1. J. White, A. McBride, E. Redhead Criticism of Methodology
Comparison of Tethering and Group-Pen Housing for Sled Dogs (Siberian Huskies)
J. White, A. McBride, E. Redhead
University of Southampton, School of Psychology
Results
Abstract
This experiment investigated the difference in recorded behaviours in Siberian husky sled dogs between two types of housing systems (tethering and group pen housing) and exercise. It was predicted that group housing would record lower levels of repetitive behaviours and other behavioural indices of stress. The experiment employed a repeated measures design with tethered/non-tethered, exercise/no exercise conditions as the independent variables and various behavioural indicators as dependent variables. Participants (n=9) were exposed to four different housing conditions and filmed to record behaviours for analysis. The conditions consisted of six-months with no-exercise/tethered (Condition A), exercise (daily running)/tethered (Condition B), four weeks no-exercise/tethered (Condition C) and no exercise/un-tethered (Condition D). The tethered housing conditions produced evidence of more repetitive behaviours and fewer social behaviours than the un-tethered housing condition. Un-tethering reduced rebound and repetitive behaviours. Tethering without exercise (conditions A and C) showed significantly greater levels of vigilance and agonistic behaviour than all other conditions, indicative of compromised welfare. It can therefore be concluded that long-term tethering without exercise is associated with abnormal activity patterns and levels of behaviours expressed.
Behavioural analyses were conducted post-hoc. Upon the first viewing of video, the most frequent behaviours where divided into inactive and active categories. Only significantly different behaviours were used in the analysis. Upon subsequent review of video footage, these categories were further divided into investigative behaviours, social-directed behaviours and self-directed behaviours. The most frequent behaviours notes by the experimenter were used to create the behavioural ethogram.
A level of significance of 5% was used. The most frequent active (pacing, fast pacing, investigative and social behaviours) and inactive behaviours (lying down/alert and sleeping) were used in the analysis. As difference in behaviour between the four independent variables was predicted but as the direction of that difference not predicted, the hypothesis was non-directional and the test was 2-tailed. One-way ANOVA was carried out with conditions A,B,C, and D as the independent variables and frequency of behaviours as the dependent variable, followed by post-hoc paired t-tests.
Most Common Behaviours
Sleeping is the most frequent behaviour in Condition C
Little evidence of sleeping in the three remaining conditions.
Lying down awake is the second most frequent behaviour in the tethered conditions (A,B,C).
Pacing is the most frequent behaviour in the un-tethered condition (D).
Sleeping
Significant difference between conditions (p<.01). The dogs remained vigilant in the non-exercise conditions the variable of un-tethering did not have an effect on sleeping.
Figure 2, The Tethered Conditions
In conditions A, B and C dogs were secured by means of a 2.3 metre chain-tether attached to each individual dog’s leather collar. Dogs were tethered in front of their individual dog-houses. The surface area was dirt. Each dog had freedom of movement within the radius of the tether, as well as on top of and inside their doghouse. The tethers permitted some physical contact with neighbouring dogs, but this was inconsistent.
Figure 3, The un-tethered condition
In the un-tethered condition (D) dogs were moved to group pens measuring 16m2 and un-tethered. This occurred one week prior to filming to allow the dogs to habituate to the new condition. Each pen was fenced with 3 metre high chain-liked fencing. The pens contained two open wooden structures with raised floors, which housed drinking and food bowls. The size of these structures varied, from 1.2m x 1m m to 3m x 1.5m. Each group pen bordered another by chain-link fencing, allowing visual and olfactory access to two other pens.
Method
Objective
The experiment considered previous research on sled dog housing and investigated whether there would be a difference in behaviour between the independent variables of exercise, pen housing and tethering.
Participants
Nine dogs (3=male; 6=female) were selected at random from a population of 300 purebred Siberian husky dogs at a commercial sled dog establishment. All dogs were born at the kennel and raised there from birth. All have been tethered continually from four months of age. The participants ranged in age from 3 to 7 years old, with a mean age of 4.5 years (SD = 1.75).
Design
Repeated measures design. All dogs participated in all conditions.
9 dogs were exposed to four different housing conditions and filmed to record behaviours for analysis. The conditions consisted of six-months with no- exercise/tethered (Condition A), exercise (daily running)/tethered (Condition B), four weeks no-exercise/tethered (Condition C) and no exercise/un-tethered (Condition D) (Table 1).
