The Effect of Day Time and Night Time on the
Natural Foraging Behaviour of Captive Fennec
Foxes (Fennecus zerda) in the Zoo
Rosalie Ursinus (11278110)
Bsc Thesis, University of Amsterdam
AbstractTo maintain and improve the welfare in captivated animals, zoos tend to keep their animals in natural-looking enclosures. For the fennec foxes in ARTIS Zoo, Amsterdam, it is important to include features that can satisfy the animals’ need to dig and forage. The aim of this study was to examine the effect of day time and night time on the natural foraging behaviour of captive fennec foxes. Two individuals were included, one male and one female fennec fox, housed in an inside enclosure. The duration and frequency of digging and exploring behaviour of the fennec foxes were observed for a period of 23 consecutive days with 90 minute’ observations at noon and at midnight. The behavioural data was analysed in R, version 3.6.1, using a negative binomial model, and the overall effect of day time and night time was assessed by comparing the full model’s deviance with that of a null model using a likelihood ratio test. The results of this study showed that the male fennec fox had a significant increase in exploring (P<0.001) and digging (P<0.01) behaviour during night time. However, no differences in the foraging behaviour of the female fennec fox were found between day time and night time. Still, since both fennec foxes manifested foraging behaviour at night while no food was provided, this study suggests that the fennec foxes clearly express a need for digging. Therefore, it is advisable for the zoo to further facilitate this behaviour as it can be beneficial for their wellbeing.
Keywords: day time, environmental enrichment, fennec fox, Fennecus zerda, natural
foraging behaviour, night time, zoo
Data repository (restricted access)
DOI: 10.5281/zenodo.3975636 Uploaded online 07 August 2020
Introduction
There is a growing interest and need for zoos to develop and implement enrichment projects in order to maintain a high standard of animal welfare. Many studies report the positive effect of enrichment on persistent stereotypic behaviour of captive animals in the zoo (for a review see Shyne, 2006). While an increasing number of zoos have improved the enclosures of their animals by giving them a more natural look and furnishing them with all sorts of natural components such as trees and rocks, it is still not always certain whether the behavioural and ecological needs of the animals are being adequately met. At ARTIS Zoo in Amsterdam, the Netherlands, I investigated whether this is the case for their fennec foxes. While their enclosure may look like a miniature desert, it is unclear whether enough stimuli are present for the fennec foxes to behave naturally and to prevent stereotypic behaviour.
The fennec fox (Fennecus zerda) is the smallest canid and can be found in the Sahara of North Africa, the Arava desert in Israel, and the Arabian desert (Sillero-Zubiri et al., 2004). These tiny foxes are well adapted to the extreme temperatures in these environments. The soles of their feet are covered with a layer of long, soft hair to protect them from the hot sand and their thick fur provides thermoregulatory advantages during cold desert nights (Dempsey et al., 2009). The most distinctive feature are their relatively big ears that can grow up to half of their body length. Apart from keeping the foxes cool by radiating body heat, the ears serve a purpose in detecting underground prey (Gauthier-Pilters, 1967). Fennecs hunt alone and mainly rely on their hearing senses to locate prey hidden in the sand (Kingdon, 1990). Besides excavating their prey, they dig underground burrows where they shelter and sleep with their family which can consist of around 10 individuals (Larivière, 2002). While it has been established that digging is a common and important natural behaviour of the fennec foxes, data from a pilot study at ARTIS Zoo in Amsterdam, the Netherlands, showed that captive fennecs in the zoo rarely dig in their enclosure (Ursinus, 2019). It is even possible that most fennec enclosures in zoos impede digging behaviour as the substrate is too thin or even impenetrable. Because of this inability to manifest their natural behaviour, captive fennec foxes can show patterns of stereotypic behaviour: functionless, repetitive behaviours induced by frustration (Mason et al., 2007). While stereotypic behaviours may have resulted from passed stressors, they often indicate that the current environment leads to uncontrollable chronic stress. The pacing of wild captive carnivores illustrates this quite well; fennec foxes pace more often when the number of zoo visitors increases (Carlstead, 1991).
