
strate toward people is important, and the factors
that influence this fear should be studied.
Genetic differences among cattle affect the extent
to which they fear people (10, 16); however, the way
cattle are handled can also affect fearfulness. When
cattle are handled aversively by people, they associate
the handling with people and develop a learned fear
(6, 11). Whether cows can distinguish among differ-
ent people and, hence, become frightened of individu-
als because of the aversive treatment they receive is
an important question. Previously, de Passille Â et al.
(6) and Munksgaard et al. (20) found that dairy
cattle readily learned to discriminate between people
who treated them differently, avoiding those who
treated them aversively and approaching those who
fed or handled them gently.
Fear of individual people may be important in the
differences in milk yield observed when cows were
milked by relief milkers (1, 15, 24). In this experi-
ment, we examined whether cows could discriminate
between different handlers and whether they used
the color of clothing of the handlers to do so. We also
examined how much fearfulness of a specific handler
who was present during milking influenced the milk
yield and behavior of the cows.
Journal of Dairy Science Vol. 82, No. 4, 1999
FEAR OF PEOPLE BY COWS AT MILKING
MATERIALS AND METHODS
Cows and Housing
Fourteen Holstein cows (9 in their second lacta-
tion, 3 in their third lactation, and 1 each in their
fourth and fifth lactations) were 5 to 9 wk in lacta-
tion when the experiment began. The cows were
housed in tie stalls from at least 4 wk before the
experiment began and were fed a total mixed diet for
ad libitum intake. The feed was delivered once daily,
and cows were milked twice daily. In the morning,
cows were milked at a fixed hour ( ±15 min) in their
stalls using a portable milking machine (Westphalia
Inc., Elk Grove Village, IL ). During the morning
milkings, the cows were always milked by the same
person in a fixed sequence. Evening milkings were
carried out in the milking parlor by the regular barn
staff at a constant time ( ±15 min) after the morning
milkings. Morning milkings took place between 0700
and 0900 h, and evening milkings took place between
1800 and 2000 h.
Experimental procedures were approved by the In-
stitutional Animal Care Committee of the Lennoxville
Research Center in compliance with the requirements
of the Canadian Council on Animal Care. Each cow
was treated by two different handlers, one male and
one female. Previous research (6) found no differ-
ences in how cows reacted to male and female han-
dlers. One handler always treated the cow aversively,
and the other handler always treated the cow gently.
Neither handler had handled the cows before the
experiment began. The two handlers wore overalls of
different color, either red or yellow. When treating
each cow, each handler consistently wore the same
color. These colors were chosen to maximize the abil-
ity of the cows to distinguish between the handlers
(20) and to distinguish the handlers from the barn
staff, who wore green or blue. Cattle can distinguish
these colors from a grey color of similar luminosity
(1). The treatments were balanced with respect to
the color of the overalls and the gender of the han-
A variety of aversive and gentle treatment proce-
dures were chosen that approximated the range of
aversive and gentle handling that dairy cattle might
experience in commercial practice. Care was taken to
ensure that the cows were not injured by the treat-
ments, and all cows were inspected by a qualified
veterinarian during and after the experiment. Aver-
sive treatments consisted of forcefully striking the
cow on the head or side with the open palm (four
times each treatment period), use of a cattle prod on
the flank of the cow (one shock was given once per
day), and striking the cow once on the horn region
and flanks with the flat surface of a plastic shovel
(once per day). All treatments were performed while
the handler shouted at the cow. We used a plastic
shovel rather than a stick, which would be more
common, to minimize any chance of injury. Gentle
treatments involved brushing the cow, offering hay
and barley mixed with molasses, and speaking in a
gentle voice. The treatments were performed from
both the front and side of the cow, except that feed
was offered only from the front. Each treatment
lasted 1 to 6 min. A variable duration of treatment
was used so that the cows would not be able to predict
when the treatment was finished.
