INTRODUCTION
Animal-dispersed plant species interact with a cohort of frugivores, whose roles regarding population dynamics depend on their specific characteristics. The number of seeds dispersed by an animal species is linked to the importance of fruits in its diet, as well as its food preferences (Schupp Reference SCHUPP1993). Both factors influence the frequency of consumption of a given plant species, and ultimately seed dispersal quantity (Schupp Reference SCHUPP1993). Seed processing in the mouth and gut might alter seed germination potential, and therefore hamper the dispersal services provided by a given frugivore (Schupp Reference SCHUPP1993). The action of granivorous species as predators, or scatter-hoarding dispersers, might also influence seed fate (Schupp Reference SCHUPP1993). As the disperser's characteristics influence seed dispersal effectiveness, fruit features determine which animals consume a plant species (Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985) and therefore its community of dispersers. It is important to study the role of mammals in the regeneration processes of animal-dispersed plant species in order to assess the impacts of the phenomenon of ‘defaunation’ (Galetti & Dirzo Reference GALETTI and DIRZO2013, Kurten Reference KURTEN2013). Large mammals are thought to be effective dispersers, but are also disproportionately affected by anthropogenic pressure (Linder & Oates Reference LINDER and OATES2011, Wright Reference WRIGHT2003). Central African great apes (western gorilla, Gorilla gorilla; chimpanzee, Pan troglodytes Blumenbach; and bonobo, Pan paniscus Schwartz) are a good example of such species. Great apes are important dispersers because of their highly frugivorous diet, large body mass and extended home range (Beaune et al. Reference BEAUNE, BRETAGNOLLE, BOLLACHE, BOURSON, HOHMANN and FRUTH2013a, Poulsen et al. Reference POULSEN, CLARK and SMITH2001, Tutin Reference TUTIN2001, Voysey et al. Reference VOYSEY, MCDONALD, ROGERS, TUTIN and PARNELL1999).
This study took place in a logged forest in Gabon. Among the dominant plant families found in Gabonese forests, species of the family Burseraceae are characterized by a high level of endemism (Aubréville Reference AUBRÉVILLE1962, Doucet Reference DOUCET2003). Those belonging to the genera Dacryodes, Canarium and Santiria, which provide important timber and/or non-timber forest products (Aubréville Reference AUBRÉVILLE1962, Doucet Reference DOUCET2003), are zoochorous and are consumed by the endangered great apes (Tutin & Fernandez Reference TUTIN and FERNANDEZ1993, Tutin et al. Reference TUTIN, WILLIAMSON, ROGERS and FERNANDEZ1991, Williamson et al. Reference WILLIAMSON, TUTIN, ROGERS and FERNANDEZ1990). But the main seed dispersers of these plant species are not known; neither is their exact influence on plant fitness.
The objective of this work was to describe the respective roles of vertebrates involved in the seed dispersal and predation of a poorly known zoochorous tree species endemic to Gabon, Dacryodes normandii Aubrév. & Pellegr. (Burseraceae), and to assess their importance in the dispersal process. We tested the following two hypotheses. (1) Considering that D. normandii fruit can be categorized in the bird-monkey syndrome (Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985), large birds and monkeys will be dispersers of its seeds. However, we expect large mammals to provide important dispersal services to this species. In particular, great apes will be the main dispersers of this species in terms of the quantity of dispersed seeds and the high quality of seed treatment in their mouth and gut. (2) Large rodents will be mainly involved in seed predation (Hecketsweiler Reference HECKETSWEILER1992), but considering their scatter-hoarding behaviour, they will occasionally disperse D. normandii seeds.
