Accepted Manuscript

Title: Effects of simulated sea motion on stepping behaviourin sheep

Author: <ce:author id="aut0005"author-id="S0168159116303793-df81211b5519ed06f0617071e1100c64"> GriselNavarro<ce:author id="aut0010"author-id="S0168159116303793-b67339302fc0ca6403667dae0c230abf"> EduardoSanturtun<ce:author id="aut0015"author-id="S0168159116303793-15d81baf56e5d5def10a5b48e942ba99"> Clive J.C.Phillips

PII: S0168-1591(16)30379-3DOI: http://dx.doi.org/doi:10.1016/j.applanim.2016.12.009Reference: APPLAN 4380

To appear in: APPLAN

Received date: 20-4-2016Revised date: 17-11-2016Accepted date: 26-12-2016

Please cite this article as: {http://dx.doi.org/

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Effects of simulated sea motion on stepping behaviour in sheep

Grisel Navarro1, Eduardo Santurtun1,2 and Clive J.C. Phillips1*

1 Centre of Animal Welfare and Ethics (CAWE), School of Veterinary Science, University of

Queensland (UQ), Australia

2 Present address: Departamento de Etología, Fauna Silvestre y Animales de Laboratorio

(DEFSAL), Facultad de Medicina Veterinaria y Zootecnia (FMVZ), Universidad Nacional

Autónoma de Mexico (UNAM), Ciudad Universitaria, 04510, Mexico City, Mexico

*Address for correspondence: Centre for Animal Welfare and Ethics, School of Veterinary Science,

Building 8134, Gatton Campus, University of Queensland, Gatton, Queensland 4343, Australia. Tel.

61754601158/0406340133; Email c.phillips@uq.edu.au

Highlights

We exposed sheep in pairs to floor movements that simulated ship transport

Sheep stepped most with their front feet, especially forwards and backwards

Sheep demonstrated laterality in their use of right fore limbs most to maintain balance

Abstract

Animals respond to a moving floor during transport by stepping in different directions to

maintain their balance, but little is known about the importance of different types of

movement. Four sheep were restrained in a crate on a platform that could be programmed to

provide the various movements that simulated the motion of a ship. They were exposed to

three movement types, Pitch, Roll, and Heave, or a Control treatment, in pairs for 30 min

periods in a changeover design. The orientation and frequency of stepping movements was

recorded from videos made during the treatments and heart rate responses were monitored.

Heave produced the biggest stepping responses, in the forelimb. Sheep stepped most

commonly straight forwards and backwards with the fore limbs, then forwards, backwards

and sideways with the hind limbs. When they did make lateral stepping movements, they

moved their feet more outwards than inwards, presumably as this maintained balance more

effectively. Stepping movements were associated with reduced high frequency heart beats,

suggesting an associated negative emotion. Sheep on the left side of the crate showed some

evidence of greater stress than those on the right, regarding their limb movement preferences

and heart rate variability responses. This may be because animals on the left side of the crate

had no other sheep in their left eye vision, signals from which are processed by the right

brain hemisphere and are associated with stress responses. In conclusion, when subjected to

simulated ship movement, sheep produced stepping and heart rate responses that were

connected and indicated negative emotions.

Keywords: Balance; Heart rate variability; Sheep; Ship motion; Stepping behaviour

1. Introduction

Livestock are increasingly transported around the world as human food resources, including

a growing number alive (Phillips, 2008; FAOSTAT, 2016), and also to satisfy a commercial

demand for breeding animals (Norris, 2005). Sheep are one of the most popular animals

exported overseas, particularly from Australia to the Middle East, with 1,959,761 head in

2015 (Meat & livestock Australia, 2016), travelling for between 14 and 23 days (Phillips,

2008). These animals have to deal with several challenges on-board, including ammonia,

high stocking density, balance maintenance, motion sickness, noise and a novel environment

(Black et al.,1994; Norris, 2005., Phillips, 2008; Santurtun and Phillips 2015).

Balance is maintained by the nervous and vestibular systems, using locomotion receptors,

input from the eyes and the biomechanical system of movement in the animals (Biewener

2003; Santurtun and Phillips, 2015). There have been few published studies of animals’

stepping behaviour during transport, but it is believed that maintenance of balance is

achieved by regular stepping movements and support from other animals and vehicle fixtures

(Santurtun and Phillips, 2015). Ships experience six major motion types (heave, sway, surge,

yaw, pitch and roll). However, heave, pitch and roll are the most relevant in terms motion

sickness (Santurtun and Phillips, 2015). By monitoring sheep on a moving platform

simulating ship motion, Santurtun et al (2015) found that roll motion required more stepping

motions by sheep than pitch. During road transportation, Jones et al (2010) found that when

sheep balance themselves against the motion of the vehicle, they widen their stance by

spreading their legs sideways, as well as taking small steps forwards and backwards and

rocking with the motion of the vehicle. The maintenance of balance can be influenced by

environmental and animal characteristics, such as the stocking density during the trip

(Broom, 2000; Cockram et al 1996; Jones et al., 2010; Randall 1993).

Lateralized behaviour in vertebrates can be determined by either somatic/visceral asymmetry,

e.g. of the otolith organs, or by functional brain asymmetry (specialisations in the hemisphere

processing) (Lychakov, 2013). Otolith asymmetry has been described in fish, reptiles and

humans (Helling et al 2005, Bisazza et al 1998, Putnam et al 1996). Laterality of limb

movements may indicate emotional responses during transport. Sheep, like others vertebrates

species, have preferences with respect to the side of the body involved in different activities

(Anderson and Murray 2013; Lane and Phillips 2004; Peirce et al 2000; Rogers 2010;

Versace et al 2007). The use of the left and right limbs in animals is contralaterally related to

the use of the right and left brain hemisphere to process the responses, which are related to

reactive and proactive behaviours, respectively (Rogers 2010). It is unclear whether sheep

respond to motion with lateralised stepping at a population level, with some authors, e.g.

