Instruction manuals

Post-Dispersal Predation and Scatterhoarding of Dipteryx Panamensis (Papilionaceae) Seeds by Rodents In Panama

Post-Dispersal Predation and Scatterhoarding of Dipteryx Panamensis (Papilionaceae) Seeds by Rodents In Panama
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Oecologia (1993) 94:255-261 Oecologia © Springer-Verlag1993 Post-dispersal predation and scatterhoarding of Dipteryx panamensis (Papilionaceae) seeds by rodents in Panama Pierre-Michel Forget* Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Republic of PanamaReceived: 5 July 1992/Accepted: 8 March 1993 Abstract. In tropical rain forests of Central America, thecanopy tree Dipteryx panamensis (Papilionaceae) fruitswhen overall fruit biomass is low for mammals. Flyingand arboreal consumers feed on D. panamensis and dropseeds under the parent or disperse them farther away.Seeds on the ground attract many vertebrate seed-eaters,some of them potential secondary seed dispersers. Thefate of seeds artificially distributed to simulate bat dis-persal was studied in relation to fruitfall periodicity andthe visiting frequency of diurnal rodents at Barro Colora-do Island (BCI), Panama. The frequency of visits byagoutis is very high at the beginning of fruitfall, but inthe area close (< 50 m) to fruiting trees (Dipteryx-rich area) it declines throughout fruiting, whereas it remainsunchanged farther (> 50 m) away (Dipteryx-poor and Gustavia-rich area). Squirrels were usually observed inthe Dipteryx-rich area. Along with intense post-dispersalseed predation by rodents in the Dipteryx-rich area, asignificant proportion of seeds were cached by rodents inthe Dipteryx-poor area. Post-dispersal seed predationrate was inversely related to hoarding rate. A significant-ly greater proportion of seeds was cached in March,especially more than 100 m from the nearest fruiting tree.This correlates with the mid-fruiting period, i.e. duringthe height of D. panamensis fruiting, when rodents seemto be temporarily satiated with the food supply at parenttrees. Hoarding remained high toward April, i.e. late inthe fruiting season of D. panamensis. Low survival ofscatterhoarded seeds suggests that the alternative foodsupply over the animal's home-ranges in May-June 1990was too low to promote survival of cached seeds. Seed-lings are assumed to establish in the less-used area of therodents' home-range when overall food supply is suf-ficient to satiate post-~tispersal predators. Key words: Dipteryx panamensis - Dasyprocta punctata- Sciurus #ranatensis - Post-dispersal seed predationScatterhoarding * Present address: Laboratoire d'Ecologie Grnrrale, Musrum Na-tional d'Histoire Naturelle, 4 Avenue du Petit Ch~teau, 91800Brunoy, FrancePredator satiation and dispersal away from a parent treecrown have been proposed as selective factors limitingthe destructive abilities of animals and promoting thesurvival of dispersed seeds (Janzen 1970, 1971, 1978;Smythe 1970; Howe and Smallwood 1982; Clark andClark 1984). There have been recent advances in knowl-edge of the fate of large seeds in some animal-disperseddominant species in tropical forest habitats (e.g. DeSteven and Putz 1984; Howe et al. 1985; Dirzo andDominguez 1986; Estrada and Coates-Estrada 1986;Hallwachs 1986; Sork 1987; Schupp 1988a, b, 1990;Forget and Milleron 1991). However, we still know littleabout the complex interaction between seed dispersaland post-dispersal seed-predation throughout seedlingestablishment, especially when seed removal by mam-mals masks secondary seed dispersal by rodents. Second-ary seed dispersal by New World rodents has not beenexplored until recently (Forget and Milleron 1991),though it may be crucial in determining the ultimatelocation of seedlings in the most suitable habitat for theirgrowth (see Schupp et al. 1989).In Central America, the seeds of Dipteryxpanamensis (Papilionaceae) are dispersed by large