Livestock Production Science 65 (2000) 131–142
www.elsevier.com / locate / livprodsci
Effect of sward height and distance between pasture and barn
on cows’ visits to an automatic milking system and other
behaviour
a ,
*
a a aC.C. Ketelaar-de Lauwere
, A.H. Ipema , C. Lokhorst , J.H.M. Metz ,
b c a
J.P.T.M. Noordhuizen , W.G.P. Schouten , A.C. Smits
a
Institute of Agricultural and Environmental Engineering(IMAG), P.O. Box 43, NL-6700 AA Wageningen, The Netherlands
b
Department of Ruminants’ Health Care, University of Utrecht, P.O. Box 80151, NL-3508 TD Utrecht, The Netherlands
c
Department of Animal Sciences, Ethology Group, Wageningen Agricultural University, P.O. Box 338, NL-6700 AA Wageningen,
The Netherlands
Received 9 February 1999; received in revised form 24 August 1999; accepted 1 November 1999
Abstract
Two experiments were conducted to study the effects of sward height (experiment I) and the distance between barn and pasture plot (experiment II) on cows’ visits to an automatic milking system (AMS) and other behaviour. The study was carried out in a situation in which grazing was combined with fully automatic milking. The cows had to visit the system voluntarily without interference of the farmer. The AMS was continuously available in the barn adjacent to the pasture. The pasture was divided into ten plots of 1 ha for the two experiments and the preceding habituation. The same group of 24 Holstein Friesian cows was used in both experiments. The cows were allowed to spend up to 15 h at pasture daily. The results of experiment I (rotational grazing) showed that the total number of AMS visits increased from 4.4 to 7.3 and that the
number of milkings increased from 2.6 to 3.0 when sward height decreased (P,0.01). The results of experiment II (strip
grazing) showed that a distance of 360 m between pasture and barn did not affect the cows’ visits to the AMS. In both experiments, cows preferred to lie in the pasture, when available, rather than in the cubicles, and tended to visit the barn group-wise and to enter the AMS in close succession. As a consequence, in the period that the cows had access to the pasture
there were peaks in AMS visits, and the AMS was unvisited for several hours. 2000 Elsevier Science B.V. All rights
reserved.
Keywords: Automatic milking; Grazing; Sward height; Distance; Cow behaviour
without human interference in an automatic milking 1. Introduction
system (AMS) with a milking robot is rising steadily (Rossing et al., 1997). Many farmers considering The number of dairy farms where cows are milked
changing from milking in a milking parlour to fully automatic milking want to be able to allow the cows
*Corresponding author. Tel.:131-317-476-543; fax:1
31-317-to graze during the summer (De Boer et al., 1994),
425-670.
as this reduces feeding costs and labour and is often
E-mail address: [email protected] (C.C.
Ketelaar-de Lauwere) seen as a period of recovery for the cows (Hopster,
1996). It is known that the behaviour of cows at experiment started one month after the first. At the pasture is affected by many factors, such as seasonal beginning of this experiment one third lactation cow variation (Krohn et al., 1992; Rook and Huckle, had to be culled for reasons of udder health. 1996), either very warm (Muller et al., 1994; Mader The experimental facilities consisted of an open et al., 1997) or very cold weather (Dunn et al., 1988; barn with a lying area with 30 cubicles and a Prescott et al., 1994), sward height (Phillips and separate feeding area with 12 feeding places (Fig. 1; Hecheimi, 1989; Rook et al., 1994), concentrate for more details, see Ketelaar-de Lauwere et al.,
supplementation (Rook et al., 1994) and forage 1999). The barn was adjacent to a pasture of
supplementation (Phillips and Leaver, 1986; Phillips approximately 10 ha, which was divided into ten and Hecheimi, 1989). However, little is known about plots of 1 ha. All plots were connected to a 4-m wide factors affecting the behaviour of dairy cows in a central alley, leading to the barn. The plots differed situation in which grazing is combined with fully in distance to the barn, which was measured from the automatic milking. The novel element in such a centre of the plot to the entrance of the barn (Fig. 2). situation is that cows at pasture have to return to the The central alley of the plots led to a collecting area
2
barn by themselves several times per day, to be of approximately 48 m in front of the barn. To
milked in the AMS. It seems likely that the amount reach this collecting area, cows had to pass through a of grass in the pasture and the distance between the one-way gate. From this area, they could enter the barn (with the AMS) and the pasture may play a role barn. Cows wishing to leave the barn had to pass in a situation in which grazing and fully automatic through a selection gate with its entrance in the milking on a voluntary basis are combined. It can be feeding area. When a cow arrived at this selection expected that cows will return to the barn more often gate, she was identified by computer. If she had been
if the sward is short, and extra forage is sup- milked more than 6 h ago, the selection gate
plemented in the barn. Conversely, if there is plenty automatically moved in such a position that the cow of grass in the pasture, cows may be reluctant to was diverted back to the lying area, giving her a new return to the barn, so there will be fewer visits to the chance to visit the AMS. If she had been milked AMS. Distance between the barn and the pasture more recently, she could leave the barn via a passage may also influence frequency of AMS visits. This is along the barn to the central alley of the plots. Water a point of concern in practice because on many farms troughs were located in the barn, but not in the plots
the barn and the pasture are not adjacent. (Fig. 1).