The existing enrichment studies on fennec foxes all focus on reducing stereotypic behaviours by either minimizing noise pollution (Carlstead, 1991), fighting poor husbandry-regimes (Mason et al., 2007), or by varying the certainty of location and timing of food arrival (Watters et al., 2011). However, no studies have ever been conducted on stimulating the natural foraging behaviour of the fennec foxes. Natural foraging behaviour can be divided into several categories (Larivière, 2002). First, the fennec fox will react on hearing the sound of a prey. This can be observed as movement of its ears and changing its posture from being relaxed to being alert. Then, the fennec fox will explore the source of the sound. This can be observed as a movement towards the sound, staring at the ground while rolling its head from side to side, and moving the sand a little with its nose. Eventually, the fennec fox will react to the sound by digging rapidly with the two front feet and capturing the prey in its mouth. Studies on fennec foxes in the wild have found that they are mainly nocturnal animals (Gauthier-Pilters, 1967) and a recent study even recorded that the peak of their activity lies between 18:00 and 07:00 (Karsenne et al., 2019). In captivity, fennec foxes have a circadian cycle of body temperature and heart rate that decreases during the day and increases at night when the animals become more active (Larivière, 2002). However, in ARTIS Zoo the fennec foxes are kept in a diurnal
rhythm where they are fed during the day. This may cause a disruption of their natural behaviour and could potentially lead to stress and poor well-being.
The purpose of this study was to examine what stimulates the natural foraging behaviour of the captive fennec foxes. In the first part of the study – which eventually had to be suspended because of the partial lockdown due to the Coronavirus in the Netherlands in March to July 2020 – food would have been introduced hidden in the substrate and would have been detectable by using the auditory cues of cricket chirping coming from a separate speaker buried with the food. Watters et al. (2011) found that introducing some uncertainty in the environment of an animal is more stimulating than full predictability and therefore reduces anticipatory behaviour. For this reason, the effect of the unpredictability of the location where the food would have been buried would have also been tested. This effect would have been examined by dividing the study in two enrichment treatments; in the first treatment the location of the buried food would have always been the same, while in the second treatment the location of the buried food would have changed every day. With this set-up, it would have been possible not only to observe whether the unpredictability of the location changes the range within which the fennec foxes will search for food, but also whether the unpredictability stimulates more frequent explorative behaviour. In order to explain how the fennec foxes use their large ears to catch prey, and whether it is a learned behaviour or not, this study would have focused on fennec foxes in captivity who had never encountered underground prey. This part of the study would have tried to answer several questions: (1) Can the introduction of underground prey that can be detected using the auditory cues of underground sound increase their natural foraging behaviour?, (2) Does their foraging efficiency increase over time?, (3) Does the certainty of the location of the food influence the frequency of their natural foraging behaviour and overall foraging activity afterwards?, (4) Is there a correlation between the time the fennecs need to find the food and the certainty of the location of the food?, and (5) How long does it take for the fennecs to forget the primary learned location of the food or the availability of underground food in general? The hypothesis would have been that (1) the fennec foxes would immediately react to the underground sound and would start digging to find the buried food, (2) they would learn that the underground sound is a cue for buried food which over time increases their foraging frequency, (3) the certainty of the location of the food would restrict the fennecs to only search for food in that particularly area and that when the location is uncertain the fennecs would perform explorative behaviour in the entire enclosure, (4) the fennecs would find the food faster when the location of the food is certain and thus show a steeper ‘learning curve’ than when the location is uncertain and random, (5) with the Temporal Weighting Rule it can be predicted that the fennecs would show a higher frequency of explorative behaviour in the learned location since they remember the high value of that location. They would be therefore expected to take longer to forget the availability of underground food after the treatment where the location is certain than after the treatment where the location is uncertain (Shettleworth, 2009). However, due to the COVID-19 pandemic in 2020, this part of the study could not be conducted and aforesaid is therefore purely hypothetical.
In the second part of this study, the effect of day time and night time on the activity and natural foraging behaviour of captive fennec foxes was examined. In their natural environment, fennec foxes hunt during night time as they are nocturnal animals. Since the fennec foxes in ARTIS Zoo were fed during the day, it may be possible that the circadian rhythm of the fennec foxes is disrupted and that they will be more active during day time. However, as mentioned before, their body temperature and heart rate still increase at night when in captivity. For this reason, it was expected that the natural foraging behaviour of the captive fennec foxes increases during night time and decreases during day time. In this study, digging and exploring behaviour were used as indicatives for natural foraging behaviour. Since these fennec foxes never received their food underground, it could be argued that their digging behaviour is not related to hunting
but perhaps indicates nesting behaviour instead. However, the results from the previous pilot study suggest that these fennec foxes do show digging behaviour as a response to underground sound and food, which may indicate that digging is an engrained behaviour that is related to searching for food.
