Learning to suckle from an artificial teat within groups of lambs: Influence of a knowledgeable partner

75

Behavioural Processes, 30 (1993) 75-82 0 1993 Elsevier Science Publishers B.V. All rights reserved 0376.6357/93/$06,00

BEPROC 00484

Learning to suckle from an artificial within Influence

teat

groups of lambs:

of a knowledgeable

partner

I. Veissier and I. Stefanova ’ INRA, Centre de Clermont-Ferrand

(Accepted

- Theix, Saint-Genk

18 February

Champanelle,

France

1993)

Abstract This work was aimed at finding evidence of observational learning in sheep. The task to be learned was that of suckling milk from a bucket provided with teats. Lambs were reared in groups of either four neonates and an older lamb accustomed to artificial suckling (14 experimental lambs) or five neonates (I 6 controls). One-hour observations were made five times a day. The lambs that learned were grouped and observed for three hours once a week. All experimental lambs learned within three days as compared to nine for the 16 controls. Experimental lambs moved and sniffed or sucked the bucket more often than the controls (2.0 + 1.5 vs. 1.5 + 1 .O% time spent moving, 7.1 + 6.2 vs. 1.6 + 1.8 sniffing/hour, 3.7 + 3.6 vs. 1.7 f 1.7% time spent sucking, P < 0.05). The time to first suckling was not related to these measurements, neither was it related to other behavioural to suckle from the teat-bucket may be a socially transmitted phenomenon, enhancement of investigation.

Key words:

Sheep;

Observational

learning;

Artificial

suckling;

traits. Learning and not only by

Social behaviour

Introduction Observational enables a naive

learning means that observation animal to acquire this task more

of a demonstrator performing a task easily when it is exposed to the same

Correspondence to: I. Veissier, INCA, Centre de Clermont-Ferrand - Theix, 63122 Saint-Gen& Champanelle, France. ’ Present address: Institute of Zootechny and Veterinary Medicine, Zootechny Faculty, Department of Farm Animal Physiology, Stara Zagora, Bulgaria.

76 conditions

(Pallaud

cultural

traits

include

sweet

termite

catching

conspecifics

Lepoivre,

emerge

1985).

distinct

accelerates

1972;

Heyes

1986).

In contrast,

the

acquisition

Dawson,

farm

animals

pigeons

a demonstrator

helps

sensitive

The

horses Baker

normal

behaviour

reported living

(Lynch

(Thorhallsdottir

these

activity

or

(Craig,

1970,

learning,

were

more

activity The

reared

that

of

their

learn;

grouped

This

either

may interfere

factor

white

nor cattle

may be that with

their

are species-accustomed

to

and reduces 1991).

leads

that was

not

to artificial

learning

and consequently

from

suckling.

of others” repertoire

artificial

teats. They

The hypotheses

partner

et of an

of observationdl

would

the knowledgeable

from

grow

behavior

in the individual’s

by a knowledgeable with

beginning

evidence

to nurse

diet

food aversion

to Thorhallsdottir the

similar

to find

close

to determine

food conditions

which

to the

of five had to learn

social relationship

conditions

acquired

According “in

dimed

of a pattern

under

a knowledgeable

to be

learn

more

partner,

the more

partner

would

be

faster.

full

(control from

It was slightly

were visited a lamb

were

separated

from

their

that is 12 to 24 h after birth.

of animals

were inserted

lambs) older

had suckled

or

artificial

was maintdined in the

milk

groups.

milk.

Milk

at about

bucket.

neonates

teat. This

than the neonates

by caretakers

different

A

40 cm above the floor

fed ad libitum

or four

an artificial

from

teats were

of the milk

of hot water

to suckling

able partner. of whether

contact

by 10 cm. Lambs

neonates

dccustomed Animals

visual

a day. Temperature

bucket

lambs

dams

Thereafter

12 to 24 h without receiving milk. Five lambs were randomly x 1 .I 5 m pen. The pens were separated by 1.2 m high walls made

five teats in each pen. The

each other

five

X Ile-de-France to ingest colostrum,

for

prevented

with

least twice second

One

which

that eat a specific

facilitation,

individual

and that lambs

in suckling

to each 2.15

provided from

between

and Methods

were

hdy.

