:35! ‘

3.3! .1 .21.}: «9.

ti.” .. .13.

.213...- 3....

1.9!. . 3......
.3 :c

3.0,:
r

.3.
E. 4:

12:1!
y’rCIi .
.11.!

.u¢ . .

K . .

. ‘ . .3! I

'4 ‘ . vt-u w I! ?
zazi

n.
t.
_ 3..
2(. 71 can. ‘ r\.-l.- )9. ‘1’
. . 2‘s... . . ‘ . u
323.31 . .5 32?, «an»; 1.3. A
Itt’la .IKI'X . V .— ($-
31. 51:52:31 ,
.. 33.5.: his;
,. .:i;.:........i.. .. E
2. is 5.2.1..

.3?

. .3. Lexy...
.5

x : If...
e 3.25:7:
k. 3:2}.

Li!

1.
r210...2z. .3:
x. 17.313134

r. 2‘ ‘ p
, fit: .39“.

. a ;. Ex.

gfiefi?

A1:

 

:50 » lIBRARY '
N Michigan State 3
University

uh

 

 

This is to certify that the

thesis entitled

Effect of early environment on stress-coping
strategies of foals.

Presented by
Christel Moons =
J
has been accepted towards fulfillment
of the requirements for

MS. degree in Animal Science

7 l
f ;
4 V , i

{ Major professor .

 

Date (Bl/[[3401

MS U is an Aflirmative Action/Equal Opportunity Institution ‘
r

0-7639

 

EFFECT OF EARLY ENVIRONMENT ON STRESS-COPING STRATEGIES OF
FOALS

By

Christel Moons

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Animal Science

2001

ABSTRACT
Effect of early environment on stress-coping strategies of foals
By

Christel Moons

The mare plays the most important role in rearing a foal and the permanent separation of
mare and foal during weaning is stressful for both. In rodent literature, it has been shown
that high levels of maternal care can facilitate coping with a stressor during early stages
in life. We tested whether repeated 10-min separations of mare and foal would lead to an
immediate and prolonged increase in maternal care. Furthermore, we assessed the effect
of these short-term separations on weaning stress using behavioral and physiological
markers. Short-term separations caused an immediate increase in tactile and non-tactile
interactions between mare and foal (P < 0.05), but we could not identify these effects
after 4 to 6 days. In addition, the immediate increase in maternal care by repeated
separations did not reduce the response to weaning as measured by heart rate, salivary
cortisol, and behavioral indicators. We did find that fillies walked longer, lay down less,
and vocalized more than colts (P < 0.05) and that the separations complicated the
weaning process for fillies more than for colts as indicated by the difference in frequency
of defecating (P < 0.05). These data show that, despite an immediate increase in maternal

care, repeated short-term separations did not appear to facilitate weaning.

In loving memory of my grandparents

Palmyre Moons-Devijver (° July 15, 1921 - 1‘ February 16, 2000)
and
Victor Moons (° April 14, 1915 — 1' March 13, 2001)

Through your unconditional love for us and for each other,
You showed me there is a time to fight and a time to let go.
I will never forget...

iii

ACKNOWLEDGEMENTS

College of Agriculture and Natural Resources, Department of Animal Science, and Dr.
Hogberg, for giving me the opportunity to perform research abroad and to be part of a
dynamic group here at Michigan State University. Fulbright Association, for financial
support. Jackie Christie and the staff of Room 103 International Center for monitoring
administration regarding my international and academic status. Dr. Adroaldo Zanella, for
the chance he gave me to enrich my life with knowledge and for insisting I live up to my
potential. My Graduate Guidance Committee, for the interest they showed from the start.
Thanks for seeing it all the way through with me! MSU Merillat Equine Center and MSU
Horse Research and Teaching Center, management and staff, for accommodating the
research and for the extra work they took on. Jennifer McGuire, Susy Turcott, Kathy
Brody, Magali de Moura, Jennifer Wells, Rebecca Chester, Cheryl Leece, Adriana
Silveira de Souza, Emilly Wilson, Kristen Seifferlein, and Camie Heleski, for their help
during my project. Jill Davidson, Dr. Renate Snider, and Dr. Kirsty Laughlin for the
advice on thesis, presentation, and other issues. Rebecca Chester, Kerrie Beckett, Amy
Shelle, Heidi Doza and Miyuki Tauchi, for their friendship and support inside and outside
the lab these past two years! Theresa Doerr for being the mom away from home, always
making sure I tied my shoes... Thanks to the graduate students, for making it easier when
times were rough. A special word of thanks to Diane, Helena, and to my office mates,
Laurie, Marci, David, for being the kind of fi'iends you only meet once in a lifetime, for
taking me as I am (Disney obsession included). And finally, to my family: you let me
explore new horizons, and trusted that I would always find my way home. I could never

repay you the love and support you’ve given me.

iv

TABLE OF CONTENTS

LIST OF TABLES

.......................................................................................................

....................................................................................................

CHAPTER I: Introduction
CHAPTER II: Literature review

Maternal care

Development of mare-foal bond afier parturition ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
Elements of maternal care in horses ___________________________________________________
Proximity __________________________________________________________________________________
Nursing ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
Definition ______________________________________________________________________
Function _______________________________________________________________________
Nursing behavior _________________________________________________________
Nursing bout frequency and duration ________________________
Bout termination __________________________________________________________
Aggression ________________________________________________________________________________
Other interactions _____________________________________________________________________
Play _______________________________________________________________________________
Allogrooming _______________________________________________________________

Communication

............................................................

Factors aflecting maternal care _________________________________________________________

Stress response

Stress ______________________________________________________________________________________________________

Behavioral response to stress ______________________________________________________________

Physiological response to stress __________________________________________________________
Anatomy of Hypothalamus-Pituitary—Adrenal gland axis _____
HPA-axis hormonal regulation ________________________________________________
Effects of maternal care on the HPA-axis ______________________________
Salivary cortisol ________________________________________________________________________
Heart rate

...................................................................................

Response to psychosocial stress __________________________________________________________

ix

16
18
20
20
21
23
26
29
31

Weaning

Introduction ___________________________________________________________________________________________ 33
Weaning methods __________________________________________________________________________________ 35
Abrupt versus gradual weaning _______________________________________________ 35
Stall versus pasture weaning ____________________________________________________ 35
Weaning foals individually or in pairs ____________________________________ 36

CHAPTER IV: Immediate and prolonged effects of short-term separations on
mare-foal interactions

Abstract 39

Introduction 40

-----------------------------------------------------------------------------------------------------

Materials and methods

Animals __________________________________________________________________________________________________ 43
Housing and feeding _____________________________________________________________________________ 43
Video recording and data collection ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 44
Short-term separations _________________________________________________________________________ 46
Evaluations of interactions initiated by foal and
mare 4-6 days after separation ______________________ 48
Results
Foal behavior during short-term separations _____________________________________ 50
Proximity before and after short-term separations ____________________________ 54
Interactions displayed by the foal before and after separation _________ 56
Interactions displayed by the more before and after separation _______ 58
Proximity during evaluations 4-6 days post-separation ____________________ 60
Evaluations of interactions initiated by the
foal 4-6 days post-separation _____________ 61
Evaluations of interactions initiated by the
mare 4-6 days post-separation _________ 63
Discussion and conclusion 65

..............................................................................

vi

CHAPTER IV: Effect of short-term separations on weaning stress in foals

Abstract

.............................................................................................................

.....................................................................................................

Materials and methods
Animals and treatment prior to weaning _____________________________________________
Housing and feeding _____________________________________________________________________________
Weaning protocol __________________________________________________________________________________
Salivary cortisol ____________________________________________________________________________________
Sampling ___________________________________________________________________________________
Analysis ____________________________________________________________________________________
Heart rate ..............................................................................................
Techniques ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
Analysis ____________________________________________________________________________________
Behavior ________________________________________________________________________________________________

Techniques ________________________________________________________________________________
Analysis ____________________________________________________________________________________

Results
Salivary cortisol ____________________________________________________________________________________
Heart rate

..............................................................................................
.................................................................................

........................................................................

Standing ____________________________________________________________________________________
Walking _____________________________________________________________________________________
Lying down _______________________________________________________________________________
Vocalizing _________________________________________________________________________________
Defecating _________________________________________________________________________________
Urinating ___________________________________________________________________________________
Eating and drinking __________________________________________________________________
Looking forward _______________________________________________________________________
Investigating _____________________________________________________________________________
Pawing _______________________________________________________________________________________

Discussion and conclusion
Salivary cortisol ____________________________________________________________________________________
Heart rate
Behavior

................................................................................................

..............................................................................................

CHAPTER V: General discussion

vii

73

74

78
78
79
80
80
82
83
83
84
85
85
86

86
89
91
91
91
91
93
96
97
98
99
101
103
104

106
107
108

APPENDIX A: Overview of nursing bout definitions ___________________________________________ 119

APPENDIX B: Foal age at first sampling ________________________________________________________________ 121
APPENDIX C: Dam, date of birth, gender and sire of foals used in the study at

MSU Merillat Equine Center _____________________________________________________________________________________ 123
APPENDIX D: Merillat observation schedule ________________________________________________________ 125
APPENDIX E: Ethogram used for post-separation observations _______________________ 130
APPENDIX F: Behaviors observed during weaning _____________________________________________ 135
APPENDIX C: Protocol during weaning ________________________________________________________________ 141
LITERATURE CITED 144

...............................................................................................

viii

LIST OF TABLES

Table 1. Frequency i SEM (per 10 min) for foals of “eating”, “pawing” and
“investigating” per foal age. __________________________________________________________________________________________ 53

ix

LIST OF FIGURES

Figure 1. Scheme of part of MSU Merillat Equine Center Barn A, not drawn
to scale. 45

...........................................................................................................................

Figure 2. Scheme of part of MSU Merillat Equine Center Barn B, not drawn

to scale. ___________________________________________________________________________________________________________________________ 45
Figure 3. Mean frequency : SEM of investigating by foals during lO-min separation for
subsequent trials. ____________________________________________________________________________________________________________ 51

Figure 4. Mean frequency i SEM of looking forward by foals during lO-min separation
for subsequent trials. 51
Figure 5. Mean frequency :t SEM of vocalizing by foals during lO-min separation for
subsequent trials. ____________________________________________________________________________________________________________ 52

Figure 6. Mean proximity score :1: SEM between mares and foals for pre- and post-
separation observation periods. _____________________________________________________________________________________ 55

Figure 7. Mean frequency i SEM of foals watching mare during 30 min for pre- and
post—separation observation periods. ____________________________________________________________________________ 57

Figure 8. Mean frequency i SEM of contact initiated by foals during 30 min for pre- and
post-separation observation periods. ________________________________________________________________ 57

Figure 9. Mean frequency i SEM of contact behaviors displayed by the mare towards the
foal during pre- and post-separation observation periods. ______________________________ 59

Figure 10. Mean frequency i SEM of watching foal during 30 min displayed by the mare
during pre- and post-separation observation periods. ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 59

Figure 11. Mean proximity scores 1- SEM between mares and foals for both genders
during post-separation evaluations. ______________________________________________________________________________ 60

Figure 12. Mean frequency i SEM of contact initiated by the foal per observation block
during post-separation evaluations. ______________________________________________________________________________ 62

Figure 13. Mean frequency i SEM of watching foal displayed by mares, per trial, during
post-separation evaluation. ____________________________________________________________________________________________ 63

Figure 14. Mean frequency i SEM of contact directed to the foal by the mare per
observation period during post-separation evaluations. ______________________________________________ 64

Figure 15. Mean frequency 2t SEM of aggressions directed towards the foal by the mare

per observation block during post-separation evaluations. _________________________________________ 64
Figure 16. Saliva collection device to collect saliva from foals. ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 81
Figure 17. Saliva collection device in place in a foal’s mouth. __________________________________ 81

Figure 18. Mean salivary cortisol concentration ’5 SEM per sampling time for each
weaning trial day. ___________________________________________________________________________________________________________ 87

Figure 19. Mean heart rate i SEM per observation block for each weaning
trial day. ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 90

Figure 20. Mean duration i SEM of walking per weaning trial day. _________________________ 92

Figure 21. Mean duration i SEM of walking per observation block for
each gender. ____________________________________________________________________________________________________________________ 93

Figure 22. Mean duration i SEM of lying down per weaning trial day for each treatment
group. ______________________________________________________________________________________________________________________________ 94

Figure 23. Mean duration i SEM of lying down per weaning trial day for
each gender. ____________________________________________________________________________________________________________________ 95

Figure 24. Mean frequency i SEM of vocalizations per time block for each weaning trial
day. _________________________________________________________________________________________________________________________________ 97

Figure 25. Mean frequency i SEM of defecation per observation block for each weaning
trial day. __________________________________________________________________________________________________________________________ 98

Figure 26. Mean frequency i SEM of eating and drinking for each weaning
trial day. __________________________________________________________________________________________________________________________ 100

Figure 27. Mean frequency 3: SEM of eating and drinking per observation
block. ______________________________________________________________________________________________________________________________ 100

Figure 28. Mean frequency i SEM of drinking for each gender. ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 101

Figure 29. Mean frequency i SEM of looking forward per gender for each treatment
group. ______________________________________________________________________________________________________________________________ 102

Figure 30. Mean frequency i SEM of looking forward per observation block for each
weaning trial day. ___________________________________________________________________________________________________________ 103

xi

Figure 31. Mean frequency i SEM of investigating per observation block for each
weaning trial day. ___________________________________________________________________________________________________________ 104

Figure 32. Mean frequency i SEM of pawing per observation block for each weaning
trial day. __________________________________________________________________________________________________________________________ 105

xii

CHAPTER I: GENERAL INTRODUCTION AND OBJECTIVES

Rearing foals to strong, healthy, and competitive animals is the goal of all commercial
horse farms. Loss of foals, or foals showing behavioral problems and physical blemishes,
results in an economic cost, since much was invested in breeding a quality mare to a
quality stallion, in carrying out foaling, and in providing the optimal environment for the
foal to develop. Many factors influence the development of foals, but the most important
influence in a foal’s neonatal life is the dam. She provides the foal with initial immunity,
protection and nourishment throughout the first months afier birth. The bond between
mare and foal is strong and studies have shown that separations of short duration of mare

and foal are stress-causing events.

Under feral conditions, mares do not wean their foals until they are about to give
birth again. However, in managed horse farms, the sooner a foal can be weaned, the
sooner both mare and foal can be prepared for routine activities. Weaning results in
maternal deprivation, ofien social isolation, and a sudden transition in the source of
nutrients. It’s been shown using behavioral and physiological indicators that weaning is
stressful for mares and foals. Because the stress of weaning puts foals at risk for
decreased immunity, reduced growth, and injury, researchers have investigated ways to
reduce the response to stress, i.e., facilitate weaning. Proposed mechanisms include
paired—stall versus individual-stall weaning, gradual versus abrupt weaning, and group-

pasture versus individual-stall weaning.

Another potential mechanism to reduce stress at weaning is to make use of the
natural bond that exists between mare and foal. In rodents, it has been shown that
intensive early maternal care can help rat pups in adulthood adapt better to stressors by
modulating the hypothalamus-pituitary-adrenal gland axis. Certain horse experts argued
that maternal care in horses can be intensified by short-term separations or by a threat to
the foal. If separations of short duration indeed lead to more intensive maternal care, this

mechanism may aid foals in coping with stressors such as weaning later on in life.

To assess stress, scientists have used markers such as concentrations of fecal
corticosteroid metabolites and urinary, plasma, serum, and salivary cortisol
concentrations. Serum and plasma cortisol concentrations to date are the most commonly
used markers, but collection of the samples involves invasive techniques that in
themselves are stressful for the foals. On the other hand, salivary cortisol measurements
are minimally invasive and are highly correlated to serum and plasma cortisol

concentrations in different species, including horses.

The overall hypothesis for this Master’s project was that increased maternal care early

in life will facilitate weaning of foals.

Following hypotheses were tested during the study:

1. Short-term separations cause an immediate increase in mare-foal interactions

and maternal care. (Chapter III)

2. Foals that have undergone repeated short-term separations from their dam
regain homeostasis faster after weaning than foals that have never been

separated fi'om their mother. (Chapter IV)

The objectives of this Master’s project were to:
1. Behaviorally assess the immediate and prolonged effects on mare-foal
interactions of short-term separations and to evaluate the response of foals to

subsequent exposures of isolation. (Chapter III)

2. Investigate the behavioral and physiological responses of foals to weaning
stress and compare the responses of foals exposed to early short-term
separations with foals that had never been separated from their dam. (Chapter

IV)

Chapters from this thesis have been or will be submitted to scientific journals. All

chapters were formatted according to the guidelines of Animal Behaviour.

CHAPTER II: LITERATURE REVIEW

Maternal Care

Development of mare-foal band after parturition

Horses and other ungulates are precocial mammals, meaning that the foal can sustain its
own therrnoregulation, has an adequate sense of smell and vision when it is born, and
develops its sensory and motor systems much faster compared to altricial mammals such
as rodents (Kendrick et a1. 1997). Nonetheless, the dam provides the foal with vital
nourishment, protection and initial colostral antibodies. The neonatal foal depends on
maternal care to survive the first months after birth. Maternal behavior can be stimulated
in the dam by separation or a threat to the foal (Crowell-Davis & Houpt 1986). The
relationship between mare and foal changes over time as the young animal gains

nutritional and psychological independence.

Less than half an hour after birth, mare and foal begin to show interest in one
another. Initially the standing foal follows any large moving object, but McCall (1991)
observed that afier the second hour postpartum, the foal is less attracted by any random,
large, moving object and becomes more attentive to its dam. As the mare recovers from
parturition, she licks the foal and fetal membranes. The purpose of this interaction is not
so much to clean the foal, but rather to establish a sense of familiarity by gaining
olfactory information (Beaver 1981). After licking, the dam appears able to distinguish
her foal fiom others and formation of the mare-foal bond has been initiated. Visual and
olfactory cues seem to be the main factors involved in the bonding process. Crowell-

Davis and Houpt (1986) reviewed that primary socialization appears to be established

within 2 h of parturition. The strong selective individual bond is of significance when
mare and foal are out on pasture with other horses because it ensures that the foal only
nurses with its own dam and that the mare protects her foal from other herd members
(Kendrick et a1. 1997). Separation of mare and foal following the bonding period results

in signs of disorientation in the foal and extreme distress in the mare.

The role of sensory systems in mare-foal social recognition was investigated
using a variety of scenarios (Wolksi et a1. 1980). Foal and mare were separated for 10-20
min and then allowed to choose between the relative and an alien horse. Vocal cues
appeared to be very important when members of a mare-foal dyad were attempting to
locate one another. However, in the tests both mares and foals responded more to the
animal with the highest vocalization rate, whether that was the relative or not (Wolksi et
a1. 1980). Finally, overt olfactory behaviors, such as sniffing, were noted only when mare
and foal were in close proximity. After parturition mare and foal are in close proximity

and it is mainly visual and olfactory cues that initiate the mare-foal bonding process.

Elements of maternal care in horses

Different aspects of maternal care in horses show that mare and foal form a strong bond
with one another, which is then disrupted at the time of weaning. The aspects of maternal
care are proximity, nursing behavior, aggression, play, mutual grooming, and

communication.

Pro unity

A close proximity, the distance between the mare and her foal, is characteristic of the
Equidae (Crowell-Davis & Houpt 1986). Mares also show a recumbency response
(Barber & Crowell—Davis 1994) and this supports the assumption that close proximity
reflects the vigilance of the mare towards the foal. The recumbency response is defined
as follows: “when her foal lies down, the mare stops engaging in any other activity and
stays near her young, by instinct on alert for predators, even though there may be no such
dangers anywhere near the property” (Anonymous 1994). Whether the mare stands close
to her foal as described in the recumbency response or lies down next to it, there is very
little physical contact between them (Crowell-Davis 1985). As the foal grows older,

resting time, nursing time, and proximity decrease and plateau at about 24 weeks of age.

Foals in general are never more than 5 m apart from their mother during the first
day after birth (Barber & Crowell-Davis 1994; Crowell-Davis 1985). Barber and
Crowell-Davis (1994) studied maternal behavior of 10 Belgian mares on pasture. They
reported that, during the first day post-partum, mare and upright foal were within 1 m of
each other for at least 95% of the time. In contrast to what we could expect, given the
recumbency response, which should draw the mare closer to her foal, this percentage was
lower for recumbent foals, namely 85%. Thus, mares did not spend as much time in close
proximity with foals that laid down compared to foals that were upright. This rather
surprising comparison possibly results from a change in behavior such as that observed
by Houpt (1992). When the foal laid down, the mare either stood over or grazed around it

in circles. When the foal was upright, the mare grazed more in a straight line and the foal

followed the mare, causing the distance to be shorter than when the foal was lying down.
The distinction in behavior of the mare toward upright and recumbent foals remained
constant throughout the development of the foal. For upright foals, the proximity to the
mare gradually decreased with age. The distance between recumbent foals and their dams
increased with age, but from 3 weeks after birth until approximately 21 weeks, proximity

remained constant (Houpt 1992).

The spatial relationship between mare and foal is mainly due to the mare’s
behavior when the foal is recumbent and due to mare as well as foal behavior when the
foal is standing. Smith-Funk and Crowell-Davis (1992) reported that, during the first days
after birth, foals laid down regardless of whether the mare was upright or recumbent.
With increasing age, foals mimicked the resting pattern of the mare, often walking
towards her and lying down beside her. While standing, a foal usually follows the mare

as she grazes, thereby maintaining close proximity.

Nursi g

Definition
A considerable amount of variation exists between nursing bout definitions across
scientific publications. Throughout consecutive studies more elements have been
included, such as pro-suck nuzzle, pre-suck pause, intra-suck nuzzle, and intra-suck pause
(Barber & Crowell-Davis 1994). Appendix A shows an overview of how nursing bout

definitions developed over 10 years.

Function
Nursing serves multiple functions: to comfort the foal (Crowell-Davis 1985), to nourish
it, and, especially during the first day of life, to passively transfer immunity from mother
to offspring. Antibodies (Immunoglobulins, Ig) are concentrated in the mare’s milk
during the last week of pregnancy (N aylor & Bell 1985). The initial Ig concentrations in
colostrum, however, drop to 15% between 4 and 8 h afler the first nursing (N aylor & Bell
1985). In order to maximize the transfer of passive immunity from mare to foal, it is
therefore important that the foal nurses as soon as possible after birth, preferably within

8h.

In the normal nursing position, the foal stands in reverse parallel next to the mare
(McCall 1991). Crowell-Davis (1985) reported that the first nursing bout takes place
within 30 to 120 min after birth. Very often, a foal bunts (bobs the head up and down
against the mare’s udder) before or during nursing. “The firnction of this bunting is to
stimulate release of oxytocin from the neurohypophysis, which in turn controls milk
ejection in the udder” (Carson & Wood-Gush 1983a). Successful nursing is dependent on
the foal’s ability to find the udder and on the mare’s willingness to stand still during the
initial nursing attempts. Because of the inexperience of primiparous mares, or because
mares with tender udders experience pain during the first nursing, foals of these dams
may take longer to nurse successfully after birth (McCall 1991). Nursing is also an
expression of the bond that has been established between mare and foal. As the foal ages,

it becomes less dependent on the dam’s milk (Carson and Wood-Gush 1983a) and joins

the mare in eating more solid food such as hay and pellets (Crowell-Davis 1985), but it

will continue to nurse until weaning.