Materials and Methods
Subjects and Housing
This study was conducted in the Small Mammal House at ARTIS Zoo in Amsterdam, the Netherlands. The study subjects were two unrelated adult fennec foxes: one male and one female. The fennec foxes were born in captivity and housed in an outside and an inside enclosure that were both accessible for visitors of the zoo. However, the fennec foxes had two cubs together that both reached sexual maternity during the study period and therefore had to be separated from their parents due to the risk of incest. The cubs were placed in the outside enclosure while the adult fennec foxes remained in the inside enclosure for the entire duration of the study period. The two enclosures were separated by a hatch that was located at the back of the inside enclosure and also prevented the adult and juvenile fennec foxes to see each other. Since the outside enclosure was too inaccessible for behavioural observations, the juvenile fennec foxes were excluded from this study.
The inside enclosure spanned an area of 9 m2 of a thin (<20 cm) layer of sand mixed with gravel that prevented intensive burrowing behaviour. The enclosure was furnished with various structures, such as multiple rocks of different sizes, a wood log, several trees, a cactus in a pot, three wooden dens, and two heat lamps hanging from the ceiling (Fig. 1). The standard temperature in the enclosure was around 25 ºC and could reach 30 ºC in the summer. Their diet existed of rodents, birds, and insects and food was offered to them in a bowl placed in a small, dark and out of sight part of their inside enclosure. The enclosure of the fennec foxes was illuminated during the day and dark at night. The caretakers fed the fennec foxes and cleaned their enclosure during day time. The food of the fennec foxes had never been hidden in the substrate and there were no living animals in the substrate. The fennec foxes were not used to sounds coming from the substrate and neither were they used to having to search and hunt for their food.
Figure 1. Photo of the inside enclosure of the fennec foxes. The fennec foxes are housed in the Small
Study 1 – Stimulating Natural Foraging Behaviour
Due to the COVID-19 pandemic in 2020, this part of the study was not conducted and the following paragraph is therefore purely hypothetical.
Auditory Enrichment
To be able to answer the question whether natural foraging behaviour of captive fennec foxes is stimulated by the sound of underground prey, the effect of auditory cues for buried food would have been tested on the exploring and digging behaviour of the fennec foxes. This enrichment study would have consisted of two parts. In the first part, the location of the buried food and the speaker would have been fixed and would be in the same location for the duration of the treatment. In the second part, the location of the buried food and the speaker would have been randomly changed every day.
Study Setup and Data Recording
To record the pattern of exploring and digging behaviour, the enclosure would have been divided into 49 grid cells of 43 x 43 cm. The grid cells would have also been used as possible locations where the food with the speaker could be buried, with exception of the grid cells that would not have been accessible for burying due to the presence of structures such as rocks or trees. Except for the certainty or uncertainty of the location, the setup of both parts would have been the same. To prevent the fennec foxes from experiencing stress if they could no longer find their food in the standard bowl, only 50% of their food for the day would have been buried in the substrate. In addition, a portable speaker playing a recorded sound of a black field cricket would have been buried with the food. The same recording of the cricket sound would have been used in all treatments. To prevent the fennec foxes from using traces of digging in their search for food, the researcher would have dug in every grid in the enclosure, instead of only digging in the place where the food and speaker would be placed. The fennec foxes would not have been able to see the researcher working in the enclosure.
Data recording would have taken place for ten weeks, starting with the first treatment where the location of the food would be certain, followed up by the second treatment where the location of the food would be uncertain. Each treatment would have started with a control measurement that would have taken one week where the behaviour of the fennec foxes would have been observed while nothing would have yet been buried in the substrate. The actual treatments would have also taken a week. To control for differences in time, each treatment would have been duplicated. To examine how long the foraging behaviour would have lasted without buried food and sound, the last treatment would have been followed up with a post-measurement that would have lasted for two weeks. A detailed overview of the phases of the study is displayed in table 1. Behavioural data would have been collected by direct observations and from analysing the recordings made by a camera that would have had a view of the entire inside enclosure. Before the start of the observations, an ethogram was made that included the behaviours important for this study (see table 7 in Appendix 1). It distinguished between active and inactive behaviour, and included the description of digging and exploring behaviour. A clear description of the fennec foxes would have helped to identify and distinguish the two individuals during the observations (see table 8 in Appendix 1). To record the behaviour of the fennec foxes, continuous sampling of both individuals for a duration of one and a half hour would have been used.
