ELECTRONIC SUPPLEMENTARY MATERIAL

Bonobos use call combinations to facilitate inter-party travel recruitment

Authors: Isaac Schamberg1*, Dorothy L. Cheney2, Zanna Clay3, Gottfried Hohmann4, Robert M. Seyfarth1

Affiliations:

1 Department of Psychology, University of Pennsylvania, Philadadelphia, PA.

2 Department of Biology, University of Pennsylvania, Philadadelphia, PA.

3 University of Birmingham, Birmingham, UK

4 Max Planck Instutite for Evolutionary Anthropology, Leipzig, Germany.

*Correspondence to:

Supplementary materials

Details of statistical models

All models were generalized linear mixed model with a logistic link function. Below we provide the variables used to create each model.

1. To test whether LHs+HHs were more likely than HHs alone to be produced during travel, we created a model with the following variables:

Fixed effect: call type (HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Caller behavior (travelling/not travelling)

2. To test whether LHs+HHs were more likely to occur prior to the initiation of travel than HHs alone, we created a model with the following variables:

Fixed effect: Fixed effect: call type (HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Post-call behavior (travel/no travel)

3. To test whether LHs+HHs were significantly more likely than HHs alone to be associated with subsequent recruitments, we created a model with the following variables:

Fixed effect: Fixed effect: call type (HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Recruitiment (yes/no)

4. To test whether callers were not more likely to approach another party after producing LHs+HHs than after producing HHs alone, we created a model with the following variables:

Fixed effect: Fixed effect: call type (HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Approach (yes/no)

5. To test whether callers who produced LHs+HHswere more likely to be on the periphery of their party than callers who produced HHs, we created a model with the following variables:

Fixed effect: Fixed effect: call type (HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Caller location (center/periphery)

6. To test whether W+HHs were more likely to result in approaches than were LHs+HHs we created a model with the following variables:

Fixed effect: Fixed effect: call type (W+HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Approach (yes/no)

7. To test whether LHs+HHs were more likely than W+HHs to result in the recruitment of others, we created a model with the following variables:

Fixed effect: Fixed effect: call type (W+HHs/LHs+HHs)

Random effect: Caller ID

Outcome variable: Recruitment (yes/no)

8. To test whether the association between vocal exchanges and subsequent recruitments was not significant, we created a model with the following variables:

Fixed effect: Fixed effect: vocal exchange(yes/no)

Random effect: Caller ID

Outcome variable: Recruitment (yes/no)

Supplementary tables

Table S1

Adults / Infants and juveniles / Total
Female / 14 / 12 / 26
Male / 7 / 6 / 13
Total / 21 / 18 / 39

Table S1. Demographic information about the Bompusa community of bonobos during this study.