Nursing behavior
In general, the foal initiates a nursing bout (Anonymous 1994). When the mare does
initiate, she first ceases other activities. She then either watches the foal and vocalizes to
it to come closer, or she approaches it herself. If the foal is immediately next to the mare,
it positions itself in a reverse parallel orientation next to the mare and places its muzzle
medial to the mare’s pelvic limb. When approaching from a distance, the foal initiates
nursing by approaching the mare and extending head and neck while bobbing the head up
and down. Another indicator of nursing intention is “crossing the bow”. The foal
approaches its mother, crosses under her neck and then moves caudally down to the
udder of the mare (Crowell-Davis 1985). This ‘crossing the bow’ actively blocks the
forward movement of the mare, therefore facilitating nursing (Barber & Crowell-Davis

1994)

Crowell-Davis (1985) reported that if Welsh pony mares continued walking after
their foals initiated nursing, the foals either aggressed against them, continued circling
after ‘crossing the bow’, ceased attempting to nurse, or nursed anyway. These types of
reactions to a resisting mare were also reported for Thoroughbred foals (Carson & Wood-
Gush 1983a). When a mare prevented her young from nursing, the foal repeated the

pattern of ‘crossing the bow’ to make the dam stand still. If unsuccessful in its efforts to

nurse, the foal also pawed the ground and occasionally kicked. If a dam was lying down,

sometimes the foal pawed near the udder until the mare stood up.

Nursing bout frequency and duration
Different studies report variation in nursing bout fi'equency. For Belgian horses, Barber
and Crowell-Davis (1994) reported an average of seven successfirl nursing bouts per hour
during the first day after birth. The rate declined by 50% during the first month. The
duration of nursing bouts ranged from 78 to 84 sec. For draft mares and their mule foals,
the nursing rate on day one was approximately 14 times per hour (Smith-Funk &
Crowell-Davis 1992). Nursing frequency decreased steadily, from a maximum of eight
bouts per hour after the first week to two bouts per hour during the seventeenth week

afier birth. Both successfirl and unsuccessful bouts followed the overall trend.

Bout termination
Crowell-Davis (1986) reported that the duration of nursing bouts varies with individual
foals and remains constant for a given foal regardless of age. Bout terminations by the
mare were very infrequent during the first day after birth in Belgian draft horses (Barber
& Crowell-Davis 1994). This may be a mechanism to insure sufficient ingestion of Ig-
rich colostrum. Barber and Crowell-Davis (1994) recorded that mares terminated bouts
less often than foals. When a mare terminated a bout, she usually walked away; overt

aggression was rare.

10

Agggession

Aggression of a mare towards her foal arises only rarely during the foal’s first
month, and is almost always nursing-related (Anonymous 1994). The mare’s udder may
be sore from being full and initial nursing may cause discomfort. Upon nursing, the
pressure in the udder decreases and the mare allows the foal to suckle. Normally a foal is
not harmed by the mother’s aggression (Crowell-Davis & Houpt 1986) and aggression is

potentially a way to regulate expression of the mare—foal bond.

In addition, at any time when a foal bunts the udder too hard or bites the teat, the
mare may intervene with some form of aggression. Crowell-Davis (1985) recorded eight
different forms of aggression during nursing: laying back of ears, threat to bite, bite,
threat to kick, kick, squeal, hunt, and tail swish. Smith-Funk and Crowell-Davis (1992)
added a ninth category: smacking, which was described as “ears being laid back, the head
turned toward the foal and the mare’s mouth opening abruptly, resulting in a loud
smacking sound”. Barber and Crowell-Davis (1994) suggested that peak aggression
occurs at a time during the foal’s development when a nutritional transition to grain and
forage is about to occur. Maternal aggression possibly encourages that change by
reinforcing grazing through resistance to nursing attempts by the foal, potentially leading

up to a natural process of weaning.

Foals exhibit a variety of reactions to maternal aggression (Crowell-Davis 1985).

The most common one is no discemable response. The second most likely reaction is a

short pause while removing the muzzle from the udder, followed by continued nursing.

11

The third most common reaction is the foal aggressing back at the mare, kicking her or
threatening to kick. For Belgian draft horses, no discemable response was exhibited by
foals in 66% of all aggressive incidents (Barber & Crowell-Davis 1994). Reciprocating
aggression (always a kick in response to a bite) towards the mare occurred in 14% of the
aggressions. No response by the foal was recorded in 100% of the cases of laying back
the ears by the mare and in 93% of the threats to bite. Welsh pony mares aggressed
against their foal during nursing, and this trend increased gradually until the foal was

about 13 to 16 weeks old and it leveled off until weaning (Crowell-Davis 1985).

Other interactions

Play
The first kind of play exhibited by the foal is solitary play, when the foal explores its
environment for example, by mouthing parts of the stall, or the mare’s tail, ears and
halter. Playfirl behavior also includes running and bucking near the mother (Anonymous
1994). When mare and foal are on pasture, playing can also be galloping away from the
mare in what appears to be thrill-seeking behavior (Houpt 1992). During the second and
third month of a foal’s life, a foal kept on pasture with other foals learns to play with herd
mates (Anonymous 1994). Three major types of play are exhibited: 1) general motor play
involving running/bucking, 2) interactive play, and 3) object manipulation. Colts seem to
engage more in interactive play whereas fillies do more running and bucking (Crowell-

Davis 1986).

12

Allogrooming
Allogrooming or mutual grooming is defined as reciprocal coat care in which the partners
nuzzle and chew on one another’s coats (Boyd & Houpt 1994). In herds it is suggested
that grooming actually occurs according to a dominance hierarchy, where the lower
ranking animal may or may not be denied access to a mutual grooming session (Moons
1999). Foals mutual-groom preferentially with other foals rather than with their mothers
when together in a herd (Crowell-Davis 1986). Fillies groom about twice as often as colts
do. In addition, allogrooming with other foals peaks at about the third month of life, and
then steadily declines (Crowell-Davis 1986). However, within a stall setting where mare
and foal are by themselves, allogrooming can provide an important indication of the

mare-foal attachment.

Communication
Communication can be divided into three major types: auditory, visual, and olfactory.
Olfactory communication, which occurs through pheromones, has the advantage that it
works even when the signaler is not present. Visual cues are used in particular for
communication over relatively short distances, while auditory cues are effective over
long and short distances (Houpt 1992). When a foal is out a mare’s field of vision, the
mare looks around to find the foal and communicates with the foal vocally. The vocal
repertoire consists of the whinny (or neigh), snort, nicker, and squeal. The whinny is a
typical separation call when horses are out of sight of one another. This rarely happens
under managed conditions as mare and foal are either kept in a stall or on a pasture where

they have visual contact with one another at all times. The snort can be an attempt to

13

clear the nostrils, but it can also indicate increased anxiety in a restrained or frightened
horse. The squeal is heard when agonistic interactions occur, and it is often difficult to

tell if the aggressor or victim is vocalizing (Houpt 1992).

Factors affecting maternal care

As demonstrated previously through the different components of maternal care, mare and
foal form a close attachment. This attachment and more specifically maternal care is
affected by several factors: breed, parity, individual variation, environment and foal age.
Dwyer and Lawrence (1998) showed that when comparing maternal care in two breeds of
sheep, Scottish Blackface and Suffolk ewes, the Scottish Blackface breed showed higher
levels of maternal care as measured in time spent grooming lambs and cooperation during
sucking attempts. In another study, Dwyer and Lawrence (2000) showed that parity had
an effect on the occurrence of rejection behaviors (withdrawal, aggression, lack of
cooperation with lamb sucking attempts). Primiparous ewes for both Scottish Blackface
and Suffolk showed more rejection behaviors towards their lambs compared to ewes that
had had at least one lamb prior to the study. Dwyer and Lawrence (2000) also found that,
regardless of parity and breed, individual ewes were consistent in their expression of
maternal care across parities. Barber and Crowell-Davis (1994) reported casual
observations done on Belgian draft horses where individual variation in maternal care

among mares occurred regardless of parity state.

A third factor that influences maternal care is the neonatal environment and

manipulation thereof. Liu at al. (1997) showed that when rat pups were handled during

14

the first 10 days after birth, the dams showed an increase in maternal care as measured by
the occurrence of licking and grooming. The dams of handled rat pups showed 155 i 21
occurrences of licking and grooming as opposed to dams of non-handled pups who
displayed 78 i 25 occurrences. In horses, Crowell-Davis and Houpt (1986) argued that if
the environment of the foal is manipulated through a threat (such as the presence of a

dog) or short-term separations fiom the dam, maternal care is stimulated.

Finally, age is another factor that influences maternal care. First of all, as the foal
grows older, it spends less time in close proximity with the darn (Tyler 1972). While
observing New Forest ponies, Tyler (1972) found that during the first week of life, the
foal spent 94 % of its time within 5 m of the mare. By week 20, this percentage decreased
to 52 % and by week 32, only 20% of the foal’s time was spent within 5 m of the dam.
Nursing-related aggression displayed by Welsh pony mares increased fi'om
approximately 0.1 i 0.03 aggressions per nursing bout during the first month to 0.3 i
0.06 aggressions during the sixth month after birth (Crowell-Davis 1985). The
association between increase in age and nursing-related aggressions is supported by the
decrease in the amount of time a foal spends nursing (Carson and Wood-Gush 1983a).
During the first week of life, Thoroughbred foals nursed for a total of 710 i 50 sec per
hour, whereas by the 24th week, total duration of nursing time per hour was 80 i 15 sec.
In addition to the decrease of time spent nursing, aging foals start to adapt to the mate’s
diet by increasing ingestion of available grass, hay, and concentrate (Crowell-Davis

1985). At 1 week of age, foals were feeding (nursing excluded) 8 i 2 % of their time

15

whereas by week 21, this percentage had increased to 46 j: 4 % of the total time budget.

Thus, as the foal ages, there are changes in the mare-foal relationship and maternal care.

Stress Response

Stress

To assess all the events occurring during a stress response, we need to clearly define
“stress”. Stressors can be physical (e. g. injury or pain), psychological (e. g. fear) and/or
psychosocial (e. g. modification of social interactions) in nature (Mason 1975). Different
definitions of stress have been developed as the science of animal behavior has evolved.
Most of the definitions for stress rely on the physiological changes within an animal with
less emphasis on changes in behavior as part of the stress response (Friend 1991). In
1956, stress was defined as “the nonspecific response of the body to any demand” (Selye
1956). In addition, the author used the term “eustress” for a beneficial or pleasant type of
stress and applied “distress” as the more serious, harmful type of stress. Later, stress was
defined as “an environmental effect on an individual which overtaxes its control systems
and reduces its fitness or appears likely to do so” (Broom & Johnson 1993). More
recently, Moberg (2000) defined stress as “the biological response elicited when an
individual perceives a threat to its homeostasis”. Homeostasis according to Broom and
Johnson (1993) is the “relatively steady state of a body variable, which is maintained by
means of physiological or behavioral regulation”. Maintaining homeostasis in body fluids
and tissues is essential for an animal’s survival. Whether the threat is real or not is of no
importance according to this definition. An animal ‘perceiving’ a certain stimulus as

threatening to its homeostasis will show a stress response.

16

Similar to the different types of stressors, there are different types of responses to
stress. Moberg (2000) describes four different kinds of responses to stress and they all
shift available energy and resources away from maintenance processes to ensure the
animal is prepared for “fight or flight”. The behavioral response is biologically the most
economical response. The animal can attempt to escape the threat simply by moving
away fiom the stressor. The second most important line of defense is the autonomic
nervous system response. Here, a number of biological systems are affected during stress,
resulting in changes in heart rate, blood pressure and gastrointestinal activity. The
biological effects are of relatively short duration. This is not the case for the third type of
response, the neuroendocrine response, which is explained below in the section
“Physiogical response to stress”. The neuroendocrine response is responsible for
profoundly altering biological functions, potentially impairing the welfare of an animal.
The fourth response is the immune response, resulting in depression of lymphocytes and
other defense mechanisms in the immune system. However, it is debatable that reduction

of immune competence during stress is a mechanism of defense.

All of the above-mentioned defense mechanisms constitute coping strategies
when an animal is faced with a stressor. Coping is defined as “having control of mental
and bodily stability” (Fraser & Broom 1990). More specifically, in his review of factors
affecting physiological coping systems during development, Plotsky (2000) defines
coping more extensively as the “utilization of available strategies and resources to
successfully reduce the impact of a stressor or to meet daily environmental challenges”.

Moberg (2000) explained why sometimes stress is harmless and why it sometimes leads

l7

to serious consequences. An animal possesses a certain amount of resources to perform
all the essential biological functions. The risk of stress is “the shift in biological functions
away from non-stress functions or activities in order for the animal to cope with the
stressor”. When an animal is exposed to a prolonged stressor, too many resources are
shifted away from the non-stress functions. It is possible that the impairment of those

biological functions poses a threat to an animal’s well being.

Behavioral response to stress

The behavioral response to stress can be assessed by the systematic techniques of
behavioral observation. Behaviors displayed by the animals of interest are recorded on
check sheets or computer files and quantified according to intensity, frequencies and
duration. Variations in these measurements can indicate a disturbance in the animal

(Broom and Johnson, 1993).

The natural behavior of animals that are not used to close human contact can be
confounded by the presence of the observer (Martin & Bateson 1999). Even if the animal
does not shy from humans, its behavior may still be altered in more subtle yet significant
ways. In order to obtain reliable data, the animals need to be desensitized to human
presence, or the observer should resort to the use of cameras, one-way screens or
camouflaged hides. Most of these solutions, however, still allow the animal to detect
olfactory or auditory cues indicating the presence of the observer. Critical evaluation of

the confounding factors of the study should not be neglected when reporting on results.

18

Behavioral responses to stress can be categorized according to the duration of the
stressor(s). Investigators have described acute and chronic stress very precisely (Moberg
1985). When an animal is exposed to one stressor for a relatively short period of time,
this is usually referred to as acute stress. Chronic stress results from consecutive acute

stressors, regardless of whether those stressors are identical.

Behavioral responses to acute stress occur when an animal experiences short-term
pain or fear, or when it is fi'ustrated (Friend 1991). Acute pain or fear can be recognized
through observation of the animal’s posture. The animal makes itself look smaller by
crouching and it may even sweat or tremble. Frustrated animals that are prevented from
performing certain behaviors may substitute the prohibited behavior for another. Pawing,
vocalizing, defecating and overall increase in activity are other behavioral responses to

acute stress (Hoffman 1995).

When responding to chronic stress, animals may initially show fierce resistance to
the stressors imposed on them as if responding to acute stress, but after a certain period of
time, the animals often adapt. Adaptation may indicate that the animals responded so
profoundly to the acute stress that the resources to fight for prolonged periods of time are
depleted. When animals cannot avoid the chronic stress, such as repeated exposure to
electric shock (Seligrnan & Maier 1967), learned helplessness can set in. In this case, the
drive to avoid the stimulus decreases to the point where the animal is apathetic. More
commonly, the animals show an intensification of drive-motivated behaviors as a

response to chronic stress. For example, a study on calves showed that animals, deprived

l9

of space through confinement, were more motivated to show locomotor behavior during
an open field test (Dellrneier et al. 1985). Similarly, when comparing distance traveled in
an open field test between horses on pasture, horses stalled with tactile access to
neighbors, and horses stalled in total isolation, researchers found that stalled horses
traveled the firrthest when compared to horses housed on pasture (Mal et al. 1989).
Another way of coping with stress occurs in the form of stereotypies, which are invariant
and repetitive behaviors that seemingly have no function (Nicol 1999). Stereotypies can
be considered displacement behaviors, but they result from chronic instead of acute

stress. They arise in captive animals faced with unsolvable problems (Mason 1991).

Physiological response to stress

Aflomy of hypothalamus-pituitarkagenal gland axis

The physiological response to stress is mediated through the hypothalamus-pituitary-
adrenal gland axis (HPA-axis). Because we wish to give a complete overview of the
HPA—axis and its hormonal regulation, this section and the next summarize the anatomy
and hormonal regulation as reported in a book chapter written by Matteri et al. (2000).
Because it is important to understand the underlying mechanisms of a stress response, we
feel summarizing the review by Matteri et al. (2000) is justified. The hypothalamus is a
bilaterally symmetric region on the floor of the brain, above the anterior pituitary gland.
From top to bottom, the hypothalamus extends from the floor of the third ventricle to the
median eminence, where the pituitary stalk is formed. From the hypothalamus, there are
two kinds of neurons traveling towards the pituitary. The axon terminals of the

parvicellular neurons release hormones into a capillary network, the hypophyseal portal

20

vasculature, that traverses the pituitary stalk from median eminence to anterior pituitary
gland. The hormones released by these axons regulate the function of the anterior
pituitary gland. The second type of neurons consists of magnocellular neurons and they
extend directly through the pituitary stalk into the posterior pituitary gland. The neurons

produce vasopressin and oxytocin, which are released from the posterior pituitary gland.

The pituitary gland consists of three regions. The anterior lobe secretes growth
hormone (GH), adrenocorticotropic hormone (ACTH), thyroid stimulating hormone
(TSH), luteinizing hormone (LH), follicle stimulating hormone (F SH), prolactin, and B-
endorphin. The posterior lobe of the pituitary gland acts as a storage site for oxytocin and
vasopressin (VP, also known as antidiuretic hormone). The third region is the

intermediate lobe and its function is species specific.

The adrenal gland consists of two regions: an inner part, medulla, and a
surrounding capsule, the cortex. The medulla releases catecholamines (epinephrine,
norepinephrine, and dopamine) into the general bloodstream. The cortex consists of three
layers: the zona glomerulosa (secretes mineralocorticoids such as aldosterone), the zona
reticularis, and the zona fasciculata. The latter two layers are responsible for production

of androgens and glucocorticoids (GC, 6. g. cortisol).

HPA-a_xis hormonal regulation

Activation of the HPA-axis can occur through stressors, circadian drives, and humoral

influences (Plotsky 2000). In what follows, we summarize a review by Matteri et al.

21

(2000) for the same reasons as stated in the previous paragraph. During a stress response,
the sympathetic nervous system is activated and stimulates catecholamines production
from the adrenal medulla. The catecholamines influence the HPA axis at many
concentrations, including stimulation of neurohorrnone release from the hypothalamus

and ACTH release from the pituitary gland.

In addition, corticotropin-releasing hormone (CRH) is released by the
hypothalamus and this hormone reaches the pituitary through local transport.
Corticotropin-releasing hormone and VP work independently as well as in concert to
cause ACTH release from the pituitary. Two other factors have been reported to induce
ACTH secretion: epinephrine and oxytocin. In the pituitary, ACTH is released into the
general bloodstream that reaches the adrenal gland. Consequently, the adrenal cortex

produces GC that travel to various regions in the body through the general circulation.

Glucocorticoids play a role in the process of gluconeogenesis: they stimulate the
liver to convert fat and protein to intermediate metabolites. These metabolites are then
transformed to glucose, a source of energy. In addition, GC'have a potentiating effect on
the synthesis of epinephrine. It should be noted that having constant, moderate
concentrations of GC in the body is necessary to maintain homeostasis. However, when
animals are faced with chronic elevations in GC, they are faced with a substantial
challenge. This situation can lead to protein catabolism, hyperglycemia, immune
suppression, and susceptibility to infection. It is because of these potential deleterious

actions that an important function of the steroids is to curtail the HPA response to stress.

22

This is effected through negative feedback inhibition. In this process, high concentrations

of GC in blood inhibit further HPA response at the brain and pituitary level.

Effects of maternal care on the HPA-axis

Before and alter parturition, the HPA-axis can be modulated during the development of
an animal. External signals interact with the central nervous system to shape the way the
HPA-axis responds to stressors and this may be beneficial or not. Factors taking
advantage of this neuroplasticity by translating outside stimuli into internal neuronal
changes are cyclic adenosine monophosphate and neurotrophic factor cascades (Plotsky
2000). The exact termination of the critical period for plasticity of the HPA-axis is

species-dependent and is not yet determined in many cases.

The mother-infant relationship is thought to be an important factor in the way the
HPA-axis develops (Plotsky 2000). The primary care giver in general is a hidden
regulator of the development of different regulatory systems in a newborn (Hofer 1994).
The dam-offspring “security of attachment”, introduced in the late 1970’s, states that both
mare and foal develop the boundaries of their relationship together and they can rely on a
certain amount of predictability when interacting with each other (Bowlby 1977). This
implies that if the care of the mother towards the infant is inadequate, the development of
the offspring is impaired (Liu et al. 1997). This impairment in an animal can lead to a
skewed perception and evaluation of the immediate environment, thereby affecting stress

coping strategies.

23

A number of studies in rodents have investigated the effect of early-life stress on
the coping strategies of developing and adult animals. The stressor can increase or
decrease HPA-axis sensitivity, depending on breed genetics, time of exposure, and type
of streSsor. The breed effect was explored in a study where two strains of rats (congenital
learned helplessness, cLH and congenital non-learned helplessness, cNLH) were exposed
to cold stress and maternal deprivation stress on day 7 or day 21 post-partum, researchers
found that the ACTH release after adult stress in cLH rats was much higher than that in
cNLH rats (King & Edwards 1999). Conversely, adrenal responsiveness was much lower
in cLH rats. These findings indicated a dissociation of the pituitary and the adrenal gland
and dysfunctional stress response in cLH rats. This study showed that breed had an effect

the responsiveness of the HPA-axis.

The effect of time of exposure was explored during an experiment where rats
were exposed daily between day 2 and 14 of age to 180 min of maternal separation and
handling (HMSISO) versus 15 min separation (HMSlS) or no separation (AFR) (Plotsky
& Meaney 1993). The animals were weaned on day 22 and housed per 3 in a cage. The
first group showed hyper-responsiveness to psychological/emotional stressors when
compared to the 15-min separation group or the control. This hyper-responsiveness
remained visible throughout the animals’ lives. In the same study, maternal care seemed
to be disrupted more for the HMSlSO rats than for the HMS l 5 or AF R. Compared to
control animals (AFR and HMSl80), neonatal rats that were handled for relatively short
periods of time showed a higher expression of corticosterone receptor proteins in the

hippocampus, an area in the brain associated with the mediation of cognitive control

24

(Toates 2000). Corticosterone is the rodent equivalent of cortisol. This in turn results in
an enhanced sensitivity of the negative feedback system regulating corticosteroid release
because more GC can bind on the hypothalamus. The rat pups that had been subjected to
‘severe stress’ (HMS l 80) showed HPA—hyper-responsiveness later on in life, as indicated

by e. g., anxiety-like behavior, compared to the ones subjected to ‘mild stress’ (HMSlS).