Data Analysis
A linear regression analysis would have been used for comparing the frequencies of the natural foraging behaviour, exploring and digging, in the control measurements before and after each treatment, as well as for comparing the frequencies of the behaviours between the first treatment
with a fixed location and the second treatment with a randomly chosen location. This would have been done to determine the effect of introducing underground food and sound to their natural foraging behaviour and overall activity, and to measure the effect of uncertainty of the location of the food on their natural foraging behaviour. The model would have also included a spatial autocorrelation term to control for the association of movement between neighbouring grid cells.
Table 1. Overview of the phases of the study. Prior to each treatment, a control measurement would
have taken place with no food and no sound yet buried in the substrate. The study would have started with the treatment with a fixed location of the buried food and the speaker (treatment 1.1), followed by a duplicate of the same treatment (treatment 1.2). Then the treatment with the randomly chosen location of food and sound would have started (treatment 2.1.), followed by a duplicate of that same treatment (treatment 2.2). The study would have ended with a post-measurement where again nothing would have been buried in the substrate.
Number of weeks (seven days per week) Phase
1 1 1 1 1 1 1 1 2 Control measurements Treatment 1.1 (fixed location) Control measurements
Treatment 1.2 (duplicate of fixed location) Control measurements
Treatment 2.1 (randomly chosen location) Control measurements
Treatment 2.2 (duplicate of randomly chosen location) Post-measurements
Study 2 – Comparing Behaviour During Day Time and Night Time
Behavioural Observations
To be able to examine whether day time and night time affect the natural foraging behaviour of captive fennec foxes, the duration and frequency of exploring and digging behaviour of the fennec foxes were observed during day time and night time for a period of 23 consecutive days.
Study Setup and Data Recording
To record the behaviour of the fennec foxes, a camera without an infrared filter was attached to a Raspberry Pi and provided a live stream that could only be accessed by the researchers. To be able to record at night, two infrared illuminators were placed in the inside enclosure. Fennec foxes are unable to perceive infrared light, making the illuminators unharmful to them. The zoo caretakers were informed about the placement of the camera and were instructed to put on hold all enrichment projects for the fennec foxes for the duration of the study.
Behavioural observations were carried out from June 10th 2020 until July 2nd 2020; at the time, the fennec foxes were ten years of age. Due to the COVID-19 pandemic, the Small Mammal House of ARTIS Zoo was closed for visitors, so the number of visitors per day was not a confounding variable. To record the behaviour of the fennec foxes, continuous sampling of both individuals for a duration of one and a half hour per observation was used. The observations were always around noon for the day observations and around midnight for the night observations. The fennec foxes were always fed in the morning, around noon, and in the afternoon, but with no fixed feeding times. In the course of this study, the fennecs were nearly always fed during a day observation, or else in the hour before a day observation. Before the start of the observations, an ethogram was made that included the behaviours important for this study (see table 7 in Appendix 1). It distinguished between active and inactive behaviour, and
included the description of digging and exploring behaviour. A clear description of the fennec foxes helped to identify and distinguish the two individuals during the observations (see table 8 in Appendix 1).
Data Analysis
All data analysis was performed in R, version R 3.6.1, with the use of the package ‘MASS’ (Venables & Ripley, 2002). To model the influence of day time and night time on the frequency and duration of exploring behaviour of both male and female fennec foxes, a negative binomial model was used. For these response variables, the standard Poisson model was overdispersed with a dispersion parameter of 2.60 for frequency exploring of the male, 1.64 for duration exploring of the male, 4.04 for frequency exploring of the female, and 2.09 for duration exploring of the female. The negative binomial model however was not
overdispersed for the response variables, with a dispersion parameter of 0.96 for frequency exploring of the male, 1.21 for duration exploring of the male, 1.06 for frequency exploring of the female, and 1.35 for duration exploring of the female. Moreover, the negative binomial model had a smaller AIC (frequency exploring male=300.22; duration exploring
male=204.17; frequency exploring female=257.11; duration exploring female=140.97) compared to the standard Poisson model (frequency exploring male=322.83; duration
exploring male=Inf; frequency exploring female=312.29; duration exploring female=Inf). The negative binomial model was tested for all response variables and was found stable as the highest value of the estimated coefficients of all response variables (frequency exploring male=0.062; duration exploring male=0.077; frequency exploring female=0.064; duration exploring female=0.093) was below the threshold (0.143 for male and 0.146 for female). The time that the individuals were active during the observations (log transformed) was included as an offset term in the count part of the model. The reason why the time that the individuals were active was chosen as the offset term instead of the time that the individuals were in sight, was because the individuals slept in the dens at night and were therefore almost always active when in sight at night. Using the time that the individuals were in sight as the offset term would have resulted in a biased comparison of the relative frequency and duration of the behaviours at day time and at night time. To perform an overall test of the effect of day time and night time on exploring behaviour, the full model’s deviance was compared with that of a null model comprising only the intercepts and the offset term using a likelihood ratio test. The sample for the male fennec fox of this model consisted of 42 observations: 21 during day time and 21 during night time. For the female fennec fox, one observation during the night
produced no data since the female was not in sight for the entire duration of the observation, resulting in a total of 41 observations: 21 during day time and 20 during night time. Of the 23 days, 2 were unavailable for observation because of technical problems.