Versace et al (2007), believing that sheep only do this for tasks that require social

coordination.

Similarly, heart rate and its variability is increasingly used in animal studies to identify

effects on the animals’ emotions (von Borell et al., 2007). In particular, the activity of the

autonomic nervous system can be illucidated by measures of heart rate variability. A healthy

heart displays irregular interbeat intervals, caused by rhythmic oscillation of the homeostatic

regulators of cardiac activity. These are moderated by vagal activity to reduce heart rate and

sympathetic activity to increase it. During anxiety, reduced vagal activity decreases the high

frequency beats, which conversely are increased during positive emotions. A high Low

frequency to high frequency ratio (LF/HF) is sometimes used as an indication of increased

sympathetic regulation of heart rate, relative to vagal activity, but it is hard to draw

inferences relating to stress with the variable results that have been obtained (Frondelius et

al., 2015).

Although much has been published about the effects of road transport on sheep behaviour

(e.g. Knowles et al 1995; Knowles 1998; Jones et al 2010), little is known about their

behavioural responses during transport by ship. The aim of this study was to investigate how

sheep respond to simulated sea motion in their stepping behaviour. Specifically, our

objectives were to simulate the various motions involved in ship transport and investigate the

stepping responses of sheep when subjected to these motions over short periods of time. We

further investigated how these related to heart rate and its variability, in order to draw some

conclusions about whether the sheep were stressed by having to step to maintain their

balance during the motions.

2. Materials and methods

The study was conducted at the University of Queensland, Australia (27.3° S, 152.2° E).

Approval for this research was obtained from the University’s Animal Ethics Committee

(SVS/443/10). An initial report of the effects of the treatments on sheep behaviour and heart

rate has been published (Santurtun et al., 2015); here we report in detail about the stepping

movements made by the sheep, and their connection to heart rate parameters.

2.1 Animals Housing and Management

Four Merino cross wethers, of approximately 24 months of age and mean weight ± SEM of

37.4 ± 0.1 kg, were used for the study from the University of Queensland’s flock. Before and

after each trial, they were kept in an outdoor pen with ad libitum water and wheaten chaff.

During the trials, they were restrained in pairs in a crate of 3 tubular steel bars (0.81 m wide

× 1.18 m long × 0.95 m high), which was bisected by a 3-barred division that prevented them

from turning around. The floor of the crate was of solid steel, diamond plate sheeting, which

provided an antislip footing for the sheep. The space provided was 0.48 m2 /sheep, 68 %

more than the space requirement for sheep of this weight in Australian standards (ASEL,

2011), to allow the sheep to express their behaviour and so that we could observe them

easily. Chaff and water were offered at all times except when the sheep were in the crate. In

the pen 1 kg of lucerne pellets was offered daily at 1600 h.

2.2 Simulating Sea Transport Motions

The design of the platform for exposing sheep to floor movement and the programming of

the movement of the platform have been described previously (Santurtun et al., 2014). In

brief, a 0.8 m wide × 1.2 m long programmable motion platform (Model T2Smp, CKAS

Mechatronics Pty Ltd Melbourne, Australia) was used that was capable of produce roll and pitch

movements in combination or independently. The platform moved in two directions, powered by

two rams at 90o to each other, with movement duration determined from computer

commands (Santurtun et al., 2014) through a BELKIN® Hi-Speed USB 2.0. An electric

forklift (Model SHR5550 series, Crown Equipment Corporation, New Bremen, OH, USA)

was used to elevate and return the platform in the heave motion. The forklift had a maximum

elevation of 3200 mm, and a lifting speed of approximately 0.2 ms-1.

The crate for restraining the sheep was placed directly on top of the platform. The

programming of the platform was conducted with Microsoft Visual Studio Solution C++

Express 2008. The entire apparatus was covered by a white cloth to prevent the sheep having

a moving image of the room during the treatments. The apparatus was contained within a

quiet room (background noise < 50 dBA) at the Queensland Animal Science Precinct.

2.2.1. Amplitude and Period. The amplitude for roll and pitch used for this experiment

represented 33% of the maximum tolerance required when a ship is converted from a cargo

to a livestock carrier (Skraastad, 1983). As the maximum tolerance does not simulate a

typical voyage, the value of 33% of the maximum was chosen to provide amplitudes and

durations that were relevant to the expected dynamic environment of a typical live export

vessel (177 x 32 m, Anon, 2016) in moderate seas (McDonald, 1993). Whilst greater values

may be experienced during high seas, a lesser value was chosen in order not to stress the

sheep too much. Heave at 67 cm is obviously much less than the 4 m maximum that is

recommended, and the 6 s period was deliberately below the normal range (McDonald, 1993,

range 7.5-20 s) because of this.

2.3 Experimental protocol

Sheep were habituated through positive reinforcement with feed pellets to the different

potential stressors they would face during the experiment, including handling, heart rate

monitor fitting and wearing, forklift noise, the ramp for loading and unloading, the

experimental rooms and the crate, over a period of 32 d. During the experiment, sheep were

exposed in pairs to four treatments, pitch, roll, heave and control in the crate for 30 min

periods. The treatments were applied in a 4 × 4 Latin Square (Table 1) over an 8 d period. In

total each sheep was exposed to 16 treatment periods, 8 in the morning and 8 in the

afternoon. Sheep experienced treatments in 4 pairs (1+2, 3+4, 2+3 and 1+4), so that pair

effects could be evaluated.