Changes in maternal care, rather than the effect of separation itself, most likely
mediates the effects of post-handling stress as described above. When rat pups were
reunited with their darn after a separation and handling session, the amount of licking and
grooming displayed by the mother towards the offspring increased and the differences
with the control group remained stable over the lO-day testing period. (Liu et a1. 1997).
The same researchers subsequently investigated the individual differences in intensity of
maternal care in relation to stress-coping strategies of the pups later on in life. The pups
that received high quality maternal care (11 = 9) actually showed a reduced response to the
stressors compared to pups that received low frequencies of licking, grooming, and
arched-back nursing. Plasma corticosterone concentrations immediately after restraint
stress was 1711.0 i 55.2 nmon for low maternal care pups as opposed to 1186.8 i 82.8
mom for pups that received high maternal care (P < 0.05). Two hours post-stressor, the
difference was still present (P < 0.05) with corticosterone concentration in low maternal
care pups being 1021.2 i 69 nmon and 496.8 i 27.6 nmon in high maternal care pups.
The greater the frequency of licking and grooming during infancy, the lower the HPA-
response to stress was in adulthood. Offspring of females showing high maternal care had

a greater suppression of plasma ACTH after corticosterone treatment when compared to

25

the low maternal care counterparts (Liu et al. 1997). In addition, the glucocorticoid
receptor mRNA expression in the hippocarnpus of the high-matemal care dams was
higher and correlated with the frequency of licking and grooming compared to low-
matemal care dams. Thus, the animals that received the most maternal care seemed to
have the most sensitive negative feedback mechanism regulating the HPA-axis response.
In addition, individual variations in maternal care and increased maternal care post-
handling sessions during early stages of life had the same dampening effect on the
response to restraint stress as adults. Therefore, the researches concluded that it was the
increase in maternal care after separation that accounted for the stress resistance of pups
in adulthood and not as much the separation and handling procedure itself. This effect of
maternal care on responsiveness of the HPA-axis of rat pups was confirmed in another
study where fearfulness to novelty in rat pups was examined (Caldji et al. 1998).
Offspring of mothers that showed high frequencies of licking and grooming showed
substantially reduced behavioral fearfulness in response to novelty when compared to

offspring of low-matemal care dams.

Saliva Cortisol

When an animal perceives a certain stimulus as threatening, a stress response occurs,
resulting in the release of GC. Cortisol, the common GC in humans, pigs, and horses is
largely bound to serum proteins, corticosteroid-binding globulin (CBG), and albumin
(Vining et al. 1983). During social stress (introduction of adult horses and foals into a
new herd), the binding capacity of these CBG has been shown to decrease (Alexander &

Irvine 1998). So in addition to neogenesis of cortisol from the adrenal gland, cortisol

26

bound to proteins is released. The bound fraction of cortisol is biologically inactive
whereas the smaller, free fraction is available to produce physiological effects
(Kirschbaum & Hellhammer 1989; Umeda et al. 1981). The unbound cortisol fraction in
non-stressed human serum is about 1 to 15% of the total amount of cortisol and is
reflected by the salivary cortisol concentrations (Vining et al. 1983). Circulating unbound
cortisol passes readily through the parotid membrane. Vining et al. (1983) found that
human salivary cortisol concentrations are one third lower than the serum unbound-
cortisol concentrations, which they attributed to conversion of cortisol to a metabolite,
cortisone, when passing through the parotid membrane. This difference was also reported
in another study where salivary cortisol concentrations were found to be 30% lower than
those of serum unbound cortisol (Umeda et al. 1981). When assessing salivary cortisol,
flow rate of saliva did not influence cortisol concentrations (Hubert & de J ong-Meyer

1989)

Salivary cortisol is a minimally invasive measure in humans to assess adrenal
gland corticosteroid output during a stress response compared to blood cortisol. One
major advantage to taking saliva samples is that we avoid the stress of venipuncture. For
blood sampling, animals must undergo capture, restraint and venipuncture all at the same
time and this procedure in itself is a potential stressor (Blackshaw & Blackshaw 1989).
Taking salivary cortisol samples requires much less restraint. Saliva sample collection for
cortisol analysis has been demonstrated to be successful in guinea pigs (F enske 1996),
tree shrews (Ohl et al. 1999), squirrel monkeys (Fuchs et al. 1997), dogs (Beerda et al.

1996; Beerda et a1. 1998), sheep (Fell et al. 1985), pigs (Parrott & Misson 1989; Parrott et

27

al. 1989; Geverink et al. 1999), calves (Fell & Shutt 1986), and horses (van der Kolk et
al. 2001). Blackshaw and Blackshaw (1989) did encounter some difficulties when
assessing correlation between salivary and plasma cortisol concentrations in pigs.
Salivary secretion in pigs is discontinuous and normally restricted to feeding. Another
study showed that when using a specialized saliva collection device (oral diffusion sink),
continuous assessment of salivary cortisol during transport in pigs was possible

(Schonreiter et al. 1999).

Blood sampling in horses creates the same difficulties as demonstrated for other
species (Lebelt et al. 1996). Recent studies investigated the correlation between saliva
and plasma cortisol during ACTH challenge and transportation in horses (Warren et al.
2001). They found the correlation in horses ranged between 0.61 and 0.87 for the ACTH
administration experiment. Similarly, Lebelt et al. (1996) found a correlation coefficient
of 0.83 and Van der Kolk et al.(2001) found r = 0.93. Salivary cortisol concentrations
averaged 6.6 i 1.3 % of the plasma cortisol concentrations in the study by Warren et al.
(2001). This value is high in comparison to a study by Lebelt et al. (1996) where salivary
cortisol concentrations were 3 to 5 % of plasma total concentrations and in the study by
Van der Kolk et al. (2001) this percentage was 1.20 :1: 0.37 %. As in other species, horses
show a circadian pattern in salivary cortisol concentrations with highest concentrations
found during the morning hours (2.76 nmon 3: 1.05) and the lowest concentrations found
during the afiemoon (1.73 nmol :1: 0.65; Lebelt et al. 1996). Van der Kolk et al. (2001)
did not find a significant diurnal variation for salivary cortisol, but plasma total cortisol

concentrations did show a circadian pattern. In contrast, in another study morning

28

concentrations of plasma total cortisol measured 251 i 115 nmon and differed from

evening cortisol concentrations, which measured 142 i 53 nmol/1. Lebelt et al. (1996)
also showed that there were great inter-and intra-individual differences in daytime

profile.

Heart rate

The relationship between behavior and heart rate has long been established and
researchers have measured heart rate to evaluate the stress response. Catecholarnines,
norepinephrine and epinephrine, released in higher concentrations by the adrenal gland
during stress than when compared to non-stress situations, enhance cardiac concentration,
heart and skeletal muscle blood flow, blood glucose concentration, lypolysis, and oxygen
consumption (Thornton 1985). In addition, catecholamines in humans have been
associated with emotionality of a stress response as measured by the response during
participation in a video game. (Cleroux et al. 1985; Niezgoda & Tischner 1995).
Niezgoda and Tischner (1995) found that 5 min into the separation of mare and foal,
concentrations of catecholamines peaked. A limitation of the use of heart rate
measurements to assess a stress response is that, concomitant with the catecholamines
increase, an increase in activity and thus an increase in metabolic rate lead to an increase
in heart rate (Broom & Johnson 1993). Heart rate shows a linear increase with increasing
exercise level (Lowe 1990; Persson & Ullberg 1974), but exercise also increases release
of cortisol (Rose & Hodgson 1994). It is therefore important to take the level of activity
of an animal into account when evaluating cardiac changes to an external stimulus that is

a potential stressor.

29

For resting foals, the average heart rate at 8 tolO weeks of age is 60 to 79 beats

min-1 whereas for 6-month old foals, this pulse rate is 60 to 71 beats min-1 (Lowe 1990).

A decrease in heart rate in Camargue horses has been associated with positive social
contact, i.e., grooming at the back of the neck, which appeared to be the preferred
grooming site (Feh & de Mazieres 1993). The measurements of heart rate for that
particular study were taken with a stethoscope. An important limitation to the use of a
stethoscope is that one can only measure heart rate when the animal is in a state allowing
manipulation. However, when intense stress occurs, the behavioral response might be of
such magnitude that handling by humans is impossible. In addition, studies have reported
that approach of humans can cause elevated heart rate in animals, such as sheep (Baldock

& Sibly 1990). This increase is even greater when a dog is involved in the approach test.

The development of telemetrical heart rate monitoring devices provided the
solution for continuous non-invasive heart rate monitoring under a variety of
circumstances. Three major types of heart rate meters have been used to investigate heart
rate in horses: Equistat (EQB, Unionville, PA), Hippocard (Isler Bioengineering, Zurich,
Switzerland), and an adaptation of Polar (Polar CIC Inc, Port Washington, NY)
equipment used for monitoring heart rate during exercise in humans. All of these heart
rate meters require electrodes being placed on the heart girth of the horse and each meter
has a data-storing unit attached to or placed near the electrodes. The Equistat monitor has
successfully been used in a study investigating the influence of a lactate-reducing
supplement, betaine, on untrained and trained horses exercising to fatigue (Warren et al.

1999). Other studies investigating the relationship between exercise and heart rate have

30

used the Hippocard monitoring device to take accurate heart rate measures (Evans et al.
1993; Snow et al. 1992). Geverink et al. (1999) conducted a study investigating the effect
of shot biopsy on heart rate in slaughter pigs. The experimenters used Polar Sport Tester
consisting of an electrode belt with built-in transmitter and a wristwatch receiver. The
data collected in the wristwatch was downloaded via a Polar interface to a PC. Very little
research on horses has been done recording heart rate using Polar technology. Rivera al.
(1999) investigated behavioral and physiological responses (including heart rate) to initial
training to compare stalled versus pastured horses. Polar Vantage NV receivers were used

and the transmitters with electrodes were held in place by a custom-made belt.

Response to psychosocial stress

Psychosocial stress refers to stress caused by a disturbance in social interactions or by the
social interactions themselves. Isolation is an example of a psychosocial stressor (Asterita
1985). The physiological responses to isolation have been studied in depth. When sheep
were isolated from the flock, a rapid rise in cortisol output was measured with the highest
values occurring 15 min after isolation (Parrott et al. 1988). The study also showed that
cortisol concentration decreased after 30 min and was elevated again 90 min after
isolation. There was no significant increase in prolactin, a hormone used as an indicator

of chronic stress in sheep, following treatment.

Although they are also herd animals, horses often undergo social isolation due to

standard operating farm protocol (SOP) or during routine husbandry procedures. Changes

in behavior and physiology of horses subjected to acute isolation stress has been reported

31

(Bagshaw et al. 1994). Walking-activity was higher for animals in isolation than for the
mares that had visual contact with conspecifics. Whinnying, blowing (defined as “high-
pitched ‘whoosh’ sound produced as the horse exhales through the nose while standing
alert”), and defecating increased with isolation as well. Since only two mares pawed

throughout the experiment, no effects of isolation on this behavior could be reported.

Mares and foals are often separated during routine veterinary procedures such as
foal-heat breeding, which usually occurs around 10 days after foaling (Harper 1999). In
light of the strong mare-foal bond described earlier in this chapter, considerable stress can
be associated with these separations. Niezgoda and Tischner (1995) investigated
physiological responses to separations, such as changes in plasma catecholamines
(epinephrine and norepinephrine) and cortisol when separating 10 Polish pony mares and
foals for 10 min on day 5, 20 and 50 post-parturition (Niezgoda & Tischner 1995). The
strongest physiological response was observed during the first separation at 5 days of age.
Epinephrine concentrations showed a maximum increase (foals: 1.21 i 0.1 to 8.3 i 0.7
nmol/l, mares: 0.98 :1: 0.1 to 6.9 i 0.9 nmol/l; P < 0.01) after 5 min, as did norepinephrine
levels (foals: 1.99 i 0.03 to 6.9 i 0.3 nmol/l, mares: 1.98 i 0.06 to 5.7 i 0.5 nmol/l; P <
0.05), and cortisol concentrations increased significantly after 10 min (foals: 133.5 i 15
to 181.2 i 15 nmol/l, mares: 116.6 i 5.5 to 164.2 i 11.0 nmol/1;P < 0.01).
Concentrations of hormones were lower in subsequent isolations, especially in mares (P <
001-005). In addition, conflicting results were seen in response to the separations on day
20 and 50. On day 20, only foals showed a significant increase in cortisol. On day 50, a

significant increase was observed in mares but not in foals. Previous research has not

32

looked at the behavioral effect of short-term separations on mare-foal interactions and

neither have reports on foal behavior during and after the separations been published.

Weaning

Introduction

As described above, mare and foal form a strong bond and this is expressed in several
behaviors associated with maternal care. Weaning disrupts this attachment between dam
and offspring. At weaning, a foal becomes independent of its mother for food, protection
and social interaction. In the wild, mares wean their foals gradually, shortly before the
birth of a new foal, l or 2 years after the weanling was born (Crowell-Davis 1986).
However, in managed herds weaning is controlled by the breeder and occurs between 3
and 6 months. Weaning causes an increase in general motor activity and vocalizations in
several species such as red deer calves (Pollard et al. 1992), pigs (V arley 1995), and
horses (Crowell-Davis 1986). Casual observations, like those reported by Houpt et a1.
(1984), show that the foal neighs, attempts to rejoin its dam and can show loss of

appetite. In addition, foals may attempt to escape and injure themselves.

The earlier the foal is weaned, the sooner the mare can be used for spring,
summer, and fall riding and show activities (Kohler 1972). In a two-year study, Kohler
investigated the effect of early weaning (at 30 days) as compared to weaning at 6 months
of age on average daily weight gain, “bloom scores” (body condition scores) and several
management procedures. They found that although early-weaned foals did not gain

weight as fast during the first 12 weeks, they were comparable in weight to foals weaned

33

at 6 months. The early-weaned foals did have better body condition scores because of
less exposure to the elements and they underwent the management procedures after
weaning more readily because they had been handled more. Weaning tended to make
foals, especially those weaned at 30 days, more sensitive to noise, rapid movements, and

other factors disturbing their environment (Kohler 1972).

Under unmanaged conditions, weaning is considered to be a gradual process and
elicits little reaction from the foal (Tyler 1972), but current management schemes require
that a foal is weaned earlier and the response to weaning can be quite extensive
behaviorally and physiologically (McCall et al. 1985 ; McCall et al. 1987; Hoffman et al.
1995, Houpt et al. 1984). Besides the time at which a foal is weaned, there are other
factors influencing the way a foal copes with the stress of weaning. Foals that had access
to concentrate (Hoffinan et al. 1995) or creep feed (McCall et a1. 1985) as well as pasture
grazing and hay tended to cope better with weaning than foals that had access to pasture
grazing and bay. The effect of diet on the way mares cope with stress has also been
investigated (Holland et al. 1997). The researchers found that serum cortisol
concentrations after weaning were lower for mares on a sugar and starch concentrate in
comparison to mares fed a fat and fiber concentrate, but both groups showed initial
depression of lymphocytes. Thus, although both groups were affected by weaning, the
animals that were fed the sugar and starch concentrate diet showed a lesser physiological

response during weaning.

34

Weaning methods

Abrupt versus gradual weaning

Abrupt weaning involves the sudden, total separation of mare and foal. The foal can no
longer see the mare, although auditory contact may still be possible. Gradual weaning
involves the separation of mare and foal by a fence, but the animals can maintain visual,
tactile and auditory contact for a while. A study on 21 Quarter Horse foals compared the
behavioral (McCall et al. 1985) and physiological (McCall et al. 1987) responses to
abrupt weaning versus gradual weaning. Gradually-weaned mares and foals were allowed
fence-line contact for 7 days and were subsequently moved out of auditory, olfactory, and
visual contact. Foals in the abrupt weaning method spent more time walking and less
time standing still. They also laid down stemally more often than laterally, indicating
they were less relaxed. Abrupt weaning led to a higher frequency of vocalizations
compared to gradual weaning. This is most likely because mare and foal were still in
contact during the first 7 days of gradual weaning. Other behaviors displayed by the
weaned foals were aggression to other foals and pawing. The adrenal response to
weaning varied greatly between animals, but the abruptly weaned animals showed a
significant positive adrenal response to weaning whereas the gradually weaned animals
did not. F rom these results we can conclude that gradual weaning causes a lesser stress

response than abrupt weaning.

Stall versus pasture weaning

Foals are often weaned individually in stalls, but the alternative of weaning foals in a

group out on pasture has encouraged researchers to compare stressfulness of both

35

scenarios (Heleski et al. In Press). In their study, Heleski et al. (In Press) weaned 6
Quarter Horse foals in individual box stalls and 6 were weaned in groups of 3 in a
paddock. Paddock-weaned foals engaged in a greater variety of activities (provided by
the opportunity to graze and engage in social interactions) and showed a lesser number of
aberrant behaviors, such as licking, biting, or kicking the stall/shed wall. Based on the
reduced occurrence of aberrant behaviors, it appeared that foals in the paddock-weaned
situation were coping better with weaning than their stall-weaned counterparts.
Conversely, in a study comparing the effects of indoor housing versus housing outdoor
after weaning of red deer calves, showed that the indoor housed calves had a better
weight gain, showed less pacing, and ate more (Pollard et al. 1992). Although the indoor-
housed animals seemed to cope better with weaning, it should be noted that these calves
were kept in groups whereas the stalled foals in the study by Heleski et al. (In Press) were
housed individually. That red deer calves were able to form social relations with their

‘herd members’, may explain the decrease of the weaning impact.

Weaning foals individually or in pair_s

It has been argued that the presence of conspecifics can decrease the stressfulness
of weaning. However, contradicting evidence for this finding has been presented. When
weaning eight ponies individually in stalls and comparing their behaviors to 14 ponies
weaned in pairs, researchers found that there were no differences between plasma cortisol
concentrations due to weaning or treatment (Houpt et al. 1984). In contrast, the foals did
vocalize (neigh) more after weaning than before and they were also more active. The

only difference between individual and paired weanlings was that individually housed

36

foals vocalized more after weaning. The researchers concluded that weaning foals in
pairs seemed less stressfirl than when weaning foals individual. However, a more recent
study showed that paired foals, although there was a tendency to be less vocal, displayed
significant aggression towards one another (Hoffman et al. 1995). It should be noted that
the occurrence of aggression may be related to the relationships, if any, existing between
the foals prior to weaning. Though it is not clearly stated, the publications of both Houpt
et al. (1984) and Hoffman et al. (1995) suggest that the foals had spent one or more
months together on pasture prior to weaning. There was a tendency for the serum cortisol
concentrations prior to ACTH challenge to be higher in paired foals than single-housed
foals. In addition, the physiological response of individually housed foals to an ACTH-
challenge was less than the response of paired foals. The researchers concluded that there
was little or no real advantage to weaning foals in pairs. Companionship during weaning
may not have an effect unless there is enough space within a confined area to avoid
aggression, or unless the companion behaves in a similar way as the dam of which a foal
was deprived. In the red deer study by Pollard et al. (1992), the presence of dry hinds in
the groups of weaned calves seemed to have beneficial effects on the stress coping
strategies of the calves. Though it was not indicated as such by the authors, the possibility
exists that the hinds behaved in ways similar to the mother that the calves were deprived

of.

Another study investigating the effect of paired versus individual weaning in
horses showed no physiological differences between the two treatments (Malinowski et

al. 1990). Cortisol concentrations were higher in weaned animals when compared to

37

control animals, but no difference was found between foals weaned individually and
weaned in pairs. Cell-mediated immunity, as measured by lymphocyte proliferation
response, was lower in foals weaned in pairs compared to foals weaned individually.
Thus, according to the physiological indicators, weaning foals individually was preferred

over weaning foals in pairs.

38

CHAPTER III: IMMEDIATE AND PROLONGED EFFECTS OF SHORT-TERM

SEPARATIONS ON MARE-FOAL INTERACTIONS

Abstract

Routine husbandry procedures often cause farm managers to separate a mare from her
unweaned foal for short periods of time. Research has shown that these are stress-causing
events for mare and foal. However, the separations in themselves may lead to intensified
mare-foal interactions and increased maternal care displayed by the mare. Exposure to
increased maternal care in rodent pups has been shown to dampen stress responses in the
animals as adults. In this study we investigated the immediate and prolonged effects of
short-term separations on mare-foal interactions. We hypothesized that short-term
separations increase the occurrence of mare-foal interactions. As expected, immediately
after a separation, both mare and foal initiated more contact behaviors compared to pre-
separation (foal: 1.1 i 0.6 to 4.1 _+_ 1.0 times per 30 min, P = 0.0105; mare: 0.3 :t 0.3 to
1.7 i 0.7 times per 30 min, P = 0.0087) and were more attentive to one another (foal: 1.5
i 0.7 to 3.9 i- 1.0, P = 0.0050; mare: 1.5 i 0.5 to 3.1 :1: 0.7, P = 0.066). They also stayed
closer to one another (P = 0.00090). Fillies tended to remain closer to their dams than
colts did (P = 0.061). However, prolonged effects of the separations could not be
identified by behaviorally evaluating mare-foal interactions 4 to 6 days post-separation.
Based on studies done in rodents, we predict that the effect of the immediate increase in

maternal care post-separation becomes apparent when submitting foals to a stressor.

39

Introduction

When carrying out routine husbandry procedures for mares and their foals, horse farm
managers can allow the foal to accompany the mare while she undergoes the procedures,
or leave the foal in the stall by itself. Both options have obvious advantages. When
leaving the foal in the stall, handlers do not need to monitor it to make sure it does not
escape, stay behind, or injure itself while running back and forth. However, if the foal is
left in the stall and the mare loses visual contact, the mare-foal bond is disrupted, which
in turn agitates the mare and makes her more difficult to handle. In addition, other
animals within hearing range of the separated pair may become aroused because of vocal
interaction between dam and foal. For the farm workers, the environment becomes quite
unpleasant to work in and, moreover, the commotion can impact other animals (personal

observation).

Casual observations such as those mentioned above have led scientists to
investigate how short-term separations impact mare and foal. A study looking into the
behavioral responses of mares and their foals to weekly five-minute separations from
birth until 12 weeks of age found that animals neighed more while alone (Houpt et al.
1981). A maximal vocal response of the foals was found at four weeks of age. Mares
were overall more active during the separation compared to when their foal was near
them. Foals were not more active during the first week, but they did show increased
activity during the separation from week 2 until week 12. The results showed that the age
of the foal plays an important role in how much short-term separations affect foals.

Separation from young foals affected mares more than when they were separated fiom an

40

older foal whereas foals less than four weeks old did not seem as affected by the

separations compared to when they were older (Houpt et al. 1981).

Other researchers investigated physiological responses to separations, such as
changes in plasma catecholamines (epinephrine and norepinephrine) and cortisol when
separating 10 Polish pony mares and foals for 10 min on day 5, 20 and 50 post-parturition
(Niezgoda & Tischner 1995). The strongest physiological response was observed during
the first separation at 5 days of age. Epinephrine concentrations showed a maximum
increase (foals: 1.21 i 0.1 to 8.3 i 0.7 nmol/l, mares: 0.98 i 0.1 to 6.9 :t 0.9 nmol/1; P <
0.01) after 5 min, as did norepinephrine levels (foals: 1.99 i 0.03 to 6.9 i- 0.3 nmol/l,
mares: 1.98 i 0.06 to 5.7 i 0.5 nmol/1; P < 0.05), and cortisol concentrations increased
significantly after 10 min (foals: 133.5 3: 15 to 181.2 i 15 nmol/l, mares: 116.6 i 5.5 to
164.2 i 11.0 nmol/l; P < 0.01). Concentrations of hormones were lower in subsequent
isolations, especially in mares (P < 001-005). In addition, conflicting results were seen
in response to the separations on day 20 and 50. On day 20, only foals showed a
significant increase in cortisol. On day 50, a significant increase was observed in mares

but not in foals.

Separations of short duration, applied by many horse farms during routine
management and veterinary procedures, are stress-causing events. However, horse
experts have argued in the past that a threat to the foal or short-term separations of mare
and foal can stimulate maternal care in horses (Crowell-Davis & Houpt 1986).