In the experiments described in this paper, the Throughout both experiments, the cows were fed effects of the sward height and the distance between at a constant level. They daily received 440 kg maize the barn and the pasture on the cows’ visits to the silage on a dry matter basis in the barn. The forage AMS and associated behaviours were tested, in order was supplied at 08:30 and 20:15 h and raked up at to determine the feasibility of combining AMS with 13:00 h. Concentrate was dispensed during milking
grazing for both the farmer and the cow. in the automatic milking system (AMS). First
lacta-tion cows received 1 kg of concentrate, and multi-parous cows 1.5 kg per milking. The system
con-2. Material and methods sisted of two milking stalls and a robot arm equipped
with sensors for teat location. When a cow entered
2.1. Animals, housing and feeding the AMS, the computer decided on the basis of the
time that had elapsed since the previous milking
Two consecutive experiments were carried out whether she should be milked. If the previous
C
.C
.
Ketelaar
-de
Lauwere
et
al
.
/
Livestock
Production
Science
65
(2000
)
131
–
142
133
The AMS was cleaned twice a day, between 07:30 and 08:15 h and between 17:15 and 18:00 h. Cows with a milking interval of 12 h or longer, and cows that were not milked due to technical failure of the milking robot were brought to the AMS and milked before the morning cleaning and around 20:30 h, when all cows were fetched from pasture.
The barn was lit between 20:00 and 06:00 h.
2.2. Experimental design
In both experiments the cows had access to one of the plots between 05:30 and 20:30 h. In this period they could choose between being indoors or out-doors. They were kept indoors between 20:30 and 5:30 h. Each time a cow was milked in the AMS her milk yield was measured with a milk meter. In the first experiment, the average daily milk yield per cow ranged from 20.0 to 31.8 kg, and in the second experiment, it ranged from 23.1 to 28.6 kg. The amount of grass was assessed with a rising-plate
¨
meter (Gabriels and van den Berg, 1993) on the day previous to the first day of an experimental period and on the day following the last day of an ex-perimental period (For more details about the rising-plate meter, see Ketelaar-de Lauwere et al., 1999). Weather conditions were recorded in two different ways. The dry bulb temperature, the humidity and the black-globe temperature were continuously regis-tered in the barn and outdoors. In addition, it was registered once every ten min during behavioural observations at pasture whether it was raining and whether the sun was shining.