Alderks, to acquire

19861,

1992).

out in sheep

1987)

experiment

accompanied

after they had the opportunity assigned

Observa-

1985;

are of great importance

and Burritt, social

by one present

Animals and procedure Thirty (Romdnov X Limousin) they

partners

et al.,

or not by a lamb accustomed

they would

Materials

social

from

in groups

lambs

that the stronger efficient

models

Provenza

result

is the transmission

Lambs

were accompanied

quickly

Lefebvre,

for

from

and Levine,

for farm animals

rats and pigeons

Social

Chapple

199Ob;

p. 75).

that

1983;

effects

increased

previously.

rapidly;

1976).

to discriminate

situations

paper was carried

Being with

et al.,

et al.,

al. (1990a),

in this

environment.

in this species.

preference

tested

(Calef,

and Crawford,

in it (Veissier,

while

that learning

in rats (Zentall

and

Examples

and the use of tools

behaviours conditions

to explain

species.

life.

experiment

to their

1965)

(Palameta

to experimental

same

It is assumed

neither

use of social cues in these environments laboratory

1979).

et al., 1983;

panel of a box to get food

are more

put forward

of the

not seem to be sufficient

and black to get food in a bucket (Baer to push the front

et al.,

of adaptive

and

does

has been

(Kawai,

in experimental

1990)

observation

watching

process

populations

(McCrew

has been reproduced and

This

wild

in Japanese macaques

in chimpanzees

learning

tasks:

in

potato washing

tional

new

and

that

The

(experimental

but of similar

at 4h20,

9hl0,

14h00,

not was

determined

18h50

was

was replaced

30-40°C

groups

lamb was referred

bucket

and separated

were

lambs)

at

by using with

a lamb

to as the knowledge-

weight. and 23h35.

by touching

a

made of

Estimation

its abdomen.

This

77 provided

a rapid

assessment which

was essential

to remove

lambs that learned

and that

could act as knowledgeable partners for the other lambs of the same group. If the lamb had not suckled, it was bottle fed a maximum of 150 ml of milk. If the abdomen of a lamb looked distended, the lamb was considered to have suckled from the teat-bucket. Such a lamb was taken away from its group as was a lamb that was observed suckling from the teat-bucket. It could be used as a knowledgeable partner for other groups. A lamb that did not suckle from the teat-bucket within four days was removed from the experiment. To maintain five animals in each pen, other newborn lambs in order to replace those that learned to suckle from the were discarded because of their lack of learning. Six 2.15 contained five neonates, except for one control pen that

were introduced into the groups teat-bucket as well as those that x 1 .I 5 m pens were used, each contained six neonates and one

experimental pen, four neonates. A total of 16 control and 14 experimental lambs were used. As the groups of five lambs were not stable, neither the number of neonates accommodated in a pen nor the treatment sizes were multiples of four or five. Afterwards, all the lambs that learned to suckle from the teat-bucket were mixed together in a 4 x 8 m pen. There they were fed milk automatically, and water, hay and concentrates were available. Measurements During the learning period in the small pens, lambs were observed for an hour before the caretakers’ visits by a first experimenter. At 5 min intervals, the position of each lamb was recorded with reference to its general activity (‘resting’ with the chin on the ground, ‘lying’ in the head-up position, ‘standing’ without moving, ‘moving’, ‘sucking’ the bucket or a teat, ‘other’ activities such as social encounters or chewing straw). In addition, social behaviour and sniffing the bucket were recorded by a second experimenter continuously from 8hl5 to 9h15 and from 13hOO to 14hOO on the first five days of the experiment. Social behaviour was classified as follows: sniffing, sucking, aggressive encounters (butts and kicks), sexual encounters (mounting and ritual approaches) and bleating (see Orgeur, 1982, for descriptions). Direct observations where preferred to video records since the spatial relationships and the behavioural patterns could not be assessed with great precision on a videotape unless two cameras per pen were used. Lambs that learned to suckle from the teat-bucket were observed in the large pen for 3 h once a week for three weeks. Their general activities were observed every 5 min. The classes were similar to those described above with the addition of ‘eating’ hay or concentrates and ‘drinking’. Each lamb was weighed at birth, on the first day of suckling from the bucket (called day I> and then once a week for four weeks (days 8, 15, 22 and 29). Data analyses The time needed by a lamb to learn to suckle from the teat-bucket was expressed as the number of caretakers’ visits before the first suckling was detected (maximum = 20). The daily weight gains were estimated from birth to day 1 and then over the following four weeks. For each lamb, the percentage of time spent in any given general activity, the mean synchronization with other lambs, the mean distance from them, the mean frequency of social encounters and the frequency of sniffing the bucket were calculated. To assess the mean synchronization, the percentage of time the lamb is doing the same activity as that of another lamb was first calculated. Then the mean of the percentages obtained for that lamb was calculated. Since the lambs did not stay the same duration of time in the small