2.4 Behaviour recording

Sheep behaviour was recorded during each treatment in real time by four video cameras

(Kobi CCD Video Camera, Model K-32HCVF, Ashmore, QLD, Australia) attached to the

bars of the crate to focus on the sheep’s sides, front and back during exposure to treatment. A

digital video recorder (Kobi H.266, Model XQ-L 900H, Ashmore, QLD, Australia) was used

to record the images, and the video data were analysed using a continuous recording of each

animal and Cowlog 2.0 software for coding of behaviours. A step was defined as a

movement made by lifting one of the limbs and setting it promptly down again. Stepping was

recorded as events and was classified as one of 9 directions: forward, backward, left, right,

the diagonals, forward left, forward right, backward left, backward right, and returning to the

same place.

2.5 Heart rate variability

Heart rate monitors (Polar S810i, Kempele, Finland) were fitted with electrodes attached to

the thorax of each sheep. Four sections of 512 beats (approximately 6 min) were extracted

from each exposure period for time and frequency domain analysis. Kubios HRV 2.1

software (Tarvainen et al., 2014) was used to analyse a number of parameters indicating

heart rate and its variability: heart rate, in beats/min (HR), standard deviation of normal to

normal intervals, in ms (SDNN), R-wave to R-wave (RR interval), square root of the mean

of the sum of the squares of differences between consecutive interbeat intervals (IBIs) (i.e.

the standard deviation of differences between successive IBIs), in ms (RMSSd), number of

differences between two successive IBIs greater than 50 ms (NN50), Fast Fourier

transformation of the high (HF) and low (LF) frequency beats in normalised units. Low (LF)

and high (HF) frequency bands widths were prescribed according to typical sheep ranges

(LF: 0.04-0.2 Hz, HF: 0.2-0.4 Hz) (von Borell et al., 2007).

2.6 Statistical Analysis

Two general linear models of the data were created, one for lateralised movement and the

second one for limb movements in the difference directions. In the models, a normal

distribution of residuals was verified with the Anderson-Darling test, including if necessary

data transformation by log10 or square.

The first general linear model was used to analyse the stepping movements of the fore and

hind limbs separately. Each movement was analysed for significance of the following

factors: sheep, left or right limbs, left or right position of sheep in the crate, treatment, day (1

to 8), as well as the interactions: left or right side position of sheep in crate (hereafter crate

position)*left or right limbs (hereafter limbs side); treatment*crate position; treatment*limbs

side.

A second general linear model was used to analyse summed left (directly left and forward

and backward diagonals to the left) and right (directly right and forward and backward

diagonals to the right) side movements for all limbs with the following factors: sheep, day,

crate position, limb side and left or right direction of limb movements, as well as the

following interactions: crate position*direction of movements; crate position*limb side; crate

position*day; direction of movements*day. Treatment factors, position of sheep, limb side

and direction of movement were all nested within sheep in the models. Fisher’s LSD

pairwise comparison tests were used to identify which means were significantly different.

For heart rate data, a general linear model of the parameters extracted from 512 beat

segments with the Kubios software (RR_interval, SDNN, HR mean, RMSSd, NN50, HF, LF,

LF/HF) was made for the following factors: sheep, day, companion and treatment. Two

commonly used methods of analysing spectral frequency components are integrals of power

spectrum density over specific bands (Fast Fourier transformation, FFT) and components

determined by autoregressive algorithms (Autoregressive method, AR) (Chemla et al., 2005).

We compared the two methods and found results to be statistically very similar, although the

FFT was slightly more closely related to limb movements. The frequency domain analysis

therefore used a Fast Fourier Transformation (FFT) obtaining high (HF) and low (LF)

frequency bands, expressed in normalized units (n.u.), and as a ratio (LF/ HF).

An initial exploratory stepwise regression was undertaken between total stepping and heart

rate variability (HRV) parameters with alpha levels to enter of 0.15. Then a Pearson

correlation matrix was prepared with correlation coefficients and P values for relationships

between each of the stepping behaviours independently and heart rate variables.

All analysis used the statistical package Minitab 17 (Minitab Inc., State College, PA, USA).

3. Results

3.1 Summary of limb movements across all treatments

Movement frequency varied both between fore and hind limbs and between directions of

movement (P< 0.001) (Table 2). The most common movements were forwards and

backwards in the fore limbs, then forwards and backwards in the hind limbs, then returning

fore and hind limbs to the same place, then fore and hind limbs sideways, then fore limbs

diagonals, and finally the least common movements were hind limbs diagonals.

3.2 Direction of movement across all treatments

There were no significant differences in overall movement (number of movements2/sheep/30

min, with squared values required to obtain normal distribution of residuals) between left and

right limbs (left [LL] = 343; right [RL] = 359, SED = 141.9, P = 0.88), left- and right-

directed limb movements (left [l] = 350, right [r] = 352 movements/ sheep/30 min, SED =

141.9, P = 0.99), or left and right position of the sheep in the crate (left [LP] = 347; right,

[RP] = 355, SED = 141.9, P = 0.94). However, there was an interaction between left and

right limbs with left and right direction of movement (LLl = 464; LLr = 222; RLl = 237; RLr

= 482 number of movements2/sheep/30 min, SED=141.9, P=0.02). Sheep therefore moved

their left and right feet more outwards than inwards. In the forelimbs this was mainly due to

differences in movement directly to the right and diagonal back right, which were

significantly increased in the right compared with the left limbs, but also diagonal back left,

which was increased in left compared with the right limbs (Figure 1). In the hind limbs it was

the forward diagonals that indicated the preference for outward movement, increased forward

left diagonal on the left limb and increased forward right diagonal on the right limb (Figure

1).