Fig. 2. Plan of the plots used in the rotational grazing experiment and the strip grazing experiment. H1, H2: plots for habituation; S1, S2, S3, S4: plots for rotational grazing experiments; D1, D2,
D3, D4: plots for strip grazing experiment. Distances are in meters 2.2.1. Experiment I: sward height (rotational
from the centre of the plot to the collecting area. grazing)
The effect of the sward height on the cows’ visits to the AMS and other behaviour was studied in this referred to as ‘non-milking visits’. Sometimes, cows experiment, which was divided into four periods of 4 that should have been milked were not, because the days. At the beginning of each period, cows were milking robot failed to attach the milking cluster; in given a fresh pasture. Grazing reduced the sward that case cows received some concentrate in the height, so after 4 days, cows were moved to another AMS until it was clear that the attachment had fresh pasture. The four pastures used in this experi-failed. These visits are referred to as ‘failed attach- ment differed as little as possible in distance to the
C.C. Ketelaar-de Lauwere et al. / Livestock Production Science 65 (2000) 131 –142 135
218 m, in the second period 288 m, in the third alley of the pastures and the collecting area, at the period 232 m and in the fourth period 296 m (plots entrance and exit of the AMS, in the passage S1, S2, S3, S4; Fig. 2). The sward height decreased between the feeding and the lying area, at the from 11.0 to 7.8 cm, from 10.8 to 7.2 cm and from selection gate and at the entrance of the central alley 12.3 to 7.3 cm in the first, second and third ex- of the pasture (Fig. 1). Every day the number and perimental period, respectively, and was 14.5 cm at duration of stays of cows in the lying and feeding the beginning of the fourth period (the decrease in areas, the AMS area and the pasture were calculated, that period was not measured because of problems using the data collected at these registration points. with electronic storage of data from the rising-plate
meter). Before the experiment started, the cows had 2.3.2. Direct behavioural observations
access to the central alley during 2 days and con- Behaviour was recorded in more detail by video or secutively to two ‘habituation pastures’ adjacent to by direct observations on the first and the fourth days the barn (plots H1, H2; in Fig. 2) for 235 days to of the second and the fourth experimental periods in
accustom them to the combination of grazing and the rotational grazing experiment and on the second AMS visits. The average black globe temperature day of each experimental period in the strip grazing between 10:00 and 17:00 h ranged from 15.2 to experiment. Direct behavioural observations started
32.18C over all periods. at 05:30 h and lasted 24 h. An observer monitored
the behaviour of the cows outdoors and indoors 2.2.2. Experiment II: distance between pasture and between 05:30 and 20:30 h, when the cows had
barn (strip grazing) access to the pasture. The behaviour was recorded by
In this experiment, which was divided into four video between 20:30 and 05:30 h, when the cows periods of 5 days, the effect of the distance between had to stay indoors. During the behavioural observa-the pasture and observa-the barn on observa-the cows’ visits to observa-the tions the following activities were recorded for each AMS and other behaviour was studied. Cows were cow once every ten min: lying in the cubicles or at given a fresh pasture at the beginning of each period, pasture, eating forage at the feeding gate or grazing, but the amount of fresh grass available was more or drinking, standing or walking in the lying, the
less the same on each day, because every day an collecting, the AMS or the feeding areas, and
electrified wire was moved, giving the cows access standing or walking in the central alley between the to a fresh strip of grass. The four pastures used in plots and the barn and at pasture. On both observa-this experiment differed in the distance to the barn. tion days at the ‘far’ plots in the strip grazing In the first period this distance was 146 m, in the experiment, a cow was in heat. Data on these cows second period 360 m, in the third period 355 m and from the days in question were discarded.
in the fourth period 168 m (plots D1, D2, D3, D4;
Fig. 2). The cows were strip grazed during 6 days in 2.4. Statistical analysis the ‘habituation pastures’ adjacent to the barn before
the experiment started (plots H1 and H2; Fig. 2). In general, differences in behaviour were assessed The average black globe temperature between 10:00 assuming that a generalised linear mixed model
and 17:00 h ranged from 17.1 to 31.88C over all (GLMM) can be used to describe the relation
periods. between differences in sward height (rotational
graz-ing) or distance (strip grazgraz-ing) and behavioural 2.3. Data collection and parameter choice response variables. The iterative reweighted residual maximum-likelihood algorithm (IRREML) (Engel
2.3.1. Indirect behavioural observations and Keen, 1994) was used to estimate fixed effects
into the model as a random effect. In the rotational variables resembles a binomial distribution (McCul-grazing experiment, the observation day was in- lagh and Nelder, 1989)
corporated into the model as a fixed effect to A log-linear model with random effects was used
estimate the effect of sward height because sward for the analysis of the distribution of the cows’ height decreased in the course of the observation activities over the barn and the pasture. In that days. In the strip-grazing experiment, the experimen- model, the total time an activity was performed tal period was incorporated into the model as a fixed outdoors was assumed to be proportional to the total effect to estimate the effect of the distance between time the activity was performed. In the rotational the pasture and the barn because this distance grazing experiment, the observation day was used as differed between experimental periods. The Wald an explanatory factor representing sward height, and statistic (VWS) was calculated to test for a 99% in the strip grazing experiment, the distance between probability level whether fixed effects were signifi- the pasture and the barn was used as an explanatory cant under the null hypothesis that there was no factor. In both experiments, the lactation number was effect. This statistic has an asymptotic chi-squared used as a covariable.
distribution (Genstat 5 Committee, 1993). Model To get an impression of possible weather effects, a selection took place by backward elimination, based relationship was calculated between the black globe
on this Wald statistic. temperature outdoors and the time the cows spent
In general, the model looked as follows: indoors between 10:00 and 17:00 h. It was assumed
that the standard deviation of the response variables increased proportionally to expected response values.