Apparatus and Materials
Three digital video cameras on tripods were used to film the dogs’ behaviour in all conditions. Tripods had been installed 10 metres away from the dogs 24 hour prior to filming to allow them to habituate to the filming equipment. Video cameras were installed each morning by the observer, 30 minutes prior to filming to habituate the dogs to her presence. For filming accuracy the participants were divided into three groups of 3 dogs. Film from each condition was digitally transferred to computer and DVDs for analysis
Procedure
After the installation of each camera, participants were filmed remotely (observer not present) in each of the four conditions for a period of 13.5 hours over three days (4.5 hours per day). Of this, 1.5 hours of filming occurred each morning, from 8:30am to 10:00am, and 3 hours each afternoon, from 2pm to 5pm prior to the evening feed. There was no interaction between the participants and the observer during filming, prior or after filming. Weather conditions were noted, as were uncontrolled variables as they occurred.
Table I
Description of Independent Variables (Conditions)
Fast Pacing
A significant difference (p< 0.05) was revealed in the level of fast-pacing activity (running in a fast-pace two beat gait and jumping on and off boxes). In the un-tethered condition, virtually no fast pacing occurs.
Aggression
Significant levels of aggression seen in non-exercise conditions A and C (p<.05).
No aggression occurred in condition B (dogs were exercised daily).
Figure 1, Sled dogs in northern Canada.
Conclusion
Although differences were found in the occurrence of a number of behaviours between the un-tethered and the tethered condition, the effect of exercise (Condition B) was found overall to produce more a significant difference, especially in the amount of time spent sleeping in Condition B, as was expected. By contrast, when exercise was not provided (Conditions A, C and D), the dogs remained alert.
Long-term tethering of sled dogs produced evidence of more alert and repetitive behaviours such as fast pacing and fewer social behaviours than the un-tethered housing condition. Placing participants in un-tethered (group housing) reduced rebound and repetitive behaviours. The variable of exercise also affected behaviour. Tethering without exercise (Conditions A and C) produced significantly more vigilance and agonistic behaviour than either tethering with exercise or group housing without exercise. It is likely that social facilitation through sled running exercise produced a calming effect, which enabled the dogs to sleep more in Condition B (Clarke, et al., 1997).
Levels of aggression were highest in the tethered non-exercise conditions of A and C and did not occur in any significant level in conditions B or D. By contrast the most aggression occurred in the condition with the least stimuli, C, when there were fewer caregivers in the kennel and no exercise. When stimuli were presented, such as a caregiver walking through the grounds, dogs would react with a frustrative response which was re-directed onto neighbouring dogs. If the tethers allowed physical contact, aggression took the form of nipping neighbouring dogs’ noses and tails. Although actual physical contact rarely occurred, there was evidence that these aggressive responses were not inhibited and that injury would have occurred more frequently had the tethers allowed more contact.
The lack of activity in the non-exercise conditions was expected to increase the number of self-directed behaviours recorded. This did not occur. Less than 1% of behaviours were self-directed, and this took the form of grooming. The most grooming took place in the exercise condition B. This is self explanatory: the dogs had been running previous to filming on a number of snow, ice and road surfaces and were therefore motivated to groom and clean their paws. These behaviours were of short-duration (less than 3 minutes per instance) and were not correlated with any individual dog or sex of dog.
There was no evidence of repetitive or stereotypical behaviours which supports with previous research on farm animal tethering and the incidences of stereotypies (Schouten, et al., 1991; Vieuillethomas, et al, 1995). In these studies, although there were no obvious behavioural indicators of stress from tethering, biobehavioural indicators, such as elevated cortisol and heart rate revealed significant stress levels. Therefore, measurement of participants’ heart rate and/or cortisol concentrations in the current study may have revealed more obvious indicators of stress from tethering.
It can therefore be concluded that long-term tethering without exercise produces abnormal activity patterns and levels of behaviours in sled dogs, which may be indicative of compromised welfare. Further research on sled dog welfare using biobehavioural indicators of welfare and motivation testing in a controlled experimental setting should be undertaken. Such research might provide evidence that the initial investment in building group housing pen facilities would be offset by the physical and psychological benefits associated with good welfare.
Introduction
This study investigated the differences in behaviour between sled dogs based on housing methods (tethering and un-tethering) and exercise (exercise/no exercise).
Research on tethering other domestic species found behavioural indicators of stress, such as increased repetitive locomotory behaviour in sheep (Wemelsfelder & Farish, 2004), excessive vocalisations in cattle (Watts & Stookey, 2000) and stereotypic pacing in pigs (Schouten, et al. 1991). However, the only study which had previously investigated sled dog behaviour (Yeon, et al. 2001), found an increase of repetitive behaviours when sled dogs that had traditionally been tethered were released from tethers and housed in small, single pens. In this case, the small size of the pens and single-dog housing could have been responsible for the increase in repetitive behaviours, findings which are supported by other studies (Clark, et al ; Hubrecht, et al. 2002). In contrast the current experiment used group housing in large pens as the alternative housing condition to tethering.
