Journal of Tropical Ecology (2010) 26:251–262. Copyright © Cambridge University Press 2010doi:10.1017/S0266467410000027

Seed dispersal of the Brazil nut tree (Bertholletia excelsa) byscatter-hoarding rodents in a central Amazonian forest

Joanne M. Tuck Haugaasen∗, Torbjørn Haugaasen1,†,‡, Carlos A. Peres‡, Rogerio Gribel∗and Per Wegge†

∗ Laboratorio de Genetica e Biologia Reprodutiva de Plantas, Coordenacao de Pesquisas em Botanica, Instituto Nacional de Pesquisas da Amazonia, Caixa Postal 478,Manaus, Amazonas State, Brazil†Norwegian University of Life Sciences, Department of Ecology and Natural Resource Management, P.O. Box 5003, 1432 A

◦s, Norway

‡Centre for Ecology, Evolution and Conservation, School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK(Accepted 6 January 2010)

Abstract: We know surprisingly little about the fate of seeds of the Brazil nut tree (Bertholletia excelsa) under naturalconditions. Here we investigate seed removal, predation and caching of Brazil nuts by scatter-hoarding rodents in thewet and dry seasons, based on an experimental approach using 900 thread-marked seeds. We tracked the fate of seedshandled by these animals to examine how seasonal food availability may influence caching rates, dispersal distancesand cache longevity. Most seeds exposed to dispersal trials were removed by scatter-hoarders during the first week inboth seasons and seeds were generally buried intact in single-seeded caches within 10 m of seed stations. Seeds wereremoved significantly faster and buried at greater distances during the dry season. The proportion of seeds buried intactwas considerably higher in the wet season (74.4%) than in the dry season (38.2%). Most (99.4%) of the 881 primarycaches monitored were recovered, but these had a significantly shorter lifetime in the dry season. Our results showthat rodents are highly skilled at retrieving buried Brazil nuts and that caching behaviour appears to be affected byseasonal resource abundance. Reduced seed availability due to intensive harvest could potentially create a dry-seasonscenario where most seeds succumb to pre-dispersal predation, thereby adversely affecting the natural regenerationof Brazil nut trees.

Key Words: agouti, Bertholletia excelsa, Dasyprocta fuliginosa, fruit abundance, Myoprocta pratti, neotropical forest,resource seasonality, secondary caches, seed germination, seed predation

INTRODUCTION

Most phenological observations in tropical plants showthat fruiting is episodic and that fruit availability variesseasonally (Chapman et al. 1999, Foster 1982, Frankieet al. 1974, Haugaasen & Peres 2005, White 1994).Many frugivorous and granivorous birds and mammalsin tropical forests cope with these seasonal fruit shortagesusing dietary shifts to non-fruit or keystone fruit resources(Peres 2000) or reducing their metabolic requirements(Dausmann et al. 2005, Ganzhorn et al. 2003). Otherswill seasonally move into second-growth (Levey 1988,Loiselle & Blake 1991), other forest types (Haugaasen& Peres 2007) or elevations (Loiselle & Blake 1991)

1 Corresponding author. Emails: torbjorn.haugaasen@umb.no;T.Haugaasen@uea.ac.uk

in search for food. However, large diurnal caviomorphrodents in neotropical forests – agoutis (Dasyprocta spp.)and acouchis (Myoprocta spp.) – are exceptions. Thesescatter-hoarders regularly bury large seeds intact inshallow caches in the soil in anticipation of periodsof fruit scarcity (Dubost 1988, Forget 1992, 1996;Hallwachs 1986, Smythe 1978). In fact, agoutis livealmost exclusively off these reserves in times of fruitscarcity (Dubost & Henry 2006, Henry 1999, Smythe1978). Seed hoarding, therefore, appears to be crucial totheir survival and reproduction.

Scatter-hoarding accrues several potential advantagesto plants. For example, scatter-hoarding involves thetransportation of seeds away from the parent, which isrequired for the colonization of new sites (Stapanian &Smith 1978). Scatter-hoarding also reduces the risk ofdensity-dependent mortality and increases the probabilityof a seed ending up in a suitable site for germination and

252 JOANNE M. TUCK HAUGAASEN ET AL.

establishment (Hammond & Brown 1998). The scatteringof seeds in spatially diffuse caches will similarly reducethe risk of other seed-eaters finding and killing the seeds(Vander Wall 1990). However, scatter-hoarding alsocarries a potential cost – seeds will only germinate andbecome established if they happen to escape consumptionby seed hoarders. Nevertheless, the net contributionof scatter-hoarding rodents to seedling establishment islikely positive in most years for large-seeded plant species(Jansen et al. 2004).

In neotropical forests, seed dispersal and recruitmentof some plant species depend heavily on large scatter-hoarding rodents such as agoutis and acouchis. Scatter-hoarded seeds are typically large and nutritious, and areusually produced by large trees. One typical exampleof such a tree species is the Brazil nut tree (Bertholletiaexcelsa Humb. & Bonpl., Lecythidaceae). Bertholletiaexcelsa produces extremely hard globose fruits with 10–25 large seeds (or Brazil nuts) that remain encasedafter fruit fall. Under natural conditions, the fruits aregnawed open by agoutis that also act as the main scatter-hoarding seed disperser of the species. These rodentsare therefore crucial to the natural regeneration of theBrazil nut tree. Given that Brazil nut trees provide one ofthe most socio-economically important non-timber forestproducts (NTFP) in Amazonia and that these nuts arecollected exclusively from natural forests, this interactionis of primary importance for the future of the Brazilnut industry which in 1998 was worth US$ 57 million(Zuidema 2003).

The fate of seeds is a critical aspect of plant populationecology (Willson 1992). However, despite the importanceof seed dispersal we know surprisingly little about thefate of Brazil nuts under natural conditions. Dispersaldistances have been measured during two previous fieldexperiments (Bouwman & van Dijk 1999, Peres & Baider1997). It has also been shown that agoutis destroy moreand cache fewer seeds when food availability is low(Bouwman & van Dijk 1999; but see Peres et al. 1997), butno study to date has followed the fate of individual seedsfrom natural caches. The proportion of rodent scatter-hoarded seeds that eventually germinate to become aseedling has yet to be quantified. This study attemptsto redress some of these shortfalls by investigating theeffectiveness of scatter-hoarding rodents as B. excelsa seeddispersers and track the fate of live seeds handled bythese animals. We also investigated how seasonal foodavailability may influence caching behaviour, focusingon caching rates, dispersal distances and cache longevity.In particular, we tested the hypotheses that more seedswould be buried during the season of plenty, andthat these caches would remain intact for longer. Wealso tested the hypothesis that seed dispersal distanceswould be longer during the dry-season food resourcebottleneck.