78

pens,

the behavioural

it learned

to suckle

Data

distributions

Student’s

were

analysed

using

Skewness

performed

when

the hypotheses

were

variances

used.

between

time

activities.

Fisher

could

Pearson

to

learn

or to

not be rejected,

Spearman

suckle

exact probabilities

(Snedecor

and Cochran,

the group

effect was included

ANOVAs

averaged over the whole

(df = (1,411

pens (n = 6). The

period

(i.e.

before

the teat-bucket).

paired tests were

of homogeneous tests

data of each lamb were from

1967).

and

Results

the

performing

to compare

section

will

number

the

to

the

and and

before

results

and Wilcoxon

assess

proportions

the

relation of

other

in each group

learning

analyses

on significant

ANOVAs distribution

frequencies

of pens for degrees

Mann-Whitney

focus

used

and

As far as the data obtained

tests.

Mann-Whitney

were

teat-bucket

calculated

by using only

Kurtosis

of a gaussian

otherwise

correlations

from were

and

are concerned,

of freedom

in the

on data averaged

by

(P < 0.05).

Results Experimental controls learned

lambs

and three within

learned

to suckle

from

of the latter did not learn

15 caretaker’s

visits,

the

(Fig.

teat-bucket

I).

In

more

quickly

the experimental

that is three days, while

in each control

out of five or six ldmbs

learned

within

three

days (Fisher

exact probability

pens

learned

within

three

days:

vs. O/3,

where

all lambs

caretaker’s and only Ten

visits, one,

either

none

whereas

most

experimental

the task (Fisher During

lambs

or several

learned

close

learned

lambs

learned

for 1 O/4

vs. 3/l

the neonates to the

bucket

3

were were

the

all lambs

pen, only three on proportion

P = 0.10). to suckle to suckle

of

Between

two

but seldom

one

one at a time.

one at a time vs. three out of the 13 controls

exact probability observed

lambs

of the experimental

the 1 -h observations,

the activities

control

3/O

than

pens,

that learned

< 0.05).

rarely

suckling

non-nutritive,

from that

the bucket. is lambs

Most of

sucked

the

neonates

(Ih

100 90

I

60..

70.. 60..

5

0 Fig. 1. Time controls)

3

no caretakers’

controls

12’13’14‘15 16.17 vis its (5 per day) experimental lambs

to learn to suckle from a teat-bucket in groups or four

neonates

and a lamb accustomed

to artificial

of lambs suckling

made

of five

(I 3 experimental

lambs).

79 TABLE

1

Behaviour and weight of lambs in six groups of five neonates (controls) accustomed to artificial suckling (experimental a gaussian distribution

or four neonates and a lamb

lambs). Results from ANOVAs

when the hypotheses of

and of homogeneous variances could not be rejected (df = (I ,4) and (I ,251

before and after learning respectively), and from Mann-Whitney

analyses in other cases (n = 6 groups

and 27 lambs). General activities, frequencies of sniffing and bleating and distances were assessed before learning. The animals suckled

General

activity

were

weighed

before

and after

learning.