METHODOLOGY
Study site
This study was carried out in a sustainably managed and certified (Forest Stewardship Council) logging concession in south-east Gabon (0°30′–1°00′ S, 12°30′–14°00′ E) from December 2013 to March 2014. The study area covers 617000 ha and the vegetation mostly comprises moist tropical evergreen forest. It is an old secondary forest as characterized by White (Reference WHITE1986), dominated by Scyphocephalium ochocoa Ward (Myristicaceae) and Aucoumea klaineana Pierre (Burseraceae) (Demarquez & Jeanmart unpubl. data, White Reference WHITE1986). In similar forests in central Gabon, it has been shown that about 60% of tree species are dispersed through endozoochory (Doucet Reference DOUCET2003). Western lowland gorilla (Gorilla gorilla gorilla Savage & Wyman) and central chimpanzee (Pan troglodytes troglodytes Blumenbach) coexist sympatrically within the study site. The climate is equatorial with an average annual rainfall of 1700 mm (Moupela et al. Reference MOUPELA, DOUCET, DAÏNOU, TAGG, BOURLAND and VERMEULEN2013), with two seasons of heavy precipitation (March–May and September–December) and two seasons of lower precipitation. The average annual temperature is 26°C (Moupela et al. Reference MOUPELA, DOUCET, DAÏNOU, TAGG, BOURLAND and VERMEULEN2013), with the highest temperatures being recorded from February to April, and the lowest in July and August.
Study species
Dacryodes normandii is a medium to large canopy tree, endemic to the evergreen forests of Gabon (Doucet Reference DOUCET2003). It provides timber and non-timber forest products (Aubréville Reference AUBRÉVILLE1962). It has intermediate light requirements, as it tolerates shade at seedling stage but becomes light-demanding when mature (Doucet Reference DOUCET2003). Dacryodes normandii is dioecious and female trees produce obovoid drupes of 3.5 cm in length, black/purple in colour, with a thin exocarp and a sweet-sugary, juicy and soft pulp (Aubréville Reference AUBRÉVILLE1962). Fruiting is observed for trees from 13 cm dbh (Doucet Reference DOUCET2003). Fruit production is defined as supra-annual, occurring at intervals longer than 1 y (Fourrier Reference FOURRIER2013). Flowering takes place from September to November and fruits are produced from October to February (Aubréville Reference AUBRÉVILLE1962, Doucet Reference DOUCET2003, Fourrier Reference FOURRIER2013, Hecketsweiler Reference HECKETSWEILER1992). Fruits are available for 1.6 mo on average (Fourrier Reference FOURRIER2013). Within the study site and during the study period, unripe fruits were observed in December, and ripened from January to February (B. Haurez, pers. obs.). The average density of Dacryodes normandii in the study site is 54.8 stems km−2 (for dbh >20 cm) (Demarquez & Jeanmart unpubl. data).
Identification and contribution of vertebrate seed dispersers
Direct observations. In December 2013, six mature trees of Dacryodes normandii with the main part of the crown visible from the ground (hereafter referred to as focal trees), were visited and examined for signs of fruiting (mainly unripe fruits in the canopy and on the ground). In order to ensure independence of observations, focal trees were located at least 1 km apart (Babweteera & Brown Reference BABWETEERA and BROWN2009). Observations were made with binoculars between 06h30 and 17h30 in periods of availability of both unripe and ripe fruits (from mid-December to end of February). Observers were hidden in a position providing the best visibility of the focal tree canopy. For all vertebrate visits to focal trees the following information was recorded: animal species; feeding behaviour when eating fruits and/or seeds (Chapman & Chapman Reference CHAPMAN and CHAPMAN1996, Kitamura et al. Reference KITAMURA, SUZUKI, YUMOTO, POONSWAD, CHUAILUA, PLONGMAI, MARUHASHI, NOMA and SUCKASAM2006); time and duration of visits; and number of animals observed (Kitamura et al. Reference KITAMURA, SUZUKI, YUMOTO, POONSWAD, CHUAILUA, PLONGMAI, MARUHASHI, NOMA and SUCKASAM2006). Sampling effort was defined as the number of observation hours.
