Table 1.
Summary of all the variables we included in our analyses with definitions where necessary. The variables are divided into those variables which were of interest to this study and those that were included in our analyses as controls. Each of the four models listed were fitted separately to natural and experimental data
| Variable | Definition | Included in models of |
|||
|---|---|---|---|---|---|
| Aggression rate1 | Proportion of aggression received1 | Intake rate2,3 | Patch residency time2,3 | ||
| Variables of interest | |||||
| Rank | Position within troop dominance hierarchy. See Methods text for details of calculation | All models | |||
| Social bond strength | Mean proportion of focal follow time an individual was observed in grooming interactions | All models | |||
| Social bond balance | Mean proportion of time spent grooming as the groomer | All models | |||
| Relatedness | Mean dyadic relatedness to all other troop members | All models | |||
| Rank × Social bond strength | All models | ||||
| Rank × Social bond balance | All models | ||||
| Rank × Relatedness | All models | ||||
| Social bond strength × Relatedness | All models | ||||
| Social bond balance × Relatedness | All models | ||||
| Social bond strength × Social bond balance | All models | ||||
| Initial intake rate | Number of bites during first 10 s of patch visit | No | No | No | Yes |
| Initial intake rate × Rank | No | No | No | Yes | |
| Initial intake rate × Social bond balance | No | No | No | Yes | |
| Initial intake rate × Social bond strength | No | No | No | Experiment only | |
| Control variables | |||||
| Individual | |||||
| Age | Age at the start of the study period, estimated by tooth eruption and wear patterns. See Methods text. | All models | |||
| Sex | All models | ||||
| Satiation | Cumulative time the individual had spent on other patches prior to visiting the current patch (s) | No | No | Experiment only | Experiment only |
| No | No | ||||
| Patch | |||||
| Quality | Natural: patch surface area (m2); Experiment: Initial density of food (g/m2) | No | No | Yes | Yes |
| Depletion | Cumulative baboon-seconds the patch had been used for prior to the current patch visit | No | No | Experiment only | Experiment only |
| Species | No | No | Natural only | Natural only | |
| Food handling time | Binary high/low (0/1). See Methods text | No | No | Natural only | Natural only |
| Number of occupants | No | No | Yes | Yes | |
| Number of occupants2 | No | No | Yes | Yes | |
| Residency time in previous patch4 | No | No | No | Yes | |
| Habitat | |||||
| Mean patch quality | Natural: Mean number of food items per patch; Experiment: Mean density of food across total feeding area (g/m2) | All models | |||
| Patch density | Natural: Number of patches per km2; Experiment: Inter-patch distance (m) | All models | |||
Random intercepts included in the models:
Focal animal ID, nested in troop ID;
Natural intake and patch residency time models: focal follow number, nested in focal animal ID nested in troop ID;
Experiment intake and patch residency time models: Focal animal ID, experimental patch ID and experiment day all cross-classified, and nested in troop ID.
Residency time in the previous patch was included in our models on patch residency time as a previous study at this site showed this can be an important predictor (Marshall et al. 2013).