frugivorous batsand suffer high post-dispersal predation by terrestrialmammals including scatterhoarders (Bonaccorso et al.1980; Glanz et al. 1982; Smythe et al. 1982; De Stevenand Putz 1984). The recruitment pattern of D. panamen-sis agrees with Janzen's (1970) and Connell's (1971)hypotheses that survival rates of seedlings decrease withincreasing density and with proximity to a conspecificadult (Clark and Clark 1984). Fast growth in gaps pro-motes long-term seedling survival and compensates forthe intense herbivory observed in the understory whereseedlings often establish (De Steven and Putz 1984;Clark and Clark 1985, 1987; De Steven 1988; see alsoColey 1988). The farther seeds are dispersed, the lowerthe conspecific seedling density. Therefore, it is less likelythat predation will occur, and seedlings will be mostlikely to survive farther from the nearest fruiting adult ina newly-formed opening (Hubbell and Foster 1986).Given the habit of bats of carrying and dropping seedsbelow feeding-roasts (Janzen et al. 1976; Morrison 1980;  256 Howe 1986), dispersal of D. panamensis seeds in openhabitats is unlikely (see Schupp et al. 1989); therefore,seed survival in the shaded understory is essential.As in some other species whose seeds are dispersed bymammals and consumed either by insects or by rodents(e.g. Wilson and Janzen 1972; Janzen et al. 1976; Wright1983 ; Howe et al. 1985; Hallwachs 1986; Traveset 1990;Forget and Milleron 1991), D. panamensis seeds that arelong-distance dispersed can still be subject to high preda-tion by mammals in moderately undisturbed forests (DeSteven and Putz 1984). Besides, the large amount ofinsect-free D. panamensis seeds that fall below the fruit-ing trees could constitute an abundant food supply thatmay satiate predators and decrease seed removal, allow-ing establishment of seeds. Among seed-eaters, satedrodents could scatterhoard seeds when they find themin excess (see Vander Wall 1990) under fruiting tress(Bonaccorso et al. 1980) as well as at bat feeding-roosts.To some extent, rodents may "kidnap" seeds from otherpost-dispersal predators, and, if they abandon or forgettheir cached food items, increase survival of seeds andseedlings growing from them (Forget 1990; Forget andMilleron 1991). The probability that D. panamensis seedsand germinating seedlings escape post-dispersal preda-tion could depend on the complex interplay betweenseasonal change in community-level seedfall, density andforaging behavior of rodents in microhabitats whereother animals are primary seed dispersers (Janzen 1971 ;Hallwachs 1986; Sork 1987; Schupp 1990; Vander Wall1990; Forget and Milleron 1991).The objective of this study is to establish the selectivefactors that determine where and how D. panamensis seeds may survive and germinate away from the tree. Inthis paper I address the following questions: (a) Doesscatterhoarding prevent D. panamensis seeds from beingdestroyed by post-dispersal predators away from theparent tree? (b) How do post-dispersal predation, andprobable scatterhoarding, fluctuate during D. panamen-sis fruiting? (c) How do rodents interact with bats so that D. panamensis seeds and germinating seedlings can suc-cessfully be established away from the parent tree? (d)How does frequency of diurnal rodents at areas near andaway from D. panamensis fruiting trees relate to thefrequency of seed predation away from parent trees ?To answer these questions, I investigated the fate of D. panamensis seeds artificially dispersed in clumps away(> 50 m) from the nearest fruiting tree (to simulate seeddispersal by bats), during three contrasting periods of D. panamensis fruiting, in relation to rodent visitation atdifferent forest microhabitats in Panama. Methods Study site The study was carried out on Barro Colorado Island (BCI), Panama(9°09'N, 79°51'W). The climate of BCI is seasonal with a well-marked dry season, of variable