Fire has shaped the evolution of both plants and animals. Animals exposed to fire throughout their evolutionary history are predicted to exhibit behavioural adaptations that enhance survival during fire. Here, we investigated whether Australian sleepy lizards (Tiliqua rugosa), a large skink from fire-prone regions, recognise and respond to cues of fire. Motivated by reports of captive sleepy lizards reacting to smoke, we conducted behavioural trials exposing wild-caught sleepy lizards to the chemosensory (smoke) and auditory (fire sounds) cues of fire. Behavioural analysis revealed that sleepy lizards exhibited increased activity and significantly greater movements in response to smoke than to water vapour. They did not, however, react aversively to auditory cues of fire, suggesting a reliance on chemosensory rather than auditory cues for fire detection. Our findings provide empirical support for the hypothesis that chemosensory cues of fire elicit escape responses in animals from fire-prone regions, suggesting an evolved, likely innate, behavioural adaptation to recognise and respond to fire cues as indicative of a threat. As climate change increases the frequency and intensity of wildfires, understanding how animals perceive and respond to fire will prove crucial for predicting the threat posed by a more fire-prone future.

Australian sleepy lizards (Tiliqua rugosa), also known as shinglebacks or bobtails, are a large (mean snout–vent length approx. 30cm [17]), viviparous, terrestrial skink, with an omnivorous diet [18]. They are widely distributed across southern Australia and inhabit a variety of mesic to arid habitats, including coastal heaths, eucalypt woodlands, Acacia scrublands and spinifex-dominated deserts [18,19]. Sleepy lizards typically shelter under shrubs and ground debris, such as logs and sheets of iron, but will seek shelter in the burrows of other animals during extreme temperatures [20]. Sleepy lizards are long-lived (at least 50 years [21]), experience low adult mortality [21], form monogamous breeding pairs [22] and produce only a few large offspring per litter (1–3 neonates [23]), which take 3−4 years to reach sexual maturity [21].

In 2018, 10 adult female sleepy lizards were collected from near Lake Victoria (−34.0, 141.3), southwest New South Wales, Australia, as part of ongoing research at Macquarie University (e.g. [24–26]). Lake Victoria is in the Murray–Darling Depression bioregion, which experiences a warm arid climate and is dominated by mallee shrublands [27]. This region experiences medium-intensity shrub fires in spring and summer with historical fire return intervals of 20−100 years [28]. Following collection, the lizards were housed at Macquarie University’s animal husbandry facility in temperature-controlled rooms and, during periods of low humidity, occasionally outdoors. All lizards were held in captivity from 2018 until the commencement of this study between September and November 2023. During this time, individuals were housed separately in open-topped plastic tubs (80 L × 60 W × 45 H cm) and individually marked for identification.

To test the behavioural response of sleepy lizards to the chemosensory and auditory cues of fire, we exposed each lizard to four treatments: a chemosensory cue (smoke), a control (water vapour), an auditory cue (recording of a wildfire) and an auditory control (white noise). Each lizard received all four treatments on three separate occasions, resulting in a total of 12 trials per individual. Trials were conducted in each lizard’s home enclosure, with all furnishings (i.e. water and food bowls, logs), except for a substrate of wood shavings, removed. All enclosures were fitted with a clear acrylic lid to which a camera was mounted. All auditory trials were conducted in a temperature-controlled (mean trial temperature ± SE = 31.34 ± 0.1°C; mean trial humidity ± SE = 39.15 ± 0.65%), relatively soundproof experimental room. To avoid setting off smoke alarms, all chemosensory trials were conducted outdoors under the shelter of a veranda awning, which prevented direct sun exposure and helped maintain consistent conditions across trials. Outdoor chemosensory trials were conducted on days predicted to be ≥30°C and at the same time of day (mean trial temperature ± SE = 30.5 ± 0.37°C; mean trial humidity ± SE = 27.86 ± 1.41%). Across all trials, humidity and temperature (°C) were recorded immediately prior to each trial.

Prior to each trial, lizards in their enclosures were placed in the experimental room (auditory trials) or outside (chemosensory trials) and were allowed to acclimatize for 1 h. After the acclimatization period, lizard pre-trial behaviour was recorded for 10 min using a GoPro Hero7 Silver (GoPro Inc., San Mateo, California, USA) mounted above the centre of each enclosure. After the pre-trial period, all lizards were exposed to one of the four treatments, and their behaviour was recorded for a further 10 minutes.