STUDY SPECIES

The Brazil nut tree represents a monotypic genusbelonging to a pantropical tree family, the Lecythidaceae,which includes approximately 200 neotropical species(Mori & Prance 1990). Bertholletia excelsa is anexceptionally large emergent tree regularly attaining 40–50 m in height, 1–2.7 m in diameter at breast height(dbh) and a crown diameter of 20–35 m. The species hasa widespread distribution in lowland Amazonia and theGuianan Shield (Mori & Prance 1990) where it naturallyoccurs in upland forests that never flood (hereafter terrafirme forest). Climatic limits for its natural distributionare reported to be a mean annual precipitation of 1400–2800 mm, a mean annual temperature of 24–27 ◦C and amean annual relative humidity of 79–86% (Diniz & Bastos1974).

Throughout its range, the Brazil nut tree is commonlyreferred to as growing in natural clusters or stands, locallyknown as castanhais (Mori & Prance 1990, Peres & Baider1997, but see Wadt et al. 2005), and it has been suggestedthat the current spatial distribution of Brazil nut treeshas been shaped by ancient Amerindian agro-ecosystempractices (Miller 1990, Mori & Prance 1990, Muller et al.1980). The tree is considered to be a gap-dependentspecialist (Mori & Prance 1990), meaning that it will onlyreach reproductive size in natural forest gaps where lightis abundant.

The Brazil nut fruit is unique among the Lecythidaceaein that the seeds remain encased after fruit fall (Mori &Prance 1990). The mesocarp is an extremely hardglobose, woody capsule (pyxidium) which drops to theground after fruit maturation. Given that the pyxidium isfunctionally indehiscent, mature seeds remain trappedinside the thick pericarps of fruits on the groundunless opened by a vertebrate seed predator. The seedsconsist of a lignified seed testa (shell) protecting alarge embryo measuring 4.0 ± 0.06 cm (mean ± SE) inlength, 2.0 ± 0.03 cm in width and weighing 6.7 ± 0.2 g,although inter- and intra-populational fruit and seed sizevariability may be extremely high (T. Haugaasen & J. M.Tuck Haugaasen, unpubl. data). The edible embryo ishighly attractive to seed predators and contains 17–25%protein and 70–72% lipids (Peres 1991).

The scatter-hoarding rodents present in our study areaare the black agouti (Dasyprocta fuliginosa Wagler) andthe green acouchi (Myoprocta pratti Pocock). Agoutis andacouchis are the only rodents that regularly bury largeintact seeds in neotropical forests (Dubost 1988, Forget1990, 1991; Smythe 1978). However, natural dispersaland predation of Bertholletia seeds is primarily performedby agoutis. These large-bodied (3.0–5.9 kg) diurnalcaviomorph rodents occur throughout the geographicrange of B. excelsa and are virtually the only rodentscapable of gnawing through the thick pericarp wall of

Seed dispersal of the Brazil nut tree 253

Figure 1. Map of the study area in the lower Rio Purus region of central Amazonia, Brazil. The triangle represents the location of the Brazil nut standin which experiments were carried out.

newly fallen fruits to access the seeds inside. The Brazilnut tree therefore relies almost entirely on these largeterrestrial rodents for the release of their well-protectedseeds. Those seeds that are not consumed within thegermination period of Bertholletia (12–18 mo; Muller1981) may germinate and seeds may remain viable for atleast 6 y (Watson 1901). However, germination successappears to be dependent on moist storage conditions(Kainer et al. 1999).

METHODS

Study area

This study was conducted at Lago Uauacu, a large, 32-km long crescent-shaped black-water lake located in thelower Rio Purus region (04◦14′S, 62◦23′W) of centralAmazonia, about 350 km south-west of Manaus (Fig-ure 1). This lake acts as the northern border of thePiagacu-Purus Sustainable Development Reserve andthe region incorporates a large interdigitated mosaic ofunflooded (terra firme) and flooded forests inundatedby white-water (varzea) and black-water (igapo) on aseasonal basis for as long as 6 mo of the year (Figure 1,Haugaasen & Peres 2006). There is a pronouncedseasonal variation in rainfall, with July-October beingthe driest months, and February–June the wettest(Haugaasen & Peres 2006). In terms of basal area, the

Brazil nut tree is the dominant species in terra firmeforest in the region (T. Haugaasen, unpubl. data) asdefined by the importance value index. The 30 extractivehouseholds inhabiting the study area rely primarily onthe collection of Brazil nuts from natural stands of Brazilnut trees, in addition to small-scale fishing and huntingfor their subsistence and income. However, hunting didnot take place in our experimental area. In general,recent subsistence hunting has been restricted to lightharvest of large-bodied game species such as ungulates.The study area therefore retains a largely intact large-mammal fauna.

Experimental design

Experiments were conducted in the wet season towardsthe end of the 2006 Brazil nut season (April) andthen repeated at the same sites during the dry season(September) when Brazil nut trees were no longer fruitingand fruiting trees in general were scarce in the studyarea (Haugaasen & Peres 2005). Seeds used in theexperiments were new (i.e. from the same year’s seedcrop) and had been cleaned and sun-dried until the outershell was dry. Embryo viability was tested using a waterimmersion technique widely performed by local Brazilnut collectors across Amazonia in which all seeds failingto sink were discarded. Seeds were subsequently markedwith a 50-cm-long piece of dental floss where one end of

254 JOANNE M. TUCK HAUGAASEN ET AL.

the thread was glued to the seed testa using transparentweather-resistant epoxy. A small piece (6 cm) of orangefluorescent flagging tape was then tied to the oppositeend of the thread and individually numbered using ablack permanent marker. In this way the flag stayedvisible despite the seed being cached by a scatter-hoardingrodent. Thread-marking techniques have been widelyused in neotropical forests (Chauvet et al. 2004, Forget1990, 1991, 1992; Forget & Milleron 1991, Jansen et al.2002, 2004; Peres & Baider 1997) and have been shownto be effective in tracking individual seeds following seedremoval by scatter-hoarding rodents (Xiao et al. 2006).The advantages and disadvantages of the method havebeen discussed elsewhere (Forget & Wenny 2005).

In total, 900 seeds were placed at 60 circular seedstations (15 seeds per station) surrounding 20 very large(>100 cm dbh) fruiting Brazil nut trees during our initialwet-season experiment, and this was repeated again 5mo later during the dry season. Our 20 focal trees werewidely spaced within a total forest area of ∼ 92 ha, andno two trees were closer than 100 m from one another(mean ± SE = 117.4 ± 5.6 m; N = 20). Seed stations weremarked with a flagged pole standing upright at thecentre of a circular patch of ground approximately 1.5 min diameter from which all leaf litter and other debrishad been removed. At each station, 15 marked seedswere arranged in a circle with their markers stretchedoutwards, roughly mimicking the average number ofseeds contained in a single Brazil nut fruit (17.4 ± 0.3;N = 195). Groups of three stations were positioned at10 m from the base of each tree, forming a triangle centredat the base of each tree. Once seeds had been laid out,stations were monitored every 2 d by four observers untilall seeds had been removed. Where seeds were missing,intensive searches by the observers were undertakento relocate the marked seeds or markers. Simultaneoussearches throughout concentric rings radiating outwardsfrom the seed station were carried out to most efficientlycover a large area. To conform to previous studies (Peres &Baider 1997), we abandoned searches only beyond40 m from the seed station, unless all translocated seedsor seed flags had already been found. Upon finding a givenseed, seed fragment or tag, we determined the seed statusand position, and measured the distance between eachseed and its seed station. If the seed was found to be intact(including those buried into the soil, and those hiddenunderneath the leaf-litter or simply lying on the ground)we marked its position with a piece of orange flaggingtape, and the fate of the seed was then followed every 10d thereafter. Since markers are known to provide cues forcache pilferage by diurnal rodents (Vander Wall 1991,2003), we carefully covered the flags attached to seedswhich were removed from the seed stations during bothexperiments. If a flagged seed disappeared subsequently,intensive searches in the same manner as described above

were mounted to determine the fate of the seed, but werestricted these searches to within 20 m of each primarycache location.