Camera trapping. Camera traps (GameSpy Moultrie M-80XT, approximate detection range 13.5 ± 1.5 m) were installed beneath 10 mature fruiting trees, located at least 1 km from each other (Babweteera & Brown Reference BABWETEERA and BROWN2009). A pile of fresh Dacryodes normandii fruits was placed in front of the camera to attract frugivores (adapted from Babweteera & Brown Reference BABWETEERA and BROWN2009, Kitamura et al. Reference KITAMURA, SUZUKI, YUMOTO, POONSWAD, CHUAILUA, PLONGMAI, MARUHASHI, NOMA and SUCKASAM2006, Moupela et al. Reference MOUPELA, DOUCET, DAÏNOU, TAGG, BOURLAND and VERMEULEN2013, Seufert et al. Reference SEUFERT, LINDEN and FISCHER2009). Cameras were monitored weekly from December 2013 to March 2014. They were automatically activated any time a movement occurred within the range of the detector. An independent visit was defined as (1) consecutive records of individuals of different species, (2) non-consecutive records of individuals of the same species, and (3) consecutive records of individuals of the same species taken more than 30 min apart (O'Brien et al. Reference O'BRIEN, KINNAIRD and WIBISONO2003). For all independent visits, the following data were recorded: animal species, feeding behaviour when applicable, time and duration of visits, and the number of animals detected (Chapman & Chapman Reference CHAPMAN and CHAPMAN1996, Kitamura et al. Reference KITAMURA, SUZUKI, YUMOTO, POONSWAD, CHUAILUA, PLONGMAI, MARUHASHI, NOMA and SUCKASAM2006). Sampling effort was defined as the number of trap days, i.e. the number of active cameras multiplied by the number of operating days (24 h).
Role of great apes and rodents. To determine the fate of fruits and seeds of Dacryodes normandii potentially removed by terrestrial rodents, rodent burrows were sought in the vicinity of four of the focal trees. Twenty-one burrows were found and excavated and seeds were counted and identified to the most precise taxonomic level. Their status was recorded (‘intact’ when the seed was undamaged, and ‘predated’ when the seed was broken or rotten and the embryo eaten).
Fresh gorilla faeces collected at nest sites were inspected to investigate the presence of Dacryodes normandii seeds during the fruiting period (January–February) in 2012 and 2014. A total of 49 faecal units were analysed. Seed viability after gut passage was assessed through germination trials under the same conditions in both years in a nursery at the study site (n = 53 gut passed seeds). A control treatment, involving seeds extracted from Dacryodes normandii fresh fruits and cleaned from pulp (n = 64 control seeds), was also realized. Seeds were sown individually in soil collected from the study site. The nursery was watered on a daily basis. Germination trials involving seeds collected in chimpanzee faeces were not performed because of the complexity of identifying chimpanzee faeces with certainty.
Data analysis
All photographed or directly observed mammal species were identified following Kingdon (Reference KINGDON1997) and birds using Serle & Morel (Reference SERLE and MOREL1993). On pictures where only parts of the animals’ bodies were visible, they were identified to the most precise taxonomic level possible or classified as ‘undetermined’. Additionally, to avoid incorrect identifications, all red duiker species (Cephalophus callipygus Peters, C. dorsalis Gray, C. nigrifrons Gray, C. leucogaster Gray, C. ogilbyi Waterhouse) were collectively classified as Cephalophus spp.
The duration of visits involving consumption was calculated as the time difference between arrival and departure of the animal directly observed, or the time between the first and last photographs of the visit. As the relative amount of fruit consumed, a component of dispersal effectiveness, is influenced by the frequency of visits and their duration (Schupp Reference SCHUPP1993), an index of fruit consumption, I FC was computed as follows (adapted from Moupela et al. Reference MOUPELA, DOUCET, DAÏNOU, TAGG, BOURLAND and VERMEULEN2013):
\begin{equation}
{I_{\rm FC}} = ({\rm N_{\rm obs.cons}}/{{\rm{N}}_{{\rm{obs.}}}})\,{\rm D}
\end{equation}
with Nobs.cons = number of observations of a given species involving consumption of fruits, Nobs = total number of observations of this species, and D = mean length of visit involving consumption of fruits. The ecological functions of different animal species were determined using the information provided by pictures and fruit remnants, as well as from existing literature. Animals were classified into four categories (adapted from Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985 and Moupela et al. Reference MOUPELA, DOUCET, DAÏNOU, TAGG, BOURLAND and VERMEULEN2013): (1) predator: those that destroy the seeds; (2) disperser: those that consume fruits while leaving seeds intact and removing them from the vicinity of the parent trees; and (3) neutral: those that show no interest in D. normandii fruits. All species displaying null I FC were considered neutral.
