Introduction

Koshima Island, Japan is a storied fieldsite in the annals of primatology. Observations of wild macaques (Macaca fuscata) began there in 1948, but four years of sustained effort failed to habituate the monkeys. In August 1952, Itani and Tokuda (1954) resorted to scattering wheat grains and sweet potatoes on oft-used paths, gradually shifting the provisions to a sandy beach at Otomari Bay, a strategy that afforded a clear view of the entire group of 22 animals. Individual identifications followed quickly, laying the foundation for decades of influential research. In September 1953, a young female named Imo gathered a sweet potato from the beach and washed it in a freshwater stream, a behavioral innovation that spread horizontally to peers and then vertically to older kin (Kawai, 1965). By 1958, sweet potato washing had become a group-wide trait with a key modification: the monkeys began using sea water instead of standing freshwater, a preference that continues today seven generations later (Hirata et al., 2001).

These events have since passed into canon as an example of socially transmitted behavior, or culture, among nonhuman primates (McGrew, 1998; Matsuzawa and McGrew, 2008; Matsuzawa, 2015). But our preoccupation with culture has overshadowed fundamental questions of motivation (Sarabian and MacIntosh, 2015; Fiore et al., 2020). Two tacit assumptions—that sand produces an objectionable sensation on teeth, and that it is prudent to minimize tooth damage—are sufficiently intuitive that formal tests are wanting (Fannin et al., 2021). In consequence, the mineral and physical properties of contaminant sands are unknown, let alone the efficiency of different cleaning behaviors (Schofield et al., 2018). Another enigma concerns the preference of some monkeys to brush food on themselves (Kawai et al., 1992), a rapid but seemingly inferior means of sand removal. Food-brushing individuals have been characterised as inept (Kawai, 1965) or subordinate (Watanabe, 1994) in part because the quartz in sand can cause severe tooth damage (Lucas et al., 2013; Towle et al., 2022). Yet, carrying food to the ocean is expected to incur energetic costs as well as opportunity costs, factors that impelled usto explore the trade-offs of mitigating sand-mediated tooth wear.

Koram Island, Thailand

The long-tailed macaques (M. fascicularis) of Koram Island, Thailand use stone tools to harvest shellfish, a phenomenon that came to light during biodiversity damage assessments following the Sumatra-Andaman earthquake and tsunami of December 26, 2004 (Malaivijitnond et al., 2007; Gumert and Malaivijitnond, 2012; Tan et al., 2015). The monkeys are now a magnet for tourists, who provide them with market-sourced fruits (cucumbers, melon, pineapple), jettisoning them onto the beach. These events produce food surfaces with considerable concentrations of sand ( = 3.7 ± 1.3 mg mm^-2) and elicit food-washing and food-brushing behaviors among the monkeys. To understand the factors driving these reactions, we examined the mineral properties of foodadhering sands (n = 758 particles), finding that 78% of our sample was composed of crystalline quartz (Figure 1A).

Figure 1:

Harder than enamel, quartz can exact a heavy toll on teeth, but the probability and degree of enamel loss is governed partly by the size and shape of individual particles, factors that determine the ‘attack angle’ during particle-enamel contact (Lucas et al., 2013). We calculated the circularity of particles as a convenient proxy for sphericity (Grace and Ebneyamini, 2021), finding that it decreased as a function of particle size (Figure 1B). This result suggests that larger particles are more angular, posing a greater risk to enamel. However, we also calculated a median Feret diameter of 25.8 μm (range: 8.7 to 644 μm), meaning that nearly half the sample existed below the human threshold (25 μm) of oral detection (Imai et al., 1995). To put 25 pm into perspective, it is one-twelfth the diameter of the period ending this sentence.