Consequently, the natural variation in maternal care that potentially exists between horses

41

could be decreased as the levels of maternal care increase. Intensive maternal care at
early stages in life has been found to dampen stress responses during adulthood in rodents
(Liu et a1. 1997). Having received maternal care fiom a mother who displayed high
licking and grooming and who facilitated nursing by standing in an arched back position,
rat pups exposed to restraint stress as adults showed a smaller increase in ACTH and

corticosterone concentrations, which was also of shorter duration.

As foals on farms are exposed to multiple stressors, ranging fiom weaning to
halter and riding training, we deemed research looking into the effects of short-term
separations on maternal care and mare-foal interactions in general was warranted. The
objective of this study was to investigate how mare-foal interactions were impacted
immediately after separations. In addition, we evaluated how the mare-foal relationship
evolved throughout the period in which separations were performed by comparing mare-
foal interactions in a separation group with mare-foal interactions in a control group
where mare and foal were never separated. We hypothesized that interactions between
mare and foal would increase after separation and that this effect would be apparent
during all six separations. Furthermore, we hypothesized that, by observing control and
separation group mare-foal interactions, we would find that separations increased mare-

foal interactions in the separation group as compared to a control group.

42

Materials and Methods

Animals

This research was conducted at MSU Merillat Equine Center fiom January through
October 2000. A total of 10 mare-foal pairs were studied (Appendix C). The animals
were divided into two groups according to a random design blocked by foal gender. Each
group consisted of three fillies and two colts. The ten foals were sired by three different
stallions as indicated in Appendix C. Each group represented a treatment. There was a
control group (n=5) in which mare and foal were never separated from one another and a
separation group (n=5) where mare and foal underwent six 10 min separations performed
every other week from the time the foals were 2 weeks old until 3 months of age. The
protocol was designed and some decisions (10 min separation) were based on our current
understanding of foal physiology and behavior, whereas others (separations every other
two weeks) were based on farm logistics. The MSU All University Committee on Animal
Use and Care reviewed and approved the experimental protocols described in this

chapter.

Housing and feeding

From January through April 2000, mares were individually housed in Barn A in 13.38 m2

box stalls. The stalls were bedded with wood shavings and had a combination feeder in

one corner. The mares had ad libitum access to water via an automated drinker. Prior to

foaling, mares were moved to one of two 26.76 m2 foaling stalls that were bedded with

straw. By May 2000, all mares had foaled and the animals were moved to the climate

43

controlled Barn B. All mares were fed twice daily according to NRC recommendations
(National Research Council, 1989). The morning feeding occurred between 08:00 and

10:00 and the afternoon feeding, between 15:00 and 17:00.

Video recording and data collection

A Panasonic CCTV WV—BP334 video camera was installed out of the animals’ reach on
the wall between two stalls housing the study subjects. All cameras were connected to a
Panasonic WV-CM146 field switcher monitor. In addition to the live observations, we
recorded the animals on VHS tape using a Panasonic AG—6740 VCR. Depending on
which dyad would be subjected to the separation, we turned the camera towards the stall

of interest (Fig. 1 and 2).

44

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Foal 8 Foal 10 Foal 4 Foal 6 Foal 2
1 '\ /' '\ /'
a a a a
L] LJ u I r
51 I?! 13‘ El
a a a H
1 \ ./ \. \.
Foal 7 Foal 3 Foal 5 Foal 9 Foal l

 

 

 

Figure 1. Scheme of part of MSU Merillat Equine Center Barn A, not drawn to
scale. Each rectangle represents a stall. The numbered boxes are cameras and the

arrows indicate which stalls were covered by each camera.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Foal 8 FM! 10 Foal 4 Foal 6 Foal 2
~ I
1 \ x I /' '\ /'
E1 E1 , 0 1:1
Ll \ l—l / ’ L-l Ll
\ ——————— J
Moved beginning of June
\ EPI\ \. l \.
Foal 3 Foal 7 Foal 9 Foal 5 Foal 1

 

 

 

Figure 2. Scheme of part of MSU Merillat Equine Center Barn B, not drawn to
scale. After May 5, 2000 all the animals on the project were moved to this barn.
Merillat management assigned foals to stalls according to availability of Dutch

doors that could be opened to allow airflow.

45

Eight cameras were used, which is the maximum number of cameras from which the
field switcher could collect data before sending them to the time lapse VCR. The data
were recorded onto VHS tapes either in 18 h or 24 h time mode. We used the same
equipment for the post-separation observations when we evaluated mare-foal interactions.
The tapes were decoded on-site using a Psion Workabout (Psion, PLC, England)
equipped with The ObserverTM behavioral software (version 3.0, Noldus Information
Technology Inc., Sterling, Virginia) or in the behavioral laboratory using personal
computers with the same software. The equipment was moved when the animals were

relocated from Barn A to Barn B.

Short-term separations

Each mare-foal dyad in the test group was subjected to six 10 min separations performed
every other week, starting when the foal was approximately 2 weeks old until 12 weeks
of age. Separations were performed during the first part of the week. An undergraduate
student working on the farm entered the stall, haltered the mare, and led her out of the
stall while making sure the foal stayed inside. The foal could not see the mare, but we
were unable to take the mare out of the foal’s hearing range when both animals vocalized

during the separation. After 10 min, the handler reunited mare and foal.

For each separation we recorded behaviors 30 min before and after separation as well
as during the 10 min of separation. To evaluate the immediate effect of separations, each
mare-foal pair served as its own control as we compared mare-foal interactions before

and after separation. The complete ethogram used is described in Appendix E. Because of

46

interference of the Panasonic field switcher with the time-lapse recording, we were
unable to reliably record durations for the behavioral categories of walking, trotting,
cantering, and nursing during separation trials and post-separation evaluations.

It should be noted that only “eating”, “urinating”, “defecating”, “vocalizing”,
“investigating”, “pawing”, and “looking forward” were discussed for the 10-minute
separations. We chose these behaviors on the basis of their occurrence and relevance in
demonstrating a stress response (Hoffman 1995; Houpt 1984; Heleski, in Press). For the
pre- and post-separation evaluation we discussed a different set of behaviors. Analysis of
behaviors performed by the mare consisted of “contact” (including “nuzzling” and
“allogrooming” events), “watch foal”, and “aggression” (including “laying back ears”,

“tail swish”, “bite threat”, “bite”, “kick threat”, and “kick”). Analyzed behaviors

displayed by the foal towards the mare were “contact”, “watch mare”, and “aggression”.

We also visually assessed the distance between mare and foal every 5 min for the
observation period pre- and post- separation on a scale of “a” to “c”; “a” denoted the
distance when the anterior part of mare and foal was closer than 1/3 the length of the
stall; “c” was scored when the animals were in opposite corners of the stall; “b” denoted
any distance between “a” and “c”. To analyze these data, we replaced the initial codes for
the distance between mares and foals by numbers (“a”=3, “b”=2, “c”=1). For the data
collection prior to and after separation, we collected seven data points per 30 min of
observation time. The values replacing the letter scores were averaged and these
‘proximity scores’ were analyzed for differences between pre— and post-separation. We

weighed the number of proximity scores we used per average.

47

The data we collected represented counted values, so we square root transformed the
data to normalize the distribution. For the 10-min separations, we used the repeated
measure of trial in mixed model procedures of SAS. Trials refer to the two-weekly
separation, which is inter-related with the age of the foal, as indicated by the graphs in the
results section. The variables used in the model for the 10-minute separation were foal,
gender, and trial. To analyze the effect of the separations on mare-foal interactions we
used period (before and after separation) in mixed model procedures of SAS. The model
for the data analysis included gender and period as main effects and the two-way
interaction between them. Significant main and interaction effects were analyzed using
multiple comparisons t-test. We used a Bonferroni adjustment to correct for the number

of comparisons.

Evaluations of interactions initiated by foal and more 4-6 days after separation

We evaluated mare-foal interactions at the end of every week following a separation,
allowing 4 to 6 days between separation and assessment of maternal care. In this part of
the experiment, we compared mare-foal interactions in the separation group (n=5) with
mare-foal interactions in an actual control group (n=5). Although we used 12 h, 18 h, or
24 h time-lapse mode to record, the videotapes were analyzed in 12 h mode. This
provided a better transition between frames and reduced observation time. The
observation schedule (test animals and corresponding control animals) listed in Appendix
D was such that no more than five pairs of animals were observed within one week. We
divided the observation day into four two-hour blocks equally distributed between 08:30

and 18:00 (08:30-10:30; 10:50-12:50; 13:30-15:30; 15:50-17:50). We used continuous

48

recording and focal sampling to record behaviors. Every 5 min we recorded behaviors for
a different mother-foal pair and continued this rotation for two hours. At the start of each
5 min observation session we visually assessed the distance between mare and foal
according to the scale described previously. For a given mare-foal pair we collected 100
min of data per post-separation observation day. The complete ethogram we used is
described in Appendix E, but the behaviors we discussed were limited to the mare-foal

interactions recorded on the day of separation, which are described previously.

The behaviors recorded represented count data, so we transformed the data using a
square root transformation to normalize the distribution. The model for the post-
separation evaluation included treatment, trial, period, and their two- and three-way
interactions. It should be noted that the class factor “period” for post-separation
evaluations represents one of four time blocks of the observation day during which data
were collected. Significant factors fi'om the mixed model were further investigated using
multiple comparisons t-test with Bonferroni adjustment. When no difference was found

between control and separation group, data were combined for further analysis. The

means reported in this study are lsmeans i SEM.

To analyze proximity, a similar procedure was used as that for separations and pre—
and post-separation observations (see Short-term separations), but per observation block
we had a maximum of 5 proximity values, which we averaged to obtain the proximity

score.

49

Results

Foal behavior during short-term separations

The frequency of “investigating” (F5,1 5:3.94, P = 0.018) and “looking forward”
(F5,15=3.05, P = 0.043) were influenced by the main effect of trial (Fig. 3, 4). However,
frequencies of “eating” (F5,15=0.42, P = 0.83), “urinating” (F5,15=0.92, P = 0.49),

“vocalizing” (F5,15=0.15, P = 0.98), and “pawing” (F5,15=1.70, P = 0.20) did not differ

between subsequent trials.

In Fig. 3, we show that the first trial, when foals were 2 weeks old, has
significantly lower investigating fi'equencies than the fifth trial, at 10 weeks of age.
Frequencies of investigating at 4 weeks of age are significantly lower when compared to
8, 10, and 12 weeks (P < 0.050). In addition, at 4 weeks of age, investigative behavior
tended to have lower frequencies than at 6 weeks as well (P < 0.10). Finally, at 6 weeks,

less investigative behavior was displayed when compared to 10 weeks of age (P < 0.050).

Figure 4 shows that the fi'equencies of looking forward for trial 3 (6 weeks of age)
and 4 (8 weeks of age) are lower than for trial 1 (2 weeks), 5 (10 weeks), and 6 (12
weeks). It should be noted that probability for the difference between trial 1 and 4 is P =

0.063.

50

P<0.l

 

 

 

Frequency of investigating by
foals

2 4 6 8 10 12
Foal age (weeks)

Figure 3. Mean frequency i SEM of investigating by foals during lO—min separation
for subsequent trials. (N=5, P < 0.05 except where indicated). Differences are
indicated by the letters (a,b,c,d).

 

 

 

 

 

P<0.l
g ab
15.
E
g, abc
37,210-
°8
2n—
“-
03'
5 5‘
c
e
:1
E
u. 0-

 

   

2 4 6 8 1 0 12
Foal age (weeks)

Figure 4. Mean frequency :1: SEM of looking forward by foals during lO—min
separation for subsequent trials. (N=5, P < 0.05 except where indicated). Differences

are indicated by the letters (a,b,c).

51

Frequency of vocalizations did not prove to be significantly different between
subsequent trials (Fig. 5). During the separations, frequency of vocalizations ranged

between 18.2 i 14.2 and 31.00 i 12.2 vocalizations over 10 min. Table 1 shows the mean

frequency i SEM for eating, urinating, and pawing per trial.

J:
O
r

foals
N
O

l

           

 

Frequency of vocalizing by

o
I

 

2 4 6 8 10
Foal age (weeks)

Figure 5. Mean frequency :1: SEM of vocalizing by foals during lO—min separation
for subsequent trials. (N=5, NS).

Foal gender did not affect the frequencies of “eating” (F1,3=0.22, P = 0.67), “urinating”
(F 1,3=0.14, P = 0.73), “vocalizing” (F1,3=0.81, P = 0.43), “investigating” (F 1,3=0.20, P =

0.68), “pawing” (F1,3=l.09, P = 0.37), and “looking forward” (F1,3=l .23, P = 0.35)

during short-term separations.

52

‘F"

 

 

Table 1. Frequency :1: SEM (per 10 min) for foals of “eating”, “pawing”, and
“urinating” per foal age. N 0 significant differences were revealed by the square root
transformed analysis.

2 weeks

4 weeks

6 weeks

8 weeks

10 weeks

12 weeks

Eat (P = 0.83)
0.13 i 0.25
0.16 i 0.37
0.092 r 0.190
0.020 r 0.098
0.25 r. 0.32

0.42 :1: 0.41

Paw (P = 0.20)
0.13 i 0.17

1.20 :1: 0.80
0.00035 i 0.00800
0.00061 i 0.01200
0.15 i 0.17

0.22 i 0.20

Urinate (P = 0.50)
0.13 i 0.16

0.0044 i 0.0350
0.035 i 0.077
0.058 i 0.110
0.035 i 0.077

0.00018 i 0.00550

 

53

Proximity before and after short-term separations

The statistical analysis revealed that proximity was affected by the main effect of period

(period = data collected 30 min before or after separation), i.e., the distance between mare

and foal was different during pre- and post-separation observations (F 1,37=13.08, P =

0.00090). In addition, there was a tendency for an effect of gender (Fl,37=3.74, P =

0.061). The interaction between these two main effects was not significant (F1,37=1 .59,
P = 0.22). As indicated previously, proximity scores were analyzed with 3 being the
closest proximity and 1 the furthest. The overall proximity score for fillies was 2.70 i-
0.05 and for colts 2.61 i 0.04, indicating that both sexes stayed fairly close to their dams
at all times. Before separation, the overall proximity score was 2.51 :1: 0.05 and after
separation, the score increased to 2.80 i 0.06. The effect of the separation on proximity is

represented in Fig. 6. After the separation, foals remained closer to their dam than before.

54

(A)
1
CT

Proximity score
N
1

 

   

._L
l

pre post
Pefiod

Figure 6. Mean proximity score i SEM between mares and foals for pre- and post-
separation observation periods. (N=5, P < 0.05). Differences are indicated by the
letters (a,b).

55

Interactions displayed by the foal before and after separation

In all cases, the foal initiated nursing immediately following 10 min of separation. Of the
four interactions displayed by the foal, “aggression” was the only one that did not differ
during pre- and post-separation observation blocks. In fact, aggressions from the foal

directed towards the mare were extremely rare; only 6 occurrences were recorded. Prior

to the separation, foals aggressed 0.0085 i 0.0130 times during 30 min whereas after the

separation, this was 0.0089 :t 0.0130 times.

Frequencies of “watching mare” and “contact” for both periods are shown in Fig.
7 and 8, respectively. The figure shows that, as a result of the separation, the frequency

with which foals watched their dams increased from 1.5 i 0.7 to 3.9 i 1.0 times per 30
min (F 1,42=8.76, P = 0.0050). The same pattern was found for the interaction in which

physical contact occurred that was initiated by the foal. The frequency increased from 1.1

i- 0.6 to 4.1 i 1.0 times per 30 min (F1,42=7.l9, P = 0.011).

56

Frequency of watching
mare by foals

OANCAJ#UIO)

   

pre post
Peflod

Figure 7. Mean frequency :t SEM of foals watching mare during 30 min for pre-
and post-separation observation periods (N=5, P < 0.05). Differences are indicated
by the letters (a,b).

 

 

 

 

.- 6— b
0
:25-
a
884—
“53'
>i-u3‘
Ow a
5,152—
3::
gai—
“ 0
pre post
Pefiod

Figure 8. Mean frequency i SEM of contact initiated by foals during 30 min for pre-
and post-separation observation period. (N=5, P < 0.05). Differences are indicated
by the letters (a,b).

Interactions displayed by the mare before and after separation
Gender of foal did not affect the fiequencies of any of the mare-initiated interactions

investigated. Aggression was the only behavior that was not affected by the separations

(F1,42=0.13, P = 0.72). This indicates that the increased contact initiated by the foal did

not lead to increased rejection by the mare, but instead was tolerated. In fact, the

frequency of contact behavior directed towards the foal increased after the separations

(F1,42=7.58, P = 0.0087), as shown in Fig. 9. Per 30 min, frequency of contact before the

separation was 0.3 i 0.3 and after separation 1.7 i 0.7 times.

Following the separations, mares tended to watch their foals more (F1,42=3.56, P

= 0.066), which is represented in Fig. 10. Frequency per 30 min increased from 1.5 i 0.5

times before separation to 3.11 i 0.72 times after separation.

58

Frequency of contact
initiated by mares
N O)
L .l—l
0"

 

 

1 a
0 mi _L-
pre post
Pefiod

Figure 9. Mean frequency i SEM of contact behaviors during 30 min displayed by
the mare towards the foal during pre- and post-separation observation periods.
(N=5, P < 0.05). Differences are indicated by the letters (a,b).

Frequency of watching foal
by mares
N

 

   

pre post
Pefiod

Figure 10. Mean frequency i SEM of watching foal during 30 min displayed by the
mare pre- and post-separations observations. (N=5, P < 0.05). Differences are
indicated by the letters (a,b).

59

Proximity during evaluations 4-6 days post-separation

Proximity did not change over subsequent trials (F5,164=1.16, P = 0.33). Similarly, there
was no effect of separation on the distance between foals and their dam (F1,164=1.77, P =
0.19). We did find an effect of gender on proximity, as shown in Fig. 11 (F1,164=3.92, P

= 0.049). The average proximity score for fillies was 2.63 i 0.04 whereas for colts this

was 2.50 i 0.05.

Proximity score

 

 

filly colt

Gender

Figure 11. Mean proximity scores :t SEM between mares and foals for both genders
during post-separation evaluations. (N=10, P < 0.05). Differences are indicated by
the letters (a,b).

60

Evaluations of interactions initiated by the foal 4-6 days post-separation

Out of all the behaviors we investigated, the occurrence of aggression was too low to
reliably detect any differences (7 total occurrences), so we removed this interaction from
the analysis. Regarding the frequency of “watching mare”, there was a significant day by
treatment (treatment = control versus separation) interaction (F 5,150=2.37, P = 0.042).
However, none of the trials showed a significant treatment difference. When comparing
trials within treatment we found that, for foals exposed to short-term separations at week
4 of age, they had a higher frequency of “watching mare” than at 2 or 6 weeks.
Frequencies of “watching mare” between 2 and 6 weeks of age did not significantly

differ.

The frequency with which the foal initiated contact with the mare was not
affected by treatment (F 1,3=0.21, P = 0.66). We did see a trend regarding the effect of
observation period. Figure 12 illustrates how frequencies of contact were significantly

lower during the last observation block when compared to the first or third block.

61

1.5-

 

 

TE. 1-
>0
0
C
0
3-
2005-
III.

0.

08:30-10:30 10:50-12:50 13:30-15:30 15:50-17:50
Observation block

Figure 12. Mean frequency :1: SEM of contact initiated by the foal per observation

block during post-separation evaluations. (N=10, P < 0.05). Differences are indicated
by the letters (a,b).

62

Evaluations of interactions initiated by the mare 4-6 days post-separation
The fi'equency with which mares watched their foals was not affected by any of the

factors in the model. Figure 13 shows the mean frequency per trial.

      

E 1-
.2
a:
.5
.c
o
in!
gA
‘7 05-
‘85
>.
o
c
o
a
8
r. 0-

 

2 4 6 8 10 12
Foal age (weeks)

Figure 13. Mean frequency :1: SEM of watching foal displayed by mares, per trial,
during post-separation evaluation. (N=10, NS).

The frequency with which the mare directed contact behaviors towards the foals

depended on the time of observation (F3,135=6.88, P = 0.0002). Figure 14 shows the

difference in frequency between the different periods. Most of the contact by the mare

was initiated during the third observation block, i.e., early afternoon.

63

P<0.l

 

 

 

 

 

 

Aoe— b
£750.7-
C «10.6-
8 2
.6 20.5—
5, >.0.4- a
g 30.34
§§02 c
,3: 20.1-
0‘

 

      

08:30-10:30 10:50-12:50 13:30-15:30 15:50-17:50
Observation block

Figure 14. Mean frequency :1: SEM of contact directed to the foal by the mare per
observation period during post-separation evaluations. (N=10, P < 0.05 except where
indicated). Differences are indicated by the letters (a,b,c).

.1)

.0
or

Frequency of aggressions
displayed by mares (h

 

 

08:30-10;30 10:50-12:50 13:30-15:30 15:50-17:50
Observation block

Figure 15. Mean frequency :t SEM of aggressions directed towards the foal by the
mare per observation block during post-separation evaluations. (N=10, P < 0.05).
Differences are indicated by the letters (a,b).

64

Aggression occurred during specific time blocks throughout the day (F3,135=3.36,

P =0.021). Figure 15 represents the main effect of observation period on the frequency of
aggressions by the mare directed towards the foal. The graph shows that aggression was
significantly higher during the first observation block when compared to block 2 and 4.

No difference was found between blocks 2, 3, and 4.

There was a significant treatment by trial interaction for the frequency of

aggression (F 5,13 5:2.41, P = 0.039). The frequency of aggression for the separated group

at age 8 weeks was 0.3 i 0.1 occurrences per 30 min. This was significantly different

from the frequency at 6 weeks (0.024 :1: 0.042; t135=-2.57, P = 0.011) and from the

frequency at week 8 within the control group (0.003 t 0.015; t135=2.56, P =0.011).

Discussion and conclusion

This study set out to investigate immediate and prolonged effects of lO-min separations
on mare-foal interactions. Supporting our selection of the duration for the separation, we
mention work by Niezgoda and Tischner (1995). When investigating stressfulness of
short-term separations in mares and foals, they showed that a significant increase in
catecholamines occurred 5 min into the separation. In addition, their results showed that
an increase in cortisol in the foal occurred 10 min into the separation. We predicted that
10 min of separation would be sufficient for the changes in cortisol to be biologically
relevant in our study. Barber and Crowell-Davis (1994) reported 9 nursing bouts per hour

during the first week after parturition and about 4 bouts per hour at one month of age.

65

Draft mule foals nursed 14 times per hour during the first week and nursed twice an hour
by week 17 according to Smith-Funk and Crowell-Davis (1992). Given the high nursing-
bout frequency in foals during the first weeks of life, we predicted that 10-min

separations would not cause nutritional deprivation.

Repeated separations caused changes in the foals’ response to short-term
isolation. The frequency of investigating was low during the first two trials (two and four
weeks of age) and then increased during subsequent trials. Houpt (1981) reported that
foals were least affected by the separations during the first month of life. This is in
agreement with the results we found for the frequency with which foals performed
investigative behavior in our study. We speculate that “Investigating” is displayed
because the foal attempts to gather olfactory cues from the mare during her absence.
Casual observations support this hypothesis in that foals were observed to frequently
investigate and paw in areas where the dam urinated or defecated. However, pawing as

well as urinating, vocalizing and eating, did not show an effect of repeated exposures.

