Activity of tjakura (great desert skinks) at burrows in relation to plant cover and predators

Abstract Increased predation where ground cover is reduced after severe wildfire is increasingly implicated as a factor causing decline of vulnerable prey populations. In arid central Australia, one species detrimentally affected by repeated wildfire is the great desert skink or tjakura (Liopholis kintorei), a distinctive lizard of the central Australian arid zone that constructs and inhabits multi‐entranced communal burrows. We aimed to test whether tjakura or predator activity at burrow entrances varied with cover and how tjakura respond to predator presence. Using time‐lapse photography, we monitored tjakura and predator activity at the largest entrance of 12 burrows ranging from high (>70%) to low (<50%) cover and at multiple entrances of two other burrows. Overall activity did not vary between burrows with high and low cover. Within burrow systems tjakura were more active at sparsely vegetated entrances, often sitting wholly or partly inside the burrow. However, consistent between and within burrow systems, skinks spent proportionally more time fully outside where cover was higher. Predators—mostly native—were detected at most burrows, with no apparent relationship between predator activity and cover. Skinks also did not appear to modify their activity in response to predator visits. Our results indicate that tjakura may spend more time outside burrow entrances when cover is higher but there was no direct evidence that this related to perceived or real predation risk. Differences in food availability, thermoregulatory opportunities and opportunities for ambush foraging associated with differences in vegetation cover or composition are other factors likely to be important in determining the activity of tjakura at burrows. Our research demonstrates the usefulness of camera traps for behavioural studies of ectothermic burrowing animals. The complex relationships between tjakura activity and vegetation cover thereby revealed, suggest outcomes of fire‐mediated habitat change on predator–prey interactions are not easily predictable.