Sled dogs are highly motivated by social facilitation (Coppinger & Coppinger, 2001), behaviour which is expressed while running with others in a dog team. The current research hypothesised that the prevention of this highly motivated behaviour would impact significantly on behaviour. Therefore “extended periods without exercise” was included as a variable for analysis, as sled dogs are often not exercised (group running) for anywhere between 4 and 6 months during the summer.
Dogs were filmed in four conditions for a period of 13.5 hours over three days (4.5 hours per day). The film was analysed and four behaviour categories noted were: locomotory behaviours (active vs. inactive); investigative behaviours (object-directed) such as scavenging or eating; social-directed behaviour (directed at another dog, human carer or visitor); and self-directed behaviours (such as self-grooming or licking fur). An ethogram was used to record each participant’s behaviours on the subsequent review at one minute intervals. Previous research had suggested a significant difference in locomotory behaviours between restricted housing environments, therefore this current experiment assessed activity levels overall (active and inactive behaviours) to provide an initial indicator of differences due to the independent variables of tethering and exercise. As a significant difference between activity levels was found between conditions, the experiment then looked at the most frequent behaviours within the active and inactive categories. For example, although the dogs would lie down, the difference in the amount of actual sleep versus lying down/alert behaviours between conditions was recorded. This experiment therefore predicted that a significant difference in behaviours would be found between the four conditions.
Criticism of Methodology
CONDITION
Mean
Temp
Tethered/
un-tethered
Housing
Exercise A
8C
Tethered.
Individual
No exercise for 6 months B
4C
Tethered
Dogs received daily exercise of a minimum of one-hour on dog sleds, except during the filming period C
11C
No exercise for 4 weeks D
24C
Un-tethered
Group pens of 25 metres sq.
No exercise for 3 months.
This study took place in a commercial sled-dog facility. A number of uncontrolled variables periodically occurred: dogs escaping from tethers or puppies wandering through kennels. Daily routines were interrupted causing additional confounding variables. In the exercise condition, the number of non-participating dogs taken out on sled runs varied over the filming period. Dogs received more one-on-one human contact during the dog sledding season. Only nine dogs were selected at random for the study as a larger number could not be accommodated in the group-pen condition furthermore, although it has been suggested that smaller group pens, for example 10 m2, could house up to five dogs (see Hubrecht et al., 1992), insufficient time prior to filming prevented the habituation of more than 3 dogs to the group kennel condition. A larger sample in the un-tethered condition may have produced different results.
This study also did not assess the sequences of behaviour patterns, which may have revealed different results. An example of this occurred in analysing “pacing”. The function and motivation of pacing in Conditions A, B, and C differed substantially from D, although in this experiment, they were both defined as “pacing”. For example, pacing was repetitive and seemingly functionless in A,B and C but allelomimetic in D, in that the dogs would follow each other around the pen.
Another criticism of this study was that data of the dogs once returned to baseline could not be included in this paper due an abnormal number of uncontrolled variables (i.e. neighbouring bitches in oestrus).
Therefore, although the results from this study indicate a change in behaviour between conditions, the inability to control variables meant that conditions could not be standardised to the degree expected for an experimental study. One aspect that could be controlled, however, was the prior experience of the dogs, as they had all been born and raised in the kennel and reared identically.
References
Clark J.D., Rager D.R., Crowell-Davis S., & Evans D.L. (1997). Housing and exercise of dogs: Effects on behaviour, immune function, and cortisol concentration, Laboratory Animal Science, 47 (5),
Coppinger R., & Coppinger, L.(2001). Dogs, A new Understanding of Canine Origin, Behaviour, and Evolution, Chicago: University of Chicago Press.
Hubrecht, R.C., Serpell, J.A., & Poole T. B (1992). Correlates of pen size and housing conditions on the behaviour of kennelled dogs. Applied Animal Behaviour Science, 34,
Schouten, W., Rushen, J., & Depassille, A.M.B. (1991). Stereotypic Behavior and Heart-Rate in Pigs. Physiology & Behaviour, 50 (3),
Vieuillethomas C., Lepape G., & Signoret J.P. (1995). Stereotypies in Pregnant Sows: Indications of influence of the housing system on the patterns expressed by the animals. Applied Animal Behaviour Science , 44 (1),
Watts, J.M., & Stookey, J.M. (2000). Vocal behaviour in cattle: the animal’s commentary on its biological processes and welfare. Applied Animal Behaviour Science, 67,
Wemelsfelder, F., & Farish, M. (2004). Qualitative categories for the interpretation of sheep welfare: a review. Animal Welfare 13 (3),
Yeon, S.C., Golden, G., Sung, W., Erb, H.N., Reynolds, A.J., & Houpt, K.A. (2001). A comparison of tethering and pen confinement of dogs. Journal of Applied Animal Welfare Science, 4(4),