The first day when

a lamb

to as dl

is referred

Control

Experimental

lambs

lambs

(n=16)

(n = 14)

Analyses

(% time)

Patterns of activity: resting chin

59.7

f

11.2

54.0

+

10.6

F = 5.26

15.8

*

9.2

14.7

i

7.8

F = 2.03

standing

4.1

*

3.1

4.8

+

5.3

u=4

moving

1.5

*

1.0

2.0

f

1.5

F=11.81*

sucking

1.7

f

1.7

3.7

f

3.6

u=o*

17.1

f

5.6

20.7

f

9.4

u=2

62

f

9

52

f

10

F = 9.39

21.8

+

21.2

41.0

f

58.8

u=2

lying in the head-up

position

other Synchronization Sniffing

between

and bleating

neonates

(per h)

bleating sniffing a neonate

3.21 +

2.06

7.62+

sniffing the bucket

1.6

1.8

7.1

0.1

0.66k

Distance

between

*

neonates

(m)

+

0.55+

Birth weight

(g)

2 640

Daily weight

gain (g)

(n = 13)(2)

k610

2 840

f

6.20

u=2

6.2

F = 20.7(l)

0.1

F = 6.03

f 500

*

F = 0.90

(n = 14)

from birth to dl

70

f

50

60

i

40

F = 0.05

from dl

to d8

60

f

50

240

*

50

F = 1.85

from d8 to dl5

260

*

70

280

i

80

F = 0.28

from dl5

to d22

310

*

50

240

*loo

from d22 to d29

380

f

80

360

f

70

F = 0.69

F = 5.20

*

*: P
after transformation

into square

roots.

(2) Three lambs that did not learn to suckle were discarded.

edge of the controls

bucket

on several

teats, were

partners

synchronised

lambs

was

(respectively correlations between

with

never

sniffing, their

them

drinking. more

than and over

those

were

the bucket

the experimental

sexual

partners.

done.

No

more

from

Therefore,

and were

Sucking

lambs

toward

I). not

between

very

seldom

comparisons

difference

no significant

the

and the

often (Table

lambs occurred

significant

of experimental In addition,

differed the bucket

each other.

encounters

h of observation). not

lambs

or sucking

in between

or

and synchronization

knowledgeable

than

were

the 10

moving

and sniffed

more

aggressive

Experimental

time

each other

not move

parameters

distance

spent

with

did

observed,

these

without

they

3 and 21 times on

and toward

a teat

less synchronised

Knowledgeable less

or

points:

was

and found

each other

correlations

were

80 found

between

able partner

the number

or any other

of caretakers’

After

all lambs

that had learned

suckling

was very

seldom

Thus

differences

be assessed.

in birth lambs

vs. 1.1 f

differences

weight

nor

(daily

1.4,

in growth

weight

and the behaviour

toward

the knowledge-

were

in the large pen,

F = 6.62,

during

a 3 h observation).

and control

lambs

more often than the controls

moved

df = (1,25), lambs

between

(Table

before

times

experimental

experimental

than controls

the teat-bucket

or two

between

rate except

gain from

from one

experimental

between

put on less weight

learning

(about

behaviour

the first week,

(% time = 3.0 k 2.3 Significant

to suckle

observed

in suckling

During

visits

activity.

could

not

P < 0.05). and controls

days

I). The

learning

lambs

were

15 and 22

lambs

detected

where

grew slower

vs. the following

before

week:

neither

experimental than after

t = 15.6,

df = 26,

P < 0.01).

Discussion In natural body

and

conditions,

licking

lamb.

teat-seeking

activity

teat-seeking

is elicited

base of the udder and the position are usually

and

gain

(Vince,

that have already lambs

were

suckled

a knowledgeable physical

learning

1988;

(Renner, and

Also

sucking was found

the

behaved

with

learning

in

probably Most

lambs the former partners.

been helped of the

lambs

may be needed.

suckle

at the

rapidly

in the group.

maturation.

same

Veissier,

Investigation

1992).

In support

According

and the time to learn to suckle

more

than

Moreover,

experimental

the

controls

measurement

is

unlikely learning

of the

a knowledgeable that

artificial

behavioural

once suckling

assumption,

one

from

learned

speed did from

of a quicker traits

of the

teat-bucket

the task

the same control

of the this

the

it.

was

has

partner.

may require

once

from

moving

and the latter

the tendency

partner

lambs

suckling

no

from

did not differ

Thus,

suckling

hypothesis,

learning

of other

of the knowledgeable

fact that several

since

in a help

However,

again,

partners

apart from

them.

a result

of this

lambs. but

partner

the knowledgeable

did in between

to this

involved

can, in turn,

the bucket

groups

Learning

it is unlikely

in

moved

suggests

of items

1992).