3.3 Treatment effects

There were significant differences between treatments in two forelimb movements (Figure

2). An increased movement of left limb back left and right limb back right occurred in the

Heave motion and for right limb back right in Roll motion, whereas there were no

differences in the Pitch or Control treatments for the forelimb or any treatments for the

hindlimbs.

3.4 Effect of the sheep position in the crate on limb movement

Sheep on the left side of the crate moved their forelimbs more in the right forward diagonal

and their hind limbs more directly forward and to return to the same place than those on the

right side (Figure 3). In the hind limbs this laterality also differed between the right and left

limbs: sheep on the left side of the crate moved their left limbs more in the forward and

backward directions more than sheep on the right side (Figure 4). The sheep on the left side

moved their right hind limbs backward less than sheep on the right side.

3.5 Heart rate Variability (HRV)

Across all treatments, sheep on the right side of the crate showed greater variation in normal

intervals between beats (which excluded premature ventricular or atrial contractions)

(RR_SDNN) and tended to show greater variation in successive interbeat intervals (RMSSd)

(Table 3). There was an interaction between position and treatment in the ratio of LF/HF

(SED = 2.08, p = 0.04). Sheep on the left side of the crate had a lower LF/HF ratio than those

on the right in the Control treatment, but not during Heave, Roll or Pitch treatments (Table

4). Sheep on the right hand side had an increased LF/HF ratio in the control treatment

compared to the roll treatment.

3.6 Correlations between stepping behaviour and heart rate variables

In the stepwise regression model, only one heart rate variable, HF, entered the stepwise

regression model of variables influencing total stepping at p < 0.05 (Equation 1).

Total stepping = 149.0-1.868 HF; 6.4%, P = 0.047

Equation 1

with Total stepping measured in movements/30 min and HF measured in beats per

minute. This indicated a negative relationship between the two variables.

There correlations between individual stepping movements and heart rate parameters that

differed for right and left limbs (Table 5). Correlations with mean heart rate were positive for

the right hind limb diagonal (combined L and R) and negative for the left hind limb diagonal

(and vice versa for RR mean as this is the reciprocal of heart rate). There was a negative

relationship between right limb forwards/backwards movements and HF beats, and a positive

relationship with LF beats and the ratio of LF/HF beats.

4. Discussion

Simulated sea transport motion had an impact on sheep stepping behaviour. The most evident

stepping responses of the sheep were to spread their limbs, by stepping in different

directions, principally forward and backward in the fore limbs (Figure 1, Table 2), but also

by lateralised limb movements, with left limbs moving more to the left and right limbs to the

right. This was probably to increase their stability in response to heave in particular and to a

lesser extent roll motion, but may also relate to ease of limb movement. This finding is

supported by Jones et al (2010) who worked with 5 different stocking densities during road

transport. They found that sheep, in their attempts to maintain their balance, spread their legs

sideways or took small steps forwards and backwards. As well as movement directly to the

left and right, the difference between left and right forelimbs was mainly in the backwards

diagonal movements, and that of the left and right hind limbs was mainly in the forward

diagonal movements. These may have been corrective movements in the event of sudden loss

of balance on the right or left side, with the inwards diagonals selected to maintain support

for the body.

We provide evidence that the stepping movements were stressful, since as they increased

they were associated with reduced HF measurements. HF beats represent primarily

parasympathetic (vagal) influences (and LF both parasympathetic and sympathetic), and a

reduced HF value represents a reduced parasympathetic influence that has been associated

with stress in humans (Hjortskov et al, 2004). Reduced HF has also been associated with

acute stress responses in livestock, for example during rectal palpation in dairy cows (Kovács

et al. 2014).

One of the major differences in lateral movement was that it was greater in the right than left

fore limb to the right, but not in the left fore limb to the left (Figure 1). Also in roll it was

also increased more in the right fore limb in the diagonal back right direction, whereas the

left limb did not respond similarly (Figure 2). Spatial processing is primarily in the right

hemisphere (Shinohara et al., 2012). Stepping movements are presumed to represent

subcortical startle reactions, as opposed to a reaction mediated by the cerebral cortex, which

approximately halves the reaction time from 150 to 80 ms (Valls-Solé et al., 1999). There is

evidence in humans that ipsilateral startle stimuli produce faster responses than contralateral

startle stimuli (Schilling et al., 2014), suggesting that the right forelimb would produce the

greater responses, as we found. This is despite the fact the left hemisphere is used for focal

attention, as opposed to global attention mediated by the right hemisphere (Robertson and

Lamb, 1991; Suavransri et al., 2012). It is also possible that the increased right sided stepping

was a learnt response, following handling of the sheep predominantly on the left hand side.

In support of the latter theory, no evidence of lateralisation of limb stepping movements has

been found in ovine neonates (Lane and Phillips, 2004).

Comparing the treatments (roll, pitch, heave and control), we found an impact mainly in the

fore limb stepping in backward diagonal directions (Figure 2). This could be because during

balance maintenance, forelimbs are fundamental to the support role in quadrupeds as they

carry 60% of the animal’s weight (Broom and Fraser 2007). It has been related in cows to the

anatomic differences between fore and hind legs and associated with the weight of the head

(Chapinal et al 2009). In goats the forces generated by the fore limbs are generally larger

than hind limbs during walking, which supports greater shoulder joint movement (Pandy et al

1988). It is also possible that reactions in quadrupeds in the forelimbs are faster than in the

hindlimbs because the latter have to travel down the spinal cord. In one study with humans,

hand movements reacted to a startle stimulus about 17% faster than foot movements (Valls-

Solé et al., 1999).