To expose lizards to smoke, we burnt dry native grass (Heteropogon contortus) through an infusion smoker (Davis and Waddell, Preston, Queensland, Australia). Smoke was blown for 5 s through a rubber tube directly into each enclosure via an opening on the top of a wall of the enclosure. To expose lizards to water vapour, we used a Hurricane 700 fog machine (Chauvet DJ, Sunrise, Florida, USA) to blow odourless water vapour for 5 s through a rubber tube using the same method as the treatment. This methodology attempted to expose lizards to similar visual cues, but only smoke provided the chemical cues of fire [13].

To expose lizards to the auditory cues of fire, we played a recording of a wildfire from Álvarez-Ruiz et al. [16]. As an auditory control, we exposed lizards to a recording of white noise from Bent et al. [29]. Auditory treatments were played through a WONDERBOOM Portable Bluetooth Speaker (Ultimate Ears, Irvine, California, USA) at full volume. The speaker was positioned in the middle of the room approx. 1 m above the ground and approx. 1 m from each lizard enclosure. We recorded the maximum decibels and average decibels of the recordings during each trial.

The order in which individuals were exposed to treatments (smoke or the sound of fire) and controls (water vapour or white noise) was randomized; however, auditory trials were typically conducted first, and the chemosensory trials were conducted when outside temperature allowed. After every individual had been exposed to all four treatments, the treatments were repeated twice more (i.e. each lizard was exposed to the four treatments three times).

We assessed lizard behavioural responses to the cues of fire by autonomously tracking their movement during the 10 min trial period using the video tracking software AnimalTA, v. 3.1.0 [30]. AnimalTA tracked the proportion of time spent moving and distance travelled (cm) by each lizard before and after exposure to treatments. Videos were tracked at a frame rate of 7.5 frames s⁻¹ and the moving threshold was set to 0.2 mm s⁻¹. This moving threshold detected stationary movements (i.e. walking on the spot against enclosure walls) without excessive detection sensitivity. Because lizards use their tongue to detect chemosensory cues, we also used BORIS v. 8.22.6 [31] to score the number of times each lizard extended and retracted its tongue during each trial. All tongue-flicks were counted by a single observer (EHVvdP), who was blinded to the treatment being scored where possible (i.e. with sound muted). Despite precautions, in some cases smoke and vapour could be differentiated.

All statistical analyses were performed using R v. 4.3.1 [32]. To test the behavioural response of sleepy lizards to the chemosensory and auditory cues of fire, we ran generalized linear mixed models (GLMMs), fitted via the glmmTMB package [33], to compare the behavioural responses of lizards (i.e. proportion of time spent moving, distance travelled and tongue-flick count) between treatments (smoke or fire sounds) and controls (water vapour or white noise). Proportion of time spent moving was modelled using GLMMs with a beta distribution and logit link function. Because the beta distribution requires values strictly between 0 and 1, we applied a bias correction transformation to this response variable. Distance travelled by lizards was analysed using GLMMs with a gamma distribution after adding 0.01 to the response variable to avoid zeros. Tongue-flicking behaviour was modelled using negative binomial GLMMs to account for overdispersion. In all models, we initially included interactive effects of treatment, temperature, humidity and—for auditory trials only—average decibel level. No interactions were significant. We then used Akaike’s information criterion (AIC) to compare models with and without non-significant fixed effects and retained models with the lowest AIC values. Ultimately, all fixed effects except for the treatment variable were excluded, as they did not improve model fit. All models included a random intercept for individual lizard to account for repeated trials.

References

13. Álvarez-Ruiz L, Belliure J, Pausas JG. 2021 Fire-driven behavioral response to smoke in a Mediterranean lizard. Behav. Ecol. 32, 662–667. (doi:10.1093/beheco/arab010)

14. Grafe TU, Döbler S, Linsenmair KE. 2002 Frogs flee from the sound of fire. Proc. R. Soc. Lond. Ser. B 269, 999–1003. (doi:10.1098/rspb.2002.1974)

15. Scesny AA. 2006 Detection of fire by eastern red bats (Lasiurus borealis): arousal from torpor. MSc thesis, Missouri State University, Springfield, MO.

