Skip to main content

Temporal patterns of visitation of birds and mammals at mineral licks in the Peruvian Amazon

Cite this dataset

Griffiths, Brian; Bowler, Mark; Gilmore, Michael; Luther, David (2021). Temporal patterns of visitation of birds and mammals at mineral licks in the Peruvian Amazon [Dataset]. Dryad.


Mineral licks are key ecological resources for many species of birds and mammals in Amazonia, providing essential dietary nutrients and clays, yet little is known about which species visit and their behaviors at the mineral licks. Studying visitation and behavior at mineral licks can provide insight into the lives of otherwise secretive and elusive species. We assessed which species visited mineral licks, when they visited, and whether visits and the probability of recording groups at mineral licks were seasonal or related to the lunar cycle. We camera trapped at 52 mineral licks in the northeastern Peruvian Amazon and detected 20 mammal and 13 bird species over 6,255 camera nights. Generalized linear models assessed visitation patterns and records of groups in association with seasonality and the lunar cycle. We report nocturnal curassows (Nothocrax urumutum) visiting mineral licks for the first time. We found seasonal trends in visitation for the black agouti (Dasyprocta fuliginosa), red howler monkey (Alouatta seniculus), blue-throated piping guan (Pipile cumanensis), red brocket deer (Mazama americana), collared peccary (Pecari tajacu) and tapir (Tapirus terrestris). Lunar trends in visitation occurred for the paca (Cuniculus paca), Brazilian porcupine (Coendou prehensilis) and red brocket deer. The probability of recording groups (>1 individual) at mineral licks was seasonal and related to lunar brightness for tapir. Overall, our results provide important context for how elusive species of birds and mammals interact with these key ecological resources on a landscape scale. The ecological importance of mineral licks for these species can provide context to seasonal changes in species occupancy and movement.


Study Site

Fieldwork took place in the Maijuna community of Sucusari and the Maijuna-Kichwa Regional Conservation Area (MKRCA), a 391,039-hectare protected area in Loreto, Peru (El Peruano, 2015). This area is about 120 kilometers north by river of Iquitos, Peru (Fig 1). The title lands of the Maijuna community encompasses 4,771 hectares and directly adjoin the MKRCA to the south. The Sucusari River is a tributary of the Napo River and terrestrial habitats include both upland terra firme primary rainforest and floodplain forest. The mean annual temperature is 26oC and an average precipitation of 3100 millimeters per year (Marengo, 1998). The wet season consists of the months November to May, while the dry season is mainly June to October in the Iquitos region (Espinoza Villar et al., 2009).

Camera Trapping

We installed motion-activated camera traps (Bushnell Aggressor, Boly Scout Guard) in the Sucusari River Basin at a sample of 52 mineral licks that were identified with the assistance of Maijuna hunters. Starting in August, 2018, we visited all mineral licks, obtained GPS coordinates, and placed camera traps in a series of four deployments, each lasting at least 60 days to achieve even coverage of the whole basin (Fig 1). We left camera traps undisturbed at mineral licks for the entire rotation period. Every 60 days cameras were removed, batteries and SD cards changed, and cameras were rotated to new mineral licks (Kays et al., 2020). During the third rotation most cameras went to previously unvisited mineral licks, but some went to mineral licks that held a camera in August but experienced camera malfunctions that prohibited the camera from gathering 60 camera-nights of data.  

The mineral licks in the Sucusari River basin are generally characterized by waterlogged mud with standing water and a face, which was often associated with a slope. The area inside the lick was generally devoid of vegetation. The number of camera traps placed in each mineral lick was determined by the size and shape of the mineral lick, with the goal of recording all animal visits to the mineral lick and meeting the assumption of perfect detection (all medium and large-bodied animals entering the lick are captured). We set cameras to record three rapid-fire images at each motion trigger with a delay of two minutes between each set of images to avoid expending the camera’s batteries. Cameras were set at a minimum of 50 centimeters from the ground, facing the active face and entrance to the mineral lick, following Tobler et al. (2009). We determined the location of the face from signs of animal activity. Camera traps at mineral licks that did not have a face were placed facing mud with signs of active animal activity.

We identified all medium and large sized mammal and bird species (weight > 1 kg) in camera trap images (Blake, 1977; Emmons & Feer, 1997), removed empty images and organized data for analyses using CameraBase v1.7 (Tobler, 2015). The number of individuals and species identity in instances where multiple individuals appeared in the same photograph was also recorded. Small-bodied birds and mammals, including bats, were removed from analyses because they could rarely be identified to species level. Mixed species flocks of parakeets were also not considered for analysis since they commonly visited in groups of several hundred individuals and could not be reliably identified to species level. Images were sorted into independent visitation events, where multiple visits by the same species within one hour of each other were considered one visitation event, following Tobler et al. (2008).

Usage notes

Birds were identified separately from mammals, in a different sheet. An independent event was considered two visits separated by one hour.


Fulbright Association