Data analysis

To account for the spatially hierarchical nature of ourdata set, we modelled (1) our short-term proxy of seedsurvivorship, (2) time-lag to seed removal, and (3)the straight-line distance (log10(x+1)) to which theywere carried from the 60 experimental seed stationsreplicated during each season using generalized linearmodels (GLMs). In these models, the outcome of seed-handling events (by dispersal agents or seed predators)was assumed to be nested within clusters (seed stations),which in turn were nested underneath the large crownsof the 20 focal Brazil nut trees over which the randomeffects varied. We collapsed a maximum of nine possiblecategories of seed status into a single binary outcome(dead or alive) depending on whether seeds had beendestroyed or were still intact at the time each seedstation was inspected and surrounding areas searched.Individual seed survival for each set of 900 seeds exposedtwice (wet and dry seasons) was therefore modelled usinga binomial error structure and logit link function, whereasother responses could be modelled using either a normalor Poisson distribution. Since all experimental focal treeswere spaced by at least 100 m from one another, theycould therefore be considered to be spatially independent.The JMP statistical software package (JMP 8.0. SASInstitute Inc.) was used in all analyses.

RESULTS

Fate of seeds at seed stations

All seed stations were rapidly discovered by seed-eatingmammals and all seeds were removed. Additionally, theminority of seeds remained unrecovered or had theirmarkers cut off (13.5% and 31% in wet and dry season,respectively) and a large number of seeds were buriedby scatter-hoarding rodents (Table 1). Preferred cachingsites included the base of arborescent palms and terrestrialbromeliads, in and around dead and rotting fallen treetrunks and at the base of trees (particularly betweenbuttress roots).

The vast majority of seeds were buried intact as diffuse,single-seeded caches during both seasons. However, therewere several significant seasonal differences in the fate ofseeds placed at seed stations. Firstly, there was a strongeffect of season on seed handling outcomes (χ2 = 305,P < 0.0001), and short-term seed survivorship (GLM:N = 1495, χ2 = 157, P < 0.001), where seeds were more

Seed dispersal of the Brazil nut tree 255

Table 1. Fate of marked Bertholletia excelsa seeds in dispersal experiments at Lago Uauacu, central Amazonia, Brazil. The experimentswere performed in the B. excelsa fruiting season (April 2006) and in the dry season (September 2006) showing the original distancethe seed or marker was carried and the initial time taken for seed removal.

Number Distance carried (m) Removal time (d)(%) of seeds Mean ± SE (range) Mean ± SE (range)

Fate of seeds Wet season Dry season Wet season Dry season Wet season Dry season

Seeds removed from stationsBuried intact 637 244 7.1 ± 0.3 10.0 ± 0.5 7.3 ± 0.3 4.8 ± 0.3

(74.4) (38.2) (0.7–50.1) (0.9–48.4) (1–51) (1–25)Not buried but hidden 38 75 4.3 ± 1.0 4.7 ± 0.3 13.8 ± 2.4 8.4 ± 0.9

(4.5) (11.7) (0.8–37.9) (1.0–13.7) (2–51) (1–27)Eaten elsewhere 14 37 8.4 ± 2.6 6.3 ± 1.0 15.9 ± 5.4 6.8 ± 1.0

(1.6) (5.8) (1.1–32.5) (0.9–30.1) (1–51) (1–25)Intact (neither buried nor hidden) 12 4 3.8 ± 0.9 2.4 ± 0.6 10.5 ± 3.0 10.0 ± 3.0

(1.4) (0.6) (1.3–12.5) (1.4–3.5) (1–32) (7–19)Markers cut off and recovered elsewhere 32 68 10.8 ± 2.0 5.2 ± 0.7 11.3 ± 2.2 8.1 ± 0.8

(3.7) (10.6) (0.7–36.0) (0.9–29.1) (1–51) (1–33)Markers cut off and recovered at station 7 77 – – 16.9 ± 8.9 8.8 ± 0.6

(0.8) (12.1) (1–51) (3–19)Markers unrecovered 77 53 – – 6.6 ± 1.2 6.7 ± 1.0

(9.0) (8.3) (1–51) (1–29)Seeds remaining at stationsBuried 11 0 (0) 0 (0) – 1.2 ± 0.2 –

(1.3) (1–3)Eaten 28 81 0 (0) 0 (0) 5.9 ± 0.9 5.6 ± 0.5

(3.3) (12.7) (1–18) (1–25)

All seeds taken 856 639 7.1 ± 0.2 7.9 ± 0.3 7.8 ± 0.3 6.5 ± 0.2(100.0) (100.0) (0.7–50.1) (0.9–48.4) (1–51) (1–33)

likely to be destroyed immediately upon detection duringthe dry season. The proportion of seeds buried intactwas nearly two-fold higher in the wet season (74.4%)than in the dry season (38.2%; Table 1). The proportionof seeds simply hidden under the leaf litter was higherin the dry season. In addition, more seeds were eatenimmediately around or away from the seed stations, anda larger proportion of seeds had their markers severed orlost, during the dry season (Table 1).

Tree dbh (or basal area) had no effect on short-termseed survivorship (χ2 = 0.088, P < 0.767), likely becauseall focal trees were large. However, the B. excelsa treedensity within the immediate neighbourhood of each focaltree had a significant effect on short-term seed mortality,with a larger proportion of seeds being destroyed wherethe number of conspecific adult Brazil nut trees within a100-m radius of each focal tree was higher (χ2 = 15.1,P < 0.001). This is likely to largely explain the significanttree identity effect on the proportion of seeds surviving tothe end of the experimental period (χ2 = 100, P < 0.001).

Removal time and dispersal distances

Bertholletia excelsa seeds exposed to dispersal trials wererapidly removed or eaten in situ (Table 1). In fact, mostseeds were removed in the first week of both experiments(Figure 2). We also found significant effects of season

Figure 2. Removal time of thread-marked seeds during the wet and dryseasons at Lago Uauacu, central Amazonia, Brazil.

and distance to the nearest conspecific B. excelsa treeon seed removal time, where seed removal by rodentswas significantly faster during the dry season (GLM:χ2 = 14.4, P < 0.0001), particularly if seeds came fromseed stations underneath more isolated trees (χ2 = 41.6,P < 0.0001).