To model the influence of day time and night time on the frequency and duration of digging behaviour of both male and female fennec foxes, a negative binomial model was also used. In contrast to the model for exploring behaviour, the standard Poisson model was only overdispersed for frequency digging of the male with a dispersion parameter of 3.42 and for frequency digging of the female with a dispersion parameter of 1.25. However, the AIC of the negative binomial model was smaller for all response variables with digging (frequency digging male=233.38; duration digging male =93.94; frequency digging female=117.92; duration digging female=60.51) as compared to a standard Poisson model (frequency digging male=267.92; duration digging male =Inf; frequency digging female=117.60; duration
digging female=Inf). Since the negative binomial model also had a small dispersion parameter for each response variable (frequency digging male=1.19; duration digging male =0.46; frequency digging female=0.97; duration digging female=0.68), this model was chosen instead of the normal Poisson model. The negative binomial model was tested for all response
variables and was found stable as the highest value of the estimated coefficients of all response variables (frequency digging male=0.061; duration digging male=0.089; frequency digging female=0.106; duration digging female=0.126) was below the threshold (0.143 for male and 0.146 for female). Again, the time that the individuals were active during the observations (log transformed) was included as an offset term in the count part of the model. To perform an overall test of the effect of day time and night time on digging behaviour, the full model’s deviance was compared with that of a null model comprising only the intercepts and the offset term using a likelihood ratio test. The sample for both fennec foxes of this model consisted of the same number of observations as that of the model of exploring behaviour: 42 for the male and 41 for the female fennec fox.
Results
Descriptive Statistics
This study included two individuals: one male and one female fennec fox. Both individuals were observed for a period of 23 consecutive days: during the day and at night. After checking for missing data, one night observation of the female was removed as the individual was not in sight for the entire duration of the observation. Different variables were collected within this study; that data is summarized in table 2. The relative values of the frequency of both exploring and digging behaviour were acquired by dividing the absolute values by the duration that each individual was active (duration active (min)). The exploring and digging time represent the percentage of the time each individual was exploring or digging during the time frame that they were active (duration active (min)).
Table 2. Summary of the data. This summary includes the data of both individuals. The relative values
of the frequency of both exploring and digging behaviour were acquired by dividing the absolute values by the duration that each individual was active (duration active (min)). The exploring and digging time represent the percentage of the time each individual was exploring or digging during the time frame that they were active (duration active (min)). For ‘time in sight’ and ‘duration active’, the numbers indicate the number of minutes within the 1.5 hour sampling period.
Male
Day (n=21) Night (n=21)
Mean Sd Mean Sd
Time in sight (min) 80.83 7.446 22.98 6.806
Duration active (min) 38.28 14.41 23.10 6.622
Relative frequency exploring per minute 0.436 0.201 1.349 0.317 Relative frequency digging per minute 0.142 0.119 0.346 0.213
Exploring time (%) 11.33 8.458 31.23 9.437
Digging time (%) 1.945 1.674 4.545 2.819
Female
Day (n=21) Night (n=20)
Mean Sd Mean Sd
Time in sight (min) 48.75 18.19 14.72 7.160
Duration active (min) 28.21 15.34 14.65 6.963
Relative frequency exploring per minute 0.457 0.349 0.592 0.248 Relative frequency digging per minute 0.058 0.062 0.049 0.064
Exploring time (%) 8.229 11.19 9.132 5.537
Effect of Day Time and Night Time on the Frequency and Duration of Exploring Behaviour
In general, there was a clear significant impact of day time and night time on both the frequency (full null model comparison: LR=48,65, P(Chi)<0.001) and duration (full null model comparison: LR=28.67, P(Chi)<0.001) of the exploring behaviour of the male fennec fox. In particular, it was at night time that both the frequency (P<0.001) and the duration(P<0.001) of the exploring behaviour of the male increased (Table 3).