In contrast, the frequency of looking forward was high at first, then decreased
until trial 4 (eight weeks of age) and subsequently showed equally high concentrations
during trial 5 (10 weeks of age) and 6 (12 weeks of age). By recording this behavior we
assume that the foal establishes a link between the direction the mare disappeared in and
the direction it needs to go to find the mare. When comparing the different frequencies
per trial, we may be observing a habituation effect. The low frequencies during the

second, third, and fourth trial are in part due to the fact that the foal is engaged in other

66

behaviors such as investigating. However, for the fifth and sixth trial, at 10 and 12 weeks
of age, the frequencies for both investigating and looking forward increase, contradicting
our previous statement regarding a habituation to separations. Because increase in
activity is related to an increase in the response to stress, we propose that the last two
separations (at 10 and 12 weeks of age) were more stressful for the foals. This result is
based on behavioral indicators and supports the finding by Houpt et al. (1981) that at a
young age foals are not as affected by the separations as compared to when they are
older. Future research may be useful to investigate whether physiological indicators

correspond with the behavioral results we presented.

However, studies using physiological markers of stress found that foals show the
strongest response to maternal deprivation during early stages in life (N iezgoda &
Tischner 1995). Cortisol and catecholamines concentrations for foals showed the biggest
increase at 5 days of age and smaller responses were seen on day 20 and 50. More
specifically, on day 20, only foals showed a significant increase in cortisol whereas on
day 50, a significant increase was observed in mares but not in foals. Based on previous
findings by Cleroux et al (1985), Niezgoda and Tischner (1995) also argued that the
decreasing ratio of epinephrine over norepinephrine indicated repeated separations were
less emotional in nature. However, no behavioral evidence supporting this finding was
found in our study. There is a significant difference in protocol between our study and
that conducted by Niezgoda and Tischner (1995). The Polish pony mares were separated
from their foals on three occasions (day 5, 20, and 50) whereas in our study mares were

separated six times, every other week, starting at approximately day 14. Niezgoda and

67

Tischner saw a decrease in response to maternal separation, but our study did not reveal
such a trend. We propose that there is a critical period prior to 14 days of age, and thus
earlier than the 30-day period as argued by Houpt et al (1981) in which the foal is most
sensitive to separations. It is then during this critical period that the actual modification of
the HPA-axis occurs, leading to a decrease in response to stress during subsequent
separations. Supporting this prediction, we mention the study by Houpt et al. (1981)
where mares and foals were separated for 5 minutes every week between 1 and 12 weeks
of age. Though the response of foals to separations did not peak until 3 to 4 weeks of age,
results showed a steady decline in vocalizations and overall activity towards 12 weeks of

age.

It should be noted that we could not make statements in our study regarding
overall activity (duration of stand, walk, trot, and nursing as measured by “muzzle at
udder”). As stated above, we could not reliably record these data, and our conclusions
therefore are based on behaviors (walking, standing, and trotting excluded) previously

shown to increase during a stress response (Houpt et al. 1984; Hoffinan et al. 1995).

When mare and foal were reunited, the foal immediately nursed, supporting the
theory that nursing serves to comfort as well as nourish the foal (Crowell-Davis 1985).
Though we did have information on frequency with which foals performed the “muzzle
at udder” behavior, casual observations suggested it is important that the distinction is
made between nursing attempts and actual suckling, based on the duration of “muzzle at

udder”. Relatively speaking, one nursing bout consists of a few short “muzzle at udder”

68

sessions and one or two longer ones. Though milk ingestion was not measured, it is likely
that the longer “muzzle at udder” occurrences represent successful nursing.
Unfortunately, we could not reliably record the true duration of each “muzzle at udder”

occurrence in this study.

When comparing mare-foal interactions immediately before and alter separations,
we find that the occurrence of aggressions was not influenced by the separations.
Moreover, we found an increase in the fi'equency with which mare and foal watched each
other and with which they initiated contact. Thus, mares show greater interaction with
their foals following separation than before and this interaction is reciprocal. In addition,
proximity scores immediately after separations were significantly higher than before,
again indicating that the short-term separations intensified mare-foal interaction. We do
acknowledge that the foals were limited in the distance they could move away from the
mare due to stall confinement. However, we attribute the increase in mare-foal
interactions to the immediate effect of short-term separations. This argument is supported
by findings in rodents. After rat pups were handled by humans for a short period of time,
dams would show increased concentrations of licking and grooming of pups and arched-

back nursing (Liu et a1. 1997).

In addition, fillies tended to stay closer to their dam compared to colts during the
separation trials as well as during post-separation evaluations. We did not see any effects
of gender on other behaviors we investigated. Very little evidence exists that maternal

care is directed differently towards colts and fillies (Crowell-Davis 1986). However, this

69

significant difference in proximity, despite the limitation to move away fiom the mare
due to stall walls, suggests that fillies are more closely associated with their dam than are
colts. Fillies engage in mutual grooming behavior twice as much as colts do and prefer to
groom with other fillies and their darn (Crowell-Davis 1986). Crowell-Davis (1986) also
argued that herd dynamics have shown that mares will stay within a herd, regardless of
the presence of a stallion. Therefore, the close association between mare and filly as
reported in our study and the higher frequency of mutual grooming by fillies potentially

serves to reinforce the female-female bond.

Analyzing data from the evaluations 4-6 days post-separation of separation group
mare-foal pairs and their age-matched controls, we did not find a clear effect of treatment
(short-term separations) on mare-foal interactions. The frequency with which mares
watched their foals was not affected by whether or not mares had been separated from
their foals previously. Similarly, no effect of treatment was found on the occurrence of
contact, but where contact initiated by foals occurred mostly during early morning and
altemoon, contact initiated by mares occurred the most during early afternoon. In fact,
aggressive behaviors displayed by the mare were noticeably higher during the early
morning than compared to observations later in the day, indicating that the interactions
from the foal were not welcomed at that time. As early morning observations coincided
with pre-feeding and sometimes feeding time, this increased aggression may be
associated with the anticipation of feeding. It should be noted that aggressions by the

mare were also increased during the early afternoon, but fiiendly contact initiated by the

70

mare was higher at this time as well. It is possible that the aggression displayed during

this time block served to restrict contact-interactions initiated by the foal.

Thus, our data differs from results reported from research on rodents in that we
saw an immediate increase in maternal care due to social isolation, but this increase
disappeared soon after the separation. We propose that the explanation lies in the
evolutionary biology. Rodents are altricial mammals, born naked with no means to
sustain essential processes such as therrnoregulation and digestion, and developing more
slowly than precocial mammals, such as the horse (Kendrick et a1. 1997). Evolutionary,
horses are a follower species and it is imperative foals get up and move with the herd

within a few hours after birth.

In sheep for example, two breeds of the same species can differ in the intensity of
maternal care (Dwyer & Lawrence 1999; Dwyer & Lawrence 2000). Since maternal
styles can vary between breeds within the same species, we assume that differences exist
between species. Compared to horses, a different intensity of maternal care is required for
the rearing of young rodents. Rat mothers give constant care through physical contact
(licking, grooming, nursing) and only leave their pups sporadically for 15 to 25 rrrinutes
(Plotsky & Meaney 1993). In addition, immediate care is given upon return to the nest. In
contrast, mares do not often let the foal get out sight and are rarely in physical contact
with their foals, showing less nursing and allogrooming bouts (Crowell-Davis 1985).
Thus, although maternal care increased immediately after separation similar to findings in

rodents when dam and pup were reunited, the increase in maternal care was not sustained.

71

This could be an expression of the fundamental differences in the regulation of maternal

behavior between horses and rodents.

In conclusion, in agreement with our prediction, we found that interactions
between mares and foals intensified immediately after separations. Contrary to the second
part of our hypothesis, we did not see a prolonged effect on mare-foal interactions based
on mare-foal interactions of separated versus control animals 4-6 days post-separation.
Based on rodent literature, future research needs to look into the impact of a stressor on
both groups of foals at an older age to investigate whether the immediate increase in
maternal care following a separation leads to a dampened stress response later on in life.
Chapter V describes the study we conducted investigating how the foals from the

separation group responded to weaning when compared to the foals from the control

group.

72

CHAPTER IV: EFFECT OF SHORT-TERM SEPARATIONS ON WEANING

STRESS IN F OALS

Abstract

Weaning is a stress-causing event that may have prolonged effects on foal welfare. Based
on results found in rodents, we anticipated that intensive maternal care during early
stages in life aid foals to cope with stressors, such as weaning, later on in life. We
previously demonstrated that maternal care is intensified immediately after short-term
separations. To verify whether these separations had a long-lasting effect, this study
investigated physiological (salivary cortisol and heart rate) and behavioral markers of
weaning stress in foals that have undergone six short-term separations from their dam
compared to control foals. Measures taken during weaning indicated that separations of
mare and foal at an early age did not affect heart rate (P = 0.85), the duration of udder-
directed behaviors -“muzzle at udder”- (P = 0.46), standing (P = 0.43), walking (P =
0.22), or the frequency of vocalizing (P = 0.66), urinating (P = 0.27), eating (P = 0.12),
drinking (P = 0.87), looking forward (P = 0.96), investigating (P = 0.0.49), and pawing
(0.90). However, based on sustained elevation of salivary cortisol during the post-
weaning day we concluded that the separation group was impacted longer by the weaning

process (separation pre-weaning: 1.82 i 0.55 nmol/1, post: 5.61 i 1.13 nmol/l; control

pre: 2.75 i 0.83 nmol/l, post: 3.43 :t 1.37 nmol/l) (P = 0.066). Furthermore, fillies in the

test group defecated more than fillies in the control group (0.8 i 0.5 h.1 versus 0.1 i 0.2
- —1
h 1; t61=2.34, P = 0.023) whereas for colts the reverse was observed (0.093 i 0.15 h

versus 0.55 i 0.36 h-l; t61=-2.05, P=0.05). We also saw a main gender difference in the

73

reaction to weaning in that fillies walked longer (P = 0.039) than colts and laid down less

(P = 0.019). Finally, fillies vocalized more on the post-weaning day than colts did (18.6 i-

5.6 and 7.9 i 4.1 vocalizations h'l, respectively, P = 0.057). Thus, we did not find that

increased maternal care through short-term separations facilitated weaning. However, our
findings indicate that fillies in general are more affected by the weaning process and that

short-term separations may complicate weaning for fillies more so than for colts.

Introduction

Weaning in foals results in maternal deprivation, often social isolation, and a sudden
transition in the source of nutrients. The stressfulness of weaning has been demonstrated
using behavioral indicators, such as increase in general motor activity and vocalizations
(Houpt et al. 1984) (Hoffman et al. 1995). Physiological measurements showed an
increase in plasma cortisol concentration and a subsequent decrease in cell-mediated
immunity (Malinowski et al. 1990). Although the stress response may facilitate the
adaptation of an animal to its environment, chronic stress causes damage to the immune
system, as well as a decrease in grth and reproduction (Malinowski et al. 1990;

Compton 1987; Welsh et al. 1979).

The impact of weaning can be modulated by affecting timing of weaning and the
type of environment in which weaning occurs. When allowed fence line contact during
gradual weaning, foals are able to maintain visual contact with the mare for a certain
period of time before the complete separation occurs (McCall et al. 1985; McCall et al.

1987). In the abrupt weaning method, foals are immediately and without further contact

74

separated fi'om their dams. McCall et al. (1985) found that foals that were abruptly
weaned spent more time walking (118.8 min versus 27.5 min) and less time standing still
(162.5 min versus 272.5 min) compared to foals that had time to habituate to the
separation fi'om their dam. The suddenly weaned foals also had a higher frequency of
vocalizations compared to gradually weaned foals (93.0 versus 16.8 vocalizations in the
first hour and 39.8 versus 2.3 vocalizations in the fifth hour post-weaning) and elevated
cortisol concentrations (McCall et al. 1987, 153.18 i 21.53 nmon versus 56.58 i 21.53
nmol/l). Thus, by prolonging the weaning process through fence line contact, the impact

of weaning can be reduced.

Research on social enrichment of the weaning environment reports conflicting
results. Houpt et al. (1984) concluded that paired foals were less impacted by weaning
than individually weaned foals as measured by a decrease in vocalizations (30.9 i 7.2
vocalizations per 15 min for animals weaned individually versus 14 i 3.7 vocalizations
per 15 min). However, Hoffman et al. (1995) found that foals weaned in pairs displayed
more aggression towards each other. In addition, the increase in serum cortisol in
response to an adrenocorticotropic hormone (ACTH) challenge was less in foals weaned
in pairs (165.6 a 11.04 mnol/l) than in individually weaned foals (190.44 a 11.04 nmol/l),
indicating depletion of adrenal reserves in foals weaned in pairs. A study by Malinowski
et al. (1990) found that foals weaned in pairs showed a decrease in lymphocyte
proliferation, representing a decrease in immune response associated with paired
weaning. Thus, contrary to what Houpt et al. (1984) concluded, it is beneficial to wean

foals without social enrichment when weaning in a stall. A major contributing factor to

75

 

this result is potentially the confinement, which limits the ability of animals to move
away when aggression is displayed against them. In fact, research has shown that
weaning groups of animals on pasture, i.e., combining social enrichment and
environmental enrichment, is preferred over weaning in an isolated, simple environment
such as a stall. Comparing six individually stall-weaned versus six pasture-weaned foals
(two groups of three foals), Heleski et al. (in press) found that paddock-weaned, group-
housed animals could engage in a broader range of activities, provided by social (e. g.,
mutual grooming) and environmental enrichment (e. g., grazing), and showed less
aberrant behaviours as measured by e. g., sniffing, licking, and kicking the stall/shed

walls.

From the aforementioned results, we conclude that the response to weaning can
be modified by several factors. The mother-infant relationship is thought to be an
important factor in the way the HPA-axis develops (Plotsky 2000). The darn-offspring
“security of attachment”, introduced in the late 1970’s, states that both mare and foal
develop the boundaries of their relationship together and they can rely on a certain
amount of predictability when interacting with each other (Bowlby 1977). This implies
that if the care of the mother towards the infant is inadequate, the development of the
offspring is impaired (Liu et al. 1997). This impairment in an animal can lead to a skewed
perception and evaluation of the immediate environment, thereby affecting stress coping
strategies. As shown in rodents, the amount of maternal care received during early stages
in the life has a long-term effect on stress coping strategies of the animals as adults (Liu

et al. 1997). It has been shown that rat pups who received intensive maternal care

76

characterized by high licking and grooming, and arched-back nursing responded less
intensely and for a shorter period of time to restraint stress as adults when compared to
rat pups that received low maternal care characterized by low licking and grooming, and
arched-back nursing. We predicted that, by reinforcing the natural mare-foal bond, we
could facilitate weaning in foals. Maternal care in horses can potentially be stimulated by

a threat to the foal or by a short-term separation (Crowell-Davis & Houpt 1986).

In chapter IH, we demonstrated that mare-foal interactions do intensify
immediately after a separation of short duration. Contrary to what we stated in our
hypothesis, we were unable to show, through post-separation evaluations of test and
control group animals, that the increase in maternal care was still apparent a few days
after separation. In research performed on rodents, researchers showed that high levels of
maternal care had an effect on rats as adults when a stressor was applied. To compliment
findings in chapter HI, we needed to apply a stressor on the foals of the separation group
and of the control group to investigate whether the animals in each treatment respond
differently. To measure concentrations of a common marker of stress, i.e., cortisol, blood
or saliva are commonly used. When a foal is weaned, taking blood samples becomes
more challenging and this implies a greater risk for animal and handler. The use of saliva
to measure cortisol in horses has been validated previously (Lebelt et al. 1996; van der

Kolk et al. 2001; Warren et al. 2001) and we prefer this less invasive method.

The objective of this study was to investigate behavioral and physiological

markers of weaning stress for foals that had undergone short-term separations versus

77

control foals that had never been separated from their mare prior to weaning. We
hypothesized that foals exposed to short-term separations early in life would regain
homeostasis faster after weaning when compared to foals that were never separated from

their mother.

Materials and methods

Animals and treatment prior to weaning

This research was conducted at MSU Merillat Equine Center from January through
October 2000. A total of 10 mare-foal pairs were studied (Appendix C). The animals
were divided into two groups blocked by gender. Each consisted of three fillies and two
colts. Each group represented a treatment. In the control group, mare and foal were never
separated prior to weaning. In the separation group, every two weeks, starting when the
foals were 2 weeks old until 3 months of age, the mares in the were separated from their
foals for 10 min. Each mare-foal pair in the test group was separated six times (Chapter
III). The MSU All University Committee on Animal Use and Care reviewed and

approved the experimental protocols described in this chapter.

Housing and feeding
From January through April 2000, mares were individually housed in Barn A in 13.38 m2

box stalls. The stalls were bedded with wood shavings and had a combination feeder in
one corner. The mares had ad libitum access to water via an automated drinker. By May
2000, all mares had foaled and the animals were moved to the climate controlled Barn B.

The climate-controlled environment was ensured by the availability of fans in the barn. In

78

addition, during the day, mares and foals remained together on pasture from the time
foals were approximately one month old. All mares were fed twice daily. The morning
feeding occurred between 08:00 and 10:00 and the afternoon feeding between 15:00 and

17:00. Horses were fed according to NRC recommendations (Council 1989).

Weaning protocol
Weaning in test and control groups occurred when the foals were 6 i 0.5 months old.
During the weaning trials we evaluated mare-foal interactions and observed the

physiological and behavioral response of foals to weaning.

One weaning trial consisted of three days: pre-weaning, weaning, and post-
weaning data collection day. The experimental timeline is described in Appendix G. Per

weaning trial we weaned a foal from the control group with a foal of similar age from the

test group. Foals were weaned in individual box stalls measuring 20.90 m2 and were

deprived of visual contact with one another.

The choice of weaning stalls was based on their location relative to the breeding
room. To place and check heart rate monitors or to take saliva samples from the foal for
cortisol measurements, the mare was led out (and the foal would follow) to the breeding
room where mare and foal were placed in a metal chute with which the animals were
familiar. Alter weaning, when the mare was no longer present, the stall door was opened

and two to three people directed the foal towards the chute. Some foals were extremely

79

difficult to handle after weaning and required a lead rope across the withers to prevent the

animal from jumping out of the chute.

During each weaning trial, we collected saliva samples for cortisol analysis,

measured heart rate and collected behavioral data as described below.

Salivary cortisol

mg

For saliva sampling, an in-house constructed saliva collection device (SCD) (figure 16)
was used. On average 17.5 cm2 of Tygon® tubing (R-3603, Norton Performance Plastics
Corporation, Akron, OH 44309) was used and 8 to 13 holes were cut in the tube. Gauze
was cut into separate pieces, rolled and, by using a metal wire, inserted into the tubes. At
both ends of the tube, a suspender clip was attached which in turn was connected to a
swivel-head clip using elastic ribbon that was adapted in length to comfortably fit the
foal’s head (figure 17). The swivel-head clips were attached to the halter and the tube was
placed inside the foal’s mouth. The animal immediately began to chew, thereby
stimulating saliva production. After collection, the gauze was removed from the tube,
folded in half and placed in 15 m1 polypropylene tubes. The cap was placed on the tubes

to keep the gauze in place and the samples were stored on ice.

80

 

Figure 16: Saliva collection device to collect saliva from foals. Gauze is placed in the
tubing with holes to provide the foal with a substrate to chew and to collect
produced saliva. The elastic band allows for adaptation of the device to different
foal head sizes.

 

Figure 17: Saliva collection device in place in a foal’s mouth. The position of the
device depended on the foal’s ability to manipulate the tube to parts of the mouth
where it might chew through the tubing.

81

Anemia

The tubes holding the saliva-soaked gauze were centrifuged in a clinical centrifuge model
CL (International Equipment Company, Needham, Massachusetts) at 4 104 x g. The
saliva was then transferred to cryovials and placed in liquid nitrogen at -l95 °C as this
was logistically more feasible than finding storage at —30 °C. To quantify the amount of
cortisol in the saliva, we used an Active Cortisol EIA (for saliva) kit prepared to analyze
human salivary cortisol by Diagnostic Systems Laboratory (DSL-10-67100, Webster,
Texas 77598). Briefly, in a 96-well microtiter plate coated with goat anti-rabbit globulin
serum, we pipetted in duplicate 50 ul of sample, 100 1.11 of enzyme conjugate solution and
100 pl rabbit anti-cortisol serum. After incubation for 45 min on a plate shaker, we
washed the wells five times with buffered saline wash solution and added 100 pl
tetramethylbenzidine chromogen solution. After another 15 to 20 min incubation time,
we pipetted 100 1.11 0.2 M sulfuric acid into the wells to stop the color reaction. The
absorbance of the wells was read at 450 nm. From previous research we found that foal
saliva has a much lower cortisol concentration than human saliva; we modified the
standards so they ranged from 0 to 4 ug/dl or 0 to 110.4 nmol/l. We plated the kit’s
quality controls (QC 1 = 0.2 ug/dl and QC 2 = 2.0 ug/dl or 5.52 nmon and 55.2 nmol/l)
in the wells following the standards and again in the last 4 wells of the plate. The samples
were plated after the first set of quality controls. We then included eight wells of pooled
foal saliva collected from different foals at the MSU Horse Research and Teaching
Center. Four wells contained saliva collected during the morning (QC 3 and QC 4) and

another four wells contained saliva collected during the aftemoon (QC 5 and QC 6). We

82

 

assayed one plate per day for 4 days and used the quality controls to determine inter-

assay variability.

After log-transformation of the data, the factors of day (pre-weaning, weaning, and
post-weaning data collection day) and time (morning: 07:00-09:00, noon: 11:00-13:00,
afternoon: 15:00-17:00) were used as repeated measures within mixed model procedures
of SAS with foal nested within treatment in the subject statement. The fixed effects in the
model consisted of gender, treatment, day, and time whereas the random effect was trial
(n=5). Significant results were further investigated using a multiple comparisons t-test.
When no difference was found between separation and control group, data were
combined for further analysis. The results are represented as lsmeans i SEM, back

transformed to the observed scale.

Heart rate

Techniques

To measure heart rate, we used a Polar Vantage XL telemetric system (Polar Electro Inc.,
Port Washington, NY 11050, USA). A transmitter was connected to a custom made
elastic belt with a plastic clip. We wet the foal’s coat where the transmitter would be on
the heart girth, and applied Meditrace conductivity gel (Graphic Controls, Buffalo, NY
14240). A receiver watch was secured to the elastic belt and programmed to record the
beats per minute every 15 sec. Up to 8 h of data could be stored on one receiver and the

receiver was replaced either after 4 or 8 h of data recording, or after the receiver had

83

accidentally stopped recording. Heart rate was recorded from 06:00 until 18:00 during

three consecutive weaning trial days.

Analysis

Per two hours of heart rate monitoring, a maximum of 480 measurements were recorded.
Because the heart rate monitors did not always stay attached to the foal or because the
transmission was hindered due to a loose belt, some measurements were lost or invalid.
We averaged the data available per two hours of observation and in the analysis we
weighted the number of values used to obtain each average. We opted to use the mean
instead of the median because we predicted that the few peaks that occurred would not
significantly influence the average of all the observations and the mean gives us easily

interpretable results.

After log-transformation of the data, the factors of day (pre-weaning, weaning, and
post-weaning data collection day) and time (morning: 07 :00-09200, noon: 11:00-13200,
afiemoon: 15:00-17:00) were used as repeated measures within mixed model procedures
of SAS with foal nested within treatment in the subject statement. The fixed effects in the
model consisted of gender, treatment, day, and time whereas the random effect was trial
(n=5). Significant results were further investigated using a multiple comparisons t-test.
When no difference was found between separation and control group, data were
combined for further analysis. The results are represented as lsmeans i SEM, back

transformed to the observed scale.