Seeds could be carried and buried at least 50 mfrom the seed station (Table 1). However, most seedswere buried within 10 m of their seed stations (77.6%

256 JOANNE M. TUCK HAUGAASEN ET AL.

Figure 3. Removal distance of thread-marked seeds during the wet and dry seasons at Lago Uauacu, central Amazonia, Brazil.

and 65.1% of the seeds in the wet and dry seasons,respectively; Figure 3). There were significant effects ofseason (χ2 = 6.88, P = 0.0087), focal tree basal area(χ2 = 32.5, P < 0.0001), the number of B. excelsa treeswithin 100 m of each focal tree (χ2 = 95.7, P < 0.0001),and nearest-neighbour distance (χ2 = 42.2, P < 0.0001)on log10(x+1)-transformed distances at which seeds wereburied intact. Seeds buried during the dry season werecarried significantly farther away than during the wetseason, particularly if they came from seed stations placedaround larger and more isolated trees.

Cache management and fate of buried seeds

We followed the fate of 881 primary caches during ourtwo experiments. Scatter-hoarding rodents were highly

adept at locating buried seeds as only five seeds remainedburied intact after the conclusion of our experiments 1 yafter they were established. Only one of these seeds wasgerminating at the time the remaining caches were lastsurveyed, approximately 1 y after it was buried.

Most buried seeds were rapidly recovered by therodents. The majority of caches had been recoveredwithin 1–3 wk after burial and few seeds remainedcached for more than 1 mo after either of the twoexperiments. Our experiment during the dry seasonshowed a significantly shorter life span for primary cachesthan our earlier experiment during the B. excelsa fruit-fallseason (F1,874 = 147; P < 0.001; Table 2).

It was difficult to accurately determine the fate of buriedseeds as 44.7% and 67.7% of the caches that we found inthe dry and wet season, respectively, simply disappeared.

Table 2. Fate of remaining Brazil nut seeds (those not eaten or lost) from a seed dispersal experiment established during the Brazil nut season andduring the dry season at Lago Uauacu, central Amazonia, Brazil. The number of days until final removal is expressed as mean ± SE (range). Thenumber (percentages) of seeds is given for each category. Marker signifies cases where only tags from seeds were found; Hidden seeds were notburied but hidden in other ways; Intact seeds were neither buried nor hidden, but found intact away from the seed station

Fate of seeds

Time (d) until final removal No. of seeds relocated Eaten Marker Unknown Still there

Wet season Dry season Wet season Dry season Wet Dry Wet Dry Wet Dry Wet Dry

Seeds removed from stationsBuried intact 38.7 ± 1.4 11.3 ± 0.9 38 3 125 99 78 34 431 109 3 2

(1.5–238.0) (0.5–116.0) (6.0) (1.2) (19.6) (40.6) (12.2) (13.9) (67.7) (44.7) (0.5) (0.8)Hidden 17.4 ± 2.3 12.5 ± 1.7 1 5 5 10 10 29 23 36 0 0

(2.5–90.0) (1.5–96.0) (2.6) (6.7) (13.2) (13.3) (26.3) (38.7) (60.5) (48.0) (0) (0)Intact 47.9 ± 16.5 5.4 ± 2.2 0 0 5 2 0 1 7 1 0 0

(5.0–118.0) (2.5–12.0) (0) (0) (41.7) (50.0) (0) (25.0) (58.3) (25.0) (0) (0)Seeds remaining at stations

Buried 29.1 ± 6.9 – 1 – 2 – 0 – 9 – 0 –(6.5–68.0) (9.1) (18.2) (0) (81.8) (0)

All seeds 37.5 ± 1.3 11.5 ± 0.8 40 8 137 111 88 64 470 146 3 2(1.5–238.0) (0.5–116.0) (5.7) (2.5) (19.6) (34.4) (12.6) (19.8) (67.3) (45.2) (0.4) (0.6)

Seed dispersal of the Brazil nut tree 257

Table 3. Fate of relocated seeds from an experiment established during the Brazil nut season (April) at Lago Uauacu, central Amazonia, Brazil. Theoriginal and new distances (m) from the seed stations are expressed as mean ± SE (range).

Relocated 1st time Relocated 2nd time

Original distance New distance Time (d) till relocated No. of seeds New distance Time (d) till relocated Removal time (d)

Seeds removed from stationsBuried intact 6.6 ± 0.8 16.4 ± 1.7 11.4 ± 2.3 – – – 45.6 ± 5.1

(0.8–20.2) (0.9–45.6) (1.5–48.0) (2.5–170)Hidden 3.1 31.9 1.5 1 (2.6) 49.8 40 90

Seeds remaining at stationsBuried 0.0 5.0 6.5 – – – 38

All seeds 6.5 ± 0.7 16.5 ± 1.7 11.0 ± 2.2 1 (2.5) 49.8 40 46.6 ± 5.0(0.8–20.2) (0.9–45.6) (1.5–48.0) (2.5–170)

Nevertheless, a large number had been eaten at the cachelocation and the number of seeds eaten was considerablyhigher in the dry season (Table 2). While searching forburied seeds that had been excavated, we also foundthat 7.2% of seeds had not been eaten but relocated tosecondary caches. These secondary caches stored duringthe wet season were far more abundant than those in thedry season (Table 2).

Scatter-hoarding rodents were, however, not the onlypredators of cached seeds. In both experiments, we foundthat leaf-litter ants destroyed a significant number ofburied seeds. Ants destroyed 8.3% and 4.3% of the seedscached during the wet and dry season, respectively.

Re-buried seeds

Distances from seed stations were greater for secondarycaches than primary caches in both seasons althoughre-buried seeds that were relocated in the wet seasonwere moved considerably farther, more than doublingthe average distance between seed stations and primarycaches (Tables 3 and 4). This trend was mirrored byseeds hidden underneath the leaf litter which wererelocated in a similar manner. We were unable todetermine the fate of most relocated seeds, althoughsome seeds were found to be relocated up to threetimes. However, these all belonged to the category ofhidden seeds. For each relocation event the distance of

re-buried seeds from the seed station increased farther(Tables 3 and 4).

Other seed predators

The fact that some seeds were concealed underneaththe leaf-litter rather than buried indicates that smallerrodents, most likely spiny rats (Proechimys spp.), visitedour seed stations.

Experimental seed arrays were also handled by arborealseed predators. At several seed stations, markers attachedto seed remains were found hanging from tree branchesat heights of 0.5 to 15 m. These were most likely left bythe brown capuchin monkey (Cebus apella) which wasregularly encountered in the area where our experimentswere conducted. In fact, 27.5% of all seeds exposed in thedry season were taken by monkeys compared with 4.8%in the wet season. The collared peccary (Tayassu pecari)also obliterated one of our seed stations during the dryseason.