| INTRODUC TI ON
Fire is a regular disturbance in most vegetation communities of the arid regions of Australia and, if severe, can dramatically reduce cover and diversity of the standing vegetation (Knuckey et al., 2016).
These changes often persist for many years and potentially increase vulnerability of many smaller animals to predation if these species rely on cover for shelter (Torre & Díaz, 2004). There has been considerable interest in untangling interactions between fire and predation risk (Doherty et al., 2022), particularly in the fire-prone ecosystems of the arid and wet-dry tropical regions of central and northern Australia, where altered fire regimes and introduced predators rate among the most serious threats for native wildlife. There is increasing evidence that the key introduced predators in these regions, cats (Felis catus) and the European fox (Vulpes vulpes), hunt more effectively in recently burnt areas either because they can navigate more easily or because the refuges of prey species are more visible (McGregor et al., 2014(McGregor et al., , 2016Nimmo et al., 2018).
Our research focussed on a threatened species that inhabits fireprone vegetation communities in arid Australia for which such interactions between predation and fire are thought to be important.
Tjakura (great desert skink; Liopholis kintorei) construct and maintain extensive, multi-tunnelled burrows that become home to adults and their young for many years and breeding seasons (McAlpin, 2001; Figure 1). They exhibit social and cooperative behaviours, and most of their activities are centred around the burrows (McAlpin et al., 2011). Burrows have multiple entrances, some of which have distinctive mounds of excavated soil adjacent to them. There is usually one entrance and mound bigger than the others, with a larger mound which appears to be the most frequently used, however, there has been little documented about the activity of tjakura in and around the burrow systems. Tjakura are omnivorous, consuming plant material and invertebrates, with termites (Order: Blattodea, Superfamily: Termitoidea) thought to be the most important prey item (Chapple, 2003;McAlpin et al., 2011). They are thought to exhibit predominantly ambush strategies to catch prey, sitting partially within or near their burrow entrance and opportunistically taking invertebrates as they pass by, and only infrequently moving away from the burrow entrance to forage actively at night (Chapple, 2003; Department of Climate Change, Energy, the Environment and Water, 2023). Tjakura have also been observed thermoregulating using variable postures and basking while half in the burrow entrance (Chapple, 2003). Wildfire and introduced predators are recognized as key threats for the species (Moore et al., 2015;. One of the known strong-hold populations of the great desert skink is in the Uluru-Kata Tjuta National Park in central Australia (McAlpin, 2001) where they are known as tjakura which is their Pitjantjatjara and Yankunytjatjara name (Indigenous Desert Alliance, 2022) and the name we use here. Pitjantjatjara is the most common language spoken by Anangu, the local Aboriginal custodians where our research took place. The species is also known by other names in other central Australian Aboriginal languages and regions including warrarna (Warlpiri) and mulyamiji (Muanyjilijarra, spoken by Martu people), tjalapa (Pintupi) and nampu (Aran, spoken by Anmatjere) (Indigenous Desert Alliance, 2022). Genetic research by Dennison et al. (2015) demonstrated that skink haplotypes at Uluru were unique to the area and that overall there is very little connectivity between the six known stronghold localities for the species (Sangsters Bore, Newhaven, Watarru, Docker River and Warburton).
Of the skinks sampled, those from Uluru were the most genetically distinctive, leading to the recommendation that the population should be considered its own delineation for the purpose of conservation management (Dennison et al., 2015). At the Uluru site, tjakura burrows tend to occur in areas with slightly higher vegetation cover compared to the surrounding landscape and often inhabit areas with sparsely scattered shrubs (e.g. Grevillea eriostachya) (Ridley et al., 2018) but it is not known whether this relates to shelter, increased abundance of invertebrate prey (Morton & James, 1988) or other structural habitat features. The risk of predation for tjakura is probably highest when they are outside their burrows, with burrows providing a refuge. However, burrow sites may also be a means by which predators locate or regularly relocate skink colonies (Moore et al., 2015). As a likely focal point for predator-skink interactions burrow entrances provide an excellent opportunity to study the behaviour of tjakura in relation to vegetation cover and predator activity.
Based on previous research suggesting increased cover is likely to favour tjakura through limiting the indirect effect of predation pressure (Moore et al., 2015), we hypothesized that; 1. tjakura would be more active at burrows that had higher vegetation cover, 2. predator activity would be less at burrows with greater vegetation cover due to decreased visibility of the tjakura and their burrow systems, F I G U R E 1 Tjakura (great desert skink, Liopholis kintorei) outside a burrow entrance.
3. tjakura activity at the burrow entrance would decrease after a predator visit, because either the predation attempt was successful or, in the short term, skinks modified their behaviour in response to a greater perceived threat of predation.
We also expected that there may be an interaction between diel activity patterns and vegetation cover, especially in relation to activity in daylight compared to dark. To gain a better understanding of how tjakura use different entrances within a burrow system, we also monitored all the entrances of two burrows in relation to specific attributes of the entrance, including vegetation cover.

| MATERIAL S AND ME THODS
This study was conducted within Uluru-Kata Tjuta National Park, Northern Territory, in sand plain habitat (25.3009° S, 130.7184° E), over one active season. The tjakura population is restricted to the sand plain areas of the park which are dominated by spinifex (Triodia spp.) and tussock grasses with a very sparse over-storey of honey grevillea (Grevillea eriostachya) and desert oaks (Allocasuarina decaisneana) and contains a mosaic of vegetation with different fire histories.

| Burrow selection
Twelve active tjakura burrows were chosen for monitoring out of a larger sample of 20 burrows that were being actively used by skinks at the time of our study. The 20 burrows were a subset of burrows that had previously been mapped and regularly monitored by the National Park rangers (Ridley et al., 2018). To select burrows for monitoring, all known burrows were checked for signs of recent activity and habitation, indicated by the presence of fresh scats and tracks. For the 20 burrows identified as active the percentage projected horizontal canopy cover of all plant species within a 2.5 m radius circle centred at the largest burrow entrance was visually estimated.
We then selected 12 of these burrows to represent two distinct cover categories (low/high) for monitoring. The six burrows with the lowest vegetation cover (all <50%) around the main entrance were classed as low cover. The six burrows with the highest vegetation cover (all >70%) were classed as high cover. The three burrows with the lowest cover (26%-36%) were in areas that had been burnt 3 years previously in a 2012 wildfire, and the other three low cover burrows (41%-49% cover) had not been burnt for 12 years.
High cover burrows were all in areas unburnt for at least 12 years.
Burrows varied in size and number of entrances (Table 1). Burrow size was estimated by measuring the distance between entrances furthest away from each other in two perpendicular directions and multiplying to derive an estimated area.