Mitigating tooth wear

Chewing undetected quartz is expected to cause severe tooth wear, but this cost can be mitigated behaviorally if food-cleaning is proficient. To test this contention, we simulated the brushing and washing actions of monkeys with cucumber slices exposed to three concentrations of sand: low ( = 0.2 ± 0.2 mg mm^-2), intermediate ( = 0.9 ± 0.1 mg mm^-2), and high ( = 1.8 ± 0.9 mg mm^-2). We found that brushing was less efficient than washing across treatments, eliminating 76 ± 7% vs. 93 ± 4% of sand particles, respectively (Figure S1). It is a modest difference, perhaps, but it is freighted with fitness consequences when extrapolated over years of life (Fannin et al., 2022). It follows that monkeys should compulsively wash sand from food whenever the opportunity avails itself, a prediction that motivated a field experiment.

Results and Discussion

Our experiment (Figure 2A) was designed to test two concepts at once. The first pivots around intentionality, a thorny problem that emerged from studies of raccoons (Procyon lotor). Celebrated food-handlers, the submersion of food objects in water is better termed ‘dousing’ for greater haptic sensation, not washing with the intention of removing surface contaminants (Lyall-Watson, 1963). If the intent of monkeys is to eliminate sand, then the time devoted to brushing or washing food should vary as a positive function of sandiness. The other concept draws on observations from Koshima Island, which alluded to rank effects on individual cleaning behaviors, a pattern that is difficult to detect without controlling access to food or distance to the ocean.

Figure 2:

We conducted 101 feeding trials, recording 1,282 food-handling events by 42 individuals. We found that monkeys were sensitive to sand on their food, responding to each treatment—low, intermediate, and high concentrations—with greater median durations (± 1 SD) of brushing (low: 0.0 ± 0.1 s; intermediate: 1.1 ± 2.0 s; high: 3.1 ± 2.0 s; Figure 2B) and washing (low: 0.04 ± 0.3 s; intermediate: 0.6 ± 2.0 s; high: 3.3 ± 4.3 s; Figure 2C). This result is important for upholding long-held assumptions of intentional cleaning. Further, we found that dominant monkeys of both sexes showed a strong propensity for food-brushing (Figure 2B; Table S3) over food-washing (Figure 2C; Table S4). This finding reverses the pattern observed on Koshima Island (Watanabe, 1994), and it raises the possibility that food-washing is an indulgence subject to diminishing returns. To explore this premise, we developed a theoretical model where the time devoted to food-cleaning is predicted to maximise the rate of sand removal as a function of handling time.

Figure 3A illustrates the fastidious nature of our study population: monkeys allocated excess time to washing and brushing—by factors of 1.5 and 3.0, respectively—beyond that predicted by the optimization of sand removal. Our model also highlights sharply divergent responses to the sunk costs of food-handling time (Figure 3B). Given the efficacy of washing (Figure S1) and time needed to carry food to the ocean ( = 22 ± 15 s; range: 5 to 78 s), there is little incentive to over-wash food (Figure 3B, region I). At the same time, the lowest- and highest-ranking monkeys abstained from washing altogether, choosing instead to minimize food-handling time by overbrushing their food (Figure 3B, region II). This tolerance for fast-diminishing returns underscores the monkeys’ strong aversion to sand; but even so, the long-term benefits of mitigating tooth wear must be balanced against urgent energetic requirements.

Figure 3:

Disposable soma

The disposable-soma hypothesis of senescence predicts investment in the immediate needs of survival or reproduction over tooth preservation (Carranza et al., 2004). Dominant monkeys face this predicament because rapid food intake rates are integral to sustaining dominance and accruing reproductive success. For dominant males, energy intake determines their ability to sustain consortships during mating (Higham et al., 2011); and, for dominant females, it affects practically every measure of fitness (Alberts, 2019; Cooper et al., 2022). Our findings suggest that dominant monkeys refrained from washing to maximize short-term energy intake (Figure 2D). In short, they prioritized pressing energetic needs over the long-term benefits of tooth preservation—a ‘live fast, die young’ life-history strategy. This view of teeth as disposable soma may explain why dominant male monkeys experience faster senescence and earlier mortality (Anderson et al., 2021). Estimating fitness consequences is beyond the scope of our study, but our findings suggest that a prolonged life is also subject to diminishing returns.