For each cow, there were 3 to 6 handling periods
per day for 5 d. Each handling period included one
gentle and one aversive treatment of the same dura-
tion separated by an interval of 15 min. The number
of handling periods per day depended on the time
available following normal barn routines. An interval
of at least 15 min was allowed between subsequent
handling periods. Aversive and gentle treatments
were balanced so that the cows received the same
daily duration of aversive and gentle treatments. The
order of treatments within the handling period was
chosen so that each treatment (aversive or gentle)
was followed by a treatment of the same type one-half
of the time and was followed by the opposite treat-
ment the other half of the time. Consequently, cows
should not have been able to determine how they
would be treated from how they were treated previ-
ously. One-half of the cows started with an aversive
treatment, and one-half started with a gentle treat-
Tests of the fearfulness of the cows were based on
those administered previously (6, 20). During the
tests, the handlers stood motionless with hands in
pockets for 60 s at 0.8 m in front of the bar to which
the cow was tethered. An observer, wearing blue over-
alls stood in front of a neighboring stall in full view of
the cow. The position of the cow was scored every 5 s
for a total of 60 s according to the following criteria:
contact with the handler (score = 1), muzzle in front
of the tie bar and within 40 cm of the handler (score =
2), muzzle in front of the tie bar but further than 40
cm from the handler (score = 3), muzzle at the level
of the tie bar (score = 4), muzzle behind the tie bar
but with the chain loose (score = 5), or muzzle behind
Journal of Dairy Science Vol. 82, No. 4, 1999
RUSHEN ET AL.
the tie bar with the chain fully extended (score = 6).
Such scores are an effective, simple way of measuring
the extent that cows are fearful of people (1, 11, 20,
21). The fear tests were performed before gentle and
aversive treatments and after 5, 10, 15, and 19 han-
dling periods. The tests were interspersed throughout
the treatments to determine whether the cows could
successfully discriminate between the gentle and
aversive handlers and when this discrimination be-
came apparent. Tests were also performed with a
third person, a male wearing green clothing, who did
not handle the cows. These tests were performed both
before treatments and after 19 handling periods.
Three further fear tests were performed on the
final day after the final experimental milkings to
determine whether the color of the clothes worn by
the handlers was used by the cows as a cue to dis-
criminate between the handlers. Two tests were per-
formed to determine whether the cows could still
distinguish between the handlers when they ex-
changed the color of their clothing, that is when the
handler that normally was dressed in red, wore yel-
low and vice versa. For one-half of the cows, the
handlers wore clothes of the usual color during the
first test and clothes of the opposite color during the
second test. For the other half of the cows, the han-
dlers wore the opposite color clothes during the first
test and the usual color clothes during the second
test. During a third test, both handlers wore green
overalls, which is a color worn by the normal barn
staff. Between tests, the cows were subjected to one
aversive and one gentle treatment; each treatment
lasted 1 min. The order of the treatments was
balanced across cows.
responses of the cows during four milkings, which
occurred after 8 (first control milking), 15 (first ex-
perimental milking), and 19 (second experimental
and second control milking) handling periods. The
first and last milkings were control milkings at which
neither of the two handlers were present. The second
and third milkings were experimental milkings. Dur-
ing these two experimental milkings, one of the two
handlers was present. For one-half of the cows, the
gentle handler was present during the first ex-
perimental milking, and the aversive handler was
present during the second experimental milking. The
reverse order was followed for the other half of the
cows. At least 1 d elapsed between subsequent milk-
ings. We carried out the final 4 treatment periods
during the day between the two experimental milk-
to treatments,we recordedthe
ings in case the cows learned to ignore the handlers
because the handlers did not treat the cows during
the milkings. These milkings were performed in the
morning in the tie stalls by the same milker and at
the same time ( ±15 min) as the regular morning
During the experimental
milker began udder preparation, the handler walked
toward the cow until he or she stood directly in front
of the cow, in contact with the front of the stall. The
handler stood in this position for 3 min with hands in
pockets, not looking at or touching the cow. After 3
min, the handler stood in the same posture by the
side of the cow, within arm's reach, for the remainder
of the milking. Once the milking was completed, the
handler left the stall.
Sixty minutes prior to milking, cows were fitted
with a heart rate monitoring system (Polar Oy, Hel-
sinki, F inland), which was set to record mean heart
rates during each 5-s period (12, 23). During the
milkings, an observer sat 2 m behind the cow and
recorded each instance of kicking or movement. A
movement was scored whenever one hoof was lifted
off of the ground. Kicking behavior was defined as a
kick in any direction so that the hoof was lifted at
least 15 cm off of the ground. These actions were
recorded separately for the periods of udder prepara-
tion and actual milking. The durations of both phases
were recorded. Detachment of the milking machine,
defecation, urination, and vocalization were also
recorded, but these behaviors were observed so rarely
that we did not include them in the analysis. Two
minutes after the end of milking 10 IU of oxytocin in
0.5 ml of saline were injected into the coccygeal vein
of the cow, and the milking machine was reattached
30 s later to obtain residual milk.