DISCUSSION

Fate of seeds at seed stations

This study tracked the fate of a large number of markedseeds exposed on the forest floor close to the parent

Table 4. Fate of relocated seeds from a seed-dispersal experiment established outside the Brazil nut season (September) at Lago Uauacu, centralAmazonia, Brazil. The original and new distances (m) from the seed stations are expressed as mean ± SE (range).

Relocated 1st time Relocated 2nd time Relocated 3rd time

Originaldistance

Newdistance

Time (d) tillrelocated

No. ofseeds

Newdistance

Time (d) tillrelocated

No. ofseeds

Newdistance

No. of days tillrelocated

Removaltime (d)

Seeds removed from stationsBuried intact 8.1 ± 2.6 12.9 ± 0.9 13.3 ± 9.4 – – – – – – 35.3 ± 1.8

(3.0–11.6) (11.2–14.3) (2.5–32.0) (32.0–38.0)Hidden 6.4 ± 0.9 9.3 ± 2.0 6.1 ± 2.1 2 (2.7) 25.4 25.0 ± 1.0 1 (1.3) 28.2 36.0 37.3 ± 15.6

(4.3–8.8) (4.1–16.1) (1.5–14.0) (24.0–26.0) (6.5–96.0)

All seeds 7.0 ± 1.1 10.7 ± 1.4 8.8 ± 3.6 2 (25.0) 25.4 25.0 ± 1.0 1 (12.5) 28.2 36.0 36.6 ± 9.4(3.0–11.6) (4.1–16.1) (1.5–32.0) (24.0–26.0) (6.5–96.0)

258 JOANNE M. TUCK HAUGAASEN ET AL.

tree. Admittedly, handling times for agoutis encounteringunprotected seeds were shorter than during an encounterwith an intact fruit, and this may have influencedthe observed caching rates and dispersal distances.Nevertheless, this is the first study to examine the fateof individual B. excelsa seeds after caching by scatter-hoarding rodents, providing further understanding onhow scatter-hoarders either disperse or predate large seedsin neotropical forests.

Conforming to other studies using the thread-markingtechnique, scatter-hoarding rodents were the first animalsto detect our seed stations (Jansen et al. 2004). However,in contrast with the Kayapo Indian Reserve located ineastern Amazonia where acouchis are absent (Peres &Baider 1997), agoutis and acouchis are sympatric inour study area. Therefore, we are unable to conclusivelyascertain the relative importance of each species as visitorsto our seed stations, although footprints and the mannerin which most seeds were buried suggest that agoutisremoved a disproportionate fraction of all seeds.

Typically seeds were buried singly which appears tobe a consistent feature of caviomorph rodent caches inthe Neotropics (Jansen & Forget 2001). Preferred cachingsites were also similar to other studies (Forget 1990, Kiltie1981, Smythe 1978, Vander Wall 1990). Equally, wewere able to recover the majority of marked seeds (86.5%and 69% in the wet and dry season, respectively). Theserecovery percentages were notably higher than those ofPeres & Baider (1997) and Forget (1990). The overall seedremoval during the first 2-wk period of 90.1% and 84.5%in the wet and dry season, respectively, was similar to thatreported by Peres & Baider (1997). However, our resultscontrast somewhat with Peres et al. (1997) who found aconsistently lower seed removal in the dry season acrosstheir experimental treatments, which they attributed todifferences in seasonal foraging behaviour.

Dispersal distances

Previous experiments have shown that Bertholletia seedsare unlikely to be dispersed over long distances by scatter-hoarding rodents (Bouwman & van Dijk 1999, Peres &Baider 1997). In fact, most seeds were buried within5–10 m of the seed stations, a trend repeated duringboth of our experiments (Figure 3). Peres & Baider(1997) attributed the clustered spatial distribution andgrove formation of adult Bertholletia trees to the limitedseed dispersal distances of the agoutis. Nevertheless,exceptionally long dispersal distances of B. excelsa seedsby scatter-hoarding rodents have been observed. Forexample, dispersal distances of 100 m were reported froma study in Madre de Dios, Peru (Peres & Baider 1997)and dispersal distances beyond 50 m were observed in thecurrent study. At Lago Uauacu, many seeds were also

dug up and re-cached, which further increased dispersaldistances. Exceptionally long dispersal distances havealso been found to occur in other large-seeded scatter-hoarded tree species (Jansen et al. 2004; P.-M. Forgetpers. comm.), although scatter-hoarders in general tendto carry seeds over short distances (Asquith et al. 1999,Forget 1990, 1992; Hallwachs 1986). The spatial effectsobserved on seed survival and dispersal are most likelyattributed to the total number of Bertholletia trees withineach experimental neighbourhood. Jorge & Peres (2005)have shown that local densities of agoutis clearly covarywith densities of Bertholletia trees, so there must be acertain amount of density-dependence built into the seedcaching/foraging behaviour of scatter-hoarding rodentsin Amazonian forests.

Cache management and fate of buried seeds

Most seeds disappeared within a few weeks after caching,which is consistent with studies of other large-seededspecies (Forget 1990, Jansen & Forget 2001). However,the fact that some seeds were still intact in their cachesmany months after first burial suggests that at leastpart of the cache were long-term reserves. Large rodentsappear to be highly skilled at retrieving buried seeds, asonly five seeds remained cached of the 881 single-seededcaches monitored during our 12-mo experiments. Onlyone of these seeds germinated. However, caution shouldbe taken in interpreting this germination success. Seedswere only monitored for up to 1 y after burial which doesnot cover the entire dormancy period of the Brazil nut(12–18 mo; Muller 1981) and some of the remainingcached seeds could still subsequently germinate. Inaddition, viable cotyledons may have been damaged bythe sun-drying treatment prior to the experiments. Inany case, germination success cannot be equated toseedling establishment and survival as many scatter-hoarding rodents eat germinating seeds with seedlingsalready protruding (Forget 1992, Jansen et al. 2006), andagoutis are known to excavate and consume germinatingB. excelsa seeds in much of the geographic range of theBrazil nut tree (C.A. Peres, unpubl. data). Conversely,seed disappearance (markers cut off) cannot be equatedto mortality. The number of seeds re-located to secondarycaches is probably higher than that suggested here giventhe large number of unrecovered cached seeds followingtheir disappearance from primary caches (Tables 3 and4), and these were apparently moved beyond the 20-msearch radius.

The excavation and re-caching of seeds appear to becommon among scatter-hoarding rodents (Jansen et al.2004, Vander Wall 2002, 2003). There are severalreasons why agoutis and other scatter-hoarders mightmanage their caches in this way. Jansen & Forget (2001)

Seed dispersal of the Brazil nut tree 259

suggest that scatter-hoarders initially hide seeds onlyprovisionally, and as quickly as possible, to decrease therisk of losing exposed seeds to competing granivores.The rapid recovery of caches and the large number ofsecondary caches encountered in the wet season supportsthis hypothesis. Rodents may also open caches to refreshtheir memory or inspect the seed condition (Jansen &Forget 2001). The fact that ants were also a significantpredator of cached seeds in this study suggests thatthis may be important and monitoring the conditionof cached seeds may be key to rodents surviving thedry-season resource bottleneck. However, Vander Wall(2002) suggests that the main reason for re-caching ofseeds is seed pilfering by other individuals transportingseeds into their own territory. It is therefore unlikely thata single factor can universally explain the occurrence ofre-caching.

