Taı Chimpanzees Use Botanical Skills To Discover Fruit .
Anim Cogn (2013) 16:851–860DOI 10.1007/s10071-013-0617-zORIGINAL PAPERTaı̈ chimpanzees use botanical skills to discover fruit:what we can learn from their mistakesKarline R. L. Janmaat Simone D. BanChristophe Boesch Received: 30 August 2012 / Revised: 28 December 2012 / Accepted: 18 February 2013 / Published online: 11 April 2013Ó Springer-Verlag Berlin Heidelberg 2013Abstract Fruit foragers are known to use spatial memoryto relocate fruit, yet it is unclear how they manage to find fruitin the first place. In this study, we investigated whetherchimpanzees (Pan troglodytes verus) in the Taı̈ NationalPark make use of fruiting synchrony, the simultaneousemergence of fruit in trees of the same species, which can beused together with sensory cues, such as sight and smell, todiscover fruit. We conducted observations of inspections, thevisual checking of fruit availability in trees, and focused ouranalyses on inspections of empty trees, so to say ‘‘mistakes’’.Learning from their ‘‘mistakes’’, we found that chimpanzeeshad expectations of finding fruit days before feeding on it andsignificantly increased inspection activity after tasting thefirst fruit. Neither the duration of feeding nor density of fruitbearing trees in the territory could account for the variation ininspection activity, which suggests chimpanzees did notsimply develop a taste for specific fruit on which they had fedfrequently. Instead, inspection activity was predicted by abotanical feature—the level of synchrony in fruit productionof encountered trees. We conclude that chimpanzees makeuse of the synchronous emergence of rainforest fruits duringdaily foraging and base their expectations of finding fruit on acombination of botanical knowledge founded on the successrates of fruit discovery, and a categorization of fruit species.Electronic supplementary material The online version of thisarticle (doi:10.1007/s10071-013-0617-z) contains supplementarymaterial, which is available to authorized users.K. R. L. Janmaat (&) S. D. Ban C. BoeschDepartment of Primatology, Max Planck Institute forEvolutionary Anthropology, Leipzig, Germanye-mail: karline janmaat@eva.mpg.de; kjanmaat@hotmail.comS. D. BanUniversité de Félix Houphouët Boigny, Abidjan, Côte d’IvoireOur results provide new insights into the variety of foodfinding strategies employed by primates and the adaptivevalue of categorization capacities.Keywords Foraging strategies Fruiting synchrony Frugivores Categorization Pan troglodytesIntroductionRipe fruits are ephemeral. They only appear at certain times inthe year, and when they do, many animals compete over thissweet and energy-rich food (Marriott et al. 1981; Diaz-Perezet al. 2000; Houle et al. 2006). Ripe fruit availability fluctuatesin time, and the percentage of rainforest trees carrying ripefruit can be as low as 0.2 % (Chapman et al. 2005). A lowpercentage of (ripe) fruit in a diet, during such fruit-scarceperiods, is shown to influence life history traits such as waitingtime to conception and breeding activity (primates: Thompsonand Wrangham 2008, rodents: Glanz et al. 1982; Milton et al.2005). These studies suggest that it would pay to discovernewly emerged fruit earlier than other foragers and to be thefirst to feed on it. Fruit-dependent foragers (frugivores), suchas primates, use spatial memory to relocate fruit-bearing trees(primates: reviewed in Janson and Byrne 2007; Zuberbühlerand Janmaat 2010, fruit bats: Holland et al. 2005). However, itis unclear how they discover it in the first place. Fruit treesbecome depleted and new fruiting seasons begin, meaningfrugivores must continuously update their knowledge of thelocations of edible fruit. Since not all rainforest trees carryfruit every year, and sometimes skip one or more years(Chapman et al. 1999; Koenig et al. 2003; Struhsaker 1997;Polansky and Boesch in press), frugivores have to learn, everyyear, which individuals produced fruit and which ones did not.The relatively short duration of fruiting seasons, which can be123
852as short as 2 weeks, in combination with the low density andwide distribution of trees of many primate fruit species,increases the challenge of discovering newly emerged fruit(Anderson et al. 2005; Chapman et al. 1999; Milton 1981;Milton et al. 2005 (pg.19); Vooren 1999).Primates could complement their use of sensory cues withother search strategies that facilitate the discovery of newlyemerged fruit. Since many rainforest tree species have areproductive strategy that causes different trees of the samespecies to fruit simultaneously within a clustered time period(i.e. fruiting season; Chapman et al. 1999; Koenig et al. 2003),primates could use this phenomenon to increase their fruitfinding efficiency (Milton 1980). We expected primates touse the