Mean distance scores (distances that cows kept
from each handler) were calculated for each cow from
the 12 scores recorded every 5 s of the 1-min fear test.
These mean scores were analyzed by PROC GLM
(23) using a model that included cow, handler (aver-
sive or gentle), and fear test number. The effect of
handler was tested using the residual error; scores on
the initial pretreatment test were included as covaria-
bles. Separate analyses were performed using the
same model except that the model also included
either the gender of the handler or color of the cloth-
ing worn. The effects of gender of the handler and the
effects of the color of the clothing were tested against
the residual error. A further analysis was used to test
the effects of a change of clothing color. The model
Journal of Dairy Science Vol. 82, No. 4, 1999
FEAR OF PEOPLE BY COWS AT MILKING
Figure 1. Mean ( ±SE ) distance score (distance that the cows
kept from each handler) for all cows on each test. One treatment
period included one gentle and one aversive treatment of the same
duration. During each test, the position of the cow was scored every
5 s for a total of 60 s as either 1 (contact with the handler), 2
(muzzle in front of the tie bar and within 40 cm of the handler), 3
(muzzle in front of the tie bar but further than 40 cm from the
handler), 4 (muzzle at the level of the tie bar), 5 (muzzle behind
the tie bar but with the chain loose), or 6 (muzzle behind the tie
bar with the chain fully extended). The mean of the 12 scores was
calculated for each cow during each test. Legend: gentle handler
(open bar); aversive handler (solid bar). NS = P > 0.10. **P < 0.01;
n = 14.
included cow, handler (aversive or gentle), color of
clothing worn (normal color, opposite color, or green),
and the interaction between handler and color of
clothing. The effect of handler, clothing color, and
interaction between the two was tested against the
residual error. Planned linear contrasts tested the
difference between the gentle and aversive handlers
separately for each level of the clothing color factor.
Marked differences were evident among cows in the
degree to which they distinguished between the two
handlers. Some cows barely discriminated between
the handlers. Therefore, a median split was done
using the scores for the fear tests performed after 15
and 19 treatment periods to divide the cows into two
equal groups (poor or good discriminators) based on
the mean distance score for the aversive handler mi-
nus the mean distance score for the gentle handler.
During the control and experimental milkings, we
recorded milk yield, residual milk, and the duration of
udder preparation and milking. Total milk was calcu-
lated as the sum of milk obtained plus residual milk.
Cows were also scored for frequencies of movement
and kicking during both udder preparation and milk-
ing. Mean heart rates during the 15 min preceding
milking and during the entire milking were calcu-
lated. Because 1 cow had mastitis during two of the
milkings, her data were excluded from the analysis.
Results are presented as least squares means ( ±SE ).
The measurements were analyzed by PROC GLM
(23) using a model that included degree of discrimi-
nation by the cows of the two handlers (i.e., the cows
were classified as poor discriminators or good dis-
criminators), cow (nested within discriminative abili-
ty), and type of milking (i.e., control milking with no
handler or experimental milking with either the aver-
sive or gentle handler present). The effect of dis-
criminative ability was tested using cow as the error
term. The effect of the type of milking and the inter-
action between type of milking and discriminative
ability of the cow was tested against the residual
error. Paired comparisons using least squares means
were performed to examine the differences between
the control milkings and the two experimental milk-
ings separately for each group of cows (good discrimi-
nators and poor discriminators). Because the be-
nonparametric tests were also performed. However,
the results were the same as those obtained using
PROC GLM (23). Therefore, we report only the latter
results. The two control milkings did not differ ( P >
0.10) for any measure. Therefore, for ease of presen-
tation, the Results section contains mean values for
the two control milkings. During the second milking,
1 cow was removed because of health reasons. There-
fore, the results at milking are based on the remain-
ing 13 cows.
Distance Scores During Fear Tests
Before treatments, cows stood an equal distance
from both handlers (F igure 1). Following the first 5
treatments, cows initially stood further from both
handlers, but, with further treatment, the cows began
to approach the gentle handler and distanced them-
selves from the aversive handler (F igure 1). After 10
handling periods, the cows stood closer ( P < 0.01) to
the gentle handlers than to the aversive handlers.
