Individual dietary variation has important ecological and evolutionary consequences. However, it has been overlooked in many taxa that are thought to have homogeneous diets. This is the case of vultures, considered merely as “carrion eaters”. Given their high degree of sociality, vultures are an excellent model to investigate how inter-individual transmissible behaviors drive individual dietary variation. Here, we combine GPS-tracking and accelerometers with an exhaustive fieldwork campaign to identify the individual diet of 55 griffon vultures (Gyps fulvus) from two Spanish populations that partially overlap in their foraging areas. We found that individuals from the more humanized population consumed more anthropic resources (e.g., stabled livestock or rubbish), resulting in more homogeneous diets. In contrast, individuals from the wilder population consumed more wild ungulates, increasing their dietary variability. Between sexes, we found that males consumed anthropic resources more than females did. Interestingly, in the shared foraging area, vultures retained the dietary preference of their original population, highlighting a strong cultural component. Overall, these results expand the role of cultural traits in shaping key behaviors, and call for the need of including cultural traits in Optimal Foraging models, especially in those species that strongly rely on social information while foraging.

GPS-tracking and study areas

We captured 65 adult griffon vultures, 30 individuals in Southern Spain between December 2014 and January 2015 and 35 individuals in Northern Spain between December 2015 and March 2016. The Southern area is a mountainous region (500–2,107 m a.s.l.) covered by Mediterranean woodlands and pasturelands, where the main human uses are traditional farming, hunting, forestry and tourism (44). The Northern area is a flat area (28–659 m a.s.l.) surrounded by mid-sized mountains (up to 1500 m a.s.l.) and highly transformed for intensive agriculture, with traditional sheep livestock being replaced by intensive farming (45, 46). This leads to more predictable ungulate carrion sources in the Northern area compared to the Southern one (30) although trophic resources can be considered abundant in both areas. In addition to these areas, both vulture populations share a second foraging area in Southwestern Spain (Figure 1 and Figure S3) where carrion from wild and domestic ungulates is also abundant (42).

We sexed vultures by molecular procedures (47). We determined age by morphological features (48). All birds were equipped with 90 g GPS/GPRS-GSM devices that include accelerometers from e-obs digital telemetry. The setting of the GPS/ACC devices varied depending on weather conditions and the power level of the batteries (see Table S1). Except for those birds that died or whose device failed (N=15), we tracked all the vultures between the capture day and December 2018.

Identification of feeding events and diet description

For each tracked vulture, we identified potential feeding events using Accelerater, a supervised learning algorithm (http://accapp.move-ecol-minerva.huji.ac.il/; see (40)) implemented with validated samples recorded in feeding stations (26, 16). We recorded 11,636 possible feeding events (Figure 1); from these, we visited 4,372 locations during fieldwork campaigns, confirming feeding by vultures in 3,338 cases (efficiency to locate feeding events: 76.35%). For each event, we recorded: i) individual identifier of the vulture/s involved, ii) coordinates of the feeding events, grouping them in grids of 10 x10 km, iii) the feeding site, classified into seven categories: random (i.e. carcasses not associated to infrastructures), hunting property, extensive farming, intensive farming, carcass dumping site, landfill, and “other” (see Figure 2 for details),  iv) the origin of the carcass, according to five groups: livestock, wild, mixed (i.e., carcasses from both livestock and wild ungulates found together), rubbish, and other (e.g., dogs, rests of human food or unknown carcasses); v) the species to which the carcass/es belonged in those cases where it was possible to identify.

In addition, when access to the feeding sites was prevented (e.g., terrain too steep or private property), we identified feeding events of GPS-tracked vultures by crossing expert opinion and official data on livestock and hunting areas from the Spanish Ministry of Agriculture, Fisheries and Food (50, 51). Thanks to this information, we identified 1,493 feeding events. Our final dataset is composed of 4,831 feeding events (Figure 1), representing 41.52% of the total feeding events identified by GPS-tracking data. This dataset excludes individuals with less than 30 feeding events recorded, being the total number of individuals studied was 55, 29 from the Northern population (15 females and 14 males) and 26 from the Southern population (11 females and 15 males).

Then, we compared the number of feeding events at different feeding sites and with different carrion origin between the Northern and Southern populations and between sexes, using chi-square tests in R (52).