Roll and heave movements resulted in a higher number of stepping movements diagonally

backwards and away from the body, compared with pitch and control (Figure 2). It was in

heave that there was the greatest difference between inwards and outwards movements,

rather than roll. This may have been because heave was more likely to lead to loss of balance

than roll, with sheep spending more time bracing themselves against the sides of the crate

and less time lying down (Santurtun et al., 2015). Heave also appears to have been more

challenging for animals to deal with, as the sheep were observed to spend less time

ruminating and more time in affiliative behaviour with each other (Santurtun et al., 2015).

There were two right side lateralities – more lateral movement in the right fore limb to the

right than left fore limb to the left (Figure 1), and more right fore limb movement in the

diagonal back right direction during roll, whereas the left limb did not respond similarly.

This coincided with a positive correlation between HR and stepping to the right side and

between right limbs’ forwards/backwards movements and LF/HF ratio, which suggests that

the extent of these right side movements may be associated with stress (Hjortskov et al,

2004; Kovács et al. 2014). Versace et al (2007) observed population level lateralities in

relation to sheep preference to pass an obstacle on the right side, which they suspected arose

so that the sheep could maintain the obstacle in its left eye field of vision, processed in the

right brain hemisphere as for other flight/fight and reactive behaviours (Rogers, 2010; Robins

and Phillips, 2010). The same argument can be made for sheep in our study, that they stepped

to the right to allow the left eye to observe any potential threat, a response that may have

been specific to this study, developed during experiential learning or a characteristic of the

species.

There was also laterality in terms of sheep position in the crate; those on the left side moved

more frequently in several directions than those on the right (Figure 4). In humans and some

other animals, there is a preference for the left part of the body to be more involved in

responsive actions when they are subjected to stress (Phillips et al 2003; Rogers 2010;

Siniscalchi et al 2008). Others authors (Lane and Phillips 2004; Versace et al 2007) did not

find any population bias in tasks unrelated to social coordination, such as stepping forward

from a standing position. However, this left/right asymmetry may have been produced by

stress, as evidenced by increased SDNN and RMSSd heart rate variability measures in sheep

on the right side (Table 3). Low RR_SDNN has been related to stress, associated with a

reduction in the vagal tone (Morh et al 2002). Sheep standing on the left side viewed the

other animal through their right eye, which in several studies has been confirmed as less

preferred by animals under stress, compared with viewing them with the left eye (Peirce and

Kendrick 2002; Robins and Phillips 2010; Siniscalchi et al 2008). The lateralised movement

bias we observed in sheep on the left side, which were not viewing a sheep on their left side,

was therefore probably associated with the left visual hemisphere (Larose et al 2006; Rogers

2009). This is related to right hemisphere dominance (Anderson and Murray 2013; Peirce

and Kendrick 2002; Peirce et al 2000; Rogers 2010). The right brain hemisphere is associated

with stress and unexpected stimuli, and it controls escape and other emergency responses

such as escape and fear (Rogers 2010). We suggest that, because sheep located on the left

hand side of the crate had not got a left-side visualization of a companion, they experienced

an additional stress.

Sheep may also recognize faces more accurately from the left than the right side (Peirce and

Kendrick 2002), which when they are those of familiar animals can reduce considerably the

heart rate, cortisol and adrenaline levels (Da costa et al 2004). We therefore suggest that

sheep on the right side of the crate were less anxious due to the presence of their companion

on the left side, reducing the level of stress. Jones et al (2010) concluded that during road

transportation sheep preferred to be close to their neighbour if given the opportunity to stand

beside them. The right brain hemisphere not only processes visual stress in vertebrates, this

hemisphere also controls key physiological responses, such as heart rate (Rogers 2010).

Surprisingly, the low LF/HF ratio of sheep in the Control treatment suggests that animal on

the left had greater sympathetic nervous activity or less vagal activity than those on the right,

but this was not the case in Heave, Pitch or Roll. However, LF/HF ratio may not always be a

very reliable tool to evaluated cardiac sympathetic and parasympathetic nervous activity

(Billman, 2013).

A limitation of the study was our ability to replicate ship motion on land. The extent of

heave, roll and pitch motions was probably less than is regularly experienced on ships, but

this is unknown. Conservative values had to be chosen to safeguard the welfare of the sheep.

Furthermore, in ships the motions are complex, containing different combinations of these

motions, and others. Further work is needed on ships to quantify the extent of the motions

and the responses of livestock.

5. Conclusions

Exposing sheep to typical movements of the three principle types on ships, roll, pitch and

heave, had major effects on stepping behaviour. This appeared to relate mainly to their

attempts to maintain their balance as they moved their feet outside rather than inside.

Forelimbs played the major role in the adjustment of the body position during the attempt to

maintain the balance. Lateralised stepping responses that depended on the positioning of the

sheep in the crate suggest that sheep benefit from being able to see other sheep in their left

eye field of vision.

Conflict of interest

The authors confirm that no conflicts of interest are associated with this publication and no

financial support was given that could have influenced the outcome of the present study.

Acknowledgements

GN and ES acknowledge support from Becas Chile (Conicyt) and Conacyt Mexico,

respectively, for postgraduate student scholarships. The authors acknowledge the assistance

of Andrew Kelly of the Queensland Animal Science Precinct, UQ, and financial assistance

from the Humane Slaughter Association and the Humane Society International. The sponsors

had no involvement in the study design, conduct, report writing or decision to submit the

article for publication.

References

Anderson, D.M., Murray, L.W. 2013. Sheep laterality, Laterality 18, 179-193.

