Orang-utans (Pongo spp.) are primarily frugivorous (Morrogh-Bernard et al. Reference MORROGH-BERNARD, HUSSON, KNOTT, WICH, VAN SCHAIK, VAN NOORDWIJK, LACKMAN-ANCRENAZ, MARSHALL, KANAMORI, KUZE, SAKONG, Wich, Utami Atmoko, Mitra Setia and van Schaik2009) and are often regarded as important seed dispersers (Corlett Reference CORLETT1998). In Tanjung Puting, Borneo, Galdikas (Reference GALDIKAS and de Boer1982) found intact seeds in 94% of faecal samples, with a median 111 seeds per defecation; and in Ketambe, Sumatra, Rijksen (Reference RIJKSEN1978) found seeds in 44% of faecal samples. Furthermore, orang-utans have large day ranges (e.g. mean = 968 m, range = 280–2834 m across adults in Sabangau; Harrison Reference HARRISON2009) and slow passage rates of digesta through the gut (Caton et al. Reference CATON, HUME, HILL and HARPER1999), and, hence, may disperse seeds far from parent trees. Many seeds are also spat out or discarded at distances up to 75 m from parent trees (Galdikas Reference GALDIKAS and de Boer1982).

Until now, however, a critical piece of evidence has been missing from claims regarding the importance of orang-utans as seed dispersers: that showing successful germination of seeds following passage through the gut. In apes, this has been demonstrated for the chimpanzee (Pan troglodytes troglodytes) and gorilla (Gorilla gorilla gorilla) by Poulsen et al. (Reference POULSEN, CLARK and SMITH2001), who report that germination was achieved for the seeds of 8/10 and 4/4 defecated species, respectively. Similar results are reported elsewhere for the chimpanzee (Wrangham et al. Reference WRANGHAM, CHAPMAN and CHAPMAN1994, but see Gross-Camp & Kaplin Reference GROSS-CAMP and KAPLIN2005), bonobo (Pan paniscus, Idani Reference IDANI1986) and Bornean hybrid gibbon (Hylobates muelleri × albibarbis) (McConkey Reference MCCONKEY2000). We therefore initiated a short-term study on the orang-utan, in which we hypothesized that, following passage through the orang-utan gut: (1) seeds would remain viable; (2) germination success would increase (or at least not decrease); and (3) length of time to germination would decrease (or at least not increase) relative to control seeds collected by the authors, which were of the same ripeness stage as those ingested by the orang-utans.

This research was conducted on a population of wild Bornean orang-utan (Pongo pygmaeus wurmbii Tiedemann), as part of the OuTrop-CIMTROP multi-disciplinary research project in the Natural Laboratory of Peat-Swamp Forest, Sabangau, Central Kalimantan, Indonesia. Data on diet composition of habituated individuals through focal-animal follows have been collected at this site since 2003, using standard protocols (Harrison Reference HARRISON2009). The data presented in this paper were collected during a 6-wk period in October and November 2008. Germination trial methods match closely those used by previous ape researchers (McConkey Reference MCCONKEY2000, Poulsen et al. Reference POULSEN, CLARK and SMITH2001).

Three different categories of seeds were collected: (1) excreted, in which faeces excreted during daily follows were collected, carried to camp and washed through a sieve, before extracting any seeds present; (2) seeds spat out by orang-utans while feeding, which were cleaned with tissue paper to remove remaining pulp; and (3) control seeds, obtained directly from fruits of equal ripeness as those eaten by the orang-utans from either the parent tree or another tree. Although seeds from four species (Ebenaceae: Diospyros bantamensis Valeton, Elaeocarpceae: Elaeocarpus mastersii King, Gnetaceae: Gnetum sp. and Tetrameristaceae: Tetramerista glabra Miq.), plus unidentified Ficus spp. (Moraceae) were collected, adequate sample size and comparable control samples were only obtained for E. mastersii (minimum 36 seeds for each category-substrate combination), which is both spat out and swallowed frequently. Results for the other species are available upon request.

