While remote photographies have been used for more than a century, as presented by O’Connell et al. (2011), the automated camera trap as it is now known came onto the market at the end of the 1980s. Savidge et al. (1988) used a film camera connected to an infrared transmitter, which was able to shoot a picture as soon as the beam was interrupted by an animal. The system was automatic; after a picture had been taken, the film was reloaded and the camera ( for example: MOULTRIE PANORAMIC 150I ) was ready to take more pictures. This technique was used to identify predators visiting bird nests. Some years later, Carthew et al. (1991) and Kucera et al. (1993) listed the advantages of the automated camera trap system for an array of different field applications such as the study of activity patterns, intra-community interactions and large carnivores populations.

The first studies using camera traps for the purpose of large mammal conservation appeared in the 1990s and focused on the tiger, Panthera tigris (e.g., Griffiths, 1993; Karanth, 1995). Following the designation of P. tigris as endangered (Chundawat et al., 2011), one of the few “flagship” species listed on the IUCN red-list as early as 1986, these studies aimed at estimating home range span and population size. In this way, the use of camera traps to estimate population size greatly helped towards the conservation strategy for the species, and more generally, the monitoring of other threatened populations and communities. This use of camera traps was highlighted in a study on the activity patterns of mammal communities in Indonesian rain forests (van Schaik et al., 1996). The aforementioned early studies of the use of camera traps clearly illustrate the major advantages of using the technique, including being able to observe cryptic or elusive animals living in difficult to access habitats such as dense tropical forests. The use of camera traps has been revolutionary for studying the behavior of carnivores, as they are difficult to observe in their natural habitat due to their solitary nature. The technique has also been the subject of many other scientific papers since the beginning of the 21^st century, revealing more about the ecology of rare, nocturnal animals, as well as those highly sensitive to the presence of humans or those living in large home ranges. A good example is the study of Moruzzi et al. (2002), which promotes the use of this technology for estimating carnivore distribution over large area and documenting species-specific habitat preferences.

A large proportion of conservation projects aim at managing threatened species, which implies to monitor populations over time and space. Thus, the majority of studies using camera traps nowadays appear to deal with the estimation of population density (e.g., Kalle et al., 2011; Garrote et al., 2012; Oliveira-Santos et al., 2012) or simply with the presence of species in given areas (e.g., Gil-Sanchez et al., 2011; Gray et al., 2011; Liu et al., 2012). Population characteristics are, to a greater or lesser extent, related to habitat use behaviors and habitat selection. Camera traps are useful for monitoring these aspects as they allow the estimation of home range size (e.g., Gil-Sanchez et al., 2011).

Some studies also deal with activity budget (e.g., van Schaik et al., 1996; Azlan et al., 2006; Gray et al., 2011; Oliveira-Santos et al., 2012) and a smaller number with more specific behaviors. For instance, Soley et al. (2011) reported the storing behavior of non-ripe fruits by a mustelideae, allowing the fruits to mature and to be consumed on future occasions; this is a specific behavior that is very hard to report without camera traps. Blake et al. (2010) studied the importance of salt licks for an animal community in a neotropical forest. Other studies have dealt with animal infant care (e.g., Charruau et al., 2012) or social interaction (Lopucki, 2007; Srbek-Araujo et al., 2012).

Camera traps are also increasingly being used to study plant-animal interactions such as seed dispersal and predation (e.g., Babweteera et al., 2010; Nyiramana et al., 2011; Campos et al., 2012; Koike et al., 2012; Pender et al., 2013). Moreover, focal observations need to be conducted in the study of the seed dispersal capacity of a given plant species, to list the frugivore species interacting with the plants and to define the quantitative contribution of each species in the process of seed dispersal. Camera traps are revolutionary in this regard, as they allow the identification of diurnal, nocturnal, and shy species that would not be seen using other methods such as direct observation. This is exemplified by the study of Nyiramana et al. (2011), who discovered that a species of rodent, the forest giant pouched rat Cricetomys emini(Wroughton, 1910), was responsible for the secondary dispersal of large seeds in an Afro-tropical forest.