Anon 2016. Ocean Dover, Livestock carrier.

http://www.marinetraffic.com/en/ais/details/ships/shipid:728318/mmsi:53800612

2/imo:9232852/vessel:OCEAN_DROVER. Accessed 15 November, 2016

Australian Standards for the Export of Livestock (Version 2.3) 2011. Minimum per area

per head for sheep and goats exported by sea. P. 90. Department of Agriculture,

Fisheries, and Forestry, Australian Government, Canberra.

Billman, G.E., 2013. The LF/HF ratio does not accurately measure cardiac sympatho-

vagal balance. Front Soc. Psychol.4, 26. doi: 10.3389/fphys.2013.00026

Bisazza, A., Rogers, L.J., Vallortigara, G. 1998. The origins of cerebral asymmetry: a

review of evidence of behavioural and brain lateralization in fishes, reptiles and

amphibians. Neurosci. Biobehav. Rev. 22, 411–426.

Black, H., Matthews, L.R., Bremner, K.J. 1994. The behaviour of male lambs

transported by sea from New Zealand to Saudi Arabia. New Zeal. Vet. J. 42, 16-

23.

Broom, D.M., 2000. Welfare assessment and welfare problem areas during handling and

transport. In: Grandin, T (ed) Livestock handling and transport 2nd, CABI

publishing: Oxon, UK, pp.43-61.

Broom, D.M., Fraser, A.F., 2007. Domestic Animal Behaviour and Welfare, 4 th Edition,

Wallingford: CABI.

Chapinal, N., de Passille, A.M., Weary, D.M., von Keyserlingk, M.A.G., Rushen, J.

2009. Using gait score, walking speed, and lying behaviour to detect hoof lesions

in dairy cows, J. Dairy Sci. 92, 4365–4374.

Chemla, D., Young, J., Badilini, F., Maison-Blanche, P., Affres, H., Lecarpentier, Y.,

Chanson, P. 2005. Comparison of fast Fourier transform and autoregressive

spectral analysis for the study of heart rate variability in diabetic patients. Int. J.

Cardiology 104, 307-313.

Cockram, M.S., Kent, J.E., Goddard, P.J., Waran, N.K., Mcgilp, I.M., Jackson, R.E.,

Muwanga, G.M., Prytherch, S. 1996. Effect of space allowance during transport

on the behavioural and physiological responses of lambs during and after

transport, Anim. Sci. 62, 461-477.

Da Costa A., Leigh A., Man M., Kendrick K. 2004. Face pictures reduce behavioural,

autonomic, endocrine and neural indices of stress and fear in sheep. Proc. Royal

Soc. B: Biol. Sci. 271, 2077–2084.

FAOSTAT, 2016. Trade statistics. http://faostat3.fao.org/home/E. Accessed 17

November, 2016.

Frondelius, L, Jarvenranta, K., Kopenen, T. and Mononen, J. 2015. The effects of body

posture and temperament on heart rate variability in dairy cows. Phys. Behav.

139, 437-441.

Helling, K., Scherer, H., Hausmann, S., Clarke, A.H. 2005. Otolith mass asymmetries in

the utricle and saccule of flatfish. J. Vestib. Res.15, 59–64.

Hjortskov N, Rissen D, Blangsted AK, Fallentin N, Lundberg U, Sogaard K. 2004. The

effect of mental stress on heart rate variability and blood pressure during

computer work. Eur. J. Appl. Physiol. 92, 84-89.

Jones, T.A., Waitt, C., Dawkins, M.S. 2010. Sheep lose balance, slip and fall less when

loosely packed in transit where they stand close to but not touching their

neighbours. Appl. Anim. Behav. Sci. 123, 16-23.

Knowles, T.G., Brown, S.N., Warriss, P.D., Phillips, A.J., Dolan, S.K., Hunt, P., Ford,

J.E., Edwards, J.E., Watkins, P.E. 1995. Effects on sheep of transport by road for

up to 24 hours. Vet. Rec. 136, 431-438.

Knowles, T.G. 1998. A review of the road transport of slaughter sheep. Vet. Rec. 143,

212-9.

Kovács L, Tőzsér J, Szenci O, Póti P, Kézér FL, Ruff F, Gábriel-Tőzsér G, Hoffmann D,

Bakony M, Jurkovich V. 2014. Cardiac responses to palpation per rectum in

lactating and nonlactating dairy cows. J. Dairy Sci. 97, 6955-6963.

Lane, A., Phillips C.J.C., 2004. A note on behavioural laterality in neonatal lambs. Appl.

Anim. Behav. Sci. 86, 161-167.

Larose, C., Rogers, L.J., Ricard-Yris, M.A., Hausberger, M. 2006. Laterality of horses

associated with emotionality in novel situations. Laterality 11, 355–367.

Lychakov, D.V., 2013. Behavioral lateralization and otolith asymmetry. Evol. Biochem.

Phys. 49, 441–456.

McDonald, M., 1993. SHF SATCOM terminal ship-motion study. Report No.

1578 Naval Command, Control and Ocean Surveillance Center, California, USA.

Meat & livestock Australia, 2016. Live export statistics. http://www.mla.com.au/Prices-

markets/Trends-analysis.

Mohr, E., Langbein, J., Nürnberg G. 2002. Heart rate variability — a noninvasive

approach to measure stress in calves and cows. Physiol. Behav. 75, 251–259.

Norris, R.T. 2005. Transport of animals by sea. Rev. Sci. Tech. 24, 673-681.

Pandy, M.G., Kumar, V., Berme, N., Waldron, K.J. 1988. The dynamics of quadrupedal

locomotion. J. Biomech. Eng. 110, 230-237.