in

help learning.

enhanced

by observation

time

Thus, could

surface

In addition,

Le Neindre,

as they

without

lamb to seek

a warm

partner.

near the bucket

earlier

the knowledge-

were

Therefore,

The

the teat-bucket

for another

investigation

the teats

and

in any behavioural

experimental

knowledgeable

stimulus

a knowledgeable partner

can promote

sniffing

frequency.

since

for the lamb.

argue that when

in the effect of the knowledgeable

moving

the experimental

time

lambs

not involved

not vary with

bucket

warm

at the

manage by itself

difficult

from

less

In addition,

as in the case in this experiment.

nibbling

partner

between

experimental

probably

dams,

contact (cattle: Veissier

the

more

no effect of contact with

of the knowledgeable

task even without

relationship

their

never observed

that the mere presence

sniffing

observed

must

her

show

is higher

that for dam-suckling

could

an adequate

(1984)

1964). which

the animal

is probably

One

by orienting

restricted

Williams,

tend to start to suckle

was present.

it provided Vince

from

suckling

the lambs

and

are

and by warmth,

suckling,

different

to suckle

movements

(Alexander

an object

artificial

partner

after a teat. However,

Nevertheless,

weight with

its offspring

whose

For artificial

Hence,

able lamb was suckling,

newborn

less

1984).

experiment,

helps

of ewes

of its head is slightly

a knowledgeable

lambs

probably

Lambs

by contact

not vertical.

In the present when

the dam

the

lambs

learned,

a certain

although group

more

started

learning

to

spread

stage of physiological

stage is achieved,

the experimental

81

lambs could learn the task from the knowledgeable partner and, because the age of the lambs were not exactly the same, they did not learn together. Since learning without a knowledgeable partner requires a longer time, control lambs may have reached the required physiological stage before one of them started to suckle. Afterwards they might have learned from this lamb and thus nearly altogether. Hence, observational learning might have occurred also in groups of lambs without a knowledgeable partner. This would explain the lack of a very clear-cut difference between experimental and control groups. This hypothesis is highly speculative. Further experiments should be designed to assess precisely how learning is spread within a group of naive animals. The strength of the link between an experimental lamb and its partner did not influence the speed of learning. This lack of effect may be a result of age since young sheep do not establish clear-cut preferential relationships (Hafez, 1962, p. 333). Thus, social behaviour per se did not have an effect on learning. Finally, the lambs that learned to feed from a teat-bucket with a knowledgeable partner had a lower growth rate than the controls. This surprising result might have come from the fact that the three controls that did not learn the task were discarded or from a higher intake of dried food in controls. This study suggests that sheep are able to learn through observation. Mere group effects are unlikely to have brought about this result. Observational learning may have occurred in the present experiment because the animals were observed in their home environment without disturbance and small groups were used in contrast to previous experiments on farm species where pairs of animals were placed in an experimental chamber for a short length of time (see references mentioned in Introduction). In this study, animal comfort was maximized in order to exclude environmental sensitivity and maximize the information gathered.

Acknowledgements The authors are grateful to the animal staff for their efforts in designing the experimental devices and care of the lambs. They also thank A. Brelurut for his advice.

References Alderks,

C.E.,

1986.

Observational

percentage of reinforcement. Alexander,

G. and Williams,

Science, 146: Baer, K.L.,

D.,

Learn. Behav., 14: 331-335.

1964.

Maternal facilitation of sucking drive in newborn

lambs.

665-666.

Potter, CD.,

horses. Appl. Anim. Baker, A.E.M.

learning in the pigeon: effects of model’s rate of response and

Anim.

Friend,

T.H.

and Beaver, B.V.,

Ethol., 11: 123-l

and Crawford,

B.H.,

1983.

Observation

effects on learning in

29.

1986.

Observational

learning in horses. Appl. Anim.

Behav. Sci.,

15: 7-13. Chapple, R.S.,

Wodzicka-Tomaszewska,

M. and Lynch, J.J., 1987.

when introduced to wheat. II. Social transmission Appl. Anim. Craig J.V.,

Behav. Sci., 18: 163-l

1981.

Management. Wellington,

Domestic Prentice-Hall

364 pp.