16. Álvarez-Ruiz L, Pausas JG, Blumstein DT, Putman BJ. 2023 Lizards’ response to the sound of fire is modified by fire history. Anim. Behav. 196, 91–102. (doi:10.1016/j.anbehav.2022.12.002)

17. Bull CM, Pamula Y. 1996 Sexually dimorphic head sizes and reproductive success in the sleepy lizard Tiliqua rugosa. J. Zool. 240, 511–521. (doi:10.1111/j.1469-7998.1996.tb05302.x)

18. Norval G, Gardner MG. 2020 The natural history of the sleepy lizard, Tiliqua rugosa (Gray, 1825)—insight from chance observations and long‐term research on a common Australian skink species. Austral Ecol. 45, 410–417. (doi:10.1111/aec.12715)

19. Cogger HG. 2018 Reptiles & Amphibians of Australia, 7th edn. Collingwood, Victoria: CSIRO Publishing.

20. Kerr GD, Bull CM, Burzacott D. 2008 Refuge sites used by the scincid lizard Tiliqua rugosa. Austral. Ecol. 28, 152–160. (doi:10.1111/j.1442-9993.2003.tb00238.x)

21. Bull CM. 1995 Population ecology of the sleepy lizard, Tiliqua rugosa, at Mt Mary, South Australia. Aust. J. Ecol. 20, 393–402.

22. Bull CM, Cooper SJB, Baghurst BC. 1998 Social monogamy and extra-pair fertilization in an Australian lizard, Tiliqua rugosa. Behav. Ecol. Sociobiol. 44, 63–72. (doi:10.1007/

s002650050515)

23. Bull CM, Pamula Y, Schulze L. 1993 Parturition in the sleepy lizard, Tiliqua rugosa. J. Herpetol. 27, 489. (doi:10.2307/1564848)

24. Szabo B, Hoefer S, Whiting MJ. 2020 Are lizards capable of inhibitory control? Performance on a semi-transparent version of the cylinder task in five species of Australian skinks. Behav. Ecol. Sociobiol. 74, 118. (doi:10.1007/s00265-020-02897-y)

25. Szabo B, Holmes ML, Ashton BJ, Whiting MJ. 2024 Spontaneous quantity discrimination in the Australian sleepy lizard (Tiliqua rugosa). Behav. Ecol. 35, arad089. (doi:10.1093/beheco/arad089)

26. Szabo B, Whiting MJ. 2020 Do lizards have enhanced inhibition? A test in two species differing in ecology and sociobiology. Behav. Process. 172, 104043. (doi:10.1016/j.beproc.2020.104043)

27. Bastin G, ACRIS Management Committee.2008 Rangelands 2008—taking the pulse. Canberra, Australia: National Land & Water Resources Audit. See https://www.dcceew.gov.au/sites/default/files/documents/rangelands08-pulse.pdf.

28. Murphy BP et al. 2013 Fire regimes of Australia: a pyrogeographic model system. J. Biogeogr. 40, 1048–1058. (doi:10.1111/jbi.12065)

29. Bent AM, Ings TC, Mowles SL. 2018 Anthropogenic noise disrupts mate searching in Gryllus bimaculatus. Behav. Ecol. 29, 1271–1277. (doi:10.1093/beheco/ary126)

30. Chiara V, Kim S. 2023 AnimalTA: a highly flexible and easy‐to‐use program for tracking and analysing animal movement in different environments. Methods Ecol. Evol. 14, 1699–1707. (doi:10.1111/2041-210x.14115)

31. Friard O, Gamba M. 2016 BORIS: a free, versatile open‐source event‐logging software for video/audio coding and live observations. Methods Ecol. Evol. 7, 1325–1330. (doi:10.1111/2041-210x.12584)

32. R Core Team. 2024 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. See http://www.R-project.org/.

33. Brooks ME, Kristensen K, van Benthem KJ, Magnusson A, Berg CW, Nielsen A, Skaug HJ, Mächler M, Bolker BM. 2017 Modeling zero-inflated count data with glmmTMB. bioRxiv 132753. (doi:10.1101/132753)