Peirce, J.W., Leigh, A.E., Kendrick, K.M. 2000. Configurational coding, familiarity and

the right hemisphere advantage for face recognition in sheep. Neuropsychologia

38, 475–83.

Peirce, J.W., Kendrick, K.M. 2002. Functional asymmetry in sheep temporal cortex,

Neuroreport. 13, 2395-2399.

Phillips, C.J.C., 2008. The welfare of livestock during sea transport. In: MC Appleby, V

Cussen, L Garcés, LA Lambert, J Turner (eds): Long Distance Transport and

Welfare of Farm Animals, CAB International, Wallingford, UK, pp.137-156.

Phillips, C.J.C., Llewellyn, S., Claudia, A., 2003. Laterality in bovine behaviour in an

extensive and intensive dairy herd. J. Dairy Sci. 86, 3167–3173.

Putnam, S., Noonan, M., Bellia, C. 1996. Direction and reliability of head tilt in humans.

Laterality 1, 153–160.

Randall, J.M. 1993. Environmental parameters necessary to define comfort for pigs,

cattle and sheep in livestock transporters. Anim. Prod. 57, 299-307.

Robertson, L.C., Lamb, M.R. 1991. Neuropsychological contributions to theories of

part/whole organization. Cognitive Psychol. 23, 299-330.

Robins, A., Phillips, C.J.C. 2010. Lateralised visual processing in domestic cattle herds

responding to novel and familiar stimuli. Laterality 15, 514-534.

Rogers, J.L., 2009. Hand and paw preferences in relation to the lateralized

brain. Philosophical Trans. Royal Soc. London B: Biol. Sci. 364, 943-954.

Rogers, J.L. 2010. Relevance of brain and behavioural lateralization to animal welfare.

Appl. Anim. Behav. Sci. 127, 1-11.

Santurtun, E., Moreau, V., Phillips, C.J.C. 2014. A novel method to measure the impact

of sea transport motion on sheep welfare. Biosyst. Eng. 118, 128-137.

Santurtun, E., Phillips, C.J.C. 2015. The impact of vehicle motion during transport on

animal welfare. Res. Vet. Sci. 100, 303-308.

Santurtun, E., Moreau, V., Marchant-Forde, J.N., Phillips, C.J.C. 2015. Physiological

and behavioural responses of sheep to simulated sea transport motions. J. Anim.

Sci. 93, 1250-1257.

Schilling, T. M., Larra, M. F., Deuter, C. E., Blumenthal, T. D., Schächinger, H. 2014.

Rapid cortisol enhancement of psychomotor and startle reactions to side-

congruent stimuli in a focused cross-modal choice reaction time paradigm. Eur.

Neuropsychopharm. 24, 1828-1835.

Shinohara, Y., Hosoya, A., Yamasaki, N., Ahmed, H., Hattori, S., Eguchi, M.,

Yamaguchi, S., Miyakawa, T., Hirase, H., Shigemoto, R. 2012. Right-

hemispheric dominance of spatial memory in split-brain mice. Hippocampus 22,

117–121.

Siniscalchi, M., Quaranta A., Rogers, L.J., 2008. Hemispheric specialization in dogs for

processing different acoustic stimuli. PloS One e3349.

Skraastad, S.T., 1983. Conversion to livestock carriers. In: D. Deere (Ed.), Animal

Transport by Sea. Marine Publications International Limited, London, pp. 41-59.

Suavansri, K., Falchook, A.D., Williamson, J.B. Heilman, K.M. 2012. Right up there:

hemispatial and hand asymmetries of altitudinal pseudoneglect. Brain Cognition

79, 216-220.

Tarvainen, M.P., Niskanen, J.P., Lipponen, J.A., Ranta-aho, P.O., Karjalainen, P.A.

2014. Kubios HRV–Heart rate variability analysis software. Comput. Meth. Prog.

Bio. 113, 210–220.

Valls-Solé, J., Rothwell, J.C., Goulart, F., Cossu, G., Munoz, E. 1999. Patterned ballistic

movements triggered by a startle in healthy humans. J. Physiol. 516, 931-938.

Versace E., Morgante, M., Giuseppe, P., Vallortigara, G. 2007. Behavioural

lateralization in sheep (Ovis aries). Behav. Brain Res.184, 72-80.

Von Borell, E., Langbein, G., Després, S., Hansen, C., Leterrier, J., Marchant-Forde, J.,

Marchant-Forde, R., Minero, M., Mohr, E., Prunier, A., Valance, D., Veissier, I.

2007. Heart rate variability as a measure of autonomic regulation of cardiac

activity for assessing stress and welfare in farm animals — a review. Physiol.

Behav. 92, 293–316.

Figure 1. Differences in movement (steps/sheep/30 minutes) between left and right fore

limbs and between left and right hind limbs of sheep in all treatments combined

Figure 2. Effects of treatment on significant differences in movement (steps/sheep/30

minutes) of left and right fore limbs, from treatment x side of body interactions

Figure 3. Significant differences in movement (steps/sheep/30 minutes) of the fore and

hind limbs (left and right combined) of sheep on the left and right sides of the

crate

Figure 4. Significant differences in movement (steps/sheep/30 minutes) of left and right

hind limbs for sheep on the left and right sides of the crate

Figure 1.

( highly significant; significant; non-significant)

* P < 0.05; ** P < 0.01, *** P < 0.001

(return to same place)

3.02 4.11

FORELIMBS 13.5 12.58

1.42 1.36 1.92 0.94

** ** 2.38 3.80 3.70 1.94

*** ***

*

* *

1.06 0.44 14.44

0.84 0.22 12.89

HINDLIMBS 4.17 4.31

RIGHT LIMBS LEFT LIMBS

0.05 0.45 0.19 0.36

***

* 3.44 4.25 4.66 3.08

0.23 0.22 0.34 0.13

5.53 5.03

**

*

***

*

**

*

4.41 5.89

Figure 2.