length and intensity from Decemberor January to April (Leigh et al. 1982). The tropical moist forest(2616 mm rainfall/year; Rand and Rand 1982) includes youngstands about 100 years old and primary forest several centuries old(Foster and Brokaw 1982; Piperno 1990). There are two peaks infruiting: in April-May (wind- and animal-dispersed species) andagain in September-October (animal-dispersed species) (Foster1982b). Plant species nomenclature follows Croat (1978). Study species On BCI the duration of D. panamensis fruiting (vernacular name:almendro fruit) is variable and lasts from late December to earlyApril [Smithsonian Environmental Science Program (ESP),D. Windsor pers. comm.] when overall fruit biomass is low (Smythe1970; Bonaccorso et al. 1980). The number of D. panamensis treesbearing fruits and the abundance of their crops varies from year toyear, sometimes with direct effects on animal populations (Foster1982a; Giacalone-Madden et al. 1990). In 1990, the crop of D. pa-namensis was delayed by 1 month relative to other years (Smith-sonian ESP, D. Windsor, pets. comm.) as a consequence of lateflowering in 1989 (K. Kitajima, pers. comm.). Fruits are largeelliptical almond-like green drupes 5-6 cm long and 2-3 cm wide,weighing 18-26 g (Bonaccorso et al. 1980), that are exposed outsidethe crown. These features fit the bat seed dispersal syndrome (Pijl1972; Howe 1986). Such fruits are likely to be removed by largefrugivorous bats (weight > 50 g) such as Artibeus lituratus, A. ja-maicensis (Stenodermatinae) or Phyllostomus hastatus (Phyllosto-minae) (Bonaccorso et al. 1980; see also Emmons and Feer 1990).Fruits are carried to a feeding roost under palm leaves or in densevegetation (Janzen et al. 1976; Morrison 1980), the juicy pulp(5-6 mm thick) is consumed and the seeds are dropped, and thusdispersed (see also Howe 1986). Arboreal mammals forage inten-sively and drop many fruits and seeds which may then be destroyedby terrestrial mammals such as peccaries (Tayassu tajacu), agoutis (Dasyprocta punctata), red-tailed squirrels (Sciurus 9ranatensis) and, possibly, spiny rats (Proeehimys semispinosus) (Hladik andHladik 1969; Bonaccorso et al. 1980; Smythe 1970, 1978; Heaneyand Thorington 1978; Smythe et al. 1982; De Steven and Putz 1984;Glanz 1984; Giacalone-Madden et al. 1990). The consumed part isthe embryo which accounts for 2 g, on average, of the fruit weight(P.-M. Forget, pers. obs.). Study areas The study area is located on the R.C. Shannon trail between the100 m and 700 m trail markers (Fig. 1). According to Foster andBrokaw (1982), the vegetation is old forest, with young forestnearby. A 1-ha area (Dipteryx-rich area) was delimited by fiveparallel trails 100 m long and 25 m apart ("Virola-poor area" inForget and Milleron 1991, and "Gustavia-poor" in Forget 1992). Itcontained relict D. panamensis tree; this area had been cleared,judging from the candelabra-shaped trunks. For logistic reasons,the trails were arranged so that the central trail extended betweentwo reproductive trees (A and B) 109 m apart. Three other adulttrees (diameter at breast height, dbh > 30 cm) occurred on the areabut bore no fruit in 1990. A third reproductive tree (C) grew 15 moutside the area. One of the five trails (ca~ossed a gap which was notcensused (due to obstruction by br~ches) and the total length ofthe five trails was 487 m. The spa~ng of trails allowed unrestrictedsurveillance of the area (see below: Diurnal rodent sightings). Asecond 1-ha area > 100 m away f~ D. panamensis trees and withan abundant population of Gustavia superba (Lecythidaceae), the "Gustavia-rich area" in Forget (1992), was also delimited (Fig. 1).In addition, I also choose two portions of 350 m and 150 m of RCShannon trail (Dipteryx-poor area) located > 50 m away from D. panamensis trees (Fig. 1). Overall seed predation From 30 November 1989 to 20 May 1990, the overall number ofseeds