f(Y )5A 1B 1U 1e
ijkl ij k l ijkl The distribution of AMS visits over the 24-h
period was analysed using a log-linear model. The 24-h period was divided into eight periods of 3 h
Y 5observed value of response variable under ijkl
(05:30–08:30 h, 08:30–11:30 h, etc.). The number of consideration (for example the number of AMS
AMS visits in a certain period is assumed to be visits) of cow l (l51 . . . 24) with lactation number k
proportional to the total number of AMS visits. A (k51 or k$2) on day i (i51 . . . 4 in rotational
generalised linear model (GLM) was fitted instead of grazing experiment and i51 . . . 5 in strip grazing
a GLMM because differences between observation experiment) of experimental period j ( j51 . . . 4);
days (sward heights) were tested per cow over all
A 5effect of observation day i (sward height) on ij
experimental periods in the rotational grazing experi-experimental period j, or effect of experi-experimental
ment and differences between distances were tested period j (distance) on day i; B 5effect of lactation
k
per cow over all experimental days in the strip number k; U5random effect of cow l;e 5residual
l ijkl
grazing experiment (McCullagh and Nelder, 1989). error; f5link function — effects are assumed to be
It was checked whether time intervals between linear on link scale
cows arriving at the AMS followed a negative Response variables monitored by on-line
registra-exponential distribution, because this would indicate tions (AMS visits and duration of the cows’ sojourns
that the arrival times were randomly distributed in the pasture, and in the lying, feeding and AMS
(Parzen, 1962). areas) were analysed according to a linear model. For
The Genstat 5 statistical package was used for all AMS visits, it was assumed that the variance of each
analyses (Genstat 5 Committee, 1993). response variable is proportional to the expected
response value. This was confirmed by residual plots. In other cases, it was assumed that the standard
3. Results deviation of the response variables is proportional to
expected response values.
3.1. Sward height Monitoring the cows’ time budget during direct
behavioural observations resulted in binary response
3.1.1. AMS visits variables. These were analysed according to a
C.C. Ketelaar-de Lauwere et al. / Livestock Production Science 65 (2000) 131 –142 137
number of milkings, the number of non-milking 6.3 and from 5.1 to 6.3 when sward height decreased visits and the total number of AMS visits increased (S.E.51.0; VWS 510.6; P,0.05).
df53 as sward height decreased (Table 1). The number of
failed attachments was not affected by sward height.
The statistical analyses confirmed that on each 3.2. Distance between barn and pasture observation day, the time intervals between
consecu-tive AMS visits were not randomly distributed (P, 3.2.1. AMS visits
0.01). This means that the cows visited the AMS in The distance between the barn and the pasture did closer succession than could be expected on the basis not affect the total number of AMS visits, nor the number of milkings, non-milking visits or failed of an exponential factor alone
attachments. The predicted mean of the number of milkings was 2.8 in each experimental period
3.1.2. The cows’ time budget (S.E.D.50.1; VWS 50.2). The predicted means
df53
Table 2 gives an overview of the cows’ time of the total number of AMS visits ranged from 5.0 to
budget. The time in the AMS area and indoors 5.2 (S.E.D.50.2; VWS 51.8), those of the
num-df53
increased in the course of the experimental periods ber of non-milking visits from 2.6 to 2.8 (S.E.D.5
as sward height decreased and the time in the pasture 0.2; VWS 51.7) and those of the number of df53
decreased. The data of the direct behavioural data failed attachments from 1.2 to 1.7 (S.E.D.50.2;
showed that in the two experimental periods where VWS 510.0). It was confirmed by the statistical df53
these behavioural observations were carried out, the analyses that the time intervals between consecutive predicted means of the percentage of time that the AMS visits were not randomly distributed (P,0.01),
cows spent at the feeding gate increased from 5.4 to meaning that the cows visited the AMS in closer
Table 1
e
Predicted means of the number of AMS visits in rotational grazing
Day 1 Day 2 Day 3 Day 4 S.E.D. VWSdf53
a a b b
Milkings 2.6 2.7 2.9 3.0 0.1 24.1
a a b c
Non-milking visits 2.0 2.5 3.4 4.6 0.3 73.0
a a a a
Failed attachments 1.4 1.5 1.5 1.2 0.3 2.4
d a a b c
Total no. of AMS visits 4.4 4.6 5.9 7.3 0.3 115.3
a,b,c
Different superscripts in a row indicate a significant difference between days (P,0.01).
d
Excluding failed attachments because these reflect the ‘behaviour’ of the milking robot instead of that of the cows.
e
Sward height decreased from day 1 to 4; standard error of difference (S.E.D.) is mentioned as indication of deviation.