For the female fennec fox, day time and night time had no significant impact on both the frequency (full null model comparison: LR=1.86, P(Chi)=1.173 N.S.) and duration (full null model comparison: LR=0.308, P(Chi)=0.579 N.S.) of its exploring behaviour. This result was confirmed by the negative binomial model where no significant differences between day time and night time for both frequency (P=0.167 N.S.) and duration (P=0.573 N.S) of the exploring behaviour of the female were found (Table 4).
Figure 2 shows the clear significant differences between the exploring behaviour of the male fennec fox at day time and night time and the lack of significant differences between the exploring behaviour of the female fennec fox at day time and night time.
Table 3. The effect of day time and night time on the exploring behaviour of the male fennec fox
Frequency of exploring behaviour
Predictor Estimate SE Z value P value
(Intercept) -0.855 0.084
Time of the day (Night time) 1.154 0.112 10.31 <0.001
Duration of exploring behaviour
Predictor Estimate SE Z value P value
(Intercept) -2.222 0.123
Time of the day (Night time) 1.038 0.159 6.522 <0.001
Table 4. The effect of day time and night time on the exploring behaviour of the female fennec fox
Frequency of exploring behaviour
Predictor Estimate SE Z value P value
(Intercept) -0.784 0.128
Time of the day (Night time) 0.260 0.188 1.382 0.167
Duration of exploring behaviour
Predictor Estimate SE Z value P value
(Intercept) -2.542 0.190
Figure 2. Frequency and duration of exploring behaviour of both male and female fennec foxes at day time and night time. The relative values of the frequency of exploring behaviour were acquired by dividing
the absolute values by the duration that each individual was active. The exploring time represents the percentage of the time each individual was exploring during the time frame that they were active (duration active (min)). Bars: median frequency or percentage; top and bottom of boxes: 75 and 25 percentiles, respectively; whiskers: highest and lowest values (no outliers); circles: outliers; significance codes: ‘***’ p<0.001, ‘**’ p<0.01.
***
Effect of Day Time and Night Time on the Frequency and Duration of Digging Behaviour
In general, day time and night time clearly influenced both the frequency (full null model comparison: LR=16.33, P(Chi)<0.001) and the duration (full null model comparison: LR=9.43,
P(Chi)<0.01) of the digging behaviour of the male fennec fox. In fact, during the night both the
frequency (P<0.001) and the duration (P<0.01) of the digging behaviour of the male increased (Table 5).
For the female fennec fox, day time and night time had again no significant impact on both the frequency (full null model comparison: LR=1.084, P(Chi)=0.298 N.S.) and the duration (full null model comparison: LR=0.474, P(Chi)=0.491 N.S.) of its digging behaviour. These findings were also confirmed by the negative binomial model where no significant differences between day time and night time for both frequency (P=0.293 N.S.) and duration (P=0.486 N.S) of the digging behaviour of the female were found (Table 6).
Figure 3 shows the clear significant differences between the digging behaviour of the male fennec fox at day time and night time and the lack of significant differences between the digging behaviour of the female fennec fox at day time and night time.
Table 5. The effect of day time and night time on the digging behaviour of the male fennec fox
Frequency of digging behaviour
Predictor Estimate SE Z value P value
(Intercept) -2.013 0.154
Time of the day (Night time) 0.955 0.209 4.573 <0.001
Duration of digging behaviour
Predictor Estimate SE Z value P value
(Intercept) -4.074 0.271
Time of the day (Night time) 1.026 0.342 3.003 <0.01
Table 6. The effect of day time and night time on the digging behaviour of the female fennec fox
Frequency of digging behaviour
Predictor Estimate SE Z value P value
(Intercept) -2.725 0.200
Time of the day (Night time) -0.380 0.361 -1.052 0.293
Duration of digging behaviour
Predictor Estimate SE Z value P value
(Intercept) -4.142 0.326
Figure 3. Frequency and duration of digging behaviour of both male and female fennec foxes at day time and night time. The relative values of the frequency of digging behaviour were acquired by dividing
the absolute values by the duration that each individual was active. The digging time represents the percentage of the time each individual was digging during the time frame that they were active (duration active (min)). Bars: median frequency or percentage; top and bottom of boxes: 75 and 25 percentiles, respectively; whiskers: highest and lowest values (no outliers); circles: outliers; significance codes: ‘***’ p<0.001, ‘**’ p<0.01.