that retained a hard husk and debris after being gnawed that  Dipterr x- poor area _ ~//~c. . Gustavia ich area / Dipteryx. ich area loo m Fig. 1. R.C. Shannon trail on Barro Colorado Island, Panama,showing the three study areas. Empty and full large stars representunreproductive (but mature) and fruiting Dipteryxpanamensis treesin 1990; letters efer to 1990 fruiting trees in the Dipteryx-rich area; squares refer to the 16 seed fate experimental sites; the small blackstar represents the 10 artificially cached seeds that survived through-out seedling establishment in August (see text)were found on the network of trails were counted biweekly in Dipteryx-rich and Gustavia-rich areas (Fig. 1). Diurnal rodent sightings From 30 December 1989 to 8 May 1990, monthly strip censuses of D. punctata and S. granatensis were conducted in the three studyareas (Fig. 1) over 2 weeks at start of each 4-week period. Eachcensus consisted of regular one-way walks (n = 10 replicates) of30.45 rain on trails, beginning at 0630.0700 h; the five 100-m trailswere considered as one trail in Dipteryx-rich and Gustavia-rich areas. Although some errors may arise when foraging individualsare counted twice in the same area, this method gives a good relativeestimation of rodent frequency in each forest area (see Eisenbergand Thorington 1973; Cant 1977). Both the surrounding understoryand the crown of the three reproductive D. panamensis trees weresurveyed. Because litter fall is high while D. panamensis is fruiting,trails were raked before each census period, thus permitting a silentwalk and minimal disturbance of rodent activity. Seed removal and seed fate Seed removal and the fate of seeds after artificial long-distancedispersal (> 50 m) were studied in the Dipteryx-poor area (Fig. 1).Seeds were collected under other fruiting trees (> 500 m from studyarea) where monkeys dropped seeds after consumption of the fleshyexocarp. Holes 1.5 mm in diameter were drilled through the woodyendocarp and a white thread (0.3 mm in diameter and 60 cm long)was attached to the seed (Forget 1990). This marking method doesnot affected animal behaviour or seedling viability, germination orestablishment. Two clumps of ten seeds were placed 3-5 m apart at100-m intervals at 0630-0700 hours, seeds being placed 25 m fromthe previous set on each of 4 consecutive days. A total of 16 depotsof 20 seeds (two sets of ten seeds 3-5 m apart) were thus placed perexperiment (under palms when possible), eight at 50.100 m and 8over 100 m from the nearest fruiting tree. Some of these sites werepresumably used by bats, judging by the various scraped fruits andseeds found below them.Seeds were regularly censused before dawn and after sunset todetermine whether or not removal occured at night. The clumpedseeds were surveyed daily at 1100-1400 hours and the thread-marked seeds were sought within 20 m of the experimental baits(depots). Caches were marked with a plastic white label 16 cm long257fixed to the ground with a metal stalk, and followed until fullyestablished (mid-June). Each experiment was repeated three timesat each site 4 weeks apart; 24 January (early fruiting period), 21February (mid-fruiting) and 21 March (late-fruiting). Survival of artificially cached seeds Because few naturally cached seeds were available after animalsretrieved them, I decided to analyse survival of seeds artificiallycached late in the fruiting season. This experiment is assumed toreflect survival of seeds that would have been cached betweenFebruary and April, and remained unretrieved by rodents as lateas April. Seeds were collected in March and conserved in a trap inthe field. By late April some had already germinated and had aradicle 0.5-2.0cm long. By 23 April, when D. panamensis hadfinished fruiting, both kind of seeds (germinated and ungerminated)were buried 1-2 cm deep and covered with soil and leaves, as doneby agoutis (Smythe 