Introduction
Understanding animal behavioural responses to environmental variation and change is a frequent goal in animal ecology and has important implications for conservation (e.g., Hopkins et al. , 2023; Kayset al. , 2015; O’Brien et al. , 2023a). However, collecting behavioural data on animals, especially in remote areas, is very time-consuming and requires a lot of effort and resources. Camera systems are increasingly used to remotely monitor wildlife because they can record behaviour over long and continuous periods and are generally non-invasive (see reviews: Cutler & Swann, 1999; Jolles, 2021; Trollietet al. , 2014). Camera systems can be commercially built systems (e.g., Meek et al. , 2014; Trolliet et al. , 2014), which are ready to use and generally compact but can be expensive and have less flexibility in program settings, battery life and data storage (Coxet al. , 2012; Prinz et al. , 2016; Reif & Tornberg, 2006). Alternatively, they can be do-it-yourself assembly, which can be more flexible and cheaper (e.g., Cox et al. , 2012), but can be time-consuming to make and harder to use.
Self-assembled camera setups generally use a microcomputer, typically Raspberry Pi (www.raspberrypi.org) or Arduino (www.arduino.cc) (Allan et al. , 2018; Greenville & Emery, 2016; Johnston & Cox, 2017). Raspberry Pi-based video recorders have been used in animal behavioural studies; for example, behavioural studies of the waggle dance of honeybees (Aiet al. , 2017), the spacing of foraging fruit flies (Churchillet al. , 2020), nematode behaviour (Nuñez et al. , 2017), monitoring of mammal populations (see review Swann et al. , 2004), respiration and pupil dilation in laboratory mice (Kallmyer et al. , 2017; Privitera et al. , 2020), avian studies assessing nest box use, parental care and other behaviours (e.g. Hereward et al. , 2021; Prinz et al. , 2016; Zárybnická et al. , 2016). The use of Raspberry Pi units to collect video data on behaviour in amphibians has, to the best of our knowledge, only been conducted on Hellbender salamanders (Cryptobranchus alleganiensis ) in the United States (O’Brien et al. , 2023a,b), where cameras were used in artificial aquatic shelters to described courtship and parental care behaviours and nest outcome. Self-assembled cameras can also be combined with other sensors; for example, temperature and humidity sensors (McBride & Courter, 2019).
Artificial shelters have been proven to be a successful tool to monitor amphibian populations. They have an advantage over labour-intensive trapping and observation methods (Sutherland et al. , 2016) and reduce disturbance to the animals and their environment (Hesed, 2012). Artificial shelters built with the option of camera observation are a good way to obtain important behavioural data while minimising impacts on animals or their habitat. Once an animal is using a shelter, it can be observed for extended periods using a small camera that is serviced without any disturbance to the animal or the shelter. However, many amphibian species are small, so cameras need to focus on small subjects at close range. Furthermore, many amphibian species live in remote areas (e.g., mountain tops), so shelters and associated camera setups need to be small and transportable and have long recording periods. Previous studies have made portable camera setups, but the minimum focal distance is 15–25 cm (Hereward et al. , 2021; O’Brien et al. , 2023a, b) and the recording time is limited to a maximum of 96 hours (O’Brien 2023a, b). These cameras are of limited use for the study of small animals under shelters, and need to be serviced regularly or recordings need to be short and/or well-spaced in time.
In this study, we describe and test a Raspberry Pi camera setup that is portable, robust to weather, can continuously record for 21 days (504 hours), and has a focal distance of 2.5 cm. The camera setup was tested on a small (approximately 3.5 cm long) microhylid frog species that lives among leaf-litter, logs and rocks on remote mountaintops in north Queensland, Australia. We also describe two artificial shelters to mount the cameras in. The combination artificial shelter/camera setup we present could be used to collect natural history and behavioural data on any small, cryptic vertebrate or invertebrate species.