84

Behavior

Techniques

Per stall, a Panasonic CCTV WV-BP334 video camera was installed out of the animals’
reach. Both cameras were connected to a Panasonic WV-CM146 field switcher monitor.
In addition to the live observations, we recorded the animals on VHS tape using a
Panasonic AG-6740 VCR. In addition, an Optimus Unidirectional Condenser microphone

was placed in each weaning stall and the audio data were also recorded using the VCR.

Each day of the weaning trial, the foals were observed during three two-hour
observation blocks (07:00 to 09:00; 11:00 to 13:00; 15:00 to 17:00). On the weaning day,
the foals were weaned at approximately 07:00. The weaning procedure was identical to
the short-term separation protocol as described in chapter IV, except that the mare was
not brought back to the stall after 10 min. Instead, the handler took her to a pasture
approximately 400 m away from the breeding barn. The ‘Dutch doors’ were shut prior to
weaning. The ethogram with the behaviors observed during the weaning trial is described
in Appendix F. The behaviors retained for analysis included: duration of “muzzle at
udder” (on pre-weaning day), stand, walk, and lie down, as well as the frequency of
vocalize, defecate, urinate, eat, drink, look forward, investigate, and paw. It should be
noted that nursing behavior was included in “muzzle at udder”. We assumed that mare-
foal interactions are reflected in whether or not a mare allows her foal to nuzzle the
udder, with or without actual ingestion of milk. The location of our observation station
made it impossible to reliably detect sucking, and thus our definition of a nursing bout is

closest to the definition by Crowell-Davis based on observations in 1979 (see Appendix

85

A). For the same reason, no distinction will be made between successful and unsuccessful
nursing bouts. Smith-Funk (1992) described successful nursing bouts as including pre-
nurse nuzzling, sucking, infra-bout pauses and intra-bout nuzzling. Unsuccessful nursing
bouts, according to the same author, did not include sucking, which again would have

been impossible to reliably detect in our experimental setup.

Analysis

Day (pre-weaning, weaning, and post-weaning data collection day) and time (morning:

 

07:00-09:00, noon: 11:00-13:00, afternoon: 15:00-17:00) were used as repeated measures l
within mixed model procedures of SAS with foal nested within treatment in the subject

statement. The fixed effects in the model consisted of gender, treatment, day, and time

whereas the random effect was trial (n=5). Significant results were further investigated

using multiple comparisons t-test. Frequencies obtained were regarded as count data and

square root transformations were performed prior to analysis. When no difference was

found between separation and control group, data were combined for further analysis.

The results are represented as lsmeans :2 SEM, back transformed to the observed scale.

Results

Salivary cortisol

A total of four plates were assayed with respective within-assay coefficients of variation
of 9.2 %, 3.7 %, 3.2 %, and 3.7 %. Between-assay variation coefficient of variation was
9.8 %. Salivary cortisol concentrations ranged from 0.17 to 101.02 nmol/l. However,

after testing for outliers, the range of salivary cortisol concentration varied between 0.17

86

and 45.13 nmol/l. The results did not reveal a main effect of treatment (ANOVA,

F1,66=0-04a P = 0.89). There was a significant day by time interaction (ANOVA,

F 6,66=7-67a P < 0.0001), as shown in Fig. 18.

 

 

 

% 30.00 -

E .-

5

"’5) Dpre

E Iwean
ost

Z: P

(U

.2

E

a)

06:20 10:20 14:20 18:20

Figure 18. Mean salivary cortisol concentration :I: SEM per sampling time for each
weaning trial day (N =10, P < 0.05). Differences as indicated by the letters (a,b; c,d;
e,f; g,h) are reported within time blocks.

87

The cortisol concentration at 10:20 was higher on the weaning day than when compared
to pre- (t66=-5.75, P < 0.0001) or post-weaning day (t66=3.04, P = 0.0034). Similarly, for
14:20 and 18:20, pre-weaning cortisol concentrations were significantly lower when
compared to weaning day concentrations (14:20: t66=-3.36, P = 0.0013; 18:20: t66=-4.71,
P < 0.0001). Pre-weaning and post—weaning concentrations were significantly different
for the 18:20 sample (t66=-2.65, P = 0.010), but not for the 14:20 measurement 066;

1.45, P = 0.15).

In addition, we investigated the presence of diurnal variation during the pre-
weaning day as represented in the day by time interaction. Only one significant
difference was found in the cortisol concentrations for sampling at different time points.

The 6:20 sample showed a significantly higher concentration than the 18:20 sample (3.71
i 1.54 nmol/l versus 1.77 i 0.77 nmol/l; t66=2.21, P = 0.030). Another trend was found
when comparing the cortisol concentrations from the 10:20 (3.10 i 1.33 nmol/l) and

18:20 sample (t66=1.81, P = 0.075). The cortisol concentrations at 06:20 for pre-weaning

and weaning day are not significantly different (166:0.71, P = 0.48). This is to be

expected, as weaning on weaning day did not occur until 07:00. All samples taken after
weaning show higher cortisol concentrations when compared to the pre-weaning sample

on the weaning day, though for the 18:20 sample, this approaches significance (10:20

t66=-5.00, P < 0.0001; 14:20 t66=-2.08, P = 0.041; 18:20t66=-1.91, 0.061). The cortisol

concentration at 10:20 is higher than those at 14:20 (t66=2.82, P = 0.0064) and 18:20

88

(t66=3.00, P = 0.0039). Finally, concentrations of cortisol did not differ between 14:20

and 18:20 (t66=0.l7, P = 0.87). On the day post-weaning, we found a high cortisol

concentration at 06:20 (12.51 i 7.21 nmol/1), which was significantly higher than the

concentration at 10:20 (t66=2.32, P = 0.024), 14:20 (t66=2.03, P = 0.046), and tended to

be higher than 18:20 (166:1 .87, P = 0.066). No differences were found between other

time blocks during the post-weaning day.

Heart Rate

Heart rate during the pre-weaning day averaged 59.60 i 3.40 BPM. This average
increased on the weaning day to 78.00 i 4.31 BPM and decreased to 71.37 i 4.51 during
the post-weaning data collection day. The main effect of treatment was not significant
(ANOVA, F156 = 0.03, P = 0.85) and no other interactions involving treatment were
found to be significant. Similar to the salivary cortisol results, there was a significant day

by time interaction (ANOVA, F456 = 6.79, P = 0.0002), as shown in Fig. 19.

89

El pre
l wean
post

Heart rate (bpm)

 

 

07:00-09:00 11:00-13:00 15:00-17:00
Observation block

Figure 19. Mean heart rate i SEM per observation block for each weaning trial day.
(N=10, P < 0.05). Differences are indicated by the letters (a,b,c).

The mean heart rate did not differ between weaning and post-weaning days except
for the first time block (t56=3.86, P = 0.0003). Weaning caused a significant increase in

mean heart rate for each time block when comparing pre-weaning to weaning heart rate

(07:00-09:00 t56=-8.l2, P < 0.0001; 11:00-13:00 t56=-3.51, P = 0.0009; 15:00-17:00

t56=~4.36, P < 0.0001). Similarly, pre-weaning and post-weaning heart rate

concentrations were different for all time blocks rate (07:00-09:00 t56=-3.89, P = 0.0003;

11:00-13:00 t56=-3.75, P = 0.0004; 15:00-17:00 t56=-4.15, P = 0.0001).

90

Behavioral results

Muzzle at udder

Because we couldn’t report on nursing in chapter IV, duration of nursing (muzzle at
udder) was analyzed during the pre-weaning day to test for pre-weaning treatment
differences. At time 07:00 on the weaning day the mare was removed from the stall. The

separations did not cause a difference in the time during which the foal touched the

mare’s udder (test: 1.76 i 0.82 min, control: 2.59 i- 0.82 min, F1,16=0.58, P = 0.46). The

other main effects and two-way interactions investigated were not found to be significant.

 

M

The duration of standing over the course of an hour on the pre-weaning day ranged from
2.85 to 60 min (Mean 47.49 i 3.08 min) whereas during the weaning and post-weaning
day, the values ranged from 22.19 to 60 min (Mean weaning: 46.04 i 4.00 min, Mean

post-weaning: 41.27 i 1.84 min). No effect of treatment was found on the time during

which foals were standing (test: 43.50 i276 min; control: 46.62 i 2.86, F1,66=0.62, P =
0.43). There was a trend for a gender by day interaction (F 2,66=2-77, P = 0.070): fillies

stood longer on day 1 when compared to day 2 (166:1.98, P = 0.0515) and 3 (t66=3.14, P

= 0.0025).

Walk
The duration of walking over the course of an hour on the pre-weaning day ranged from

0.61 to 37.10 min (Mean: 3.00 i 1.11 min). After weaning, the duration of walking

91

ranged from 0.42 to 58.60 min (Mean weaning day: 9.01 i 1.91 min, Mean post-weaning

day: 14.73 i 1.62 min). No main effect of treatment was found for the duration of

walking (F1,66=1-49a P = 0.23). However, a main effect of day was present, indicating the

effect of weaning (F 2,66=16-03a P < 0.0001). Figure 20 shows the main effect of day. All

days proved to be significantly different from one another, with increasing duration

between consecutive days.

 

 

 

   

g 18 c
g 1
2’
g; 12 -
i
E
o
1: 6 "
.2
E
3 0 4 .
pre weaning post
Day

Figure 20. Mean duration :1: SEM of walking per weaning trial day. (N=10, P <
0.05). Differences are indicated by the letters (a,b,c).

92

The interaction between gender and time also showed to be significant (F 2,66=3.42, P =

0.039) and is represented in Fig. 21.

 

 

 

   

 

 

 

 

 

 

 

7:00-9:00 11:00-13:00 15:00-17:00
Observation block

1':~

'g Elfilly
E15 — a c e lcolt
.E 'r

x

a 10 ~ '1'

3

"6 5"

C

.3 0 . .

E

3

D

Figure 21. Mean duration i SEM of walking per observation block for each gender.
(N=10, P < 0.05). Differences as indicated by the letters (a,b; c,d; e,f) are reported
within observation block.

There is no difference for the first time block between genders (fillies: 9.96 i 1.19
min, colts: 7.33 i 1.44 min, P = 0.15), but for both subsequent time blocks, fillies walk

more than colts do (fillies: 13.07 :t 1.19 min, colts: 6.18 i 1.44 min, P = 0.0003 and

fillies: 11.59 i 1.19 min, colts: 5.33 i 1.44 min, P = 0.0010, respectively).

Lfl'gg down
During the day before weaning, foals laid down between 0 and 20.33 min (Mean: 8.07 i
2.05 min). During the day of weaning, the foals laid down between 0 and 42.75 min per

hour (Mean: 0.31 i 0.67 min). The range on the post-weaning day was 0 to 32.35 min per

hour (Mean: 2.30 i 1.29 min). The short-term separations showed a tendency to increase

93

the time during which foals laid down (F1,66=3-44, P = 0.068). However, there was a

dominating day by treatment interaction (F2,66=8.24, P = 0.0006), which showed that
only during the pre-weaning day there was a significant higher duration for the test group

when compared to the control animals (t66=3.47, P = 0.0009). This interaction is

represented in Fig. 22.

 

E

E.

: ltest
; 15 a

o Elcontrol
'u

a) 10 b

.E

5' 5

° fl

8 0 1 T _T_ 1

“E pre wean post

8 Day

Figure 22. Mean duration 5: SEM of lying down per weaning trial day for each
treatment group. (N=10, P < 0.05). Differences, as indicated by the letters (a,b) are
reported within day.

94

In addition, there was a difference between genders when looking at the duration

oflying down per day (F2,66=4-22, P = 0.019), Fig. 23.

 

 

 

 

 

 

E

E

E P < 0.1 Elfilly

B 15 ‘ Icolt

8 u
a) 10 i

.5

.7: 5 ‘ c d _
g o , +i

"E pre wean post

3 Day

Figure 23. Mean duration 1 SEM of lying down per weaning trial day for each
gender. (N=10, P < 0.05, except where indicated). Differences, as indicated by the
letters (a,b and c,d) are reported within day.

As the graph shows, fillies lie down less during the pre-weaning day and tend to
lie down less on the weaning day as well. No difference between genders was found on

the post-weaning day.

95 i

Vocalizing

The range of vocalizations on pre-weaning day was 0 to 7 vocalizations per hour (Mean:

0.040 i- 0.11 hul). On the weaning day the number of vocalizations per hour ranged from
2 to 181 (Mean: 36.15 i 7.99 h-1) and during the post-weaning day the number ranged
from 0 to 63 vocalizations per hour (Mean: 12.68 i 4.08 h-l). The main effect of
treatment did not affect the frequency of vocalizations (F 1,61=O-19, P = 0.66). We did

find a trend in the gender by treatment interaction (F 1,61=2.89, P = 0.094) and after

investigation we found that within the control group, fillies vocalized more than colts

(t61=2.58, P = 0.012). There was also a gender by day interaction (F2,61=4.08, P =

0.022) indicating that fillies tended to vocalize more than colts in particular on the post-

weaning day (t61=l .94, P = 0.057). Finally, there was a significant day by time

interaction (F4,61=4.50, P = 0.0030), which is represented in Fig. 24.

96

[:1 pre

 

'0- ‘:‘715O “ Iwean
o 5 c
5‘ 5500 - lpost
g .s
g E
2 3
o
u. >

 

7:00-9:00 11:00-13:00 15:00-17:00
Observation block

Figure 24. Mean frequency i SEM of vocalizations per time block for each weaning
trial day. (N=10, P < 0.05). Differences, as indicated by the letters (a,b,c; d,e; f,g) are
reported within observation block.

There were significantly more vocalizations on the weaning day than when
compared to pre- and post-weaning days for all time blocks. Results from the first
observation block (07:00-09:00) showed the only weaning versus post-weaning

difference of all the time blocks.

Defecating

The frequency of defecating ranged from 0 to 1 per hour on the pre-weaning day (Mean:

0.09 i 0.13 h-l). For the weaning day, defecating frequency ranged from 0 to 6 (Mean:
0.56 i 0.33 h-l). For the post-weaning day, defecating fiequency ranged from 0 to 3
defecations per hour (Mean: 0.46 i 0.32 h-l). There was a significant gender by treatment
interaction (F1,61=7.87, P = 0.0067): For fillies, the test group animals defecated more
often than the control animals (161:2.34, P = 0.023), whereas for colts, the reverse was

97

 

true (t61=-2.05, P = 0.0045). In addition, within the test group, fillies defecated more than

colts (t61 = 2.33, P = 0.023), but in the control group fillies defecated less than colts (t61

= -2.13, P = 0.037). The results also showed a significant day by time interaction, as

represented in Fig. 25.

NOD-P.

Frequency of
defecating (h'1)

 

 

0

7:00-9:00 11:00-13:00 15:00-17:00
Observation block

Figure 25. Mean frequency :1: SEM of defecation per observation block for each
weaning trial day. (N=10, P < 0.05). Differences, as indicated by the letters (a,b) are
reported within observation blocks.

U_rirra_tin5g

The range of urinating frequency on pre-weaning day is 0 to 5 per hour (Mean: 0.18 i

0.12 h'l). During the weaning day, the frequency ranged between 0 and 4 (Mean: 0.65 i
0.30 hJ) and for the post-weaning day, the frequency ranged fiom 0 to 3 per hour (Mean:

0.43 i 0.19). There was no significant main effect of treatment (F1,66=1 .26, P = 0.27).

No other interactions involving treatment were significant. The main effect of time did

98

 

show significance (F2,65=4.25, P = 0.018). The highest frequency of urinating occurred

during 15:00-17:00 and this number was significantly higher than when compared to

07:00-09:00 (t66=2.75, P = 0.0077) and to 11:00-13:00 (166:2.21, P = 0.030).

Egting and Drinkflg

We did not collect data on the duration of time spent eating and drinking. Hence, the
frequencies of eating and drinking, as indicated by the definitions in Appendix F,
represent the number of visits to the feeder or drinker. Data on nursing, as recorded by
“muzzle at udder” on the pre-weaning day, was collected separately from eating and

drinking data. Frequency of eating on the day prior to weaning ranged from 0 to 26 per

hour (Mean: 3.36 i 1.81 h-1). The frequency on the weaning day ranged between 0 and
69 (Mean: 10.15 i 3.19 h-1) whereas for the post-weaning observations, the count ranged
between 1 and 57 (Mean: 18.95 i 3.59 H1). For drinking, the ranges for the frequency
were 0 to 10 for pre-weaning (Mean: 0.48 i 0.27 h-l), 0 to 7 for weaning (Mean: 0.39 i
0.25 h-l), and between 1 and 8 for post-weaning observations (Mean: 2.95 i 0.54 h-l).
Both behaviors showed a main effect of day, but no effect of treatment (Eat: F1,66=2-52,

P = 0.12; Drink: F1,66=0-03a P = 0.87). The main effect of day is represented in Fig. 26.

For both behaviors, a significant main effect of time was also present for both behaviors

and this is represented in Fig. 27.

99

   

A 30 w, Elpre
'35 Iwean
E. 20 Ipost
g 10

8

h

‘L 0

Figure 26. Mean frequency i SEM of eating and drinking for each weaning trial
day. (N=10, P < 0.05). Differences, as indicated by the letters (a,b,c and d,e) are
reported within each behavior.

.7200 to 9:00
11:00 to 13:00
[115200 to 17:00

N
O

 
 
  

.3
01

Frequency (h'1)
01 'o‘

0

Eating Drinking

Figure 27. Mean frequency :1: SEM of eating and drinking per observation block.
(N=10, P < 0.05). Differences, as indicated by the letters (a,b and c,d), are reported
within each behavior.

100

yr .fi‘!“ 4 .5

Finally, for drinking there was a significant main effect of gender (F1,66=8.94, P
= 0.0039), which is represented in Fig. 28. Overall, fillies drink more often than colts do

066:2.99, P = 0.0039).

 

 

 

   

 

*7“ 2.5 -
5
2’ 2 -
%
'c 1.5 —
3 b
O _
> 1
8
o 0.5 ~
3
u-
s o— .
filly colt
Gender

Figure 28. Mean frequency 1: SEM of drinking for each gender. (N=10, P < 0.05).
Differences are indicated by the letters (a,b).

Looking forward
The frequency for looking forward ranged from 0 to 26 per hour on the day before

weaning (Mean: 5.91 i1 .76 h-l). During weaning day, this range was 0 to 118 (Mean:

29.81 :t 6.25 h-1) and during post-weaning observations, the frequencies ranged between

7 and 81 per hour (Mean: 32.31 i 4.96 h'l). Overall, separations did not affect the

101

 

frequency of looking forward (F1,61=0.00, P = 0.96), but there was an interaction

between gender and treatment (F “51:17.93, P < 0.0001) as represented in Fig. 29.

 

 

 

 

 

 

 

g ltest
‘5 A 40 — Elcontrol
o ‘7
:6 5 30 — d
a, g 20 ‘
g g 10 4
3 to-
g 0 ' l m
u. filly colt
Gender

Figure 29. Mean frequency 3: SEM of looking forward per gender for each
treatment group. (N=10, P < 0.05). Differences, as indicated by the letters (a,b and

c,d), are reported within gender.

Test group fillies look forward more frequently than the control animals, but colts
that have been separated look forward less than the animals in the control group.

In addition to the significant gender by treatment interaction, we found a significant day

by time interaction (F4,61=3.77, P = 0.0083). This interaction is represented in Fig. 30.

102

Pub“ L116

 

 

 

Frequency of looking
forward (11")

7:00-9:00 11:00-13:00 15:00-17:00
Observation block

Figure 30. Mean frequency :1: SEM of looking forward per observation block for
each weaning trial day. (N=10, P < 0.05). Differences, as indicated by the letters (a,b;
c,d; e,l), are reported within observation blocks.

Weaning caused a significant increase in the frequency of looking forward for the
first and third time block between pre-weaning and weaning day. Weaning and post-

weaning days did not show significant differences for all three time blocks.

Investigating

The frequency-range for investigating on the pre-weaning day was 0 to 35 (Mean: 11.03

i 2.32 h-1). During the weaning day, the frequencies ranged between 1 and 78 per hour
(Mean: 23.33 i 3.98 h-l), whereas for the post-weaning day, values ranged from 2 to 49
per hour (Mean: 18.95 i 2.79 h-l). No significant effect of treatment was found for the

frequency of investigating (F1,61=O.56, P = 0.46). As for other events, a significant day

by time interaction was present (F4,61=10.02, P < 0.0001), see Fig. 31.

103

 

1:] pre
I wean
post

 

 

 

7:00-9:00 11:00-13:00 15:00-17:00
Observation block

Frequency of investigating
(h 1)
N J:- O) on
o o o o c

Figure 31. Mean frequency i SEM of investigating per observation block for each
weaning trial day. (N=10, P < 0.05). Differences, as indicated by the letters (a,b), are
reported within each observation block.

Weaning caused a significant increase in the frequency of investigating for the
first time block compared to pre— and post-weaning values. Pre- and post-weaning

concentrations were not significantly different.

Pawin

Pawing frequencies ranged between 0 and 40 occurrences per hour for the pre-weaning

observations (Mean: 1 .41 i 1.15 h-l), whereas for weaning observations values ranged
between 0 and 279 per hour (Mean: 19.40 i 6.22 h'l) and for post-weaning frequencies
between 0 and 254 (Mean: 5.17 i 5.00 h'l). Short-term separations had no effect on the
frequency of pawing (F1,61=0.O2, P = 0.98). Fig. 32 shows a significant day by time

interaction F4,61=10.93, P < 0.0001) for the fi'equency of pawing.

104

 

 

 

.7“ Upre
5 120 q b Iwean
g Ipost
g 80 -

I.—

2, 40 «

2

g- 0 __ I 1' t
E 7:00-9:00 11:00-13:00 15:00-17:00

Observation block

Figure 32. Mean frequency 1 SEM of pawing per observation block for each
weaning trial day. (N=10, P < 0.05). Differences, as indicated by the letters (a,b), are
reported within each observation block.

Similar to results found for investigating, during the first observation block

following weaning, pawing increased fi'om pre-weaning to weaning day (4.71 1|: 3.48 h.1
versus 85.20 i 21.46 h-l, t61=-6.02, P < 0.0001) and decreased on post-weaning day
(7.77 i 10.07 h], t61=2.77, P = 0.0075) to a frequency not significantly different from

pre-weaning values (t61=-0.30, P = 0.76).

105

Discussion and Conclusion

Salivary cortisol

Cortisol concentrations showed a significant increase due to weaning and this is in
agreement with other studies who investigated changes in plasma or serum total cortisol
(Malinowski et al. 1990, McCall et al. 1987), except for a preliminary study conducted by
Houpt et al. (1984). No changes in plasma cortisol were found after weaning, but it
should be noted that the researchers relied on four jugular blood samples (0 h, 6 h, 24 h,
and 1 week post-weaning). Due to the circadian pattern, only the 0 h and 24h should be
compared, but nonetheless, confounding factors associated with blood sampling may

have masked the effect of weaning.

In addition, we did not find a gender difference in the salivary cortisol response to
weaning. These results are in agreement with Hoffman et al. (1995) who, despite a pre-
ACTH gender difference in plasma cortisol (fillies having higher concentrations than
colts), did not see a difference between colts and fillies in the adrenal response to an

ACTH-challenge.

