DISCUSSION
In this paper, we have described and demonstrated the successful building and testing of a Raspberry Pi-based camera that produces high quality close-up (2.5 cm) videos and is portable and weatherproof. The camera setup can record continuously for up to 21 days in remote locations, and we have demonstrated how they can be used in natural settings or paired with artificial shelters. The cameras could be used for many different research topics, in many different species and settings. Collecting long-term continuous behavioural data, especially in the field and from small species, is very intensive and often logistically not possible (for example, at night or when the animal is in their shelter). This camera setup will enable behavioural data to be collected on the natural history of small and cryptic species.
A benefit of the camera is that you have control over programming when to record. In this trial, we programmed the camera to continuously record for 21 days. We recommend starting with this method to get an idea of the species daily time budget. However, the camera can be coded to record whatever data is required for the specific research questions; for example, for 30 minutes every four hours, or only day- or night-time. Specific recording periods will increase battery life and increase the camera data collection time. The self-assembled camera setup could be combined with sensors such as temperature and humidity loggers (McBride & Courter, 2019). We deployed the cameras on a mountain in the tropical rainforest which has high rainfall (average annual rainfall of 2534.7 mm) and humidity during the summer months of >90% 24 hours a day. We therefore did no attach temperature and humidity sensors because adding these would have required extra holes to be drilled, increasing the chance of moisture entering the setup. Adding sensors would also decrease the battery life. The goal of this study was to build a camera setup that is robust, extremely weatherproof, with as long as possible recording time, and relatively cheap to make. We did use separate temperature and humidity sensors (Hygrochron iButtons), which are robust and don’t connect to the camera setup or battery (Fig. 5).
The artificial shelters proved to be very successful because they were voluntarily occupied by the frogs in high numbers and enabled filming at optimal distance. Importantly, they also enabled extended behavioural observations without influencing the frog’s behaviour. Obtaining sheltering and breeding information in a species like this would otherwise require regular turning of logs and other cover, disturbing (and potentially injuring) the frogs and damaging the micro-habitat. Due to this success, we now have increased the number of shelters and expanded the research to other upland areas, with the objective of recording the breeding behaviours and breeding success of Critically Endangered microhylid frog species. While both artificial shelter types (concrete and log) were successful, filming and maintenance of the concrete shelters was easier, due to the sturdier and stable material, and the predrilled holes. In addition, the log artificial shelters will disintegrate and fall apart with time and will need to be replaced. In contrast, the concrete shelters will last and will increasingly ‘integrate’ with the environment (e.g., moss growing on them, leaf-litter falling on them).
During our trial many other species of small animals occupied the chambers and were filmed. Of particular interest were cockroaches (Periplaneta australasiae ) that occupied the shelters for short periods of time. We discovered that the infrared LED light used in the camera setup illuminate through the exoskeleton of the cockroaches, showing the internal organs (Fig. 4C). In one set of filming, the infrared LED light showed that one of the cockroaches had a gut parasite that could clearly be seen moving internally. This shelter/camera setup could be used to study, for example, the prevalence of gut parasites in cockroaches in different habitats, the development of the gut parasite, and the effect on the cockroach.
A strength of this shelter/camera setup is that the animals choose to take up residence and were then filmed with almost no disturbance. Servicing of the batteries and storage cards does not impact the camera end or shelter. We suggest these shelters and cameras could be used to film a range of behaviours, for example, parental care behaviour, shelter use, species interactions, and courtship behaviour, in many cryptic amphibian species and also other small vertebrates and invertebrates, globally.