Seasonal differences

Janzen (1971) proposed that food abundance affects theseasonal rate of caching, thereby influencing whethera scatter-hoarding rodent is a seed predator or seeddisperser. Assuming that caching is primarily inducedby food satiation, one would expect more caching at thetime of high fruit abundance. Indeed, our experimentsconformed to our initial hypothesis and showed that thenumber of seeds buried intact was significantly higherin the wet season (75%), when alternative fruits wereabundant, than in the dry season (38%), when fruitingtrees were scarce in the study area. These results agreewith other studies in more seasonal neotropical forestsshowing that when food is abundant, agoutis cacheconsiderably more than they eat (Bouwman & van Dijk1999, Forget et al. 2002, Smythe 1978). The sametrend has been found for acouchis (Jansen et al. 2004).The peak fruiting period in terra firme forests at LagoUauacu is January–April (Haugaasen & Peres 2005). Thissuggests that B. excelsa has evolved a fruiting pattern thatmaximizes seed dispersal by agoutis encountering singlemulti-seeded fruits by producing fruits synchronously andfruiting during the peak of the community-wide fruitingseason (Smythe 1970). However, variation in fruitabundance may also occur inter-annually. Bertholletiatypically has large between-year fluctuations in crop sizes(Kainer et al. 2007; T. Haugaasen unpubl. data) and weare yet to understand how this may affect foraging andcaching behaviour, although it may function in a waysimilar to that described herein between seasons.

The smaller number of secondary caches found in thedry season almost certainly reflects the need for rodentsto almost immediately draw on these resources for foodin times of food shortage. The immediate consumption oflarge numbers of seeds during the dry season also suggests

that, under natural conditions, seeds from fruits openedat times of fruit scarcity will suffer higher predation ratesand thus have less potential for survival. However, inaccordance with our expectations, we found that seedswere buried farther away from the parent tree at this timeof the year which has been shown to decrease offspringmortality (Jansen et al. 2008). Moreover, in a differentinvestigation we found a number of germinating seedsinside 1-y-old fruits (T. Haugaasen, unpubl. data). It istherefore reasonable to assume that viable seeds in thefruits opened during the dry season will have advancedconsiderably towards germination. Being buried at thistime of the year may therefore enhance the chances ofseed germination before the hoarder returns to consumeit, although cache longevity decreased dramatically atthis time of the year. Further studies are clearly needed toinvestigate the effect of seasonality on the probability ofBrazil nut seedling establishment.

Seeds encountered by scatter-hoarders in the dryseason may be of higher value to consumers due to foodshortages at this time of year. Seeds may therefore be takenfarther from the seed station to make caches more spatiallydiffuse and thereby more difficult to find for competitorsdespite the increased foraging effort by terrestrial seedpredators during this period. This is supported by studiesshowing that larger, and thus more valuable, seeds tend tobe hidden at greater distances and in lower densities thansmaller seeds (Forget et al. 1998, Jansen et al. 2002).

Other seed predators

Smythe (1970) showed that agoutis increased theirforaging effort in times of fruit shortage. The fact that seedswere removed quicker in the dry season also suggests thatscatter-hoarding rodents in our study area increased theireffort in search for food items at this time of the year. Thegeneral community-wide scarcity in food resources maysimilarly account for the increased seed predation at seedstations by other animals such as primates and peccaries,assuming that these vertebrates were also affected by foodshortages. Peres & Baider (1997) found that their markerswere generally cut off by terrestrial echimyid rodentsrather than agoutis, suggesting that the increased rateof seed loss in the dry season during this study is due tothe activity of these smaller rodents. The larger numberof seeds hidden underneath the forest leaf litter in the dryseason, a pattern consistent with caching by spiny rats,Proechimys spp. (Forget 1991), also suggests an increasedactivity of these rodents at our seed stations at this timeof year. These observations are nevertheless of limitedvalue for B. excelsa since seeds under natural conditionswould be protected by a thick case and on the whole beunavailable to spiny rats. However, unconfirmed localreports state that pacas (Agouti paca) may also open a

260 JOANNE M. TUCK HAUGAASEN ET AL.

limited number of fruits. The brown capuchin monkey(Cebus apella) and to a lesser extent the white-frontedcapuchin (Cebus albifrons) are also known to crack openolder fruits (1–2 y old) with a more fragile pericarpby bashing them on a suitably hard surface (Baider2000, Peres & Baider 1997, J. M. Tuck Haugaasen &T. Haugaasen, pers. obs.).

Management implications

The Brazil nut harvest inevitably decreases the availabilityof seeds to scatter-hoarding rodents. Intensive harvest ofBrazil nuts may therefore create a dry-season scenarioin which agoutis destroy more and cache fewer seedsupon finding a Brazil nut fruit. Consequently, harvestmay adversely affect natural Brazil nut regeneration,suggesting that a management plan to contend withreduced recruitment should be considered in intensivelyharvested natural stands of Brazil nut trees. Similarly,since large caviomorph rodents are the main dispersersof B. excelsa, seedling recruitment may be limited inthe absence of these animals. Forget & Jansen (2007)showed that hunting limited the quantitative dispersalin Carapa procera (Meliaceae), another commerciallyvaluable non-timber forest product in Amazonian foreststhat depends on scatter-hoarding for seed dispersal andseedling establishment. Given that subsistence huntingusually accompanies the seasonal collection of Brazilnuts, the natural regeneration of the Brazil nut treeand the continued harvest of this non-timber forestproduct in the future may also depend on adequatemeasures to protect seed dispersers (Peres et al. 2003).However, it is still unknown how between-site andbetween-year community-wide fruit abundance, rodentpopulation density and hunting pressure, and localvariation in harvesting systems interact to influencesecondary dispersal and seedling establishment of Brazilnut trees, adding further challenges to the sustainablemanagement of this extractive industry and globallysignificant wild seed crop.

ACKNOWLEDGEMENTS

Special thanks to Evanir, Evandro and Mariene deAlmeida Damasceno for their invaluable help in the fieldand all the local Brazil nut collectors at Lago Uauacufor sharing their knowledge and experience with us.This research was supported by the Research Council ofNorway (grant no. 165879/V10), the Brazilian ResearchCouncil (CNPq; grant no. 152077/2005-7), the BritishEcological Society and the Rufford Foundation. Thedental floss was kindly donated by BDF Healthcare Ltd.International. We thank Pierre-Michel Forget and one

anonymous reviewer for constructive comments on anearlier version of this manuscript.

LITERATURE CITED

ASQUITH, N. M., TERBORGH, J., ARNOLD, A. E. & MAILEN RIVEROS,

C. 1999. The fruits the agouti ate: Hymenaea courbaril seed fate when

its disperser is absent. Journal of Tropical Ecology 15:229–235.