1978, 1989). Paired seeds were buried 1 m apartat 5-m intervals (160 seeds: 2x 80 seeds at 50-100 and > 100m,respectively) in the Dipteryx-poor area. Emerging seedlings wereregularly censused in mid-June, after the cotyledons dropped (lessthan 2 months after burial), and again in August (after about 14weeks). Data analysis Comparison of the mean number of rodent sighting counts weretested with one-way ANOVA for the effect of month (Januarythrough May). The seed fate experimental design conforms to atwo-factor repeated-measures split-plot design (Kirk 1969). Datafor the percentage of seeds consumed or scatterhoarded wereanalysed with a fully crossed ANOVA with two fixed effects of"distance" (50-100 m, more than 100 m) and "fruiting" (early, mid,late), and a random effect of depot nested within "distance". Theeffects of treatment "distance" are described as between-block (sub-ject) effects, while the effects of treatment "fruiting" and interaction"distance x fruiting" are described as within-block effects. By incor-porating the random effect [depot(distance)I, the appropriate errorterm was used for calculation of the F-statistic. Rodent counts weresquare-root transformed and the percentage of seeds consumed orscatterhoarded was arcsine-root transformed to normalize the data.A two-way contingency table of survival of artificially cached seedsas a function of distance to the nearest fruiting trees was analysedusing G-test of independence (Sokal and Rohlf 1981). Results Overall seed predation Of seed predation observed in D. panamensis (n = 366seeds collected on trails) 56% occurred during the first 4weeks (February) of the fruiting period (14 weeks). Bythe end of March, no fruits remained intact under treeA. By April, the fallen crop under tree B was totallydepleted. During the 6-month study, D. panamensis ac-counted for 92% overall of the totally destroyed seeds(n = 398) found on the trails in the Dipteryx-rich area.The remainder were mostly seeds of Astrocaryum stand-leyanum (3.8 %) and ScheeIea zonensis (2.8 %). In contrast,in the Gustavia-rich area, three D. panamensis seeds werefound gnawed, in February only. Rodents rely upon Gnetum leyboldii (Gnetaceae), S. zonensis and A. stand-leyanum which accounted for 43 %, 31% and 11%, respec-tively, of all gnawed seeds (n = 83) recorded. Other less  258 e,,luo ~g o z 13,12, 11• 10.9- 8- 7.6. I s: 3-2.1, i|! Jan I A i i i B Ito, r o I-lo i I I i Feb Mar Apr May Fig. 2A-C. Mean number (+ 1 SD) of Dasyprocta punctata (emptybars) and Seiurus 9ranatensis (hatched bars) observed duringmonthly censuses (n= 10 per month) on the three study areas(A Dipteryx-rich; B Dipteryx-poor; C Gustavia-rich) from Januaryto May 1990 Table 1. Seed removal rate and fate of marked Dipteryxpanamensis seeds within 28 days after seed placement at R.C. Shannon trail (Dipteryx-poor area) on Barro Colorado Island, Panama, by fruit-ing period (early, mid and late) and distance treatment (near:50-100 m from the nearest D. panamensis fruiting tree; far:> 100 m)Fruiting Early Mid LateDistanceNear Far Near Far Near Far% Removed 96.2 100 93.1 98.1 99.4 100(SD) (6.9) (14.4) (5.3) (1.8)% Gnawed 76.6 80.6 65.5 26.2 74.3 66.3(SD) (22.1) (13.2) (26.2) (19.0) (19.9) (31.1)% Cached 8.7 6.8 10.6 41.5 11.9 18.8(SD) (24.8) (11.0) (12.9) (15.6) (17.1) (19.0)Percentage (SD) of gnawed and cached seeds (retrieved within 20 mfrom the srcin) are of the number of seeds removedTable 2. ANOVA results of percentage of seed predation and scat-terhoarding of D. panamensis for the effect of distance treatment(50-100 m and > 100 m from the nearest fruiting tree) and fruitingperiod (early, mid and late) at 28 days after seed placementSource df Predation Scatterhoardingof variationMS F MS FBetween blockDistance 1 1080.1Depot (Distance) 14 247.5Within blocksFruiting 2 2244.8Distance x Fruiting 2 981.6Fruiting x Depot 28 293.2(Distance)4.367.65**3.34*1657.0 