| Monitoring tjakura activity
Tjakura activity was monitored using 10 Reconyx HC600 and 2 Reconyx HC800 cameras. One camera was set up at each burrow at what appeared to be the main entrance. Every burrow had multiple entrances and the main entrance was determined as the biggest one associated with the biggest soil mound (a distinctive feature at burrow entrances created from excavated soil), where this was obvious.
Another distinctive characteristic of tjakura burrows is the presence of a 'latrine' or scat pile which is consistently used as a defecation site by burrow inhabitants. If there were two entrances and mounds of similar size within a burrow system, the entrance closest to the latrine or most centrally located among the other entrances was chosen.
Cameras were installed 1 m from the burrow entrance, mounted on platforms angled downwards at 60° to the ground and screwed onto metal garden stakes, 60 cm above the ground. This setup produced a 1 m 2 field of view centred on the burrow entrance. A small plywood shelter, attached directly above each camera, shaded it from the extreme heat typical of the central Australian summer.
Cameras were set to the time-lapse mode to take one photo every minute for two 18-day periods in early summer (from 12:00 midday 27 November to 12:00 midday 15 December 2014), and mid-late summer (from 12:00 midday 24 January to 12:00 midday 11 February 2015). Batteries and SD cards in cameras were replaced every 5-6 days, but battery failure led to the loss of 8 h of data from one burrow and 12 h each from two other burrows.
We used the number of images with tjakura visible as an index of their activity at the burrow entrance. Although we refer to these periods as 'active', most of the time when lizards were partially or fully emerged, they remained quite stationary (i.e. were in the same place in multiple consecutive images). When there were continuous TA B L E 1 Characteristics of the 12 burrows chosen for monitoring, ordered from lowest to highest vegetation cover and categorized as having low cover (L1-L6) or high cover (H1-H6). periods (i.e. consecutive images) with tjakura present, 'active bouts' was used to describe this activity. When there were no tjakura visible on an image, we assumed the residents had either retreated inside the burrow or were actively foraging (or engaging in another activity) away from the burrow entrance.

| Data analysis
The images were initially sorted into three categories: one or more tjakura visible, other vertebrate species visible or no vertebrate visible. The images with skinks present were then divided into three position categories: fully inside the burrow entrance, half emerged with front legs visible or fully emerged from the burrow entrance, with hind legs visible outside ( Figure 2). There were three clearly distinct size classes of tjakura in the images including very small, medium and large or adult size, so we were able to assign individuals into three approximate age classes; juvenile (current year hatchlings), sub-adult or adult. Where there were two or more tjakura in the image, each was classified separately for size and position.
We were not always able to distinguish among different individuals from the same burrow. However, obvious differences in size, colour, the presence of distinguishing features (e.g. part of the tail missing) or the presence of multiple individuals in single images allowed us to determine a minimum number of individuals present at each burrow during each recording period.

| Activity at burrows in relation to burrow characteristics and vegetation
Repeated measures analysis of variance was used to determine whether the minimum number of individuals varied between the two recording periods (within subjects effect) or between burrows with high and low vegetation cover. We then used Pearson correlations to test how tjakura activity at the main burrow entrance related to the minimum number of lizards occupying the burrow, the number of entrances in the burrow system and vegetation cover at the burrow.