3.2. Various technical aspects

More than a decade ago, Cutler et al. (1999) reviewed the advantages and disadvantages of using different film camera trapping equipment depending on the research objectives. Given the rapid advances in such technology, and the great variety of camera trap brands and digital models existing on the market nowadays, film cameras are competed. We present here the most important characteristics to take into account when choosing digital equipment. Characteristics such as trigger speed, detection zone, recovery time, night detection and battery consumption can vary greatly and have a significant impact on the types of data to be collected, such as the number of species detected and photographic rates (Hughson et al., 2010). Therefore, the choice of the most appropriate equipment is an important consideration.

Trigger speed. Trigger speed is the time delay necessary for the camera to shoot a picture once an animal has interrupted the infrared beam within the camera’s detection zone. This delay can vary from between 0.197 seconds for the Reconyx HC500 model to 4.206 seconds for the Stealth Cam Rogue IR model. Given the relatively narrow field of view of most camera trap lenses (42 mm), a slow trigger speed does not allow the photographing of fast moving animals (Scheibe et al., 2008). Thus, depending on the study goals and the target animal species, this time delay could be a crucial characteristic to consider. For example, if a camera is set up at a random location for a wildlife survey (Pereira et al., 2012), fast moving animals are likely to pass in front of the camera trap without stopping. In this case, a very reactive camera (with a fast trigger speed) would be necessary so it could shoot pictures of the detected animal before it left the camera’s field of view. In their comparative study of motion-activated cameras for widlife investigation, Hughson et al. (2010) showed that some camera models (such as the fast Reconyx) can detect up to 86% more animal species. If the trigger speed is too slow, the camera may frame only a part of the animal or may even take empty pictures (pictures not showing what the beam has detected). Hughson et al. (2010) observed that, in comparison with other models, Leaf River cameras took the highest percentage of empty pictures. In the case of a camera installed in front of a bird nest, a bait, or a lure, visiting animals are more likely to stay longer (to either depredate the nest or interact with the bait) and to trigger more photographs (Garrote et al., 2012; Trolle et al., 2003) even if the camera has a relatively long time delay (low reactivity). Using lures to attract large carnivores can also allow a better identification of individuals (Gil-Sanchez et al., 2011). This risk of taking empty pictures does not only depend on the speed of the camera in taking a picture; the detection zone as well as the field of view are also primary criteria to consider.

The detection zone. The detection zone is the zone covered by the camera’s infrared beam in which movement can be detected. The zone varies in width and depth, depending on the model (Table 1). This criterion is probably the most important in determining detection rate (Rowcliffe et al., 2011) and therefore the number of pictures that will be taken in a given event.

The field of view. The field of view is the zone covered by the camera lens, and which appears on the pictures. The field of view is generally 42° but there are rare exceptions such as with the Leupold brand, which goes up to 54° (Table 1) and the Moultrie panoramic model, which covers an angle of 150°. The detection zone can vary greatly according to the brand and the model. We thus find models with a detection zone wider than the field of view (e.g. DLC Covert Extreme) and models with the detection zone narrower than the field of view (e.g. Cuddeback Ambush). Where the detection zone is wider than the field of view (Figure 1a), the advantage lies in being better able to capture fast moving animals. The limitation in this case is that the camera is also likely to take empty pictures when animals enter the detection zone (thus passing through the infrared beam and triggering the camera) but without making it into the field of view. Where the detection zone is narrower than the field of view (Figure 1b), the detection zone is centered relative to the field of view of the camera, and so the advantage can be seen in gaining well centered pictures. This can be very useful for the identification of large mammals. However, the limitation in this case is that relatively fewer pictures per visit can be shot, as animals are likely to occupy the field of view without crossing the detection zone. As presented in table 1, the detection zone can be described with a given width (angle) and a given distance from the camera at which it will detect an animal. The detection distance of a camera is an important aspect to consider when focusing on animal species of either large or small body mass. Larger animals will be more easily detected at further distances than smaller animals. However, speed of movement seems to be less correlated with detection distance (Rowcliffe et al., 2011).