***
Discussion
Influence of Day Time and Night Time on Natural Foraging Behaviour
The current study examined the influence of day time and night time on the natural foraging behaviour of captive fennec foxes. The results show a strong difference between the behaviour of the male and female that were observed in this study. This study supports the hypothesis that natural foraging behaviour substantially increases at night for the captive male fennec fox. The male fennec fox showed significantly more exploring and digging behaviour during night time than during day time. However, day time and night time apparently have no impact on the natural foraging behaviour of the captive female fennec fox. Unfortunately, sex could not be implemented as a predictor variable in the model since only two individuals were available for this study and this is insufficient for running a generalized linear mixed model. Therefore, it is difficult to interpret the results in general, but it is still possible to speculate with some caution about explanations for the substantial difference between the behaviour of the male and the female fennec fox. For instance, it has been observed that male fennec foxes in the wild provide the food and therefore hunt more often at night than the females (Larivière, 2002). While in the wild fennec foxes hunt at night due to the extreme temperatures in their environments during the day, their body temperature and heart rate still increase at night when in captivity, even if the temperatures during the day are not that high in the zoo (Maloiy et al., 1982). This could explain why the male fennec fox showed an increase in its foraging behaviour at night.
In contrast, it has been observed that female fennec foxes cache more frequently in the wild than male fennec foxes (Gauthier-Pilters, 1967). Since the fennec foxes in the zoo received their food during the day, it may be that the female cached some of that food and did not need to search for new food at night. Another explanation for the lack of difference between the foraging behaviour of the female during day time and night time, could be the huge differences in temperature and harshness of the sun between the natural environment of the fennec foxes and the enclosure in the zoo. The thermoneutral zone of fennec foxes ranges from 23 to 37 ºC. As temperatures in their natural environment can exceed 40 ºC, fennec foxes avoid the harshness of the desert climate by withdrawing during the day to underground burrows (Maloiy
et al., 1982). The temperature in the enclosure of the captive fennec foxes was around 25 ºC
and therefore lies in the thermoneutral zone. This means that there is actually no need for the fennec foxes to withdraw during the day and hunt at night, which could explain the lack of difference in foraging behaviour of the female between day time and night time. However, this hypothesis would not explain the results of the male fennec fox and the question why the foraging behaviour of the male increases at night would still remain unanswered.
To be able to study whether the effect of day time and night time on the natural foraging behaviour of captive fennec foxes significantly differs between sexes, future studies need to include a larger sample size with more fennec foxes of both sexes.
Animal Welfare Implications
Little is known about the effect of diurnal rhythm on nocturnal animals in captivity and its implications for their welfare. In this study the fennec foxes showed digging behaviour at night, despite the fact that no food was provided at night and, in particular, no food was buried in the substrate. Since it is not possible for the fennec foxes to make a burrow as the substrate layer is too thin, the digging is clearly an indication of their foraging behaviour. This finding suggests that the fennec foxes express a need for digging. It may be a good idea to facilitate this behaviour as it can be beneficial for their wellbeing. The easiest way to do this can be to put a thicker layer of substrate in their enclosure for the fennec foxes to dig in. Moreover, the urge of the fennec foxes to search for food at night can be facilitated by providing automatic feeders.
This is all the more important for the zoo because previous studies have shown that stereotypic behaviour can arise when zoo enclosures lack certain stimuli that facilitate the natural behaviour of animals (Mason, 1991). Introducing animals in captivity to both structural and feeding enrichment throughout their life is crucial since it has been proven that housing conditions affect brain development and behaviour (Würbel, 2001). It is also important for zoo personnel to always keep track of the animals’ responses to the introduced enrichment projects and to monitor the general behaviour of the animals when the enrichment lasts for a longer period of time. This could help determine the most promising type of enrichment that continues to stimulate the behaviour of the captive animals for a longer period of time (Shyne, 2006). While environmental enrichment is often seen as the most adequate solution to improve the welfare of animals in captivity (Mason et al., 2007), feeding enrichment may be the most suitable method to stimulate natural foraging behaviour (Kistler et al., 2009).