The effect of passage in gorilla gut on seed germination was evaluated through the comparison of germination success of gut-passed seeds with that of control seeds, with a χ² test.
RESULTS
Identification and contribution of seed dispersers
In 172 h of direct observations, three species of Dacryodes normandii disperser were identified: one arboreal primate, the putty-nosed monkey (Cercopithecus nictitans nictitans), and two large birds, the white-crested hornbill (Tropicranus albocristatus) and the great blue turaco (Corythaeola cristata) (Table 1).
Table 1. Animal species involved in consumption events during direct and indirect observations of Dacryodes normandii trees in a Gabonese logging concession from December 2013 to March 2014, ordered by IFC. Nobs.cons = number of observations of a given species involving consumption of fruits, Nobs = total number of observations of this species and D = mean length of visit involving consumption of fruits, I FC = (Nobs.cons/Nobs.) D.
In a total of 796 camera-trap days, 572 vertebrate detection events occurred (97.2% involving mammals, 1.7% birds and 1.1% undetermined). Twenty-six species or groups of species were identified from camera-trap pictures, accounting for 71.3% of detection events. The remaining 28.7% involved undetermined animals, mostly Muridae (24.0%). Three of the identified species acted predominantly as seed dispersers: two great apes, the western lowland gorilla and the central chimpanzee, and the African forest elephant (Loxodonta cyclotis, Proboscidea) (Table 1). Both great ape species frequently consumed D. normandii fruits and their consumption events were relatively long, resulting in the highest I FC observed (74 and 13 respectively). Elephants less frequently consumed D. normandii fruits and their visits were much shorter (I FC of 0.13).
The remaining species (n = 8) observed to interact with D. normandii fruits were likely to act mainly as predators. They were involved less frequently in consumption events and their visits were consistently shorter than 10 min (I FC = 0.86–0.0073) (Table 1).
Role of great apes and rodents. No D. normandii seeds were found in the 21 prospected rodent burrows (mean depth ± SD = 36 ± 10 cm, range = 20–70 cm). Two destroyed seeds of another species of Burseraceae (Santiria trimera (Oliv.) Aubrév.) were found (Appendix 1). These results support the evidence that terrestrial rodents are mainly predators of D. normandii and that dispersal events through scatter-hoarding are likely to be rare.
We observed that 18% of fresh gorilla faeces analysed (n = 49) contained D. normandii seeds (number of seeds = 170, mean number of seeds per faecal sample ± SD = 18.9 ± 14.4, range = 0–49), which exhibited a germination success of 68% (n = 53 seeds sown). The germination success displayed by seeds passed through the gorilla gut was not significantly different from that of seeds extracted from fresh fruits (n = 164 seeds sown, mean germination success = 73%) (df = 1, χ² = 0.74, P = 0.39).
DISCUSSION
Owing to direct and indirect observations, the present study identified six disperser species of Dacryodes normandii, among which primates were the most important. In particular, supported by I FC high values, great apes are likely to be the main dispersers of this species. Rodents, squirrels and duikers are mainly predators of D. normandii seeds.