BAIDER, C. 2000. Demografia e ecologia de dispersao de frutos de

Bertholletia excelsa Humb. & Bonpl. (Lecythidaceae) em castanhais

silvestres da Amazonia oriental. PhD thesis, University of Sao

Paulo.

BOUWMAN, M. & VAN DIJK, S. 1999. Removal and fate of Brazil nut

seeds (Bertholletia excelsa) exposed in two contrasting seed densities in a

Bolivian moist forest. M.Sc. thesis, Utrecht University.

CHAPMAN, C. A., WRANGHAM, R. W., CHAPMAN, L. J., KENNARD,

D. K. & ZANNE, A. E. 1999. Fruit and flower phenology at two sites

in Kibale National Park, Uganda. Journal of Tropical Ecology 15:189–

211.

CHAUVET, S., FEER, F. & FORGET, P.-M. 2004. Seed fates of two

Sapotaceae species in a Guianan rain forest in the context of escape

and satiation hypothesis. Journal of Tropical Ecology 20:1–9.

DAUSMANN, K. H., GLOS, J., GANZHORN, J. U. & HELDMAIER, G.

2005. Hibernation in the tropics: lessons from a primate. Journal of

Comparative Physiology B: Biochemical, Systemic, and Environmental

Physiology 175:147–155.

DINIZ, T. D. A. S. & BASTOS, T. X. 1974. Contribuicao ao conhecimento

do clima tıpico da castanha do Brasil. Boletin Tecnico IPEAN 64:59–

71.

DUBOST, G. 1988. Ecology and social life of the red acouchi, Myoprocta

exilis: a comparison with the orange-rumped agouti, Dasyprocta

leporina. Journal of Zoology 214:107–123.

DUBOST, G. & HENRY, O. 2006. Comparison of diets of the acouchy,

agouti and paca, the three largest terrestrial rodents of French

Guianan forests. Journal of Tropical Ecology 22:641–651.

FORGET, P.-M. 1990. Seed dispersal of Vouacapoua americana

(Caesalpiniaceae) by caviomorph rodents in French Guiana. Journal

of Tropical Ecology 6:459–468.

FORGET, P.-M. 1991. Scatterhoarding of Astrocaryum paramaca by

Proechimys in French Guiana: comparison with Myoprocta exilis.

Tropical Ecology 32:155–167.

FORGET, P.-M. 1992. Seed removal and seed fate in Gustavia superba

(Lecythidaceae). Biotropica 24:408–414.

FORGET, P.-M. 1996. Removal of seeds of Carapa procera (Meliaceae)

by rodents and their fate in rainforest in French Guiana. Journal of

Tropical Ecology 12:751–761.

FORGET, P.-M. & JANSEN, P. A. 2007. Hunting increases dispersal

limitation in the tree Carapa procera, a nontimber forest product.

Conservation Biology 21:106–113.

FORGET, P.-M. & MILLERON, T. 1991. Evidence for secondary seed

dispersal by rodents in Panama. Oecologia 87:596–599.

FORGET, P.-M. & WENNY, D. G. 2005. How to elucidate seed fate?

A review of methods used to study seed removal and secondary

dispersal. Pp. 379–393 in Forget, P.-M., Lambert, J., Hulme, P. E.

Seed dispersal of the Brazil nut tree 261

& Vander Wall, S. B. (eds.). Seed fate: seed predation, seed dispersal and

seedling establishment. CABI Publishing, Wallingford.

FORGET, P.-M., MILLERON, T. & FEER, F. 1998. Patterns in post-

dispersal seed removal by neotropical rodents and seed fate in relation

to seed size. Pp. 25–47 in Newbery, D. M., Brown, N. & Prins,

H. H. T. (eds.). Dynamics of tropical communities. Blackwell Science,

Oxford.

FORGET, P.-M., HAMMOND, D., MILLERON, T. & THOMAS, R. 2002.

Seasonality of fruiting and food hoarding by rodents in Neotropical

forests: consequences for seed dispersal and seedling recruitment.

Pp. 241–253 in Levey, D. J., Silva, W. R. & Galetti, M. (eds.).

Seed dispersal and frugivory: ecology, evolution and conservation. CABI

Publishing, Wallingford.

FOSTER, R. B. 1982. The seasonal rhythm of fruit fall on Barro Colorado

Island. Pp. 151–172 in Leigh, E. G., Rand, A. S. & Windsor, D. M.

(eds.). The ecology of a tropical forest: seasonal rhythms and long-term

changes. Smithsonian Institutional Press, Washington D.C.

FRANKIE, G. W., BAKER, H. G. & OPLER, P. A. 1974. Comparative

phenological studies of trees in tropical wet and dry forests in the

lowlands of Costa Rica. Journal of Ecology 62:881–919.

GANZHORN, J. U., KLAUS, S., ORTMANN, S. & SCHMID, J. 2003.

Adaptations to seasonality: some primate and non-primate examples.

Pp. 132–148 in Kappeler, P. M. & Pereira, M. E. (eds.). Primate life

histories and socioecology. University of Chicago Press, Chicago.

HALLWACHS, W. 1986. Agoutis (Dasyprocta punctata): the inheritors

of guapinol (Hymenaea courbaril: Leguminosae). Pp. 285–304 in

Estrada, A. & Fleming, T. H. (eds.). Frugivores and seed dispersal.

Dr W. Junk Publishers, Dordrecht.

HAMMOND, D. S. & BROWN, V. K. 1998. Disturbance, phenology and

life-history characteristics: factors influencing frequency-dependent

attack on tropical seed and seedlings. Pp. 51–78 in Newbery, D. M.,

Brown, N. & Prins, H. H. T. (eds.). Dynamics of tropical communities.

Blackwell Science, Oxford.

HAUGAASEN, T. & PERES, C. A. 2005. Tree phenology in adjacent

Amazonian flooded and unflooded forests. Biotropica 37:620–630.

HAUGAASEN, T. & PERES, C. A. 2006. Floristic, edaphic and

structural characteristics of flooded and unflooded forests in the

lower Purus region of central Amazonia, Brazil. Acta Amazonica 36:

25–36.

HAUGAASEN, T. & PERES, C. A. 2007. Vertebrate responses to plant

phenology in Amazonian flooded and unflooded forest. Biodiversity

and Conservation 16:4165–4190.

HENRY, O. 1999. Frugivory and the importance of seeds in the diet

of orange-rumped agouti (Dasyprocta leporina) in French Guiana.

Journal of Tropical Ecology 15:291–300.

JANSEN, P. A. & FORGET, P.-M. 2001. Scatterhoarding rodents and

tree regeneration. Pp. 275–288 in Bongers, F., Charles-Dominique,

P., Forget, P.-M. & Thery, M. (eds.). Noruages: dynamics and plant-

animal interactions in a neotropical rainforest. Kluwer Academic

Publishers, Dordrecht.