Paleo matter(s)

The full extent of sand-mediated tooth wear is unknown for our study population, but it is probably extreme among the highest-ranked individuals. If affirmed, the findings could affect our views of the hominin fossil record by challenging the assumption that dietary variability is the principal cause of variable dental wear. Some species, notably Paranthropus boisei, had ready access to water, which raises the possibility that they—like many primate species—assiduously washed their food, an essential behavior if their diet featured gritty underground plant tissues (Wrangham et al., 2009; Fannin et al., 2021). Other species, notably P. robustus, have extremely variable levels of tooth pitting (Peterson et al., 2018), which could reflect, at least partially, the absence of extensive wetlands (Herries et al., 2010) coupled with interindividual variation in food-cleaning behaviors. Tellingly, the dental wear observed on the macaques of Koshima Island bears striking similarities to the hominin fossil record (Towle et al., 2022), suggesting that populations of food-cleaning monkeys are a valuable model system that warrant further study.

Significance

Our study leverages a new method in ecological research to provide the first analysis of siliceous particles on primate foods. Our experiment probes the behavioral economics of wild monkeys, revealing a strong aversion to sandy foods. Yet, the monkeys behaved irrationally when cleaning their foods, allocating excess time than predicted by an optimization model. Some individuals fell victim to the sunk cost fallacy by over-washing their foods, whereas dominant monkeys abstained from washing altogether, seemingly sacrificing their teeth at the altar of high rank, a social status that relies rapid food intake. Our results support the disposable-soma hypothesis for senescence and kick the tires of a treasured assumption in paleoanthropology.

Methods and Materials

Study site and population

Koram Island (12.242°, 100.009°) lies ~1 km offshore in the Gulf of Thailand and within Khao Sam Roi Yot National Park, Prachuap Khiri Khan, Thailand. It has an area of 0.45 km² and a coastline of 3.5 km. The habitat—limestone karst blanketed with a dense flora of dwarf evergreen trees and deciduous scrub, and encircled by rocky shore and sandy beaches—supports a population of ca. 75 long-tailed macaques described as hybrids at the subspecies taxonomic level (Macaca fascicularis aurea x M.f. fascicularis)(Gumert et al.,2019). The animals are well habituated to human observers due to regular tourism and sustained study since 2013 (Tan et al., 2018).

Rank determination

Macaques form multi-male multi-female (polygynandrous) social groups with individual dominance hierarchies. Among females, this hierarchy is strictly linear and stable through time (van Noordwijk and van Schaik, 1999). To determine the rank-order of adults, we recorded dyadic agonistic interactions and their outcomes (i.e., aggression, supplants, and silent-bared-teeth displays of submission) during 5-min focal follows of individuals based on a randomized order of continuous rotation (Tan et al., 2018). In some cases, these data were supplemented with ad libitum observations. This protocol existed during five years (2013-2018) of continual observations before we conducted our experiment in July-August 2018. To determine the effects of dominance rank on individual foodcleaning propensities, and to standardize the data by sex, we followed the methods of Levy et al. (2020). We calculated ordinal ranks between 1 (highest) and n (lowest), where n is the number of animals aged ≥ 5 years in each hierarchy. We analyzed male (n = 8) and female (n = 16) dominance hierarchies together in the same statistical models. We chose ordinal ranks because they best approximate competitive regimes during density-dependent competition (Levy et al., 2020). This type of feeding competition reflects those of our experimental trials and the natural conditions on Koram Island, where preferred resources are limited (Luncz et al., 2017).