Table 2
d
Predicted means of the cows’ time budget in the rotational grazing experiment based on on-line computer registrations (434 24-h periods)
Time spent (min Day 1 day 2 day 3 day 4 S.E.D. VWSdf53
per cow per day):
a a b b
In lying area 500.7 512.5 563.5 553.3 14.7 26.0
In feeding area 152.1 147.6 167.9 164.6 11.3 4.5
a a b b
In AMS area 34.0 33.1 38.6 39.1 1.7 20.3
a b c c
Indoors 684.0 719.8 796.0 784.2 10.1 170.9
c b a a
Outdoors 755.2 719.8 643.7 653.8 10.7 147.2
a,b,c
Different superscripts in a row indicate a significant difference between days (P,0.01).
d
succession than could be expected on the basis of an gate were 5.1 when the cows were on a ‘far’ pasture
exponential factor alone. for the first time and 6.2 when they were on a ‘far’
pasture for the second time (S.E.51.0; VWS 5 df53
3.2.2. The cows’ time budget 11.9; P,0.01).
The distance between the barn and the pasture did
not affect the time in the lying area (Table 3). In the 3.3. Other results second experimental period, when the cows were on
a ‘far’ pasture for the first time, the cows spent less 3.3.1. Distribution of activities over the barn and time in the feeding area and indoors and more time the pasture
outdoors than in the other experimental periods, When given a choice between being indoors or
when the cows were on ‘near’ pastures or on a ‘far’ outdoors, cows spent most of their time outdoors. If pasture for the second time (Table 3). The time in they could choose between lying indoors or out-the AMS area was higher in out-the fourth experimental doors, they mostly lay in the pasture. If they had the period than in the other experimental periods (Table opportunity to eat forage indoors or outdoors, they
3). spent most of the time grazing. This was found in
The data of the direct behavioural observations both experiments.
showed that the predicted means of the percentages In the rotational grazing experiment (sward
of time spent lying were 41.1 and 39.7 on ‘far’ height), the cows spent between 84.7 and 86.3% of pastures and 47.5 and 44.3 on ‘near’ pastures (S.E.5 the 15 h that they could be outdoors at pasture when
1.2; VWS 556.5); the predicted means of the they were on a fresh pasture (i.e. on first days of df53
percentages of time spent lying at pasture were 25.9 experimental periods), and between 68.2 and 73.4% and 26.5 on ‘far’ pastures and 32.6 and 31.1 on of these 15 h when they were on these pastures for ‘near’ pastures (S.E.51.6; VWS 527.0); and the the last time (i.e. on fourth days of experimental
df53
predicted means of the percentages of time spent on periods) (predicted means). The cows appeared to the central alley between the pasture and the barn spend a significantly bigger part of the time that they were 7.3 and 5.5 on ‘far’ pastures and 1.6 and 2.5 on could be outdoors at pasture at higher sward heights
‘near’ pastures (S.E.51.2; VWS 5125.1). The (S.E.D.51.0; VWS 5122.5; P,0.01). The same
df53 df53
‘far’ pastures differed significantly from the ‘near’ significant effect of sward height was found for the pastures in all these cases (P,0.01). The predicted time spent grazing: the cows spent between 93.5 and
mean of the percentage of time spent grazing was 95.5% of the time that they were eating forage 45.5 when the cows were on a ‘far’ pasture for the between 5:30 and 20:30 h grazing when they were first time and 38.8 when they were on a ‘far’ pasture on a fresh pasture, and between 86.7 and 87.7% of for the second time (S.E.51.1; VWS 541.0; P, this time when they were on these pastures for the
df53
0.01). These figures for the time spent at the feeding last time (predicted means) (S.E.D.51.0; VWS 5 df53
Table 3
c
Predicted means of the cows’ time budget in the strip grazing experiment based on on-line computer registrations (435 24-h periods)
Time spent (min per Period 1 Period 2 Period 3 period 4 S.E.D. VWSdf53
cow per day): (146 m) (360 m) (355 m) (168 m)
In lying area 519.2 532.7 559.7 541.5 14.3 8.3
b a b b
In feeding area 154.5 126.7 140.7 150.7 7.0 21.0
a a a b
In AMS area 37.4 38.0 40.8 47.9 2.4 20.7
b a b b
Indoors 729.2 700.5 738.5 738.0 11.3 15.6
a b a a
Outdoors 709.2 741.3 699.5 700.7 11.4 16.8
a,b
Different superscripts in a row indicate a significant difference between experimental periods (P,0.01).