Acknowledgements
I thank Karline Janmaat for the supervision of this project, and the staff of ARTIS ZOO, in particular Monique Versloot, Dennis de Haan, Eddy van Everdingen, Kevin Varekamp, Ans du Bois, Nathalie Casius, Danielle de Vre, Iris de Kruijf, and Stephan and Roy of the technical services for their efforts and great help during this study. I am grateful to Peter Roessingh for helping me with the setup of the observational equipment.
References
Carlstead, K. J. (1991). Husbandry of the Fennec fox: Fennecus zerda: environmental conditions influencing stereotypic behaviour. International Zoo Yearbook.
Dempsey, J. L., Hanna, S. J., Asa, C. S., & Bauman, K. L. (2009). Nutrition and behavior of fennec foxes (Vulpes zerda). Veterinary Clinics of North America: Exotic Animal
Practice, 12(2), 299-312.
Gauthier-Pilters, H. (1967). The fennec. African Wild Life, 21, 117–125.
Karssene, Y., Chammem, M., Li, F., Eddine, A., Hermann, A., & Nouira, S. (2019). Spatial and temporal variability in the distribution, daily activity and diet of fennec fox (Vulpes zerda), red fox (Vulpes vulpes) and African golden wolf (Caniss anthus) in southern Tunisia. Mammalian Biology, 95(1), 41-50.
Kingdon, J. (1991). Arabian mammals: a natural history. Academic Press.
Kistler, C., Hegglin, D., Würbel, H., & König, B. (2009). Feeding enrichment in an opportunistic carnivore: The red fox. Applied Animal Behaviour Science, 116(2-4), 260-265.
Larivière, S. (2002). Vulpes zerda. Mammalian Species, 714, 1-5.
Maloiy, G. M. O., Kamau, J. M. Z., Shkolnik, A., Meir, M., & Arieli, R. (1982). Thermoregulation and metabolism in a small desert carnivore: the Fennec fox (Fennecus zerda)(Mammalia). Journal of Zoology, 198(3), 279-291.
Mason, G., Clubb, R., Latham, N., & Vickery, S. (2007). Why and how should we use environmental enrichment to tackle stereotypic behaviour?. Applied Animal Behaviour
Science, 102(3-4), 163-188.
Mason, G. J. (1991). Stereotypies and suffering. Behavioural Processes, 25(2-3), 103-115. Shettleworth, S. J. (2009). Cognition, evolution, and behavior. Oxford university press. Shyne, A. (2006). Meta‐analytic review of the effects of enrichment on stereotypic behavior
in zoo mammals. Zoo Biology: Published in affiliation with the American Zoo and
Aquarium Association, 25(4), 317-337.
Sillero-Zubiri, C., Hoffmann, M., & Macdonald, D. W. (Eds.). (2004). Canids: foxes, wolves,
Ursinus, R. (2019). Pilot study on locating underground prey in Fennecus zerda. Amsterdam, The Netherlands.
Venables, W. N., Ripley, B. D. (2002). Modern Applied Statistics with S (4th ed). New York: Springer.
Watters, J. V., Miller, J. T., & Sullivan, T. J. (2011). Note on optimizing environmental enrichment: A study of fennec fox and zoo guests. Zoo biology, 30(6), 647-654.
Würbel, H. (2001). Ideal homes? Housing effects on rodent brain and behaviour. Trends in
Appendix 1 – Ethogram and Description of the Fennec Foxes
Table 7. Ethogram of the recorded behaviours
Table 8. Description of the fennec foxes observed in this study
BEHAVIOUR DESCRIPTION
DIGGING Moving the sand with its nose or one/two feet while staring intensively at the sand EXPLORING Moving around while staring at the sand and keeping its nose close to the ground,
nostrils may be flaring
INACTIVE Stationary position with its body lying on the ground, possibly moving head around ACTIVE Standing upright, moving around or any other behaviour where the animal is not in
a stationary position with its body lying on the ground
NOT AVAILABLE Animal is out of observers’ sight, in a den or in the separated part of the inside enclosure
INDIVIDUAL DESCRIPTION PHOTO
MALE Normal-sized body for an adult fennec fox, darker coat than the female fennec fox, dark lines running from eyes to snout on both sides of the snout, dark stripe on its back running from neck to tail
FEMALE Normal-sized body for an adult fennec fox, lighter coat and smaller head than the male fennec fox, dark spot on both sides above back legs