Measuring sand

To quantify the amount of sand on food surfaces, whether provisioned by tourists (cucumbers, melon, pineapple) or used in our experiment (sliced cucumbers), we applied a quick-drying liquid polymer—granulated plastic (Pioloform BL 16; Wacker-Chemie GMBH, Munich, Germany) mixed with ethanol (18% plastic to 82% ethanol, by weight)—to each food item. When dried, we peeled and stored the sand-infused film for analysis. The advantages of this method are twofold: the removal of exogenous particulate matter is extremely thorough; and the plastic does not detach biogenic silica such as trichromes or phytoliths (Hinton et al., 1996). In the lab, we dissolved each peel in ethanol and separated the sand by centrifugation, producing a pellet. We dried and weighed the pellet, dividing the mass by the surface area of the food object, which we calculated from digital photographs imported into ImageJ v. 1.52. This method produces an estimate of exogenous particulate mass per area (mg mm^-2), allowing direct comparison of apples and oranges.

To measure the elemental and physical properties of sand, we dispersed and filtered the pellets in water using a 0.2-μm isopore membrane filter, which we submitted for scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS). To establish the parameters for multi-field and bulk analysis, we imaged a representative area of the filter at multiple magnifications and performed discrete particle analysis. 50× magnification allowed for statistically significant representation of particle number and size range (allowing a 5 μm lower particle size range in analysis). All discrete particle analyses indicated silicon rich particles and composition distribution bins were established to include dominant accompanying elements. After establishing these parameters, we initiated multi-field automated analysis using six fields of view of the debris field (at 50×). A composition classification was assigned to each particle and data sorted by composition classification and particle size (particle sizing was binned using standard Feret maximum parameter). The sizing bins are standard ISO-16232 size classes.

Experimental design

To elicit food handling behaviors and determine individual cleaning preferences, we put three cucumber slices in each of three trays (20×30×10 cm) and manipulated the amount of contaminating sand. In the low-sand treatment, we put cucumber slices in a tray without sand; however, the presence of some aeolian sand was unavoidable. In the intermediate-sand treatment, we lined the tray with sand and put the cucumber slices on the surface. In the high-sand treatment, we buried cucumber slices completely (Figure 2A). Trays were placed 1.5 m apart and 15 m from the ocean, and we randomized the color and sequence of trays across trials.

Trials began when one or more monkeys approached the trays and ended when the animals finished every cucumber slice or abandoned the experiment (range: 10 s to 14 min). We used video recordings to determine the onset and offset of individual food-handling bouts, beginning from initial contact with a cucumber slice and ending when the final slice, in its entirety, entered the mouth. Within each bout, we determined the duration of brushing and washing behaviors, defining each from the onset of serial stereotypical forelimb movements to the moment of oral ingestion. We estimated energy intake rates by calculating the number of cucumber slices consumed during each food-handling bout, and multiplying each slice by 1.1 kcal (source: U.S. Department of Agriculture, FoodData Central, 2019) and 4.186 kJ (Hargrove, 2007). We performed 101 trials over 5 weeks and recorded 1,282 food-handling bouts by 42 individual monkeys. To minimize the potential confounding effects of dominance interactions, we analyzed trials with ≤ 3 monkeys. Thus, 935 food-handling bouts were included in GLMM statistical models, which included data on individual rank, sex, and sand treatment. Ifa monkey consumed a cucumber slice without brushing or washing it, the zero-second duration was included in both GLMMs.

Behavioral analyses

To model experimental variance in brushing and washing behaviors as a function of experimental treatment, sex, and rank, we fit generalized linear mixed models (GLMMs) using the glmmTMB package in R version 4.2.3 (Brooks et al., 2017). In each GLMM, we modelled sand treatment (a categorical variable with three levels), sex (a categorical variable with two levels), and ordinal rank (a discrete variable ranging from 1 to 18) as fixed effects. We also incorporated two additional interaction terms as fixed effects: sand treatment χ ordinal rank and sand treatment χ sex. To account for experimental variance among individuals and control for pseudoreplication (because the number of feeding bouts per individual varied widely; Bolker et al. (2009)), we included individual ID as a random intercept. The brushing and washing data sets were whole-number counts (seconds) with means < 5. The distributions were right-skewed with high concentrations of biologically-meaningful zeros (Martin et al., 2005) (i.e., instances of food-handling without any cleaning behavior). Thus, we fit four separate models in glmmTMB to account for these non-normal spreads: standard GLMMs with a log-link function and either a (1) Poisson or (2) negative binomial error distribution (default = nbinom2 in glmmTMB); or zero-inflated generalised linear mixed models (ZIGLMM) with a logit-link function, a single zero-inflation parameter applying to all observations, and either a (3) Poisson or (4) negative binomial error distribution. We then determined the best fit model using delta AIC values in the bbmle package in R.