c
C.C. Ketelaar-de Lauwere et al. / Livestock Production Science 65 (2000) 131 –142 139
70.5; P,0.01). The cows spent between 71.3 and indoors by the cows between 10:00 and 17:00 h
90.8% of the time that they were lying between (temperature effect50.4247; S.E.50.1; P,0.01).
05:30 and 20:30 h at pasture (predicted means). This The average black globe temperature between 10:00
was not related to sward height. and 17:00 h on these days ranged from 18.9 to
In the strip grazing experiment (distance), the 32.18C. No such effects were found on the other days
cows spent between 79.5 and 86.0% of the time that of the rotational grazing experiment and in the strip they could be outdoors at pasture, between 85.9 and grazing experiment.
91.2% of the time that they were eating forage On one of the observation days of the rotational between 05:30 and 20:30 h grazing and between 91.1 grazing experiment lying time was much shorter than and 96.3% of the time that they were lying between on all other observation days (including those of the 05:30 and 20:30 h at pasture (predicted means). strip grazing experiment). This day was the rainiest Although some significance were found between of all observation days: it rained for 61.1% of the these figures, they could not be related to the time that the cows could be outdoors.
distance between the pasture and the barn.
3.3.2. Time intervals between cows entering the 4. Discussion
barn
In both experiments cows tended to enter the barn 4.1. Sward height in close succession. In the rotational grazing
experi-ment, 83.3% (2.7) of the time intervals between The rotational grazing experiment showed that the cows entering the barn were shorter than 6 min, and sward height affects the cows’ visits to an AMS if 6.7% (0.8) of these time intervals ranged from 6 to fully automatic milking on a voluntary basis is 16 min (predicted means with S.E. in brackets). In combined with grazing. At lower sward heights, the strip grazing experiment, 78.4% (2.4) of the time cows paid more visits to the AMS, and were milked intervals between cows entering the barn were more often. This means that a farmer will have to shorter than 6 min, and 8.2% (0.8) of these time pay more attention to the daily milking frequency of intervals ranged from 6 to 16 min (predicted means his cows when he has offered them a fresh pasture,
and S.E. in brackets). because in such a situation they probably will be
more reluctant to return to the barn. According to Phillips (1993), there is an increase in grazing 3.3.3. Distribution of AMS visits over the 24-h
intensity when a fresh pasture is offered to cattle.
period
This would also account for the low percentage of In the rotational grazing experiment, cows
distrib-AMS visits between 08:30 and 11:30 h found on the uted their AMS visits differently over the 24-h
first days of experimental periods in our experiment, period at different sward heights. The cows had a
and the fact that cows spent more of their time eating low percentage of AMS visits between 08:30 and
forage at pasture on the first days of experimental 11:30 h and between 14:30 and 17:30 h, especially
periods than on fourth days. However, total time on the first days of the experimental periods, when
spent grazing did not differ between different sward they had been given a fresh pasture. In the strip
heights, which is in agreement with findings of grazing experiment, the distance between the pasture
Phillips and Hecheimi (1989). Those authors never-and the barn had no effect on the distribution of
theless did find higher biting rates at lower herbage AMS visits over the 24-h period.
heights, which is reported also by Phillips and Leaver (1986) and Rook et al. (1994). Unfortunately, we did not measure bite size in our experiment, but 3.3.4. Weather
we did find an indication that cows spent more time On the third days of the rotational grazing
experi-at the feeding gexperi-ate experi-at low herbage heights; this, too, ment, a positive relationship was found between the
(1989), who found that cows spent more time eating lying times found in our experiment were in the
silage at low herbage heights. range of those reported by others (Krohn and
Munk-Lying times were not affected by sward heights, sgaard, 1993; Singh et al., 1993). In our experiment but if cows had access to the pasture they clearly it was striking that when the cows were on a ‘far’ preferred lying in the pasture rather than in the pasture for the first time, they spent less time in the cubicles. This confirmed earlier findings (Ketelaar-de feeding area and indoors and more time outdoors and
Lauwere et al., 1998). grazing than when they were on ‘near’ pastures or on
The decreasing sward height was probably not the a ‘far’ pasture for the second time. This may be a only factor causing the extra time spent indoors (and matter of habituation. Probably, it is a bit more probably related to this the extra AMS visits) on difficult to adapt to the AMS environment when it is third days of the rotational grazing experiment. High combined with a ‘far’ pasture (Broom, 1996), and temperatures on these days may also have been therefore more time is needed for habituation.