( highly significant; significant)

ROLL

B

PITCH

HEAVE

CONTROL

2.31a

Left Right

** * * **

0.38bc 1.00abc 1.50ab

Right Left

**

** ** *

0.63bc 0.31abc 0.19c 0.13c

Left Right

** ** * *

1.06ab 1.38a 0.13c 0.13c

Right Left

** ** * *

0.25bc 0.31bc0.31abc 0.25bc

Means for each limb moving in a specific direction that have different superscripts between

treatments are significantly different at * P < 0.05 (*) or P < 0.01 (**).

Figure 3.

( highly significant; significant)

* P < 0.05; ** P < 0.01, *** P < 0.001

**

* *

***

1.41 1.64

***

******

**

4.24 4.08

5.15 4.94

Sheep left side Sheep right side

Fore limbs

Hind limbs

Figure 4. ( highly significant; significant)

CRATE

* P < 0.05; ** P < 0.01

Right

*

0.41a

*

4.38a *4.22a

**

**

2.66a

0.13ab

* 6.81a 5.06ab

*

*

**

6.25a 4.09ab

0.13ab0.06b

Sheep -Left Sheep-Right

4.94a

2.38b 4.25ab

6.00ab 3.25b

** ** **

** ** *

Left Left Right

*

Table 1 Allocation of sheep pairs to treatment for the eight periods

Day Sheep 1&2 Sheep 3&4 Sheep 2&3 Sheep 1&4

1 Pitch Roll Heave Control

2 Roll Pitch Control Heave

3 Heave Control Pitch Roll

4 Control Heave Roll Pitch

5 Pitch Roll Heave Control

6 Roll Pitch Control Heave

7 Heave Control Pitch Roll

8 Control Heave Roll Pitch

Table 2.Frequency of the movements in the different directions for fore and hind limbs (P<

0.001)

Movement

Mean stepping rate

(steps/ sheep/30

minutes)†

SEM

Coefficient

of variation

Fore Limb

Forward

13.1a

0.99

0.85

Backward 13.7a 0.99 0.82

Right 2.8bcdef 0.55 2.21

Left 3.0bcde 0.61 2.24

Same place 3.6bcd 0.47 1.50

Diagonal

Forward left 1.2cdef

0.21 1.99

Diagonal

Forward right 1.6cdefg

0.31 2.14

Diagonal back

left 0.6def

0.12 2.16

Diagonal back

right 0.6def

0.11 1.98

Hind Limb

Forward

4.2b

0.46

1.23

Back 5.3b 0.55 1.17

Right 3.7bcd 0.68 2.09

Left 4.0bc 0.72 2.01

Same place 5.1b 0.63 1.40

Diagonal

Forward left 0.2g

0.04 2.50

Diagonal

Forward right 0.3fg

0.06 2.07

Diagonal back

left 0.2fg

0.06 2.40

Diagonal back

right 0.2g

0.04 2.74

† Means with different superscripts differ significantly by Fisher’s LSD pairwise comparison

test

Table 3. Significant (P<0.05) or close to significant (P<0.10>0.05) effects of left or right side

position (LP and RP) of sheep in the crate on two HRV parameters, standard deviation of

normal to normal intervals in R waves, and square root of the mean of the sum of the squares

of differences between consecutive interbeat intervals (IBIs) (i.e. the standard deviation of

differences between successive IBIs) (RMSSd)

HRV

parameter

LP RP SED P-value

SDNN

(ms)

35.8 43.0

4.15 0.026

RMSSd

(ms) 40.6 49.6

6.86 0.090

Table 4. Interactions between position of sheep (L = left; R = Right) and treatment (Control,

Heave, Pitch and Roll) on the Fast Fourier transformed ratio of low to high frequency heart

beats (FFT LF/HF)

Treatment Position FFT_LF/HF†

Control L 2.62 b

Control R 5.99 a

Heave L 4.05 ab

Heave R 5.00 ab

Pitch L 3.33 ab

Pitch R 4.36 ab

Roll L 5.37ab

Roll R 2.51b

† Means with different superscripts differ significantly by Fisher’s LSD pairwise comparison

test

Table 5. Significant (P< 0.05) associations between stepping motions of the left and right

fore and hind limbs and the heart rate parameters, heart rate (HR), R-wave to R-wave interval

(RR), square root of the mean of the sum of the squares of differences between consecutive

interbeat intervals (IBIs) (i.e. the standard deviation of differences between successive IBIs)

(RMSSd), Fast Fourier transformation of the high (HF) and low (LF) frequency beats. Top

numbers are correlation coefficients and bottom numbers are P values.

Left hind

limb diagonal

(combined

left and

right)

Right hind limb

diagonal

(combined left

and right)

Right fore limb

forward

Right fore limb

backward

Right hind limb

forward

Right hind limb

backward

HR mean

(beats/min)

-0.267 0.428

(0.036) (0.001)

RR interval (ms) 0.295 -0.373

(0.020) (0.003)

RMSSd -0.242 -0.249

(0.059) (0.051)

FFT-HF, n.u. -0.359 -0.317 -0.314 -0.338

(0.004) (0.012) (0.013) (0.007)

FFT-LF, n.u. 0.358 0.317 0.313 -0.337

(0.004) (0.012) (0.013) (0.007)

FFT-LF/HF 0.335 0.260 0.310 0.029

(0.008) (0.041) (0.014) (0.020)