This difference in how the cows reacted to the two
handlers was maintained following the remaining 9
handling periods (F igure 1). No effect of gender of
the handler or of color of the coveralls was observed;
there was no interaction between these factors and
the treatment applied by the handler ( P > 0.10).
Distance scores to the third person, who wore green
and did not handle the cows, were not different ( P >
0.10) after treatments (3.13 ± 0.29) compared with
the distance scores before treatments (2.86 ± 0.26).
For the tests after the milkings, the distance scores
were higher for the aversive handler than for the
gentle handler (4.1 vs. 2.7 ± 0.3; P < 0.05) when the
Journal of Dairy Science Vol. 82, No. 4, 1999
RUSHEN ET AL.
TABLE 1. Least squares means and significance of effects of treatment (presence of handler) and of
the interaction between presence of handler and the ability of cows (n = 13) to discriminate between
the two handlers on milk yield, behavioral measures, and heart rate at milking.
a,b,cMeans within rows with different superscripts differ ( P < 0.05).
1H = Presence of handler; H × D = interaction of the presence of the handler and the ability of the
cow to discriminate between the gentle and the aversive handler.
2Pooled standard error.
3P > 0.10.
*P < 0.05.
**P < 0.01.
***P < 0.001.
HH × D
Total milk, kg
Milk yield, kg
Residual milk, kg
Heart rate change, beats/min
Figure 2. Mean ( ±SE ) residual milk yield during control milk-
ings or when the aversive or gentle handlers were present for cows
that discriminated well between handlers and cows that discrimi-
nated poorly between handlers. Columns with different super-
scripts (a, b) differ ( a = 0.05). Legend: gentle handler present
(open bar), aversive handler present (solid bar), no handler
present (lined bar) (n = 13).
handlers wore clothes of the usual, assigned color.
However, there was no difference in distance scores
for the aversive handler and the gentle handler when
the handlers exchanged the color of the clothes worn
(2.8 vs. 2.7 ± 0.4; P > 0.10) or when both handlers
wore green overalls (3.0 vs. 3.1 ± 0.4; P > 0.10).
Total milk yield (milk yield obtained at milking
plus residual milk) did not differ between the milk-
ings (Table 1). Milk obtained during milking tended
to differ between milkings; less milk was obtained
during the experimental milking when the aversive
handler was present than during control milkings
when no handler was present (Table 1). Residual
milk was significantly affected by the presence of a
handler; more residual milk was obtained when the
aversive handler was present than when the gentle
handler or when no handler was present (Table 1).
No differences ( P > 0.10) in milk yield between the
good discriminators and the poor discriminators were
observed. However, a significant interaction did exist
between the presence of the handler and the ability of
the cow to discriminate (Table 1). For cows that
discriminated poorly, there was no difference in
residual milk between milkings (F igure 2). However,
for cows that discriminated well, residual milk was
higher in the presence of the aversive handler than in
the presence of the gentle handler or no handler
(F igure 2).
Heart Rate During Milking
No differences between milkings ( P > 0.10) in
mean heart rate were detected before milking (80.4 ±
Journal of Dairy Science Vol. 82, No. 4, 1999
FEAR OF PEOPLE BY COWS AT MILKING
Figure 3. Mean ( ±SE ) change in heart rate from premilking
baseline during control milkings or when the aversive or gentle
handlers were present for cows that discriminated well between
handlers and cows that discriminated poorly between handlers.
Columns with different superscripts (a, b) differ ( a = 0.05).
Legend: gentle handler present (open bar), aversive handler
present (solid bar), no handler present (lined bar) (n = 13).
Figure 4. Mean ( ±SE ) rate of movement by the cows during
control milkings, or when the aversive or gentle handlers were
present for cows that discriminated well between handlers and
cows that discriminated poorly between handlers. Columns with
different superscripts (a, b) differ ( a = 0.05). Legend: gentle
handler present (open bar), aversive handler present (solid bar),
no handler present (lined bar) (n = 13).
1.59 beats/min). However, the change in heart rate
during milking from the premilking baseline heart
rate was significantly affected by the presence of the
handlers (Table 1). During control milkings when no
handler was present, heart rates were not higher
than those recorded prior to milking ( P > 0.10).
