Supplementary Information: Long-sightedness in old wild bonobos during grooming

Transcription

1 Supplementary Information: Long-sightedness in old wild bonobos during grooming Heungjin Ryu, Kirsty E. Graham, Tetsuya Sakamaki, Takeshi Furuichi Figure S1. Measurement of ear size and grooming distance. (A) We measured ear length by counting the number of the pixels from the top to the bottom of an ear; a. (B) We counted the number of pixels for 10cm on a ruler, which was located at the same distance of the bonobo in (A); b. (C) This diagram shows the relative positions of the ear, the ruler, and the camera (photographer). Note that the distance α and β are almost identical (Supplemental Methods). (D) We measured the ear length c as described previously, and the grooming distance d by counting the number of pixels between fingertips and eyes, both of which we can then convert to centimeters.

2 Table S1. Coefficients of the linear (LMM) and generalized linear (GLMM) mixed effect models. The grooming distance increased exponentially with age (GLMM-1), and amplitude of accommodation (diopters) decreased linearly with age (LMM-1). Bonobos over 40-years-old had significantly a greater grooming distance than younger bonobos but there was no sex difference (LMM-2). Ear length did not vary in relation to age (LMM-3). All models for which significant effect of independent variables were found, were significantly better than null models. Interaction between age and sex is not indicated since none of the models showed any significant interaction for these factors. We confirmed that there were no violations of the assumptions of a linear model except LMM-1. In LMM-1, there is some concern over heteroscedasticity as the diopter (because of its definition) allows less variation with increasing focus distance in old individuals. However, we concluded that the LMM-1 was reliable since the linear relationship between age and diopters was solid and was not caused by any extreme values (Maximum Cook s distance was 0.7). models predictors estimates se df t p GLMM-1. log-linked grooming distance with age (Figure 1B) LMM-1. amplitude of accommodation (diopters) with age (Figure 1C) LMM-2. grooming distance between age classes LMM-3. ear length with age intercept <0.001 age <0.001*** male: female intercept <0.001 age <0.001*** male: female intercept <0.001 young: old <0.001*** male: female intercept <0.001 age male: female Note: df for GLMM-4 is not available in the package of lme4 because of issues on reliability. Sex was coded as male and female for all models. Age was coded as a number except in LMM-2, where age was coded as old (over 40) and young (under 40) categories.

3 Supplemental Methods and Data Analysis Study site and subjects Data were collected from January 2015 to April 2015 on 14 individuals in a free-ranging wild bonobo group (E1), at Wamba in the Luo Scientific Reserve, DR Congo (see [S1] for details). Among the subjects, two old males, TN (45-years-old) and TW (41-years-old), were identified in 1976 when they were relatively young: 6 and 2 years old, respectively. One old female, No (44-years-old), was identified in 1983 and the other old female Ki (41-years-old) was identified in 1984 when they had just immigrated into the group. We estimated No to be 12 years old and Ki to be 11 years old when they immigrated, based on the timing of their first birth. Therefore, we are confident that these four old individuals were over 40-years-old in 2015 and their estimated ages are close to their real ages. The final of our old individuals, DI, was first seen in 2004 and we identified him as in Here we took the average, considering him to be 32 years old in 2007, which would make him 40 years old during our study in The other subjects of the study were Yk, JD, LB, NB, Nv, Ot, Fk, JR, and KT. Please refer to [S2] for more information on these individuals. For comparison between humans and bonobos, we used Duane s work on accommodation of the human eye published in 1922 [S3]. We chose Duane s work since it was the first quantitative study which had a large data set of binocular accommodation from more than 500 human subjects. We converted Duane s measurement -the human diopter- to the distance in nearest focus (1m / diopter) and also converted the bonobo grooming distance to diopters (1m / grooming distance in meters). We assumed that bonobos groom at the nearest focus distance, although they can groom at a further distance. Bonobos likely groom at the nearest focus distance because it provides them with the shortest working distance to remove dirt or ectoparasites using their mouth, and visibility increases greatly at a shorter focus distance. Increasing visibility at the shorter focus distance seems particularly important, since ectoparasites are small in size - the average length of a louse egg and an adult louse found in wild chimpanzees in Mahale was 0.9mm and 3mm respectively [S4] - and grooming performance is probably related with their ability that how clearly they can see the ectoparasites. Based on this assumption, we decided to use grooming distance as a proxy for accommodation ability and converted it into diopters. There are some confounding factors that may influence grooming distance, such as refractive status of individuals and level of luminance at the location of grooming interactions. Since we could not examine the refractive status of wild bonobos, we assumed that all bonobos are emmetropic (6/6 m or 20/20 ft vision). One study, conducted on an individual in captivity, showed that the refractive status of chimpanzees is comparable to that of an emmetropic human eye [S5]. In addition, since we could not control the level of luminance, we only took pictures when bonobos groomed in an open space (usually formed by a fallen tree or human influence), which is where more than 1/3 of grooming interactions take place in the wild [S6]. Taking photographs to measure ear length and grooming distance We used a Sony NEX-6 interchangeable-lens mirrorless camera with a Tamron mm lens for Canon and an adapter (Techart EOS-NEX). All photographs were taken in the raw format (ARW) by HR with help