4.94*335.11380.5 7.26**663.2 3.49*190.0*, P<0.05; **, P<0.01Data were arcsine-square root transformedP=- 0.63) nor at Gustavia-rich area (F4,45 = 1.85, P= 0.13)(Fig. 2).frequently consumed items were hoarded seeds such as Gustavia superba, Entada gigas (Mimosaceae) and Hyle-naea praecelsa (Hippocrateaceae) between Decemberand February, or freshly dispersed seeds such as Anacar-dium occidentale (Anarcadiaceae) in May. Diurnal rodent sightings Agoutis accounted for 75%, 90% and 91% of all rodentssighting counts in the Dipteryx-rich (n = 268), Dipteryx- poor (n = 95) and Gustavia-rich areas (n = 75), respective-ly, squirrels being particularly rare at the two other areas(Fig. 2). Months differ significantly in the number ofagoutis observed at Dipteryx-rich area (F4,4s = 17.59,P< 0.001) but neither at Dipteryx-poor area (F~,45 = 0.64, Seed removal and seed fate Virtually all seeds were removed by the end of eachexperiment (Table 1). The number of remaining seedswere 6 (2%), 14 (4%) and 1 (0.3%) in each experiment,respectively• Of removed seeds, the average percentage ofseeds cached by rodents (8 %, 26 % and 15 %) was inverselycorrelated with that of seeds found gnawed (79 %, 46 %and 70%) (Table 1). The group of"lost seeds" (14%, 28%and 15%) includes seeds that may have been carried> 20 m away as well as seeds in arboreal caches (severaloccurences observed at 1-2 m above ground). Seeds werehidden at the base of trees, especially against buttresses,at the base of palms, near logs and at the base of lianas[see also Smythe (1978); Heaney and Thorington (1977)].Distance had no effect on level of seed predation butthere was an interaction with fruiting period; in mid-  259fruiting period the proportion of seeds consumed wasgreater at 50-100 m than at over 100 m away from thenearest parent tree, whereas there was no difference in thetwo other periods (Tables 1 and 2). Both distance andfruiting period had some effect on the proportion of seedsscatterhoarded, and there was an interaction betweenthose two factors; hoarding was significantly greater atover 100 m than at 50-100 m, and was greater duringmid- and late-fruiting periods (Tables 1 and 2).Most seeds were recovered rapidly by animals, on thesame day or within 1-3 weeks, and none remained cachedfor more than 1 month after the February and Aprilexperiments. In contrast, some seeds and germinatingseedlings (2%) arising from the March experiment werestill alive toward the end of April, but none survived tomid-June, being cut after cotyledons dropped. Survival of artificially cached seeds A greater proportion of artificially cached seeds ger-minated at 50-100 m (49%) than at over 100 m (29%)from fruiting trees (G-test=6.80, df=l, P<0.01). Ofthese germinating seeds (n = 62), 48 successfully estab-lished but suffered many traumas within 2 months. Seed-ling mortality was essentially due to digging up andconsumption of cotyledons and terminal meristem bygrazing animals. Seedlings already independent of coty-ledons may also be killed by digging animals. By mid-August, there were only 10 healthy seedlings with fullydeveloped leaves (see Fig. 1), sometimes with little her-bivory. There is a slight trend for early overall survivalof seeds and young seedlings to be greater at 50-100 m(8 live seedlings) than farther away (2 live seedlings)(G-test=4.09, df= 1, P<0.05). Discussion While overall food supply is low on BCI, sporadicallyfruiting Dipteryx panamensis trees attract rodents be-tween February and April. As soon as D. panamensis fruits are dropped by arboreal/flying mammals in Febru-ary, agouti crowding is regular in the Dipteryx-rich area;for instance, up to 13 agoutis were counted around treeA in February. The later decrease in rodent visitation inthe Dipteryx-rich area reflects two effects: first, the cum-ulative effects of the increasing number of fruiting D. pa-namensis trees both at the study area and