Heart rates increased during experimental milkings
when either of the handlers was present; the largest
increase occurred when the aversive handler was
present (Table 1). However, the interaction of the
presence of handlers and the ability of the cows to
discriminate between the handlers was nearly signifi-
cant ( P = 0.06). For cows that discriminated poorly,
heart rates rose during milking and were signifi-
cantly higher when either of the two handlers was
present compared with the heart rate when no han-
dler was present (F igure 3). For cows that discrimi-
nated well between the handlers, heart rate rose dur-
ing milking and was highest when the aversive
handler was present compared with that when the
gentle handler or no handler was present (F igure 3).
Behavior During Milking
During udder preparation, the kicking behavior of
cows was lessened when the aversive handler was
present compared with that when the gentle handler
or no handler was present (Table 1). During actual
milking, the kicking behavior of the cows was les-
sened when either the gentle or aversive handler was
present compared with that when no handler was
present. The amount of movement by the cow during
udder preparation was not affected by the presence of
either handler. Cows moved less during milking when
no handler was present than when either the gentle
or the aversive handler was present (Table 1).
No overall differences ( P > 0.10) during milkings
between good discriminators and poor discriminators
were detected. However, there was a significant inter-
action (Table 1) between the presence of the han-
dlers and the ability of the cows to discriminate be-
tween the handlers. Cows that discriminated poorly
moved less when either handler was present (F igure
4). However, cows that discriminated well moved
more during milking when the aversive handler was
present than when the gentle handler was present.
Our results show that milk yield and behavior of
cows at milking depended on which handler was
present at the milking. The presence of an aversive
handler during milking reduced milk yield because of
increased residual milk and lead to elevated heart
rates. These effects varied according to the ability of
the cows to distinguish between different handlers
and were most apparent among the cows that could
best distinguish between the two handlers. These
effects on milk yield are of sufficient magnitude to
indicate that the reduced yield for cows that were
handled aversively (24) and some of the differences
in productivity between different milkers on commer-
cial dairies (1, 15) may be due to cows that were
frightened of the individuals who handled or milked
them. The results confirm our previous findings (6,
Journal of Dairy Science Vol. 82, No. 4, 1999
RUSHEN ET AL.
20), which showed that dairy cattle can distinguish
between different people.
The increase in residual milk when the aversive
handler was present suggests an inhibition of oxyto-
cin secretion, leading to a reduced or delayed milk
ejection (4), which is a common mechanism that
underlies the inhibition of milk yield induced by
stress (4, 5). A fear of people has been shown to
increase milk ejection failure in goats (19). The
amount of residual milk was much lower than that
reported following oxytocin block in Swiss Brown
cows (5) and slightly lower than that found for Cana-
dian Holsteins (7), suggesting a partial reduction of
oxytocin secretion rather than total inhibition (4).
Increased heart rate during milking may be a sign of
increased sympathetic nervous
release of catecholamines, which is another mechan-
ism by which stress can block milk secretion (2, 9,
17). Increased heart rate at milking has been as-
sociated previously with increased residual milk
(26). Measures of heart rate may be useful noninva-
sive measures of acute stress reactions in dairy cattle
Some disagreement exists in the literature as to
the relationship between the fearfulness of cows and
milk yield. Some studies (13, 24) have found strong
negative correlations across herds between fearful-
ness and milk yield; other studies (8, 21) have found
no such relationship. Some of the discrepancies may
be due to differences in the measures used. Although
all studies used the distance between cows and han-
dlers to measure fearfulness, Purcell et al. (21)
measured the closest distance between cow and han-
dler, and Hemsworth et al. (13) measured the mean
distance over a longer period of time. In addition, for
this experiment, we used the measurements to exam-
ine the acute reactions of cows to the presence of a
specific handler at a milking; Purcell et al. (21)
examined the effect of the temperament of the cow on
milk yield over the entire lactation. Similarly, Lyons
(19) found no differences in overall milk yield of
goats associated withincreased
although more milk was obtained as residual milk,
suggesting more failed milk ejections.
Marked differences exist among handlers in the
ease with which they can handle cattle (12, 15, 16).
Some of these differences may be due to the degree of
fear that cows have of specific individuals. The degree
of movement of the cow during milking was affected
by the presence of the handlers; more movement was
observed when the aversive handler was present for
those cows that could discriminate well between the
handlers. For cows that discriminated poorly, there
was more movement in the presence of either han-
dler. Kicking behavior during milking reduces the
efficiency of milking and is a possible source of injury.