¨ influential. Similar results were found by Krotzl and Hauser (1997) and Ketelaar-de Lauwere et al.
4.3. Time intervals between cows entering the barn (1998). On fourth days of the rotational grazing
experiment, the temperature did not contribute to the
In both experiments, cows returned to the barn extra time spent indoors (and probably to the extra
several times in the period that they had access to the AMS visits), because no relationship was found
pasture. They always did this more quickly after one between the time spent indoors and the black globe
another than could be expected on the basis of an temperature outdoors on these days.
exponential factor alone. This phenomenon has been Another factor which may affect the cows’
be-described previously in comparable situations haviour at pasture is rain. The shortest lying time
(Ketelaar-de Lauwere et al., 1999). Connected with was found on the rainiest day of both experiments.
the group-wise entering of the barn, the cows also Phillips (1993) has stated that cattle are reluctant to
reported to the AMS quickly after one another. For a lie on wet grass.
farmer who combines grazing with fully automatic milking in an AMS, this means that in the period that 4.2. Distance
cows have access to the pasture there will be peaks in AMS visits and the AMS will be unvisited for The cows’ visits to the AMS in this study were not
several hours. Thus ‘group arrival’ of cows from affected by the distance between the pastures and the
pasture will negatively affect the throughflow of the barn. However, some indirect effects of the distance
AMS. This group behaviour is a result of the natural were found. Cows on ‘far’ pastures spent less time
habit of cows to act as a co-ordinated social unit lying on the pasture and more time standing or
(Hurnik, 1994). Farmers should be aware of this walking in the central alley between the pastures and
phenomenon when deciding about what capacity the barn. In addition, it appeared that if they had
AMS to install. access to the pasture cows preferred lying in the
pasture rather than in cubicles, irrespective of the distance to the barn; this is similar to the findings of
the experiment about sward height. On both days of 5. Conclusions direct behavioural observations on the ‘far’ pastures
C.C. Ketelaar-de Lauwere et al. / Livestock Production Science 65 (2000) 131 –142 141 than six months; in Dutch with English summary, IVO Report
Distances between barn and pasture up to a
B-405, ID-DLO, Zeist, p. 152.
maximum of 360 m do not affect the number of
Hurnik, J.F., 1994. Ethological approach to the management of
AMS visits. It is not known what would happen at dairy parlours. In: Bucklin, R. (Ed.), Proc. 3rd. Int. Dairy
longer distances. Housing Conf, Dairy Systems for the 21st Century, ASAE,
Michigan, pp. 43–48.
Irrespective of the sward height and the distance
Ketelaar-de Lauwere, C.C., Ipema, A.H., Metz, J.H.M., 1998. Free
between the barn and the pasture, the cows preferred
access to forage and grazing as conditions for the application of
lying in the pasture rather than in the cubicles, and fully automatic milking. In: Veissier, I., Boissy, A. (Eds.), Proc. tended to enter the barn group-wise and, as a 32nd. Congress of the International Society for Applied
Ethology, Chaumeil Repro, Clermont Ferrand, p. 59.
consequence, to visit the AMS in close succession.
Ketelaar-de Lauwere, C.C., Ipema, A.H., van Ouwerkerk, E.N.J.,
The latter means that in the periods that cows have
Hendriks, M.M.W.B., Metz, J.H.M., Noordhuizen, J.P.T.M.,
access to the pasture there will be peaks in AMS Schouten, W.G.P., 1999. Voluntary automatic milking in combi-visits, and also periods when the AMS will be nation with grazing of dairy cows; milking frequency and unused for several hours. This means that the AMS effects on behaviour. Appl. Anim. Behav. Sci. 64, 91–109.
Krohn, C.C., Munksgaard, L., 1993. Behaviour of dairy cows kept
capacity will not be used optimally; this should be
in extensive (loose housing / pasture) or intensive (tie stall)
taken into account when calculating what capacity of
environments. II. Lying and lying-down behaviour. Appl.
the AMS is required. Anim. Behav. Sci. 37, 1–16.