4 from a research assistant. To determine the grooming distance, we had to set up a scale for photographs. Instead of putting a scale in each photograph, which would have been too disruptive to the bonobos, we used the groomer s ear length as a scale to measure grooming distance. To measure each individual s ear length, we took photographs of the side of their face (ca. 90º between the camera and the direction of the individual s sight; figure S1A) at a minimum of 5m distance in order not to disturb the bonobos. We first manually focused on the subject s ear and took photographs with the allowed minimum f-number of the lens (ranged from depending on the focal length). Second, we maintained the manual focus from the first step (the focus ring was fixed from the first step), and took photographs of a ruler (Lifelex 3.5m LFX ; figure S1B) which was held by an assistant who stood at the same distance from the camera as the camera was to the subject s ear from the previous photo (figure S1C). Again, we did not change the focus and zoom ring on the lens. The camera s manual focus tools, including focus peaking and x9.6 digital magnification, allowed us to keep the bonobo and the ruler at almost the same distance from the camera without disturbing the bonobos. To measure grooming distance, we took photographs when a bonobo was grooming another individual, not when it groomed itself (figure S1D). We positioned the camera to create a ca. 90º angle between the camera and the straight line from the subject s fingertips to their face. We only took photographs when bonobos were on the ground or on a fallen tree where the height from the ground was less than 1.5m to avoid possible errors. Selecting and processing the photographs and measurement We only chose photographs with perpendicular head views of the subjects to reduce error. We corrected distortions of all selected photographs using the lens profile preset in Adobe Photoshop CC. We also increased the brightness of images whenever necessary for better recognition of the edges of ears, eyes, and fingertips. We then combined a photograph of the subject s ear (figure S1A) and a photograph of ruler (figure S1B), which were taken at the same distance to measure ear length. We first counted the number of pixels from the bottom to the top of the ear ( a in figure S1A), and then counted the number of pixels from 1cm to 11cm (10cm in total) of the ruler ( b in figure S1B), both using measurement tool in the Photoshop program. We counted the number of pixels 5 times per photograph and discarded the biggest and smallest values to calculate the mean number of pixels to minimize errors. We then used the following proportional equation (Equation 1) to calculate the ear length: Equation 1. x (ear length in cm): a (mean pixels) = 10cm: b (mean pixels) After we determined each individual s ear length, we calculated the grooming distance (Equation 2). We first counted the number of pixels for the ear of the groomer ( c figure S1D) and then counted the number of pixels from their fingertips to their eye ( d figure S1D). We also measured the number of pixels 5 times and