JANSEN, P. A., BARTHOLOMEUS, M., BONGERS, F., ELZINGA, J. A.,

DEN OUDEN, J. & VAN WIEREN, S. E. 2002. The role of seeds size in

dispersal by a scatter-hoarding rodent. Pp. 209–225 in Levey, D. J.,

Silva, W. R. & Galetti, M. (eds.). Seed dispersal and frugivory: ecology,

evolution and conservation. CABI Publishing, Wallingford.

JANSEN, P. A., BONGERS, F. & HEMERIK, L. 2004. Seed mass and mast

seeding enhance dispersal by a neotropical scatter-hoarding rodent.

Ecological Monographs 74:569–589.

JANSEN, P. A., BONGERS, F. & PRINS, H. H. T. 2006. Tropical rodents

change rapidly germinating seeds into long-term food supplies. Oikos

113:449–458.

JANSEN, P. A., BONGERS, F. & Van Der MEER, P. J. 2008. Is farther

seed dispersal better? Spatial patterns of offspring mortality in three

rainforest tree species with different dispersal abilities. Ecography

31:43–52.

JANZEN, D. H. 1971. Seed predation by animals. Annual Review of Ecology

and Systematics 2:465–492.

JORGE, M. L. & PERES, C. A. 2005. Population density and home range

size of red-rumped agoutis (Dasyprocta leporina) within and outside

a natural Brazil nut stand in southeastern Amazonia. Biotropica

37:317–321.

KAINER, K. A., MATOS MALAVASI, M., DURYEA, M. L. & DA SILVA,

A. R. 1999. Brazil nut (Bertholletia excelsa) seed characteristics,

preimbibition and germination. Seed Science and Technology 27:731–

745.

KAINER, K. A., WADT, L. H. O. & STAUDHAMMER, C. L. 2007.

Explaining variation in Brazil nut fruit production. Forest Ecology

and Management 250:244–255.

KILTIE, R. A. 1981. Distribution of palm fruits on a rain forest floor: why

white-lipped peccaries forage near objects. Biotropica 13:141–145.

LEVEY, D. J. 1988. Spatial and temporal variation in Costa Rican fruit

and fruit-eating bird abundance. Ecological Monographs 58:251–

269.

LOISELLE, B. A., & BLAKE, J. G. 1991. Temporal variation in birds and

fruits along an elevational gradient in Costa Rica. Ecology 72:180–

193.

MILLER, C. J. 1990. Natural history, economic botany, and germplasm

conservation of the Brazil nut tree (Bertholletia excelsa Humb and

Bonpl.). M.Sc. thesis, University of Florida, Gainesville.

MORI, S. A. & PRANCE, G. T. 1990. Taxonomy, ecology and economic

botany of the Brazil nut (Bertholletia excelsa Humb. and Bonpl.:

Lecythidaceae). Advances in Economic Botany 8:130–150.

MULLER, C. H. 1981. Castanha-do-Brasil: estudos agronomicos.

Documentos 1:1–25.

MULLER, C. H., RODRIGUES, I. A., MULLER, A. A. & MULLER, N. R. M.

1980. Castanha do Brasil: resultados de pesquisas. Miscelania 2:1–

25.

PERES, C. A. 1991. Seed predation of Cariniana micrantha

(Lecythidaceae) by brown capuchin monkeys in central Amazonia.

Biotropica 23:262–270.

PERES, C. A. 2000. Identifying keystone plant resources in tropical

forests: the case of gums from Parkia pods. Journal of Tropical Ecology

16:287–317.

PERES, C. A. & BAIDER, C. 1997. Seed dispersal, spatial distribution

and population structure of Brazil nut trees (Bertholletia excelsa) in

southeastern Amazonia. Journal of Tropical Ecology 13:595–616.

PERES, C. A., SCHIESARI, L.C. & DIAS-LEME, C.L. 1997. Vertebrate

predation of Brazil-nuts (Bertholletia excelsa, Lecythidaceae), an

agouti-dispersed Amazonian seed crop: a test of the escape

hypothesis. Journal of Tropical Ecology 13:69–79.

262 JOANNE M. TUCK HAUGAASEN ET AL.

PERES, C. A., BAIDER, C., ZUIDEMA, P. A., WADT, L. H. O., KAINER,

K. A., GOMES-SILVA, D. A. P., SALOMAO, R. P., SIMOES, L. L.,

FRANCIOSI, E. R. N., VALVERDE, F. C., GRIBEL, R., SHEPARD, G. H.,

KANASHIRO, M., COVENTRY, P., YU, D. W., WATKINSON, A. R. &

FRECKLETON, R. P. 2003. Demographic threats to the sustainability

of Brazil nut exploitation. Science 302:2112–2114.

SMYTHE, N. 1970. Relationships between fruiting seasons and seed

dispersal methods in a neotropical forest. American Naturalist

104:25–35.

SMYTHE, N. 1978. The natural history of the Central American agouti

(Dasyprocta punctata). Smithsonian Contributions to Zoology 257:

1–52.

STAPANIAN, M. A. & SMITH, C. C. 1978. A model for seed

scatterhoarding: coevolution of fox squirrels and black walnuts.

Ecology 59:884–896.

VANDER WALL, S. B. 1990. Food hoarding in animals. University of

Chicago Press, Chicago. 445 pp.

VANDER WALL, S. B. 1991. Mechanisms of cache recovery in yellow

chipmunks. Animal Behavior 41:851–863.

VANDER WALL, S. B. 2002. Secondary dispersal of Jeffrey Pine seeds by

rodent scatter-hoarders: the role of pilfering, reaching and a variable

environment. Pp. 193–208 in Levey, D. J., Silva, W. R. & Galetti, M.

(eds.). Seed dispersal and frugivory: ecology, evolution and conservation.

CABI Publishing, Wallingford.

VANDER WALL, S. B. 2003. Effects of seed size of wind-dispersed pines

(Pinus) on secondary seed dispersal and the caching behaviour of

rodents. Oikos 100:25–34.

WADT, L. H. O., KAINER, K. A. & GOMES-SILVA, D. A. P. 2005.

Population structure and nut yield of a Bertholletia excelsa stand in

Southwestern Amazonia. Forest Ecology and Management 211:371–

384.

WATSON, W. 1901. Germination of seeds of Bertholletia excelsa. Annals

of Botany 15:99–102.

WHITE, L. J. T. 1994. Patterns of fruit-fall phenology in the Lope Reserve,

Gabon. Journal of Tropical Ecology 10:289–312.

WILLSON, M. F. 1992. The ecology of seed dispersal. Pp. 61–85 in

Fenner, M. (ed.). The ecology of regeneration in plant communities. CABI

Publishing, Wallingford.

XIAO, Z., JANSEN, P. A. & ZHANG, Z. 2006. Using seed-tagging methods

for assessing post-dispersal seed fate in rodent-dispersed trees. Forest

Ecology and Management 223:18–23.

ZUIDEMA, P. A. 2003. Demography and management of the Brazil

nut tree (Bertholletia excelsa). PROMAB Scientific series 6. Utrecht

University, the Netherlands. 112 pp.