dAIC values indicated that, for both brushing (Table S1) and washing (Table S2), negative binomial GLMMs without zero-inflation were the best-fitting models. We validated each model by calculating dispersion statistics (χ² /degrees of freedom) (Zuur and leno, 2016). Dispersion statistics for the brushing model (χ² = 541.9; residual degrees of freedom = 564; χ² /rdf = 0.96, p = 0.74, one-sided test) and the washing model (χ² = 174.2; residual degrees of freedom = 351; χ² /rdf = 0.50, p = 1.00, one-sided test) failed to detect overdispersion in either case. We report the fixed effects tests for each GLMM in Tables S3 and S4 as Analysis of Deviance Tables (Type II Wald chi square tests, one-sided) along with χ² values, degrees of freedom, and p-values (one-sided tests). For all statistical analyses, α = 0.05.

Optimal cleaning time model

To model the optimization of sand removal before consumption, we accounted for two distinct temporal periods: handling time h, which includes an assessment time and pre-cleaning time, and the cleaning time t. Assessment time (set as a constant 1 second) includes visual fixation on a food object and forelimb extension before contact, whereas pre-cleaning time represents all handling activities that precede cleaning. During brushing, the pre-cleaning time was essentially nil (zero seconds), but washing required travel from the experimental treatments to the ocean, requiring longer pre-cleaning times ( = 22 ± 15 s; range: 5 to 78 s). We assumed that the proportion of sand removed from each cucumber follows the saturating relationship g(t) = t/(c +1), where c is the halfsaturation constant associated with brushing or washing. As c increases, so does the inefficiency of a given cleaning behaviour. Given our observations that brushing removes 75% of grit in 2.97 s, and washing removes 93% of grit in 3.53 s (Figure S1), we obtain the constants c_brushing = 0.99 s and c_washing=0.26 s, such that washing (without considering handling costs) is the most efficient strategy. The rate of grit removal is then given by R(t) = g(t)/(h + t), which reaches a maximum at the optimal cleaning time . For brushing and washing cleaning strategies, we obtain the expected optimal cleaning times t*_brushing = 0.98 s, and t*_washing = 2.39 s (Figure 3a), respectively. These optimal cleaning times are defined exclusively with respect to maximising the rate of grit removal, without considering the potentially cascading effects of these strategies on fitness.

Acknowledgements

We are extremely grateful for the guidance and practical assistance of C. Hobaiter, J. Hua, L. Kaufman, W.C. McGrew, D. Pornsumrit, and Z.M. Thayer. This research was approved by the Institutional Animal Care and Use Committee of Dartmouth College (protocol no. 00002099), the National Research Council of Thailand (permit nos. 0002/3740 and 0002/3742), and the Department of National Parks, Wildlife and Plant Conservation of Thailand. Funding was received from the National Science Foundation (BCS-SBE 1829315 to N.J.D.; GRFP 1840344 to L.D.F.) and Dartmouth College, including awards to J.E.R. (Claire Garber Goodman Fund; Mark A. Hansen Undergraduate Research, Scholarship, and Creativity Fund; Student Experiential Learning Fund) and N.J.D. (Scholarly Innovation and Advancement Award).

Additional Declarations:

The authors declare no competing interests.

Supplemental Materials

Fig S1.

Fig S2.

Table S1

Table S2

Table S3

Table S4