Sumatran orangutan mothers differ in the extent and trajectory of their expression of maternal behaviour
T RevatheiD, Roger MundryiD, Sri Suci Utami-AtmokoiD, Tazkia Umaira Aprilla, Maria A. van NoordwijkiD, Marlen FröhlichiD, Paul‐Christian BürkneriD, Caroline SchuppliiD
Abstract
Mothers play a crucial role in the early development and survival of mammalian offspring, and differences in maternal care may differentially affect offsprings development. Whereas previous research has primarily focused on biological and socioecological factors to understand population-level variation in maternal behaviour, the individual as a source of variation remains understudied. We investigated between-individual variation in the average expression of, and plasticity in, six maternal behaviours in Sumatran orangutans, using 15 years of behavioural data. We found that mothers differed substantially in the average expression of four maternal behaviours, even after controlling for socioecological conditions, biological state characteristics, and the offsprings influence on these behaviours. Furthermore, not controlling for these confounding effects exaggerated or masked between-individual variation. Mothers also substantially differed in how they adjusted three of the maternal behaviours during offspring development, meaning that mothers differed in behavioural plasticity. Our results suggest that Sumatran orangutan mothers are constrained in the average expression of maternal behaviours and their plastic responses, potentially resulting in consistent differences among mothers, otherwise called maternal personality. Our findings highlight that individual variation around the population mean in maternal behaviour is more than noise and presents opportunities to study novel evolutionary processes that shape maternal behaviour.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 5
- FWCI
- 4.02
- Field percentile
- top 6%
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (31 Mar 2025) · CC BY 4.0
After controlling for confounders, 15 wild Sumatran orangutan mothers differed consistently in average expression of four of six maternal behaviours and in age-related plasticity for three.
Research on primate maternal behaviour has emphasized biological and socioecological drivers, leaving individual mothers understudied as a source of variance, and earlier estimates rarely controlled for all known confounders. Applying a behavioural reaction norm mixed-model framework to long-term wild orangutan data, this work separates mother, offspring, and within-individual variation and shows how omitting confounders distorts between-mother estimates.
Key results
ShownMothers differed consistently, across their own offspring, in average expression of carrying, feeding in close proximity, contact termination, and close proximity termination once confounders and offspring identity were modelled.
Mother random-intercept SD 0.32 to 0.9 across the four models; nested offspring-intercept SD 0.47 to 0.71; largest mother variation for contact termination and carrying; LOO-CV showed no predictive gain from adding mother effects for contact initiation or close proximity initiation. · Between-individual variation in the average…; Statistical support for random effects; Tables
ShownMothers differed in how carrying, feeding in close proximity, and close proximity termination shifted as offspring aged, yielding mother-specific trajectories; carrying showed the largest slope variation, while contact termination showed no substantial plasticity differences.
Mother random-slope SD for offspring age 0.16-0.69; nested offspring-identity slope SD 0.17-0.72; models retained when LOO-CV ΔELPD ± 2 × SE did not overlap zero. · Between-individual variation in maternal…; 4. Discussion; Statistical support for random effects; Tables
ShownOmitting the known confounders (offspring age and sex, food availability, association size, male presence) raised estimated between-mother variation for three behaviours, lowered it for the other three, and changed the uncertainty of the estimates.
Unadjusted models more than doubled the mother SD for body contact initiation and termination; shifts were moderate for feeding in close proximity and close proximity termination and negligible for carrying and close proximity initiation; 95% CIs were wider unadjusted for three behaviours, narrower for three. · Influence of confounding effects; Influence of confounding effects; 4. Discussion
SuggestedRestricting analyses to mothers with more than one offspring again recovered significant between-mother differences in both average level and age-related plasticity in carrying and in feeding in close proximity.
Subset analysis run only for behaviours with multi-offspring data on at least five mothers; six of the 15 mothers had more than one dependent offspring. · (b) Sample size; Between-individual variation in maternal…
SuggestedMother-level intercepts were largely uncorrelated across behaviours, implying distinct investment dimensions, except a negative link between carrying and feeding in close proximity, which the authors attribute to early carrying possibly speeding feeding skill acquisition.