5 discarded the minimum and maximum values and used the mean of the three measurements as we did for the ear length measurement. Equation 2. x (ear length in cm): c (mean pixels) = y (grooming distance in cm): d (mean pixels) Statistical analysis All tests were done in R Revised [S7] with R studio [S8] with packages including lme4 [S9], lmertest [S10], and lmtest [S11]. To test whether or not bonobos increased their grooming distance with age and sex, we used linear models with mixed effects. These linear mixed models accounted for within individual variations as a random effect. The dependent variable for each model was grooming distance and ear length which were both coded in centimeters. The independent variables were age and sex (male vs female) or age category (old; over 40 vs young; under 40) and sex, respectively (Table S1). We included sex as an independent variable to examine potential sex-dependent aging differences. We also checked interactions between sex and age in all of the models by adding an interaction term in the model. After we ran the model, we checked the residual distributions and used visual inspections and the Shapiro test on the residuals to confirm whether or not there were any violations of the assumptions of linear regression. We also compared the models (with all dependent variables) with a null model (only with the random variables) using a log likelihood test, when we found significant effects (p<0.05) of the dependent variables in the models. We could not conduct statistical comparison between our data and the human data since the human paper only reported the average of diopters for each age. Research Ethics We adhered to the guidelines for the care and use of nonhuman primates from Primate Research Institute of Kyoto University. The methodology we used here was permitted by the Wamba committee for Bonobo Research and Ministry of Scientific and Technological Research of the D.R. Congo (MIN.ESURS/SG- RST/026/2014). Author contributions Conceptualization, H.R., K.E.G., T.S., and T. F.; Methodology, H.R.; Formal Analysis, H.R.; Investigation, H.R. and K.E.G.; Writing Original Draft: H.R.; Writing Review & Editing, H.R., K.E.G., T.S., and T.F; Project Administration, T.S. and T.F.; Funding Acquisition, T.F. Supplemental References S1. Furuichi, T., Idani, G., Ihobe, H., Hashimoto, C., Tashiro, Y., Sakamaki, T., Mulavwa, M.N., Yangozene, K., and Kuroda, S. (2012). Long-term studies on wild bonobos at Wamba, Luo scientific reserve, D. R.

6 Congo: Towards the understanding of female life history in a male-philopatric species. In Long-Term Field Studies of Primates, P. M. Kappeler and D. P. Watts, eds. (Berlin, Heidelberg: Springer Berlin Heidelberg), pp S2. Sakamaki, T., Behncke, I., Laporte, M., Mulavwa, M., Ryu, H., Takemoto, H., Tokuyama, N., Yamamoto, S., and Furuich, T. (2015). Intergroup transfer of females and social relationships between immigrants and residents in bonobo (Pan paniscus) societies. In Dispersing primate females: life history and social strategies in male-philopatric species Primatology Monographs., T. Furuich, J. Yamagiwa, and F. Aureli, eds. (Tokyo, Japan: Springer Japan), pp S3. Duane, A. (1922). Studies in monocular and binocular accommodation, with their clinical applications. Trans. Am. Ophthalmol. Soc. 20, S4. Zamma, K., and Nakamura, M. (2015). Grooming: its hygienic and social aspects. In Mahale Chimpanzees: 50 years of research, M. Nakamura, K. Hosaka, N. Itoh, and K. Zamma, eds. (Cambridge: Cambridge University Press), pp S5. Matsuzawa, T. (1990). Form perception and visual acuity in a chimpanzee. Folia Primatol. (Basel) 55, S6. Sakamaki, T. (2013). Social grooming among wild bonobos (Pan paniscus) at Wamba in the Luo Scientific Reserve, DR Congo, with special reference to the formation of grooming gatherings. Primates 54, S7. R Core Team (2016). R: A language and environment for statistical computing. (Vienna, Austria: R Foundation for Statistical Computing). S8. RStudio Team (2015). RStudio: Integrated development environment for R. (Boston, MA: RStudio, Inc.). S9. Bates, D., Maechler, M., Bolker, B., and Walker, S. (2015). Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, S10. Kuznetsova, A., Brockhoff, P.B., and Christensen, R.H.B. (2016). lmertest: tests in linear mixed effects models. S11. Zeileis, A., and Hothorn, T. (2002). Diagnostic checking in regression relationships. R News 2, 7 10.