Animal Inc.,

The learning behaviour of sheep

of wheat feeding and the role of the senses.

72. Behaviour:

Causes and Implications

London/Sydney/Toronto/New

for

Animal

Care and

Delhi/Tokyo/Singapore/

a2 Galef, B.G.Jr.,

1976.

Social transmission

learning in vertebrates. York/San

Francisco/London,

Hafez, E.S.E., E.S.E.

Cairns, R.B.,

Hafez (Editor),

of acquired behavior: A discussion

In: Advances in the Study of Behaviour pp. 77-l

Hulet,

of tradition

and social

6. Academic Press

Inc., New

00.

C.V. and Scott, J.P., 1969.

The Behaviour of Domestic Animals.

The behaviour of sheeps and goats. In: Bailliere, Tindall

& Cassell, London, pp.

2966348. Heyes,

C.M. and Dawson,

C.R.,

1990.

A demonstration

bidirectional control. Q.J. Exp. Psychol., Kawai, M., 1965.

Newly-acquired

Koshima islet. Primates, Lynch,

pre-cultural

J.J., Keogh, R.C.,

Edwin, R.L.,

Prod., 36: 175-I

McGrew, W.C.,

Tutin,

P., 1982.

learning in rats using a

behavior of the natural troop of Japanese monkeys on

Green,

C.C.

and Mottershead

B.E.,

1983.

Effects of early

acceptance of whole grain wheat by fine-wool

Merino

lambs.

83.

C.E.G. and Baldwin,

ural comparisons of Senegal, Tanzania, Orgeur,

of observational

59-71.

6: I-30.

experience on the post-weaning Anim.

42B:

P.J., 1979.

Chimpanzees,

tools, and termites:

cross-cult-

and Rio muni. Man, 14: 1855214.

OntogenPse du comportement sexuel m8le chez les ovins domestiques (Ovis Aries

L.) Effet de I’environnement

social. Thesis

Palameta, B. and Lefebvrc, L., 1985. what is learned? Z. Anim. Pallaud, B. and Lepoivre,

of the University

The social transmission

of Tours,

France, 108 pp.

of a food-finding technique in pigeons:

Behav., 33: 892-896.

H.,

1985.

Apprentissage d’inversion

spatiale par un groupe de babouins:

dans une situation de discrimination

interactions entre facteurs cognitifs et facteurs sociaux. Cah.

Psychol. Cogn., 5: 527-551. Provenza,

F.D.

and Burritt,

E.A.,

1991.

food aversion in lambs. Appl. Anim. Renner,

M.J., 1988.

Socially

induced diet preference ameliorates

Learning during exploration:

the role of behavioral topography during explo-

ration in determining subsequent adaptive behavior. Int. J. Comp. Psychol., Snedecor,

C.W.

and Cochran, W.G.,

Ames, 593 pp. Thorhallsdottir, A.C.,

Provenza,

F.D.

1967.

Statistical Methods.

and Balph, D.F.,

1990a.

The

A.C.,

Provenza, F.D. and Balph, D.F.,

aversions in sheep. Appl. Anim. Veissier,

I. and Le Neindre, I., Le Neindre,

Appl. Anim. Veissier,

Ability of lambs to learn about novel

Reactivity of Aubrac heifers exposed to a novel environment

P. and Trillat,

1984.

Observational

178:

T.R.

Behav. Sci., 33: 11-I

C., 1989.

5.

Weaning in calves: its effects on social organization.

learning in cattle. Appl. Anim.

Teat seeking or pre-sucking

maternal skin temperature. Zentall,

Behav. Sci., 25: 45-50.

199Ob. Social influences on conditionned food

Behav. Sci., 24: 43-54.

I., 1992.

Vince, M.A.,

Press,

Behav. Sci., 25: 25533.

P., 1992.

alone or in groups of four. Appl. Anim. Veissier,

2: 43-56.

Iowa State University

foods while observing or participating with social models. Appt. Anim. Thorhallsdottir,

conditioned

Behav. Sci., 31: 229-236.

Anim.

and Levine, J.M., 1972.

1220-1221.

Behav. Sci., 35: 235-243.

behaviour in newly-born

lambs:

possible effects of

Behav., 32: 249-254. Observational

learning and social facilitation in the rat. Science,