When investigating diurnal variation on salivary cortisol, we found that during the
pre-weaning day the 6:20 and the 18:20 sample showed different cortisol concentrations,
and that concentrations at 10:20 and 18:20 tended to differ as well. This result is in
agreement with Lebelt et al. (1996) who reported that the highest concentrations of
salivary cortisol were found in the morning and that the lowest concentrations occurred

during late afternoon. Similar results were reported in our validation study described in

106

chapter 111. Though results were similar, Van der Kolk et al. (2001) did not find a
significant difference in cortisol concentrations between 06:00 and 18:00. Lebelt et al.
(1996) indicated that minimal environmental disturbance may cause a disruption in the
circadian pattern of cortisol in horses. Our results support this finding in that on the
weaning day, cortisol concentrations remained elevated after weaning compared to the
pre-weaning 06:20 sample. In addition, the concentration of cortisol measured at 06:20
on the post-weaning data collection day was remarkably higher than the concentrations at
10:20, 14:20 and 18:20. We speculate that this high cortisol concentration is due to two
factors. First of all, only 5 samples out of 10 (2 from separation group, 3 from control
group) could be used for the analysis and second, we measured high concentration of
cortisol for one animal at 06:20 on the post-weaning day. This caused the average to shift
towards higher values, but the standard error indicates that the actual average can be as
low as 5.3 nmol/l. Because this one influential value was not considered to be a statistical

outlier, we chose not to remove it from the analysis.

Heart rate

Our results show that average heart rate was fairly constant between time blocks within
days, the only exception being the first observation block on the weaning day. Overall,
average heart rate increased from pre-weaning to weaning day. The immediate effect of
weaning became apparent in the decrease in heart rate from weaning to post-weaning day
during the first observation block, whereas for subsequent time blocks, weaning and post-
weaning concentrations did not differ. Niezgoda and Tischner (1995) found a maximal

increase in catecholamines 5 min after separation of mare from foal and this may explain

107

the increase in heart rate caused by weaning. However, the concentration of activity
increased as well due to weaning and it is possible to attribute the increase in heart rate to
behavioral changes. To the best of our knowledge, no other research has measured heart
rate in foals during the weaning process.

Behavior

The duration of “muzzle at udder” between test group and control group animals did not
differ. Duration for the test group animals was 1.76 i 0.82 and for control animals 2.59 i
0.82. Nursing bouts in our study were longer than those found in Belgian mares (Barber
& Crowell-Davis 1994). We believe that the discrepancy is due to the difference in the
definition of start and end of nursing bouts, which does not allow us to compare duration
of nursing bouts. However, in agreement with the nursing definition of Barber and
Crowell-Davis (1994) we saw that each nursing bout usually consisted of several “muzzle
at udder” occurrences of short duration (< 10 sec) and one or two longer ones. Casual
observations suggested that the mare controlled the duration of the attempts and that
perseverance by the foal led to longer, successfiil nursing bouts. We believe further
research focusing on the behaviors displayed by the mare during nursing bouts is
warranted, as these behaviors control nursing by the foal and therefore may allow us to

create an index assessing the mare-foal relationship.

Weaning had an effect on the frequencies and durations of most behaviors
investigated. We observed an increase in the frequency of vocalizing, defecating, eating,
drinking, looking forward, investigating, and pawing. Weaning did not cause an effect in

the frequency of urinating, contrary to findings in the study by Houpt et al. (1984). We

108

did see an increase in the duration of walking, but not of standing. Finally, we found a
decrease in the duration with which foals laid down on weaning and post-weaning day

when compared to pre-weaning durations.

Houpt et al. (1984) found an increase in vocalizations with an average of 20 i 4
during the first 15 min after weaning. In our study, during the first two hours, foals
vocalized 88.5 i 0.086 times per hour. This is approximately 22.13 vocalizations per 15
min, if we assume vocalizations during the first two hours are randomly distributed.
Similarly, McCall (1985) found a total of 93 vocalizations during the first hour after
weaning. During the fifth hour post-weaning, this frequency decreased to 39.8
vocalizations. The corresponding time block during our study is the 11:00 to 13:00 where
on average foals vocalized 21.84 i 7.04 times per hour. This is somewhat lower than the
findings by McCall (1985), but it should be noted that the frequency in our study is
derived from a total of one-hour observation time per foal during a two-hour observation
block' and it is possible that there is a significant difference in frequency of vocalizations

between the fifth and sixth hour post-weaning.

In addition to the frequency of vocalizing, we found an increase in the duration of
walking on subsequent weaning days and an increase in defecation. Houpt et a1. (1984)
found similar results, but in addition they found an increase in frequency of urination.
The latter finding was not seen in our study, perhaps due to the low occurrence of

urinating in general.

109

Eating frequency increased on subsequent days and the frequency of drinking was
increased during the post-weaning day. The latter could be associated with the increase in
locomotor activity caused by weaning. It should also be noted that casual observations
indicated that during the pre-weaning day, foals followed the mare’s eating pattern. If the
mare was eating, the foal joined. But after weaning, the pattern changed and foals made
an increased number of short, seemingly random visits to the feeder. We propose that this
change to an erratic feeding pattern caused an increase in eating frequency, but we
emphasize that this does not inherently imply an increased food intake. Hoffman et al.
(1995) did find an increase in time spent eating between 0 to 24 h and 48 to 72 h after
weaning, but we do not have the data to support this finding. The fact that nutritional gain
through nursing was possible during the pre-weaning day may affect the amount of hay
and concentrate foals ingested on this day. Consequently, as nursing was no longer
possible during weaning and post-weaning day, we would expect foals to ingest more hay
and concentrate. However, the observation that feeders still contained feed at the time of
the next feeding leads us to speculate that the immediate impact of weaning was such that

foals did not compensate for the loss of nutrients, previously acquired through nursing.

The increase in frequencies of looking forward, investigating, and pawing were
most apparent during the first observation block during the weaning day. Investigating
and pawing are behaviors that are interrelated as they often occurred almost
simultaneously in ‘exploration bouts’. The frequency of looking forward increased with
weaning and we attribute this to foals observing humans or animals moving up and down

the aisles between the stalls.

110

Hoffman et al. (1995) found that 24 to 48 h post-weaning, foals vocalized less and
spent more time standing when compared to 0 to 24 h. In addition, all behaviors changed
over time indicating foals were adapting to the weaning process. In our study, the
duration of standing did not change with weaning and the duration for walking increased
on the subsequent weaning day, indicating that adaptation may not have occurred at the
end of the weaning trial. In addition, the duration of lying down did not increase
significantly during the post-weaning day again suggesting that activity levels were
indeed still higher than during the pre-weaning day. Frequencies of vocalization and

defecation did decrease but vocalization rate had not refined to pre-weaning day levels.

No effect was found for the frequencies or durations of any behaviors based on
separated versus control foals, except for defecation. Fillies in the test group showed
higher frequency of defecating than fillies in the control group. Contrary, colts in the test
group defecated less than colts in the control group. As increase in defecation is a
behavioral indicator of stress, this could indicate that short-term separations complicate
fillies’ stress coping strategies during weaning, but that they also facilitate weaning for
colts. Houpt et al. (1984) found a numerical, though statistically not significant increase
in frequency of vocalization for foals that had been separated from their dam a few times
prior to weaning. This observation and the knowledge that vocalizations are a good
indicator for assessing weaning stress support our prediction that short-term separations

prior to weaning complicate adaptation to weaning.

111

A potential explanation for the finding that separations had no effect on weaning
stress in foals, apart from the small sample size used, is that the duration of our post-
weaning study was limited by logistic factors. McCall (1987) found that pre-weaning
cortisol concentrations were significantly lower than concentrations measured 2 days
post-weaning, whereas no difference was found when compared to cortisol measured 9
days post-weaning. Thus, the impact of weaning appears to extend beyond 36 hours after
weaning and it is possible that the immediate intense response to weaning masked

treatment differences. Continuing observations for a longer period of time could have P“

 

revealed such differences.

From our results, it seems that gender plays an important role in the display of
certain behaviors during stressful events. We do not have data to support findings by
Hoffman et al. (1995) that colts spend more time eating. However, in agreement with
results from the study by McCall et al. (1985) we did find that fillies have a higher level
of activity and lie down less compared to colts. Potentially associated with the higher
level of activity in fillies is the fact that they drink more than colts do. In our study we
also showed that fillies vocalized more during the post-weaning day than colts did,
indicating that fillies may be more affected by the weaning process. In addition, within
the test group: fillies defecated more than colts, but for the control group the reverse was
true. The same differences exist in the frequency of looking forward. This suggests that
separations aid colts in adapting to weaning stress whereas for fillies short-term

separations during early stages in life may not be beneficial.

112

In conclusion, we did not find a difference based on separated versus control
foals. However, despite a small sample size, we did find gender-related differences in the
response to weaning. Fillies seemed more affected by weaning stress than colts,
especially the fillies in the separation group. On the other hand, colts in the separation
group tended to have a less profound and long lasting impact of weaning. We argue that

these findings warrant further investigation into gender-related reactions to stressors.

113

I'"

 

CHAPTER V: GENERAL DISCUSSION

This study assessed the immediate and prolonged impact of short-term separations on
maternal care through behavioral indicators. In addition, we combined physiological
(salivary cortisol and heart rate) and behavioral measures to investigate the effect of

increased maternal care by short-term separations on weaning stress in foals.

When weaning occurred at MSU Merillat Equine Center, heart rate levels of foals
increased from 59.086 i 3.66 BPM on the day before weaning to 86.47 i 6.032 BPM on
the day of weaning during the first observation block. It should be noted that using mean
heart rate measurements provides a rough estimate compared to when one compares
certain bursts of activities with corresponding heart rate data. Heart rate can increase
because of low oxygen concentrations in the blood stream, as a result of increased
activity, or because of an increase in catecholamines associated with the psychological
response to matemal deprivation. Investigating whether increase in heart rate is related to
increase in activity or increase in catecholamines is important. If it seems that high and
low periods of heart rate correspond to similar bursts of activity, underlying physiological
factors may be responsible for the variations in heart rate. In our study we did not have
enough detail to the data to perform such an analysis. We propose future research be done
investigating the correlation between activity levels and heart rate during psychosocial

stressors during precisely timed trials.

114

 

Proximity scores during separation trials and post-separation evaluations were on
average greater than 2, with a score of 3 indicating greatest proximity and 1 smallest.
Thus, foals stayed close to their mothers at all times. However, because of the
confinement of the stall, the distance between mare and foal was limited. The overall
finding was that fillies stayed closer to their mothers than colts. This may be an indication
of a closer association between fillies and their dams when compared to colts. Supporting
this finding, we saw an effect of gender on several behaviors during the weaning trial
suggesting that fillies had a more profound and longer lasting response to weaning than
colts. Under feral conditions, there do not seem to be any differences in nursing behavior
and both colts and fillies almost always leave the native harem before becoming sexually
mature (Tyler 1972; Carson & Wood-Gush 1983b). By being exposed to close maternal
contact, fillies may become acquainted with maternal behavior from their mothers. It is
likely that this promotes reproductive fitness and insures survival of genes into firture
generations through successful foalings and rearing of offspring to healthy, competitive
animals. In this respect, colts, who do not aid in the rearing of foals would not benefit
from a close association with the dam. In fact, the sooner they are independent, the faster
they can breed dams and pass on their genes into firture generations. It should be noted
that scientists have not yet concluded the discussion on how much of maternal care
displayed by mares is innate and how much is learned from the dam and other mares, or

how much was developed through experience.

Stressors can be physical, psychological, and/or psychosocial in nature (Mason

1975). The psychological impact of weaning has been demonstrated in other studies

115

(Houpt et al. 1984; Hoffman et a1. 1995) and also during short-term separations, stress
has been postulated to be emotional in nature (Niezgoda & Tischner 1995). We
confirmed that weaning caused an increase in most behaviors investigated and an
increase in overall activity of foals. We could not investigate duration of states (e. g.,
walking) during the separations. Technical difficulties using the Panasonic field switcher
and 24- or 12- hour time lapse recording on a VCR resulted in unreliable timing when
decoding videotapes. It seems there is a trade-off in the clarity of the tapes and the
sequence of frames between the number of cameras and the time-lapse interval. For
future research using multiple cameras and time-lapse recording, we recommend to limit
the time-lapse recording to six-hour intervals as this still renders a clear image with

continuous motion.

Contrary to frndings in rodents, early increase in maternal care by short-term
separations did not lead to a dampened stress response of the HPA axis in response to
stressors later on in life. In fact, separation group animals showed a sustained elevation of
cortisol compared to the control group. Ten-minute separations led to immediate increase
in maternal care but this effect could no longer be identified 4-6 days post-separation. We
conducted this research using a small number of animals and acknowledge that this might
have been an influential factor. However, we speculate that there are alternative
explanations for the differences between rodents and horses. It is possible that by
applying the stressor, i.e., lO-min separations, at 2 weeks of age we missed the window
of time in which the HPA axis can be modulated. It is not known whether there is a

critical period during the development of foals in which the HPA axis is sensitive to

116

external stimuli. Second, in rodent studies, the response to a stressor was tested when the
pups had become mature (Liu et a1. 1997). Weaning in our study was performed at 24 :t 2
weeks of age, when foals were still maturing. The anticipated dampened stress response
may not have been apparent at weaning because of the different rate of development in
rodents and horses. The HPA-axis was still developing and, as a result, weaning in itself
may have had a modulating effect on the HPA-axis. To verify this, it would have been
beneficial for us to take physiological measures to support behavioral data. In addition,
because it is permanent by definition, weaning has a longer lasting impact compared to
restraint stress as applied by Liu et al. (1997). We propose the complimentary
explanation that the nature and frequency of the test stressor, and the age at which it is
applied are important when evaluating the effect of increased maternal care. This
prediction could be tested by developing protocols with different types of stressors

applied at different ages.

Though we did not find an effect of increase in maternal care by short-term
separations, and despite the limitation of a small sample size in our study, we found that a
distinction should be made between fillies and colts. Fillies in particular seem to be
affected negatively by the separations when investigating the gender by treatment
interactions for defecating and looking forward. When observing behavioral indicators,
colts showed a lesser response to weaning when they had been exposed to repeated short-

terrn separations from their dam prior to weaning.

117

In conclusion, repeated short-term separations of mares from their foals resulted
in an immediate increase in mare-foal interactions. This difference was no longer
apparent after four to six days when comparing the test group mare-foal interactions to
those of age-matched controls. To test the longevity of the impact of increased maternal
care through separations on maternal care, we investigated weaning stress in foals that
had been exposed to short-term separations versus foals that had never been separated
from their dam. Contrary to our prediction, certain indicators suggested that separating

mares fiom their foals during early stages in life complicated the adaptation to weaning, *

 

especially for fillies. We therefore recommend minimizing the number of times foals are

separated from their dam prior to weaning, especially where fillies are concerned.

118

Appendix A

Overview of nursing bout definitions

119

 

 

1983: “The time foal first nuzzled the mare’s udder for at least 3 seconds until
the final time the foal removed its head from the medial side of the mare’s
hind limb” (Crowell-Davis, 1985; based on observations in 1979)

1983: Start of a nursing bout was recorded when the foal actually sucked for the
first time. (Crowell-Davis, 1985; based on observations in 1980-1981)

1983: “A period of nursing activity delimited by intervals of non-nursing activity
lasting for 27 seconds or longer” (Carson & Wood-Gush, 1983)

1992: Including pre-nurse nuzzling, sucking, intra-bout pauses and intra-bout
nuzzling (Smith-Funk & Crowell-Davis, 1992)

1994: Including pre-suck nuzzle, pre-suck pause, suck, intra-suck nuzzle and

intra-suck pause (Barber & Crowell-Davis, 1994).

Ingestion of milk -referred to as sucking- was recognized by distinctive characteristics
that were difficult to observe when the horses were on pasture, but fairly obvious when
they were placed in stalls. Holding the head still, audible sucking sounds, and the
movement of fluid down the esophagus were the indicators of sucking (Barber &

Crowell-Davis, 1994).

120

 

 

Appendix B

Foal age at first sampling

121

 

Age at first sampling

Foal # Date of Birth (days)
Group 1 1 04-Apr-00 97
2 15-Apr-00 86
3 11-Apr-00 90
4 18—Apr-00 83
5 12-Apr-00 89
Group 2 6 17-Apr-00 85
7 27-Apr-00 75
8 26-May-00 46
9 17-May-00 55
10 18-hfiy-OO 54

 

 

122

Appendix C

Treatment group, date of birth, gender, and sire of foals used in the study at MSU

Merillat Equine Center

123

 

 

Foal
N“?
2

3(*)

5(*)

7(*)

9(*)

10

Date of birth
January 19
February 1
February 14
February 4
March 4
March 7
March 24
March 14
April 21

April 24

Gender foal
Filly
Filly
Filly
Filly
Colt
Filly
Filly
Colt
Colt

Colt

Sire

“QNMNNN%NVJ

 

(*) = test group. Sires are Shadow (S), Zippos (Z), Other (0).

124

 

Appendix D

Merillat observation schedule

125

 

 

 

 

Legend: underlinedihitth
bold = separation
italic = post-separation observation. Though the control animals were not

separated, they were observed at ages similar to the test group
Separated: 1-3-5-7-9 Control: 2-4-6—8-1 O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Jan-00

MON TUE WED THU FRI SAT SUN

1 2

3 4 5 6 '7 8 9

10 11 12 13 14 15 16
' 17 18 19 20 21 22 23

1

24 25 26 27 28 29 30

 

 

 

 

 

 

 

126

 

Fob-00

 

 

 

 

 

 

 

 

 

 

 

fl “ “u
N 3 0 .7
T
a .5 n m m
m 4 n m a
U
m 3 m n a
D 6 3
W.— 2 9 1 1 2
E
5 2 9
W 1 8 1 2 2
N
4 8
m —7 1 3 WI. 2

 

 

 

 

 

 

 

 

 

 

 

 

Mar-00

 

 

 

 

 

 

 

 

24 24
N
w k5 m w w
135 13 56
T
I
ma. 4 H w :2.
=
m a m n u. 7:4.
U
m 2 9 m m w
.1
D 5 2 9
W. 13 8 1 3 2 2
E
4 8
TU! #7 1 R: m 2
N
3 O 7
W 6 1 1 2 2
L

 

 

 

 

 

 

 

 

 

 

 

127

Apr-00

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MON TUE WED THU FRI SAT SUN
1 2
1 2
3 4
3 4 5 6 j! 6 9
5 5
6
6
10 11 12 13 14 W 16
1 3 '7 1 2
3 4
7 l
17 16 19 20 21 22 23
5 9 5
6
6
24 75 26 fi 26 29 30
1 3 f 1 2
3 4 ll
19 f 7
May-00
MON TUE WED THU FRI SAT SUN ll
1 2 3 4 5 6 :7 ll
5 5
6
6
6 9 10 1 1 12 13 14
9 3 7 3 9
7 1o
—15 16 177% 16 19 20 21
5 5
6
6
22 23 24 25 26 W 26
7
9 7 9
1o
29 30 31
5

 

 

128

 

.JunHDO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MON TUE WED THU FRI SAT SUN Jl
1 2 3 4 fl
5
6
.7 8
5 6 7 8 9 10 _ 11
9 7 7
9
10
12 13 14 15 16 17 18
8
19 20 2] 22 23 24 _ 25
9 7 7
9
10
26 27 28 29 30
====
JULOO
MON TUE WED THU FRI SAT SUN
1 2
3 4 5% 6 7 6 9
9 9
10
10 11 12 13 14 15 16
17 18 19 20 21 22 23
9 9
10
24 25 26 27 28 29 30
31

 

 

 

 

 

 

 

 

 

 

129

 

Appendix E

Ethogram used for post-separation observations

130

E denotes an event whereas S denotes a state

Eat (E)

Drink (E)

Stand idle (S)

Lie Down (S)

Walk (S)

Trot (S)

Lope (S)

Pawing (E)

Animal stands by feeder, ingesting either hay or concentrate

Animal stands by waterer, ingesting water

No forward or backward movements shown and all four legs

support weight of the animal. Also removal of foal’s muzzle from

udder when foal is not lying down or in motion.

No forward or backward movements shown but animal is not

supported by its legs. Also removal of muzzle from udder by the

foal when foal is not standing idle or in motion.

Slow, forward motion in a four-beat gait.

Forward motion, faster than walk, in a two-beat diagonal gait

Forward motion, faster than walk, in a three-beat diagonal gate

Striking motion on the ground with one front leg

131

 

I‘ll

 

Rear (E) While standing the animal lifts its front legs off the ground, raises
upper body off the ground and the body weight is supported by the
hind legs. Next, the front legs are placed back on the ground and

all four legs support the animal’s body

Defecate (E) Animal raises its tail and excretes feces

Urinate (E) Animal spreads hind legs a bit further apart than when standing

idle, raises tail, points dock back and in a slight squatting position

excretes urine

Nuzzle (E) Animal touches other animal by probing with lips and potentially smelling

scent, excluding when foal shows the “muzzle at udder” behavior

Investigate (E) Animal probing object with lips and potentially smelling scent

Allogroom (E) Animal uses lips and teeth to touch, nibble and pull on other

animal’s coat

Self-groom (E) The animal uses lips and teeth to touch, nibble and pull on own

coat

132

Watch Foal (E)

Watch Mare (E)

Look Forward (E)

Muzzle at Udder (S)

Ears back (E)

Kick threat (E)

Kick (E)

Mare turns head towards foal or if head is already turned towards

foal, mare points her ears towards foal

Foal turns head towards mare or if head is already turned towards

mare, foal points its ears towards mare

Animal is present in front half (closest to the aisle) of stall, and
turns its head to look outside of stall or if head is already directed

that way, points ears forward.

Foal is standing next to mare using muzzle to make contact with

udder to nurse

Animal points ears backward potentially accompanied with a head

swing in the direction of another animal

Animal positions hind towards other animal, squats down a bit and
points ears back, potentially lifting hind leg, but no striking or

physical contact occurs

Animal positions hind towards other animal, squats down a bit and

points ears back, lifts one or both hind legs and then strikes at the

other animal

133

Bite threat (E)

Bite (E)

Julnp (E)

Mount mare (E)

Tail swish (E)

Lick (E)

Cross the bow (E)

Animal swings head towards other animal and stretches neck,
points ears back, opens mouth a bit, potentially bearing teeth, but

no physical contact occurs

Animal swings head towards other animal and stretches neck,

points ears back, opens mouth a bit, potentially bearing teeth, and

bites

While in motion or at the start of motion, the animal thrusts its

body forward and upward while arching its back and potentially

swinging its head or kicking with the hind legs

Foal lifts its front legs off the ground as if rearing, but leans on any

part of its mother’s body (shoulders, rump, hind...)

Mare swishes its tail left to right in the direction of the foal. This

behavior is a direct result of contact initiated by the foal

Stroking object or other animal with tongue

Foal approaches the mare from the front, passes underneath her
neck and continues towards the stifle as part of a nursing attempt

134

Appendix F

Behaviors observed during weaning

135

Eat (E)

Drink (E)

Stand idle (S)

Lie Down (S)

Walk (S)

Trot (S)

Lope (S)

Pawing (E)

Animal stands by feeder, ingesting either hay or concentrate

Animal stands by waterer, ingesting water

No forward or backward movements shown and all four legs

support weight of the animal. Also removal of foal’s muzzle from

udder when foal is not lying down or in motion. 1
No forward or backward movements shown but animal is not

supported by its legs. Also removal of muzzle fiom udder by the

foal when foal is not standing idle or in motion.