The handling of primiparous cows during calving is
reported to reduce kicking (12), but we found no
evidence that a high degree of kicking was related to
the degree of fearfulness of the cows. In fact, the
presence of either handler tended to reduce the inci-
dence of kicking, even though heart rates and
residual milk were elevated. The similar response of
the cows to both the aversive handler and the gentle
handler may be because both handlers were relatively
unfamiliar to the cows compared with the people
normally present at milking. Kicking behavior may
be more associated with the aggressiveness of the
cows than with fearfulness, although we might have
found different results if the handler had milked the
cows. Some behavioral measures of agitation in cattle
are persistent over time and can be used as criteria
for culling cattle (10). However, our results show
that cow behavior during milking can vary greatly
according to the handlers that are present.
Most of the cows distinguished between the han-
dlers, supporting our previous observations (20),
although there were marked differences among cows.
Using a median split to place the cows into two
groups, we showed that cows that were best at dis-
criminating between the two handlers in our distance
test were also the cows for which residual milk and
movement during milking were most affected by the
handler present. At present, we do not know why
some cows discriminated better than others, but this
may be the result of individual differences in learning
performance, which are often reported for cattle (1,
12). Differences in fearfulness (10) or in the extent
that the treatments were perceived as aversive or
gentle could also have been involved. Hemsworth et
al. (11) found that cattle behave similarly with
different people. However, other studies (6, 20) have
shown that cattle can discriminate between different
people. Color of clothing is one cue that cattle use to
discriminate between people. When the handlers both
wore green, the cows no longer appeared to dis-
criminate between them, which shows the importance
of the visual system in discrimination by cows (22).
However, the color of the clothing is not the only cue.
When the aversive and gentle handlers exchanged the
color of their clothing, the cows did not consequently
avoid the gentle handler and approach the aversive
handler. Instead, the cows did not appear to dis-
criminate between the handlers, which is in agree-
ment with our previous results (20). Discrimination
of people by cows likely involves a mix of visual,
Journal of Dairy Science Vol. 82, No. 4, 1999 Download full-text
FEAR OF PEOPLE BY COWS AT MILKING
olfactory, and auditory cues, as was demonstrated for
pigs by Tanida and Nagano (25).
Accumulating evidence suggests that aversive han-
dling of dairy cows can substantially reduce produc-
tivity and ease of handling. Our results show that
cows can recognize individual people and that aver-
sive handling of cows by certain people causes cows to
become fearful of those people. The presence of these
people at milking can substantially increase residual
milk and thus reduce milk yield. Care should be
taken not to invoke fear in cows and to ensure that
the people responsible for the cows find the least
aversive means of handling them.
This research was supported by a grant from the
Dairy Farmers of Canada (Ottawa, Ontario, Canada)
to J . Rushen and A. M. de Passille Âand a travel grant
from the Organization for Economic Cooperation and
Development (Brussels, Belgium) to L. Munksgaard.
We thank Marjolaine St. Louis, Isabelle Blanchet,
and all of the staff at the dairy barn for their help.
1 Albright, J . L., and C. W. Arave. 1997. The Behaviour of Cattle.
CAB Int., Wallingford, UK.
2 Blum, J . W., D. Schams, and R. Bruckmaier. 1989. Cathechola-
mines, oxytocin and milk removal in dairy cows. J . Dairy Res.
3 Boissy, A., and M.-F. Bouissou. 1988. Effects of early handling
on heifers' subsequent reactivity to humans and to unfamiliar
situations. Appl. Anim. Behav. Sci. 20:259±273.
4 Bruckmaier, R. M., and J . W. Blum. 1998. Oxytocin release and
milk removal in ruminants. J . Dairy Sci. 81:939±949.
5 Bruckmaier, R. M., D. Schams, and J . W. Blum. 1993. Milk
removal in familiar and unfamiliar surroundings: concentra-
tions of oxytocin, prolactin, cortisol and beta-endorphin. J .
Dairy Res. 60:449±456.
6 de Passille Â, A.M.B., J . Rushen, J . Ladewig, and J . C. Petherick.
1996. Dairy calves' discrimination of people based on previous
handling. J . Anim. Sci. 74:969±974.
7 de Passille Â, A.M.B., J . Rushen, and P.-G. Marnet. 1997. Effects
of nursing a calf on milk ejection and milk yield during milking.
J . Dairy Sci. 80(Suppl. 1):203.(Abstr.)