Importance of great apes in seed dispersal of Dacryodes normandii
Among primate species, the western lowland gorilla is considered a very effective seed disperser (Haurez et al. Reference HAUREZ, PETRE and DOUCET2013, Petre et al. Reference PETRE, TAGG, HAUREZ, BEUDELS-JAMAR, HUYNEN and DOUCET2013, Tutin et al. Reference TUTIN, WILLIAMSON, ROGERS and FERNANDEZ1991) and an essential actor in forest dynamics and maintenance of forest structure (Tutin Reference TUTIN2001). Seeds dispersed by gorillas are generally viable (Petre et al. Reference PETRE, TAGG, BEUDELS-JAMAR, HAUREZ, SALAH, SPETSCHINSKY, WILLIE and DOUCET2015, Poulsen et al. Reference POULSEN, CLARK and SMITH2001, Voysey et al. Reference VOYSEY, MCDONALD, ROGERS, TUTIN and PARNELL1999). The high germination success displayed by seeds of D. normandii collected in gorilla faeces (68%), which does not differ from that of unpassed seeds, attests to their preservation during gut passage. In other sites, consumption of D. normandii fruits has already been reported (Fourrier Reference FOURRIER2013, Tutin & Fernandez Reference TUTIN and FERNANDEZ1993, Tutin et al. Reference TUTIN, WILLIAMSON, ROGERS and FERNANDEZ1991, Williamson et al. Reference WILLIAMSON, TUTIN, ROGERS and FERNANDEZ1990). Since gorillas exhibit (1) a high frequency of visiting fruiting D. normandii trees, (2) long visit durations in cases of consumption, and (3) the consumption of a large number of fruits at each visit, they may provide reliable dispersal services for this tree species. Furthermore, more than 50% of gorilla dung is deposited at nest sites (Todd et al. Reference TODD, KUEHL, CIPOLLETTA and WALSH2008) and gorillas preferentially select open habitats for nesting (Haurez et al. Reference HAUREZ, PETRE, VERMEULEN, TAGG and DOUCET2014, Petre et al. Reference PETRE, TAGG, BEUDELS-JAMAR, HAUREZ, SALAH, SPETSCHINSKY, WILLIE and DOUCET2015, Tutin et al. Reference TUTIN, PARNELL, WHITE and FERNANDEZ1995, Willie et al. Reference WILLIE, PETRE, TAGG and LENS2012), which offer suitable conditions for seed germination and seedling growth (Rogers et al. Reference ROGERS, VOYSEY, MCDONALD, PARNELL and TUTIN1998, Tutin et al. Reference TUTIN, WILLIAMSON, ROGERS and FERNANDEZ1991, Voysey et al. Reference VOYSEY, MCDONALD, ROGERS, TUTIN and PARNELL1999). Dacryodes normandii has intermediate light requirements (Doucet Reference DOUCET2003), therefore suggesting that its seeds and seedlings may benefit from the light conditions characteristic of gorilla nest sites (Rogers et al. Reference ROGERS, VOYSEY, MCDONALD, PARNELL and TUTIN1998, Tutin et al. Reference TUTIN, WILLIAMSON, ROGERS and FERNANDEZ1991, Voysey et al. Reference VOYSEY, MCDONALD, ROGERS, TUTIN and PARNELL1999). As a consequence, gorilla may provide directed dispersal for D. normandii (Howe & Smallwood Reference Howe and Smallwood1982), as has been observed for other plant species with different degrees of shade tolerance (from high tolerance to shade intolerance) (Petre et al. Reference PETRE, TAGG, BEUDELS-JAMAR, HAUREZ, SALAH, SPETSCHINSKY, WILLIE and DOUCET2015, Voysey et al. Reference VOYSEY, MCDONALD, ROGERS, TUTIN and PARNELL1999). However, the fate of gorilla-deposited seeds of D. normandii is still largely unknown. A study encompassing all stages from faecal deposition to sapling recruitment may help to reliably assess the likely importance of gorillas in the regeneration of D. normandii.
The chimpanzee is considered a high-quality disperser (Gross-Camp & Kaplin Reference GROSS-CAMP and KAPLIN2011), displaying seed swallowing, spitting and wadging behaviours, depending on the plant species (Chapman & Russo Reference CHAPMAN, RUSSO, Campbell, Fuentes, MacKinnon, Panger and Bearder2005, Gross-Camp & Kaplin Reference GROSS-CAMP and KAPLIN2005, Reference GROSS-CAMP and KAPLIN2011; Lambert Reference LAMBERT1999). Chimpanzees generally disperse viable seeds (Wrangham et al. Reference WRANGHAM, CHAPMAN and CHAPMAN1994), although their impact on germination potential depends on the taxon considered (Gross-Camp & Kaplin Reference GROSS-CAMP and KAPLIN2005). Dacryodes normandii pulp is a valuable food resource for the chimpanzee in Gabon, and its seeds are commonly observed in chimpanzee faeces (Tutin & Fernandez Reference TUTIN and FERNANDEZ1993). However, data on the disperser effectiveness of chimpanzee for this species are lacking.