Slow, forward motion in a four-beat gait.

Forward motion, faster than walk, in a two-beat diagonal gait

Forward motion, faster than walk, in a three-beat diagonal gate

Striking motion on the ground with one front leg

136

Rear (E) While standing the animal lifts its front legs off the ground, raises
upper body off the ground and the body weight is supported by the
hind legs. Next, the front legs are placed back on the ground and
all four legs support the animal’s body

Defecate (E) Animal raises its tail and excretes feces

Urinate (E) Animal spreads hind legs a bit further apart than when standing
idle, raises tail, points dock back and in a slight squatting position

excretes urine

Nuzzle (E) Animal touches other animal by probing with lips and potentially smelling

scent

Investigate (E) Animal probing object with lips and potentially smelling scent

Allogroom (E) Animal uses lips and teeth to touch, nibble and pull on other

animal’s coat

Self-groom (E) Animal uses lips and teeth to touch, nibble and pull on own coat

Vocalize (E) Production of sound by animal

137

 

Watch Foal (E)

Watch Mare (E)

Look Forward (E)

Muzzle at Udder (S)

Leave (E)

Approach (E)

Ears back (E)

Mare turns head towards foal or if head is already turned towards

foal, mare points her ears towards foal

Foal turns head towards mare or if head is already turned towards

mare, foal points its ears towards mare

Animal is present in front half (closest to the aisle) of stall, and

turns its head to look outside of stall or if head is already directed

that way, points ears forward.

Foal is standing next to mare using muzzle to make contact with

udder to nurse

While in motion, the animal increases the distance between itself

and the other animal

While in motion, the animal decreases the distance between itself

and the other animal

Animal points ears backward potentially accompanied with a head

swing in the direction of another animal

138

 

 

Kick threat (E)

Kick (E)

Bite threat (E)

Bite (E)

Squeal (E)

Jump (E)

Animal positions hind towards other animal, squats down a bit and
points ears back, potentially lifting hind leg, but no striking or

physical contact occurs

Animal positions hind towards other animal, squats down a bit and
points ears back, lifts one or both hind legs and then strikes at the

other animal

Animal swings head towards other animal and stretches neck,
points ears back, opens mouth a bit, potentially bearing teeth, but

no physical contact occurs

Animal swings head towards other animal and stretches neck,
points ears back, opens mouth a bit, potentially bearing teeth, and

bites

A hi gh-pitched vocalization that’s given in the event of an

agonistic interaction, by aggressor or victim
While in motion or at the start of motion, the animal thrusts its

body forward and upward while arching its back and potentially

swinging its head or kicking with the hind legs

139

 

 

Mount mare (E) Foal lifts its front legs off the ground as if rearing, but leans on any

part of its mother’s body (shoulders, rump, hind...)

Null (E) Indicates that there’s no change in state by an animal after the

other animal lies down

Tail swish (E) Mare swishes its tail left to right in the direction of the foal. This

behavior is a direct result of contact initiated by the foal

Lick (E) Stroking object or other animal with tongue

Cross the bow (E) Foal approaches the mare fiom the front, passes underneath her

neck and continues towards the stifle as part of a nursing attempt

Bunt (E) A forceful thrust of one animal’s muzzle into another animal

Events during weaning

Weaning Handler walks in stall, halters mare and leads her out, leaving foal

in the stall by itself. Mare is taken to outside pasture.

140

Appendix G

Protocol during weaning

141

Friday:

6:00

7:00

9:00

10:20

11:00

13:00

14:20

15:00

17:00

18:20

Place heart rate monitor + saliva sample
Start observations

End observations

Saliva sample + check HRM

Start observations

End observations

Saliva sample + check HRM

Start observations

End observations

Saliva sample + remove HRM

Saturday

6:00

7:00

9:00

10:20

11:00

13:00

14:20

15:00

17:00

18:20

Place heart rate monitor + saliva sample
Weaning + Start observations

End observations

Saliva sample + check HRM

Start observations

End observations

Saliva sample + check HRM

Start observations

End observations

Saliva sample + remove HRM

142

 

Sunday

6:00

7:00

9:00

10:20

11:00

13:00

14:20

15:00

17:00

18:20

Place heart rate monitor + saliva sample
Start observations

End observations

Saliva sample + check HRM

Start observations

End observations

Saliva sample + check HRM

Start observations

End observations

Saliva sample + remove HRM

143

 

LITERATURE CITED

Alexander, S.L., Irvine, CH. 1998 The effect of social stress on adrenal axis activity in
horses: the importance of monitoring corticosteroid-binding globulin capacity. Journal of
Endocrinology. 157:425-432.

Anonymous. 1994. Foal Behavior. Equine Practice, 16, 33-36.
Asterita. 1985. The physiology of stress. New York: Human Science Press.

Bagshaw, C. S., Ralston, S. L. & Fisher, H. 1994. Behavioral and physiological effect
of orally administered tryptophan on horses subjected to acute isolation stress. Applied
Animal Behaviour Science, 40, 1-12.

Baldock, N. M. & Sibly, R. M. 1990. Effects of handling and transportation on the heart
rate and behaviour of sheep. Applied Animal Behaviour Science, 28, 15-39.

Barber, J. A. & Crowell-Davis, S. L. 1994. Maternal behavior of Belgian (Equus
caballus) mares. Applied Animal Behaviour Science, 41, 161-189.

Beaver, B. V. 1981. Maternal behavior in mares. Veterinary Medicine, Small Animal
Clinician, 76, 315-7.

Beerda, B., Schilder, M. B. H., Janssen, N. S. C. R. M. & Moi, J. A. 1996. The Use of
Saliva Cortisol, Urinary Cortisol, and Catecholarnine Measurements for a Noninvasive
Assessment of Stress Responsses in Dogs. Hormones and Behavior, 30, 272-279.

Beerda, B., Schilder, M. B. H., van Hooff, J. A. R. A. M., de Vries, H. W. & Mol, J.
A. 1998. Behavioural, saliva cortisol and heart rate responses to different types of stimuli
in dogs. Applied Animal Behaviour Science, 58, 365-381.

Blackshaw, J. K. & Blackshaw, A. W. 1989. Limitations of salivary and blood cortisol
determinations in pigs. Veterinary Research Communications, 13, 265-271.

Bowlby, J. 1977. The making and breaking of affectional bonds. 1. Aetiology and
psychopathology in the light of attachment theory. An expanded version of the F iftieth
Maudsley Lecture, delivered before the Royal College of Psychiatrists, 19 November
1976. Brittish Journal of Psychiatry, 130, 201 -210.

Boyd, L. E. & Houpt, K. A. 1994. Activity Patterns.: State University of New York
Press.

Broom, D. M. & Johnson, K. G. 1993. Stress and Animal Welfare. London: Chapman &
Hall.

144

 

Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P. M. & Meaney, M.
J. 1998. Maternal care during infancy regulates the development of neural systems

mediating the expression of fearfulness in the rat. Proceedings of the National Academy
of Sciences of the United States of America, 95, 5335-40.

Carson, K. & Wood-Gush, D. G. 19836. Behaviour of thoroughbred foals during
nursing. Equine Veterinary Journal, 15, 257-62.

Carson, K. & Wood-Gush, D. G. M. 1983b. Equine Behaviour: I. A review of the
literature on social and dam-foal behaviour. Applied Animal Ethology, 10, 165-178.

Cleroux, J ., Perronet, F. & J., (I. C. 1985. Sympathetic indices during psychological
and physical stimuli: before and after training. Physiology and Behavior, 35, 271-275.

Compton, M. M., L.M. Caron & J.A. Cidlowski. 1987. Glucocorticoid Action on the
Immune System. Journal of Steroid Biochemistry, 27, 201-208.

 

Crowell-Davis, S. L. 1985. Nursing Behavior and Maternal Aggression among Welsh
Ponies (Equus caballus). Applied Animal Behaviour Science, 14, 11-25.

Crowell-Davis, S. L. 1986. Developmental behavior. Veterinary Clinics of North
America. Equine Practice, 2, 573-90.

Crowell-Davis, S. L. & Houpt, K. A. 1986. Maternal behavior. Veterinary Clinics of
North America. Equine Practice, 2, 557—71.

Dellmeier, G. R., Friend, T. H. & Gbur, E. E. 1985. Comparison of four methods of
calf confinement. II. Behavior. Journal of Animal Science, 60, 1102-1109.

Dwyer, C.M. & Lawrence, A.B.1998. Variability in the expression of maternal
behaviour in primiparous sheep: Effects of genotype and litter size. Applied Animal
Behaviour Science, 58, 31 1-330.

Dwyer, C. M. & Lawrence, AB. 1999. Does the behaviour of the neonate influence the
expression of maternal behaviour in sheep? Behaviour, 136, 367-389.

Dwyer, C. M. & Lawrence, A. B. 2000. Maternal behaviour in domestic sheep (ovis
aries): constancy and change with maternal experience. Behaviour, 1-23.

Equine Research Publications 1977 What is the function of the mouth?. In: Vale
Illustrated veterinary encyclopedia for horsemen (ed. Wagener M.M. and Wagener D.M.
pp 478-479.

145

Evans, D. L., Harris, R. C. & Snow, D. H. 1993. Correlation of racing performance
with blood lactate and heart rate after exercise in thoroughbred horses. Equine Veterinary
Journal, 25, 441-5.

Feh, C. & de Mazieres, J. 1993. Grooming at a preferred site reduces heart rate in
horses. Animal Behaviour, 46, 1191-1194.

Fell, L. R. & Shutt, D. A. 1986. Adrenocortical response of calves to transport stress as
measured by salivary cortisol. Canadian Journal of Animal Science, 66, 637-641.

Fell, L. R., Shutt, D. A. & Bentley, C. J. 1985. Development of a salivary cortisol
method for detecting changes in plasma "free" cortisol arising from acute stress in sheep.
Australian Veterinary Journal, 62, 403-406.

Fenske, M. 1996. Measurement of salivary cortisol in guinea pigs. Journal of
Experimental Animal Science, 38, 13-9.

Fraser, A. F. & Broom, D. M. 1990. Farm animal behaviour and welfare. London:
Bailliere Tindall.

Friend, T. H. 1991. Symposium: Response of Animals to Stress. Behavioral Aspects of
stress. Journal of Dairy Science, 74, 292-303.

Fuchs, E., Kirschbaum, C., Benisch, D. & Bieser, A. 1997. Salivary cortisol: a non-
invasive measure of hypothalamo-pituitary-adrenocortical activity in the squirrel
monkey, Sairniri sciureus. Laboratory Animals, 31, 306-11.

Geverink, N. A., Ruis, M. A., Eisen, R., Lambooij, E., Blokhuis, H. J. & Wiegant, V.
M. 1999. The effect of shot biopsy on behavior, salivary cortisol, and heart rate in
slaughter pigs. Journal of Animal Science, 77, 1614-9.

Harper, F. 1999. Stress - good or bad? In: Arabian Horse Times, pp. 562.

Heleski, C. R., Shelle, A. C., Nielsen, B. D. & Zanella, A. J. In Press. Comparison of
behavior and welfare in weanling horses under two industry-practiced weaning methods.

Hofer, M. A. 1994. Early social relationships as regulators of infant physiology and
behavior. Acta Paediatrica (suppl), 397, 9-18.

Hoffman, R. M., Kronfeld, D. S., Holland, J. L. & Greiwe_Crandell, K. M. 1995.
Preweaning diet and stall weaning method influences on stress response in foals. Journal
of Animal Science, 73, 2922-30.

Holland, J. L., Kronfeld, D. S., Hoffman, R. M. & Harris, P. A. 1997. Weaning stress

assessment in mares. Proceedings of the Fifteenth Equine Nutrition & Physiology
Symposium, Ft. Worth, Texas, 219-222.

146

Horning M G, Brown L, N owlin J, Lertratanangakoon K, Kellaway P and Zion T E
1977 Use of saliva in therapeutic drug monitoring. Clinical Chemistry 22:868-878.

Houpt, K. A. 1992. Horse Behavior. In: Horse Breeding and Management (Ed. by
Evans, J. W.), pp. 63-83. College Station: Elsevier.

Houpt, K. A., Hintz, H. F. & Butler, W. R. 1984. A Preliminary study of two methods
of weaning foals. Applied Animal Behaviour Science, 12, 177-181.

Houpt, K. A., Hintz, H. F. & Pagan, J. D. 1981. The response of mares and their foals
to brief separation. Journal of Animal Science, 53, 129.

Hubert, W. & de Jung-Meyer, R. 1989. Emotional Stress and Saliva Cortisol Response.
Journal of Clinical Chemistry and Clinical Biochemistry, 27.

Kendrick, K. M., Da Costa, A. P., Broad, K. D., Ohkura, S., Guevara, R., Levy, F. &
Keverne, E. B. 1997. Neural control of maternal behaviour and olfactory recognition of
offspring. Brain Research Bulletin, 44, 383-95.

King, J. A. & Edwards, E. 1999. Early stress and genetic influences on hypothalamic-
pituitary-adrenal axis functioning in adulthood. Hormones and Behavior, 36, 79-85.

Kirschbaum, C. & Hellhammer, D. 1989. Response Variability of Salivary Cortisol
under Psychological Stimulation. Journal of Clinical Chemistry and Clinical
Biochemistry, 27, 234.

Kohler, P. H. 1972. The management of weanling foals at 30 days and 180 days of age.
Stud Manager's Handbook, 8, 7-18.

Lebelt, D., Schonreiter, S. & Zanella, A. J. 1996. Salivary cortisol in stallions: the
relationship with plasma levels, daytime profile and changes in response to semen
collection. Pferdeheilkunde, 12, 411-414.

Lin, D., Diorio, J., Tannenbaum, B., Caldji, C., Francis, D., Freedman, A., Sharma,
S., Pearson, D., Plotsky, P. M. & Meaney, M. J. 1997. Maternal care, hippocarnpal
glucocorticoid receptors, and hypothalamic-pituitary—adrenal responses to stress. Science,
277, 1659-62.

Lowe, J. E. 1990. The Horse. New York: W.H. Freeman.
Mal, M. E., Friend, T. H. & Lay, D. C. 1989. Effects of confinement and isolation on

the immune system of mares and distance traveled during open-field tests. Journal of
Animal Science, 67, 107-108.

147

Malinowski, K., Hallquist, N. A., Helyar, L., Sherman, A. R. & Scanes, C. G. 1990.
Effect of different separation protocols between mares and foals on plasma cortisol and
cell-mediated immune response. In: 11th Symposium of the Equine Nutrition and
Physiology Society.

Martin, P. & Bateson, P. 1999. Measuring Behaviour. Cambridge: University Press.
Mason, G. 1991. Stereotypy: a critical review. Animal Behaviour, 41, 1015-1038.

Mason, J. W. 1975. A historical view of the stress field. Journal of Human Stress, 1, 6-
12 contd.

Matteri, R. L., Carroll, J. A. & Dyer, C. J. 2000. Neuroendocrine Responses to Stress.
In: The biology of animal stress (Ed. by Moberg, G. P. & Mench, J. A.), pp. 43-76. Oxon:
CABI Publishing.

McCall, C. 1991. Neonatal Foal Behavior. Equine Practice, 13, 19-22.

McCall, C. A., Potter, G. D. & Kreider, J. L. 1985. Locomotor, vocal and other
behavioral responses to varying methods of weaning foals. Applied Animal Behaviour
Science, 14, 27-35.

McCall, C. A., Potter, G. D., Kreider, J. L. & Jenkins, W. L. 1987. Physiological
responses in foals weaned by abrupt or gradual methods. Equine Veterinary Science, 7,
368-374.

Meyer J.S., Bowman RE. 1972 Rearing experience, stress and adrenocorticosteroids in
the rhesus monkey. Physiology and Behavior 8:339—343.

Moberg, G. P. 1985. Biological Response to Stress: Key to Assessment of Animal Well-
Being. In: Animal Stress (Ed. by Moberg, G. P.), pp. 27-49. Bethesda, Maryland:
American Physiological Society.

Moberg, G.P. 2000. Biological Response to stress: Implications for Animal Welfare. In:
The biology of animal stress (Ed. by Moberg, G.P. & Mench, J .A.), pp.l-21. Oxon: CABI
Publishing.

Moons, C. 1999. Inleidende Studie naar: Dominantierelaties in Przewalski-kudden in
gevangenschap.

National Research Council 1989. Nutrient requirements of horses. Washington, DC:
National Acadamy Press.

Naylor, J. M. & Bell, R. 1985. Raising the orphan foal. Veterinary Clinics of North
America. Equine Practice, 1, 169-78.

148

 

Nicol, C. J. 1999. Stereotypies and their relation to management. In: BE VA Specialist
days on behavior and nutrition (Ed. by Harris, P. A., Gomarsall, G. M., Davidson, H. P.
B. & Green, R. E.), pp. 11-14: Equine Veterinary Journal Ltd.

N iezgoda, J. & Tischner, M. 1995. Intensity of Stress Reactions during Short-Term
Isolation of Mothers fiom Foals. Biol Reprod Mono, 1, 107-111.

Ohl, F., Kirschbaum, C. & Fuchs, E. 1999. Evaluation of hypothalamo-pituitary-
adrenal activity in the tree shrew (Tupaia belangeri) via salivary cortisol measurement.
Laboratory Animals, 33, 269-274.

Palme, R., Fischer, P., Schildorfer, H. & Ismail, M. N. 1996. Excretion of infused
14C-steroid hormones via faeces and urine in domestic livestock. Animal Reproduction
Science, 43, 43-63.

Palme, R. & Moestl, E. 1997. Measurement of cortisol metabolites in faeces of sheep as

a parameter of cortisol concentration in blood. Journal of Mammalian Biology,
Proceedings Suppl. 11, 62, 192-197.

Parrott, R. F., Houpt, K. A. & Misson, B. H. 1988. Modification of the Responses of
Sheep to Isolation Stress by the use of Mirror Panels. Applied Animal Behaviour Science,
19, 331-338.

Parrott, R. F. & Misson, B. H. 1989. Changes in pig salivary cortisol in response to
transport simulation, food and water deprivation, and mixing. Brittish Veterinary Journal,
145, 501-505.

Parrott, R. F., Misson, B. H. & Baldwin, B. A. 1989. Salivary cortisol in pigs following
adrenocorticotrophic hormone stimulation: comparison with plasma levels. Brittish
Veterinary Journal, 145, 362-366.

Persson, S. G. B. & Ullberg, L. E. 1974. Blood volume in relation to excercise tolerance
in trotters. Journal of South African Veterinary Association, 45, 293-299.

Plotsky, P. M. 2000. Intrinsic and Extrinsic Factors Modulating Physiological Coping
Systems During Development. In Press.

Plotsky, P. M. & Meaney, M. J. 1993. Early, postnatal experience alters hypothalarnic
corticotropin-releasing factor (CRF) mRNA, median eminence CRF content and stress-
induced release in adult rats. Molecular Brain Research, 18, 195-200.

Pollard, J. C., Littlejohn, R. P. & Suttie, J. M. 1992. Behaviour and weight change of

red deer calves during different weaning procedures. Applied Animal Behaviour Science,
35, 23-33.

149

I

Riad-Fahmy D., Read G.F., Walker RF. 1983 Salivary steroid assays for assessing
variation in endocrine activity. Journal of Steroid Biochemistry 19:265-272.

Rivera, E. 1999. Behavioral and physiological responses of horses to initial training: the
comparison between pastured versus stalled horses.

Rose, R. J. & Hodgson, D. R. 1994. Hematology and Biochemistry. In: The Athletic
Horse (Ed. by Hodgson, D. R. & Rose, R. J .), pp. 63-77: W.B. Saunders Company.

Schonreiter, S., Zanella, A. J. & Unshelm, J. 1999. Technique for continuous
monitoring of salivary cortisol concentration in pigs. Laboratory Animal Science, 49,
429-32.

Scott, E. M., McGarrigle, H. H. & Lachelin, G. C. 1990. The increase in plasma and
saliva cortisol levels in pregnancy is not due to the increase in corticosteroid-binding
globulin levels. Journal of Clinical Endocrinology and Metabolism, 71, 639-44.

Seligman, M. E. P. & Maier, S. F. 1967. Failure to escape to traumatic shock. Journal
of Experimental Psychology, 74, 1-9.

Selye, H. 1956. The Stress of Life. New York: McGraw-Hill Book Company, Inc.

Smith-Funk, D. E. & Crowell-Davis, S. L. 1992. Maternal behavior of draft mares
(Equus caballus) with mule foals (Equus asinus x Equus caballus). Applied Animal
Behaviour Science, 33, 93-119.

Snow, D. H., Harris, R. C., MacDonald, I. A., Forster, C. D. & Marlin, D. J. 1992.
Effects of high-intensity exercise on plasma catecholamines in the thoroughbred horse.
Equine Veterinary Journal, 24, 462-7.

Thornton, R. J. 1985. Hormonal responses to excercise and training. Equine Practice, 1,
477-496.

Toates, F. 2000. Multiple Factors Controlling Behaviour: Implications for Stress and
Welfare. In: The Biology of Animal Stress (Ed. by Moberg, G. P. & Mench, J. A.), pp.
199-226. Oxon: CABI Publishing.

Tyler, S. J. 1972. The Behaviour and Social Organization of the New Forest Ponies, 5.
Umeda, T., Hiramatsu, R., Iwaoka, T., Shimada, T., Miura, F. & Sato, T. 1981. Use
of saliva for monitoring unbound free cortisol levels in serum. Clinica Chimica Acta,
110, 245-53.

van der Kolk, J. H., N achreiner, R. F., Schott, H. C., Refsal, K. R. & Zanella, A. J.

2001. Salivary and plasma concentration of cortisol in normal horses and horses with
Cushing's disease. Equine Veterinary Journal, 33, 211-3.

150

 

Varley, M. A. 1995. Behavioural patterns of the weaned piglet. Pig Journal, 35, 71-97.

Vining, R. F., McGinley, R. A., Maksvytis, J. J. & Ho, K. Y. 1983. Salivary cortisol: a
better measure of adrenal cortical function than serum cortisol. Annals of Clinical
Biochemistry, 20 (Pt 6), 329-35.

Walker, R. F. 1989. Salivary Corticosteroids: Clinical and Research Applications.
Journal of Clinical Chemistry and Clinical Biochemistry, 27, 234-235.

Warren, L. K., Cook, N. J., Schaefer, A. L., Burwash, L., Anderson, M., Baron, V.
& Jensen, C. 2001. The use of salivary cortisol as an index of stress in horses.
Proceedings of the Seventeenth Equine Nutrition and Physiology Society, Lexington,
Kentucky, 353-354.

Warren, L. K., Lawrence, L. M. & Thompson, K. N. 1999. The Influence of betaine
on Untrained and Trained Horses Exercising to Fatigue. Journal of Animal Science, 77,
677-684.

Welsh, T. H., McCraw, R. L. & Johnson, B. H. 1979. Influence of corticosteroids on
testosterone production in the bull. Biology of Reproduction, 21, 755-63.

Wolksi, T. R., Houpt, K. A. & Aronson, R. 1980. The role of the senses in mare-foal
recognition. Applied Animal Ethology, 6, 121-138.

151

IIIIIIIIIIIIIIIIIIIII

lull11111111111111111117111