| Response by tjakura to predator activity at burrows
All vertebrates known or suspected to be predators of tjakura were included in analyses. The number of discrete 'visits' by predators to burrows was determined based on the sequence of images. If the same predator was visible in consecutive images, this was counted as a single visit. We then tested for correlations between vegetation cover at burrows and the number of predator visits.
To determine whether tjakura reduced their activity in response to predator visits, graphs were produced to show the timing of their active bouts in relation to predator visits. The activity of skinks before and after predator visits was compared based on (a) the time interval (min) between a tjakura being visible and the arrival of a predator and the interval (min) between a predator leaving and a tjakura becoming visible again, and (b) the number of active bouts by tjakura in the 24 h prior to predator arrival and post-predator departure. The null hypothesis (no effect of predators) was that there would be no difference in the 24 h before and after a predator visit in the time interval between tjakura activity and predator activity or the number of tjakura active bouts. Time data were log transformed prior to analyses.

| Daily activity patterns of tjakura
To investigate patterns in daily activity we plotted the time tjakura spent outside the burrow against the average activity of tjakura in the early and late active season. The early active season was considered as the time between tjakura first emerging during spring until F I G U R E 2 Positions of tjakura at the main burrow entrance, coded as, from left to right; in the burrow, half in the burrow, and outside the burrow (counted as outside when front and back legs were visible outside of the entrance).

| Tjakura activity at multiple entrances within a burrow system
Two additional burrows were selected from the remaining eight of the 20 active burrows; one (burrow A) with four entrances and one (burrow B) with seven entrances. Reconyx cameras were set up at every burrow entrance in the same way as described previously.
Width, height, opening, diameter and height (cm) of the mound, as well as percentage cover of vegetation in a 1 m 2 quadrat centred over each entrance were measured and the main entrance was identified, according to previous protocols.
Activity at burrow entrances was recorded at all entrances simultaneously for 4 days in early February 2015. Cameras were set to time lapse taking one picture per minute. One camera (at burrow A, Entrance 1) malfunctioned, and that entrance was excluded from all analyses. We used correlations to test whether the level of activity at burrow entrances was associated with burrow size, minimum number of individuals using the system, overall activity and vegetation cover at each entrance.  Table 2).

| Activity at burrows
There was no effect of vegetation density ( There were no significant correlations between vegetation cover and the total number of tjakura images or images of tjakura inside or outside the burrow, however, the proportion of time tjakura spent outside of the burrow was positively correlated with higher vegetation cover at the burrow site (r = .637, p = .026).

| Response by tjakura to predator activity at burrows
Four predator species and several unidentified species were detected at the burrow entrances. The most common predator species observed was the sand goanna (tinka in Pitjantjatjara; Varanus gouldii; 30 visits across all burrows), followed by the mulgara (murtja in Pitjantjatjara; Dasycercus blythi; eight visits across all burrows) and woma python (kuniya in Pitjantjatjara; Aspidites ramsayi; eight visits across all burrows) (Figure 3). The European fox (Vulpes vulpes) was only detected twice and was the only introduced predator observed.
Predators were recorded at every burrow expect for H3 and H5, two of the high cover burrows. Despite this, overall there was no significant correlation between the number of predator visits at low and high vegetation cover burrows (r = −.322, p = .307). The same predators were detected at low and high cover burrows, except the fox which was only detected at low cover burrows.
Activity of tjakura was highly variable between the 12 burrow systems and there was also much variation between the number and species of predators visiting each of the burrows (Figure 4a,b). There was no obvious decline in tjakura activity following predator visits TA B L E 2 The number of tjakura (mean, standard error and range) detected at burrows with low and high vegetation cover each summer. Analyses were conducted on all predators grouped together because most predatory species were detected at burrows too infrequently for independent analysis of their impact on tjakura activity. Sand goannas were the exception and were therefore analysed separately. Comparing activity before and after a visit, we did not detect any significant impact of predator visits (all predators or goannas) on the length or the number of active bouts (Table 3). There also were no significant interactions between time and vegetation, suggesting the response (or lack of response) to predators did not differ between burrows with high or low vegetation cover. Tjakura activity differed significantly among burrows but this did not appear to relate to predator activity or vegetation cover (Table 3).