8 Dickson, D. P., G. R. Barr, L. P. J ohnson, and D. A. Wiekert.
1970. Social dominance and temperament of Holstein cows. J .
Dairy Sci. 53:904±907.
9 Gorewit, R. C., and M. C. Aromando. 1985. Mechanisms in-
volved in the adrenaline-induced blockade of milk ejection in
dairy cattle. Proc. Soc. Exp. Biol. Med. 180:340±347.
10 Grandin, T. 1997. Assessment of stress during handling and
transport. J . Anim. Sci. 75:249±257.
11 Hemsworth, P. H., E. O. Price, and R. Borgwardt. 1996. Be-
havioural responses of domestic pigs and cattle to humans and
novel stimuli. Appl. Anim. Behav. Sci. 50:43±56.
12 Hemsworth, P. H., J . L. Barnett, A. J . Tilbrook, and C. Hansen.
1989. The effects of handling by humans at calving and during
milking on the behaviour and milk cortisol concentrations of
primiparous dairy cows. Appl. Anim. Behav. Sci. 22:313±326.
13 Hemsworth, P. H., K. Breuer, J . L. Barnett, G. J . Coleman, and
L. R. Matthews. 1995. Behavioral response to humans and the
productivity of commercial dairy cows. Pages 175±176 in Proc.
29th Int. Congr. Int. Soc. Appl. Ethol., Exeter, UK. Univ. Fed.
Anim. Welfare, Potters Bar, UK.
14 Hopster, H., J . M. O'Connell, and H. J . Blokhuis. 1995. Acute
effects of cow-calf separation on heart rate, plasma cortisol and
behaviour in multiparous dairy cows. Appl. Anim. Behav. Sci.
15 Knierem, U. 1991. The effect of human-animal interactions in
the milking parlour on the behaviour, heart-rate and milk yield
of dairy cows. M.Sc. Thesis, Univ. Edinburgh, Edinburgh, Scot-
16 Le Neindre, P., G. Trillat, J . Sapa, F. Me Ânissier, J . N. Bonnet,
and J . M. Chupin. 1995. Individual differences in docility in
limousin cattle. J . Anim. Sci. 73:2249±2253.
17 Lefcourt, A. M., and R. M. Akers. 1984. Small increases in
peripheral noradrenalin inhibit the milk-ejection response by
means of a peripheral mechanism. J . Endocrin. 100:337±344.
18 Lefcourt, A. M., and R. Barfield. 1995. Acclimation to milking of
heifers raised as calves under conditions of intensive or res-
tricted handling. J . Dairy Sci. 78(Suppl. 1):178.(Abstr.)
19 Lyons, D. M. 1989. Individual differences in temperament of
dairy goats and the inhibition of milk ejection. Appl. Anim.
Behav. Sci. 22:269±282.
20 Munksgaard, L., A. M. de Passille Â, J . Rushen, K. Thodberg, and
M. B. J ensen. 1997. Discrimination of people by dairy cows
based on handling. J . Dairy Sci. 80:1106±1112.
21 Purcell, D., C. W. Arave, and J . L. Walters. 1988. Relationship
of three measures of behavior to milk production. Appl. Anim.
Behav. Sci. 21:307±313.
22 Rehkamper, G., and A. Gorlach. 1997. Visual discrimination in
adult dairy bulls. J . Dairy Sci. 80:1613±1621.
23 SAS/STATUser's Guide, Version 6, Fourth Edition, Volume 2.
1990. SAS Inst., Inc., Cary, NC.
24 Seabrook, M. F. 1994. Psychological interaction between the
milker and the dairy cow. Pages 49±58 in Dairy Systems for the
21st Century. Am. Soc. Agric. Eng., St. J oseph, MI.
25 Tanida, H., and Y. Nagano. 1996. The ability of miniature pigs
to distinguish between people based on previous handling. Page
143 in Proc. 30th Int. Congr. Int. Soc. Appl. Ethol., Guelph, ON,
Canada. Ctr. Study Anim. Welfare, Guelph, ON, Canada.
26 van Reenen, C. G., J .T.N. van der Werf, and H. J . Blokhuis.
1996. Heart rate and behaviour of primiparous dairy cows
during milking. Page 130 in Proc. 30th Int. Congr. Int. Soc.
Appl. Ethol., Guelph, ON, Canada. Ctr. Study Anim. Welfare,
Guelph, ON, Canada.