Other dispersers and predators of Dacryodes normandii
Medium- and small-sized primates, for example Cercopithecus spp. (Gautier-Hion et al. Reference GAUTIER-HION, EMMONS and DUBOST1980, Lambert Reference LAMBERT1999), are important dispersers (Chapman & Russo Reference CHAPMAN, RUSSO, Campbell, Fuentes, MacKinnon, Panger and Bearder2005, Poulsen et al. Reference POULSEN, CLARK and SMITH2001), dispersing seeds through defecation and spitting (Chapman & Russo Reference CHAPMAN, RUSSO, Campbell, Fuentes, MacKinnon, Panger and Bearder2005). The present study highlights a role in D. normandii seed dispersal for the putty-nosed monkey (C. n. nictitans). This species is known to consume the fruit pulp of five species of Burseraceae that share fruit characteristics with D. normandii (Gautier-Hion et al. Reference GAUTIER-HION, EMMONS and DUBOST1980, Poulsen et al. Reference POULSEN, CLARK and SMITH2001, Reference POULSEN, CLARK, CONNOR and SMITH2002), and Santiria trimera seeds exhibit enhanced germination after dispersal by this monkey (Poulsen et al. Reference POULSEN, CLARK and SMITH2001).
The role of the African forest elephant in seed dispersal is well documented (Blake et al. Reference BLAKE, DEEM, MOSSIMBO, MAISELS and WALSH2009, Campos-Arceiz & Blake Reference CAMPOS-ARCEIZ and BLAKE2011). The elephant consumes D. normandii fruits, without destroying their seeds, as evidenced by the presence of whole seeds in dung (Feer Reference FEER1995a, Fourrier Reference FOURRIER2013), and can therefore be considered as a disperser of this species (Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985). However, considering their low IFC (0.13) and the fact that, based on camera-trap observations, they appear to only consume D. normandii opportunistically while moving, their role is likely to be of lesser importance.
Dacryodes normandii belongs to one of the three most important families for specialized avian frugivores (Snow Reference SNOW1981). The present study has suggested a dispersal role for the white-crested hornbill (Tropicranus albocristatus) and the great blue turaco (Corythaeola cristata), even though the overall number of observations of birds feeding on D. normandii was surprisingly low. Turacos and hornbills are endozoochorous dispersers, as passage through their gizzards does not destroy seeds (Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985).
Three squirrels (Funisciurus pyrropus, F. isabella and Epixerus ebii), two species of terrestrial rodents (Atherurus africanus and Cricetomys emini) and one undetermined Muridae were observed to handle D. normandii fruits. All these taxa are predominantly seed predators (Beaune et al. Reference BEAUNE, BRETAGNOLLE, BOLLACHE, HOHMANN, SURBECK and FRUTH2013b, Emmons Reference EMMONS1980, Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985), but may contribute to short-distance dispersal through food hoarding and seed dropping during transport to and from the cache (Emmons Reference EMMONS1980, Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985, Moupela et al. Reference MOUPELA, DOUCET, DAÏNOU, TAGG, BOURLAND and VERMEULEN2013). However, as no seed of D. normandii was found in the prospected burrows, these dispersal events are thought to be rare.
Fallen fruits are an important food resource for duikers (Seufert et al. Reference SEUFERT, LINDEN and FISCHER2009). Cephalophus spp. have been shown to act both as seed dispersers (mostly for species with medium-sized fruit and hard stones) and predators (for those with less resistant seeds) (Beaune et al. Reference BEAUNE, BRETAGNOLLE, BOLLACHE, HOHMANN, SURBECK and FRUTH2013b, Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985). The red duikers spit out hard seeds (e.g. Antrocaryon klaineanum) during rumination (Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985), but the seed size and the weak coat resistance of D. normandii suggests a vulnerability to crushing and destruction during mastication (as with Santira trimera, Feer Reference FEER1995b, Gautier-Hion et al. Reference GAUTIER-HION, DUPLANTIER, QURIS, FEER, SOURD, DECOUX, DUBOST, EMMONS, ERARD, HECKETSWEILER, MOUNGAZI, ROUSSILHON and THIOLLAY1985).