| Daily activity patterns of tjakura
Diel activity of tjakura was characterized by low activity during the middle of the day, and peak activity in the dawn and dusk crepuscular periods. This pattern was consistent between low and high vegetation cover burrows ( Figure 5). On average, using total images (i.e. tjakura visible either inside or outside the burrow) as an index of activity, tjakura in burrow systems with low vegetation cover were more active, with this activity more obvious during the late active season. Average activity at burrows with high cover did not increase noticeably between the early and late active periods whereas activity at burrows with low cover was notably higher in the late active season ( Figure 5).
Tjakura spent the most time outside burrows 2-3 h after dawn as well as just before and in the 3 h after dusk ( Figure 6). Although also frequently observed at the burrow entrance in the early morning before dawn ( Figure 6) tjakura tended to be inside the entrance rather than active at the surface at this time. Time tjakura spent outside the burrow was 20% greater on average at burrows with high cover for most times of the day, however, variation between burrows was high and this was not a significant effect in any analyses.
Time of day significantly affected total images per hour (RMANOVA: df = 3,3, F = 8.05, p < .001) and outside images per hour (RMANOVA: df = 3,3, F = 12.23, p < .001) and the post-hoc tests showed total images and outside images differed significantly between day (when tjakura were least active) and dusk (when tjakura were most active) (total images: t = 7.449, p < .05, outside images: t = 7.020, p < .05). Tjakura were moderately active at dawn and night, but activity (total and outside images) was highly variable at these times and did not differ significantly from dusk or day (Figure 7).

| Tjakura activity at different entrances within burrow systems
There was a large amount of variation in vegetation cover and activity at different entrances within a burrow ( Table 4). All entrances were used by multiple individuals and, somewhat unexpectedly, the main entrance of burrow B, identified prior to monitoring, had the lowest activity (total images) compared to other entrances at

F I G U R E 3 Predatory species recorded at tjakura burrows included sand goannas and woma pythons.
F I G U R E 4 (a) Daily activity (# active bouts) of tjakura at the six burrows with low cover (percent cover indicated on each graph) from two 18-day periods of camera monitoring. Predator visits over this time are displayed by coloured bars. Goannas = blue, woma pythons = yellow, mulgara = green, fox = orange and unidentified snake = purple. (b) Daily activity (# active bouts) of tjakura at the six burrows with high cover (percent cover indicated on each graph) from two 18-day periods of camera monitoring. Predator visits over this time are displayed by coloured bars. Goannas = blue, woma pythons = yellow and mulgara = green.

F I G U R E 4 (Continued)
that burrow. Also, at least two juveniles and two sub-adults were detected at burrow B, but only one individual of each subclass was detected at the main entrance. At burrow B there was no significant correlation between entrance diameter or mound height and tjakura activity.
Overall, activity at the burrow entrance (total images) was strongly negatively correlated with vegetation cover (r = −.932, p < .001), with most activity detected at entrances with less cover ( Figure 8a). However, when tjakura were active (either inside or outside the entrance), the proportion of time spent outside was positively correlated with vegetation cover (r = .845, p = .002; Figure 8b).
Consistent with this trend, there was also a negative relationship between total images inside the burrow and vegetation cover (r = −.928, p < .001; Figure 8c). These results show that tjakura were more active, in or near entrances, where vegetation cover was low but were more likely to be wholly outside the entrance when they were active at high cover entrances.  . Cats may be able to locate skink burrows more easily in open areas where they are more visible and, once a colony has been located, the clustered distribution of burrows within a colony increases their vulnerability to detection and repeat visits (Moore et al., 2015). Foxes and cats have been observed sitting at burrows, waiting for the skinks to emerge (Chapple, 2003); however, there was no evidence of this behaviour during our study. Surprisingly, we detected no feral cats at the burrows we surveyed. We also recorded only two single images of foxes (each at a different time and burrow location) which suggests the foxes spent less than a minute at the monitored burrow entrance on these occasions, although they may have moved to other entrances or remained in the vicinity of the burrow, but out of the field of view of the camera.