Significant negative correlation between mother intercepts from the carrying and feeding-in-proximity models, none among other behaviour pairs; carrying peaks in the first two years of life and feeding in proximity at 5-6 years. · Between-individual variation in the average…; 4. Discussion
Methods, data and reproducibility
Bayesian generalized linear mixed models (brms/Stan) fit each behaviour's scan proportions with beta-binomial or beta errors, known confounders as fixed effects, random intercepts for mother and nested offspring identity, and random offspring-age slopes; leave-one-out cross-validation (LOO-CV) compared fixed-effects, random-intercept, and full models.
- Focal follows with two-minute scan sampling of wild mother-offspring pairs at Suaq Balimbing, Gunung Leuser National Park, Indonesia (2007-2022), six maternal behaviours · Sumatran orangutan (Pongo abelii) · n = 15 mothers; 622-868 follows, 5853.7-7915.1 h per behaviour
- Subset of mothers with more than one dependent offspring, reanalysed for carrying and feeding in close proximity · Sumatran orangutan (Pongo abelii) · n = six multi-offspring mothers overall; at least five per analysed behaviour
- Methods & tools
- brms (Bayesian GLMMs in R), Stan 2.32.5, R 4.2.2, loo 2.7.0 (leave-one-out cross-validation), bayes_R2 (Bayesian R^2 goodness of fit), Beta-binomial and beta GLMMs with logit link, Behavioural reaction norm framework, Focal follows with two-minute scan sampling
- Reagents & environment
- R 4.2.2; Stan 2.32.5 via brms; loo 2.7.0
- Reproducibility gaps
- Data and code are reported as deposited at Dryad, but no DOI or URL is given in the text; the only linked URL is the field data-collection protocol, not the dataset.
Limitations (6)
- Contact initiation and close proximity initiation are rarely performed by mothers, so data were mostly zeros and neither mother-level average nor plasticity variation could be resolved for them. · stated by the authors
- Only 15 focal mothers were available, and sample size and offspring-age coverage varied considerably among them; ideal data would span full offspring development with several offspring for every mother. · stated by the authors
- Unknown or unquantified biological and socioecological factors may still confound estimates; male presence was coded only as present or absent, although males probably vary in how much danger they represent for immatures. · stated by the authors
- Repeatability, the share of behavioural variance attributable to individuals, could not be computed because each model included several random slopes alongside nested intercepts. · stated by the authors
- The confounder comparison relied on visual inspection of posterior SDs and intervals without a formal test, and consistency across offspring rests on just six mothers with multiple offspring. · noted in this summary
- No offspring developmental or fitness outcomes were linked to maternal differences, so the adaptive value of the inferred maternal personality remains untested. · noted in this summary
Who should read:Read this if you study maternal care or animal personality in long-lived species and want a Bayesian reaction-norm template that handles uneven field sampling and confounders.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Maternal behavior in Sumatran orangutans (Pongo abelii) is modulated by mother-offspring characteristics and socioecological factors
T Revathe, Roger Mundry, … Caroline Schuppli·International Journal of PrimatologyPeer-reviewed·May 31, 2024
- Maternal ranging strategies facilitate offspring social play at energetic cost in the most solitary ape
Odd T. Jacobson, Alison M. Ashbury, … Maria A. van Noordwijk·Proceedings of the National Academy of SciencesPeer-reviewed·Oct 6, 2026
- Mothers letting go: postnatal maternal investment shapes sex-specific social development in wild vervet monkeys
Josefien A. Tankink, Nokubonga Philile Mirandah. Dlamini, Erica van de Waal·bioRxiv·Feb 17, 2026
- Eight years of social and individual learning synergistically shape orangutan diet profiles
T Revathe, Marie-Theres Weidling, … Caroline Schuppli·bioRxiv·Aug 27, 2025
- The development of social attention in orangutans: comparing peering behaviour in wild and zoo-housed individuals
Paulina Kukofka, Richard Young, … Caroline Schuppli·iSciencePeer-reviewed·Dec 6, 2024
- Maternal glucocorticoids and behavior shape offspring developmental trade-offs in wild baboons
Sam K. Patterson, Katie Hinde, … Joan B. Silk·bioRxiv·Nov 18, 2025