Krohn, C.C., Munksgaard, L., Jonasen, B., 1992. Behaviour of dairy cows kept in extensive (loose housing / pasture) or intensive (tie stall) environments. I. Experimental procedure, facilities, time budgets – diurnal and seasonal conditions. Appl.
Acknowledgements
Anim. Behav. Sci. 34, 37–47. ¨
Krotzl, H., Hauser, R., 1997. Ethologische Grundlagen zum
Dr M.M.W.B Hendriks (Centre for Biometry, Platzbedarf, zur Gestaltung und zum Betrieb von Laufhofen bei¨ Wageningen) is acknowledged for her useful com- Kuhen im Laufstall [Exercise yard for cows kept in a cubicle¨
ments on earlier drafts of this paper. house – ethological bases concerning size, design and manage-ment; in German; with English summary]. Agrartechnische Forschung 3 (2), 141–150.
Mader, T.L., Fell, L.R., McPhee, M.J., 1997. Behavior response of non-Brahman cattle to shade in commercial feedlots. In:
References Bottcher, R.W., Hoff, S.J. (Eds.), Livestock Environment V, Vol.
II, ASAE, Michigan, pp. 795–802.
Albright, J.L., Arave, C.W. (Eds.), 1997. The Behaviour of Cattle, McCullagh, P., Nelder, J.A. (Eds.), 1989. Generalized Linear CAB International, Oxford, p. 306. Models, Chapman and Hall, London, p. 511.
Broom, D.M., 1996. Animal welfare defined in terms of attempts Muller, C.J.C., Botha, J.A., Smith, W.A., Coetzer, W.A., 1994. to cope with the environment. Acta Agric. Scand. Sect. A, Production, physiological and behavioural responses of lactat-Anim. Sci. Suppl. 27, 22–28. ing Friesian cows to a shade structure in a temperate climate. De Boer, P.B., Metz, J.H.M., de Pater-Huijsen, F.L. (Eds.), 1994. In: Bucklin, R. (Ed.), Proc. 3rd. Int. Dairy Housing Conf, Dairy Volautomatische melksystemen [Automatic milking systems; in Systems for the 21st Century, ASAE, Michigan, pp. 579–588. Dutch, Program Technology Assessment (TA), Vol. 4, Ministry Parzen, E. (Ed.), 1962. Stochastic Processes, Holden-Day, San of Agriculture, Nature and Fisheries, Den Haag, p. 90. Francisco, p. 324.
Dunn, R.W., Havstad, K.M., Ayers, E.L., 1988. Grazing behavior Phillips, C.J.C. (Ed.), 1993. Cattle Behaviour, Farming Press, responses of rangeland beef cows to winter ambient tempera- Ipswich, p. 212.
tures and age. Appl. Anim. Behav. Sci. 21, 201–207. Phillips, C.J.C., Leaver, J.D., 1986. The effect of forage supple-Engel, B., Keen, A., 1994. A simple approach for the analysis of mentation on the behaviour of grazing dairy cows. Appl. Anim.
generalized linear mixed models. Statistica Neerlandica 48, Behav. Sci. 16, 233–247.
1–22. Phillips, C.J.C., Hecheimi, K., 1989. The effect of forage supple-¨
Gabriels, P.C.J., van den Berg, J.V., 1993. Calibration of two mentation, herbage height and season on the ingestive be-techniques for estimating herbage mass. Grass Forage Sci. 48, haviour of dairy cows. Appl. Anim. Behav. Sci. 24, 203–216. 329–335. Prescott, M.L., Havstad, K.M., Olson-Rutz, K.M., Ayers, E.L., Genstat 5 Committee, 1993. In: Genstat 5 Reference Manual: Petersen, M.K., 1994. Grazing behavior of free-ranging beef Version 3, Clarendon Press, Oxford, p. 796. cows to initial and prolonged exposure to fluctuating thermal Hopster, H. (Ed.), 1996. Effecten van huisvesting en verzorging environments. Appl. Anim. Behav. Sci. 39, 103–113.
Rook, A.J., Huckle, C.A., Penning, P.D., 1994. Effects of sward Singh, S.S., Ward, W.R., Lautenbach, K., Hughes, J.W., Murray, height and concentrate supplementation on the ingestive be- R.D., 1993. Behaviour of first lactation and adult dairy cows haviour of spring-calving dairy cows grazing grass-clover while housed and at pasture and its relationship with sole swards. Appl. Anim. Behav. Sci. 40, 101–112. lesions. Vet. Rec. 6, 469–474.