| DISCUSS ION
It was unclear whether the three native predator species detected at burrows were actively seeking to prey on the tjakura or whether they were successful because time-lapse photography did not provide a continuous record of activity. In contrast to foxes, sand goannas and woma pythons were occasionally present for more than 10 min with one sand goanna visible at a burrow for 60 consecutive minutes. On some occasions goannas and pythons were recorded entering or exiting the burrow, strongly suggesting encounters that may have led to predation. We also captured multiple consecutive images of sand goannas sitting at the burrow entrance with front legs splayed and ventral surface of body in contact with the ground, exhibiting what appeared to be thermoregulatory behaviour, as well as sitting just inside the burrow (see Figure 3). This suggests they may have been using the burrow as shelter (Indigenous Desert

TA B L E 3
Results from three-factor repeated measures ANOVA testing for differences between tjakura activity (measured as log of minutes active and the number of active bouts) 24 h before and after a predatory visit (all predators and goannas) at burrow entrances with low and high vegetation cover; <.05*, <.01**, <.001***, NS, no significance. F I G U R E 7 Tjakura activity (mean images ±SE per hour) during dawn, day, dusk and night; (a) is total images with tjakura visible at the entrance and (b) images where the tjakura is outside the burrow entrance. . Mulgara, medium-sized dasyurid predators, have also been known to use tjakura burrows for shelter (McAlpin, 2001). While capable of preying on tjakura, there is limited evidence to suggest that mulgara are major predators of the species.

F I G U R E 5
The interaction between predation and fire-where skinks in open habitat are more exposed to the threat of predation-is thought to pose the greatest threat to tjakura populations (McAlpin, 2001;McGregor et al., 2014;Wilson, 2014). However, our hypothesis that predator activity would be greater at burrows with lower cover, was not supported. We also found no conclusive evidence that tjakura altered their level of activity at the burrow entrance following a visit by a predator, although low activity at two burrows after predators were detected suggests that successful predation may have occurred in these instances. The relatively low number of predator visits to the monitored burrows overall limited our ability to test for these effects.
Only limited evidence from our study supports the hypothesis that tjakura are disadvantaged by low cover. In fact, tjakura were observed at the surface, albeit partially or fully within the burrow, more frequently at burrows with low cover. However, the proportion of time tjakura spent outside at burrows with low cover was significantly less than at high cover burrows, both between and within burrows. Pygmy blue-tongue lizards (Tiliqua adelaidensis), exhibit somewhat comparable behaviour, spending less time foraging and basking in burnt areas (Fenner & Bull, 2007). The tendency to spend less time exposed in open areas may be linked to an increased risk of predation and aligns with suggestions that tjakura burrows in wildfire affected areas are frequently abandoned (Wilson, 2014).

F I G U R E 8
Correlations between vegetation cover at burrow entrances and (a) total tjakura images, (b) total number images with tjakura inside the burrow, (c) total number of images with tjakura outside the burrow and (d) proportion of time tjakura are outside (ns).

TA B L E 4
Minimum number of individuals, age classes (j = juvenile, sa = sub-adult and a = adult) and total images at each entrance of burrow A (A2-A4) and burrow B (B1-B7) (Entrance A1 camera malfunctioned). Open areas may, however, provide better opportunities for skinks to ambush prey from within the burrow entrance, where they are concealed from both prey and predators, and potentially better opportunities for basking while partially or fully within the burrow, as discussed further below.