CONCLUSIONS AND PERSPECTIVES
This study identified six taxa as dispersers of Dacryodes normandii seeds and eight as seed predators, out of 14 vertebrate taxa observed consuming the fruits. As some frugivores may fulfil multiple roles, complementary and specific studies (faecal analysis and germination experiments, seed displacement monitoring, identification and germination trials of seeds in rodents' burrows, rumen and stomach content characterization, etc.) are required to contribute to a fuller understanding of the extent of seed predation and dispersal of D. normandii.
The main dispersers identified in this study are great apes: the western lowland gorilla (Gorilla gorilla gorilla) and, to a lesser extent, the central chimpanzee (Pan troglodytes troglodytes). African great apes are known to disperse the seeds of many species some of which provide timber and non-timber forest products (this study, Doucet Reference DOUCET2003, Fourrier Reference FOURRIER2013, Petre et al. Reference PETRE, TAGG, BEUDELS-JAMAR, HAUREZ, SALAH, SPETSCHINSKY, WILLIE and DOUCET2015, Tutin & Fernandez Reference TUTIN and FERNANDEZ1993, Williamson et al. Reference WILLIAMSON, TUTIN, ROGERS and FERNANDEZ1990) or are endemic to the Gabonese forests (this study, Tutin et al. Reference TUTIN, WILLIAMSON, ROGERS and FERNANDEZ1991). In particular, the western lowland gorilla is thought to provide incomparable dispersal services because of seed dispersal directed to open-canopy habitats (Petre et al. Reference PETRE, TAGG, HAUREZ, BEUDELS-JAMAR, HUYNEN and DOUCET2013, Rogers et al. Reference ROGERS, VOYSEY, MCDONALD, PARNELL and TUTIN1998, Voysey et al. Reference VOYSEY, MCDONALD, ROGERS, TUTIN and PARNELL1999).
Gorillas and chimpanzees are endangered; threatened by habitat loss and modification, epidemic diseases (such as Ebola haemorrhagic fever), commercial hunting (for bushmeat, the pet market and trophies) and persecution as a result of human–wildlife conflicts (Walsh et al. Reference WALSH, TUTIN, OATES, BAILLIE, MAISELS, STOKES, GATTI, BERGL, SUNDERLAND-GROVES and DUNN2008). The depletion of great apes would result in changes in seed dispersal and seedling recruitment which would in turn impact upon forest ecosystem dynamics and regeneration (Effiom et al. Reference EFFIOM, NUÑEZ-ITURRI, SMITH, OTTOSSON and OLSSON2013, Petre et al. Reference PETRE, TAGG, HAUREZ, BEUDELS-JAMAR, HUYNEN and DOUCET2013, Wright et al. Reference WRIGHT, ZEBALLOS, DOMÍNGUEZ, GALLARDO, MORENO and IBÁÑEZ2000).
Therefore, this study encourages the implementation of great ape conservation strategies as critical to ensuring the preservation of tropical forest ecosystems. The maintenance of great ape dispersal services may also participate in the sustainable management of timber and non-timber forest product species.
ACKNOWLEDGEMENTS
The authors are thankful to the Fonds pour la Formation à la Recherche dans l'Industrie et l'Agriculture-Fonds National pour la Recherche Scientifique (FRIA-FNRS, Belgium) for financial support provided to Barbara Haurez through a PhD grant. The field data collection was financially supported by the FNRS, the University of Liège, the Fonds National pour la Recherche Scientifique, Fonds Léopold III pour l'Exploration et la Conservation de la Nature, Fonds pour la Formation à la Recherche Scientifique en Afrique, Wallonie-Bruxelles International and the École Régionale Post-Universitaire d'Aménagement et de Gestion Intégrés des Forêts et Territoires Tropicaux (ERAIFT). Partner structures (Precious Woods Gabon, Millet and Nature +) offered technical and logistical support during the field work of this PhD study. We genuinely thank all field assistants who helped in data collection, especially Armand Boubady, Noël Endahoyi and Aristide Kouba. We are also grateful to one anonymous reviewer who provided relevant comments about earlier drafts of the manuscript.
Appendix 1. Results of excavation of rodent burrows in the search for Dacryodes normandii seeds in a Gabonese logging concession