All seeds present were identified, numbered and laid out on germination trays using two substrates: (1) peat soil or (2) tissue paper. Elaeocarpus mastersii (seed size: 10 × 4 × 4 mm) was tested on both peat and tissue paper to determine the most suitable treatment for achieving germination. All planting was conducted in controlled conditions in the Sabangau Seedling Nursery. Seed trays were moved every 24 h to expose them to equal amounts of light and shade in different parts of the nursery. Seeds were kept moist and checked for germination (i.e. emergence of the radicle) daily until germination occurred.

A total 28 faecal samples were collected from eight orang-utans, of which 20 (71.4%) contained at least one intact seed. Orang-utans were observed feeding on fruits of 11 different species during the study period, but only the aforementioned four species and figs were found in the faeces. The mean (± SD) number of seeds per faecal sample was 123 ± 188 including figs, and 110 ± 168 excluding figs; the mean number of species per sample was 1.10 ± 0.96 (range = 0–4).

Germination was achieved in at least one category (excreted, spat-out or control seeds) for three of the five taxa studied (D. bantamensis, E. mastersii and Ficus spp.), all of which exhibited germination of excreted seeds. For E. mastersii, significant differences in germination success were found between categories using both the peat (control 86.1% seeds germinated, spat out 97.6% and excreted 25.6%; Kruskal–Wallis, H = 86.1, df = 2, P < 0.001) and tissue substrates (control 86.0%, spat out 98.0% and excreted 48.0%; H = 38.1, df = 2, P < 0.001). A Dunn's post hoc test (α = 0.05) revealed that, for both the peat and tissue paper substrates, germination success was lower for excreted seeds than for spat-out or control seeds.

Significant differences in length of time until germination between categories existed for E. mastersii, on both the peat (mean control = 21.0 d, spat out = 18.2 d and excreted = 14.1 d; H = 33.4, df = 2, P < 0.001) and tissue (control = 22.7 d, spat out = 20.0 d and excreted = 18.9 d; H = 10.4, df = 2, P = 0.006) substrates. Dunn's post hoc tests revealed that excreted and spat-out seeds germinated faster than controls when grown on tissue, and that all three categories differed when grown on peat.

Our results demonstrate that seeds can remain viable following passage through an orang-utan gut, supporting our first hypothesis. This provides the first confirmation of the defecation of viable seeds by orang-utans, matching findings for other ape species (McConkey Reference MCCONKEY2000, Poulsen et al. Reference POULSEN, CLARK and SMITH2001, Wrangham et al. Reference WRANGHAM, CHAPMAN and CHAPMAN1994). It is unknown how many species orang-utans disperse in Sabangau, but orang-utans at this site are known to consume fruits of at least 118 species (Harrison Reference HARRISON2009) and, hence, the total number of dispersed species is likely to be high. Combined with observations of consistent prevalence of intact seeds in orang-utan faeces (Galdikas Reference GALDIKAS and de Boer1982, Rijksen Reference RIJKSEN1978, this study), the large number of species ingested (Russon et al. Reference RUSSON, WICH, ANCRENAZ, KANAMORI, KNOTT, KUZE, MORROGH-BERNARD, PRATJE, RAMLEE, RODMAN, SAWANG, SIDIYASA, SINGLETON, VAN SCHAIK, Wich, Utami Atmoko, Mitra Setia and van Schaik2009), large day range (Harrison Reference HARRISON2009), slow digesta passage rates (Caton et al. Reference CATON, HUME, HILL and HARPER1999) and ability of orang-utans to handle even the largest fruits that other frugivores cannot manipulate (Corlett Reference CORLETT2009), this supports the belief that orang-utans are important seed dispersers in Bornean and Sumatran rain forests (Corlett Reference CORLETT1998, Galdikas Reference GALDIKAS and de Boer1982).