| General natural history and behavioural observations
Prior to our study, annual surveys of tjakura burrows at Uluru-Kata Tjuta National Park were regularly completed during late summer (February/March), when tjakura appear to be most active and burrows are most visible due to accumulated activity over the active season. Activity commences in spring, around September or October, and in the Newhaven population mating activity has been recorded in September with females giving birth approximately 10 or 11 weeks later (Moore et al., 2015). This is broadly consistent with our observations of juveniles from the beginning of the late November survey. We recorded increased numbers of individuals and higher activity at the burrow entrances in the late active season whereas the minimum number of identifiable juvenile lizards only increased at two burrows, so our results seem to reflect mainly a seasonal increase in activity, previously documented, rather than new recruitment. Because juveniles are much smaller and mostly active during the night, they tend to also be less visible which may also make it more difficult to confirm their presence. During the late active season many more images had multiple individuals present together, often of similar size, which enabled us to positively confirm that there were at least two or three sub-adults present at individual burrows.
Tjakura have previously been described as predominantly nocturnal (Chapple, 2003), however our data, consistent with other more recent work , shows primarily crepuscular activity. Activity was consistently lowest during the middle of the day, and highest around dusk and this did not vary with respect to vegetation cover. We observed some interesting behaviour, such as tjakura sitting in the burrow entrance, sometimes for hours at a time, often in the pre-dawn period. This may be a thermoregulatory strategy, documented in some other burrowing desert lizards, where individuals come up close to the surface in the early morning and only exit their burrows once they have reached an optimal temperature (Heatwole & Taylor, 1987), although surface temperatures are usually lowest during the hours just before dawn.
Alternatively, skinks sitting at burrow entrances may have been foraging from the burrow entrance. Tjakura use both active and ambush foraging and may modify their strategy depending on cover, potentially using ambush foraging more where vegetation is sparse. The congeneric Slater's skink (Liopholis slateri) appears to rely on direct line of sight to forage effectively from burrow entrances (McKinney et al., 2014). For tjakura the abundance and type of prey available may also determine the foraging strategy used. Wildfire can temporarily reduce invertebrate abundance (Chapple, 2003), however, we did not measure prey abundance in our study or study very recently burnt burrows. Potential interactions between the foraging behaviour of tjakura and the abundance and composition of their prey following fire would be a useful further area of investigation. entrance of their burrows and use other entrances, which-ever is in closest proximity, when returning from feeding or if the preferred entrance becomes shaded (Fenner et al., 2012). At the two burrows we monitored, all entrances were used by multiple individuals and some entrances (those in more open areas) were used more, or much more, than others. Other than vegetation cover, we did not observe any external features around highly used entrances that could assist in their a priori identification. Additional monitoring over longer periods and at more burrows would help to determine whether levels of relative usage among entrances within a burrow system remain stable through time, and whether the use of entrances varies based on the type of activity, for example whether a skink is foraging or basking.

| Comments on methods and suggestions for further research
Considering the variability in activity and vegetation cover at entrances within a burrow, it is likely that our ability to compare activity among entire burrow systems in relation to vegetation cover was constrained by only measuring activity at one entrance.
Nevertheless, using proportional time active outside or inside a burrow entrance as a response variable has provided new insights about tjakura activity relative to vegetation, suggesting they may reduce activity outside the burrow when there is little cover. We writing -review and editing (equal). Christine A. Schlesinger: Conceptualization (equal); formal analysis (supporting); methodology (equal); project administration (lead); supervision (lead); writing -original draft (supporting); writing -review and editing (equal).

ACK N O WLE D G E M ENTS
We acknowledge the traditional owners of the land in which this research was undertaken, the Anangu people. We recognize the importance of tjakura and the jointly managed Uluru-Kata Tjuta National Park to Anangu and thank them for permission to conduct this research on their traditional lands. We are grateful for the inkind support provided by Uluru-Kata Tjuta National Park and permission to work in the park. The project was approved by the CDU Animal Ethics Committee (project number A14009). National Park staff provided monitoring data and locations of tjakura burrows, and advice and support during field work. We also thank Andrea Hinwood and Jonric Ridley for field assistance. We also acknowl-

FU N D I N G I N FO R M ATI O N
There was no external funding for this research. The research was supported financially by Charles Darwin University, including Honours student funding allocation.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data for this paper are available on Dryad https://doi.org/10.5061/ dryad.vmcvdnczh.