around it;second, the progressive increase of community-wide foodsupply towards April (Smythe 1970; Foster 1982b). Thelack of squirrels in the Dipteryx-poor and Gustavia-rich areas is likely to be correlated with lack of D. panamensis food supply, which account for 78 % of the squirrels' diet(Glanz 1984; Glanz et al. 1982; Giacalone-Madden et al.1990). Seed fate away from parent trees might be mostlydependent on agoutis.Almost 100 % seed removal during the late part of theexperiment suggests that, whatever the amount of timeagoutis spent in the Dipteryx-rich area between Januaryand April, rodents forage regularly areas away from D. panamensis trees. There is thus a high probability thatall spaces would be foraged, and that seeds remaining onground throughout April would be encountered byagoutis. This is even more likely given the high popula-tion of this rodent on BCI (see Glanz 1991).Evidence for D. panamensis seed scatterhoarding isconsistent with the general observation that "...foodhoarding often appears to be associated with situationsin which food is available in excess..." followed byperiods of food scarcity (Vander Wall 1990). The in-cidence of scatterhoarding in D. panamensis appears tobe negatively correlated with that of post-dispersal seedpredation. Hoarding of D. panamensis seeds by agoutismay be a short-term regulated mechanism which seemsto result from a temporary satiating effect followingfeeding. The more often rodents forage at fruiting treesand the closest areas (<50 m), where most seeds arelikely to be dropped by arboreal mammals and dispersedby bats (see Janzen et al. 1976), the more likely they willbe satiated before moving farther away (> 50 m) overtheir regular territories. The greater proportion of seedscached farther than 100 m from the tree suggests that thislevel of satiation is not evenly distributed away fromfruiting trees.Moving from one food source to another after ex-hausting the D. panamensis seed shadow, agoutis con-sume D. panamensis germinating seeds that they havepreviously hoarded. Naturally cached seeds were thusrapidly used by rodents, and almost none remainedcached until the end of the fruiting period in 1990.Hoards of D. panamensis seeds were used on a short-termscale (1-2 months) when overall food supply was increas-ing. Hoarding D. panamensis seeds may be an adaptationof agoutis to compete with other potential post-dispersalpredators such as squirrels and peccaries (see Smythe1986), and to allow short-term feeding on a daily ormonthly cycle when food supply decreases.The two instances of established seedlings in May-June arising from seeds cached in March, however, sup-port the possibility of efficient secondary seed dispersalby rodents in D. panamensis. The zero survival in thisstudy is probably due to the combined effects of the smallsample of seeds used on the trail in comparison to anatural seed shadow, to a lower number of fruiting treesin 1990 (P.-M. Forget, pers. obs.) compared with otheryears (ESP, D. Winsor, pers. comm.) as well as to per-manent dense populations of grazing animals (rodents,deer, peccaries) on BCI (Glanz 1991). Simulating intensepost-dispersal survival in caches through April (experi-mental seed caching) confirms, however, that buriedseeds can establish and survive throughout August.During this study, in a year with a little production,distance from parent tree did not seem to much increasethe likelihood of escaping predation. Despite greaterhoarding farther away than 100 m, survival of artificialcached seeds was greater within 50-100 m. Beside, some1990 seedlings were observed in the Dipteryx-rich area inJune between 10m and over 50 m from the nearestparent, attesting some natural recruitment in 1990 closeto fruiting trees, due to scatterhoarding near the parenttree in March-April (P.-M. Forget, pets. obs. ; see also
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!