discovery of fruit in one tree as an indicator for thepresence of fruit in other trees of the same species. Afterdiscovering fruit in one tree, they can switch to an ‘‘inspect allstrategy’’ and start approaching and inspecting other trees ofthe same species. Indications for the use of such a synchronicity-based inspection strategy were first found in Japanesemacaques and later in grey-cheeked mangabeys (Macacafuscata, Menzel 1991; Lophocebus albigena, Janmaat et al.2012). In this study, we investigated whether chimpanzeesfollow a similar strategy and what type of botanical information they use while doing so. Not all tree species in thechimpanzee’s territory emerge fruit simultaneously (Boeschet al. 2006; Goné Bi 1999, 2007). The probability that otherrainforest trees carry fruit at the time of fruit discovery canvary extensively and depends on each species’ reproductivestrategy, for example, whether individual trees produce eachyear or at variable times of the year, such as most fig trees(Koenig et al. 2003; Van Schaik et al. 1993). We predictedthat chimpanzees will increase their success rate of inspections by knowing these differences and by especially activating an ‘‘inspect all’’ strategy for those species that havehigh compared to low synchrony levels.We recorded feeding and tree inspection behaviour ofadult chimpanzee females in the rainforest of the Taı̈National Park, Côte d’Ivoire. We analysed the females’inspections of empty trees that did not carry fruit, so to say‘‘mistakes’’. This innovative approach in field researchenabled us to exclude the possibility that inspections wereguided by the use of sensory cues emitted by the fruitthemselves and provided us with unique insights into thebotanical parameters that influenced the females’ expectations about fruit finding.MethodsData collection and analysesWe followed five adult chimpanzee females from 16 April2009 to 30 August 2011 for continuous periods ranging123Anim Cogn (2013) 16:851–860from 4 to 8 weeks, totalling 330 days, within the fruitscarce period of April–August (Anderson et al. 2005).Their territory (south community) was located in thelargest remaining tract (5,363 km2) of primary lowlandrainforest in West Africa: Taı̈ National Park, Côte d’Ivoire(5 500 2000 N, 7 190 1600 W; territory size: 26.5 km2; Boeschet al. 2008; Kouakou et al. 2011; N’Goran et al. 2012).K. Janmaat and S. Ban alternated days following eachfemale from the moment the target female woke up to theevening sleeping nest. They recorded the duration andlocation of each activity using a combination of a G.P.S.(Garmin 60 CSx) and voice recorder. Activities wererecorded using continuous focal sampling (Martin andBateson 2007). We marked all trees in which the targetfemale fed, or for which the crown was inspected, withbrightly coloured paint spray. Inspection was defined as amovement of the target female’s head combined with afixed gaze in the direction of a tree crown (see supplementary materials for video-recordings of inspections). Themajority of recorded inspections occurred after the femalecame to a halt (95 %). We tested for potential observerdifferences in the recording of inspections (Kappa 0.7–0.8; Martin and Bateson 2007) and controlled for thisin the final statistical model. The next day trained assistantsrelocated the marked tree, identified the species and measured whether fruit was present, by checking its crownfrom all wind directions using binoculars. Kappa coefficients for agreements on species identity ranged between 1and 0.99 (N 81). Observers were unaware of the synchrony level of the inspected trees at the moment of datacollection. From the analyses we excluded inspections (1)for which the gaze was not clearly directed at one singletree crown (e.g. distant inspections), (2) of trees belongingto species for which the fruit was only eaten on the ground,(3) for sleeping locations, which occurred after the femaleshad emitted a nest grunt (Nishida et al. 2010), (4) duringwhich monkeys or other chimpanzees were present in thetree (e.g. during hunting) and (5) that occurred prior tofeeding on food that grew in the inspected tree.To exclude the use of sensory cues as an alternativeexplanation for the observed behavioural patterns, we ran ourcomparative tests using inspections of empty trees only (treesthat were not carrying ripe nor unripe fruit). Even in highlysynchronous species, individual trees can fail to produce(Polansky and Boesch in press; supplementary materials).We therefore expected chimpanzees to make ‘‘mistakes’’,that is, to inspect trees that had in fact an empty crown. Toavoid pseudo-replication, we only considered first observedinspections to the marked trees in our analyses. To calculatethe synchrony level of each food species, we used phenology(tree life cycle) data collected monthly on 173 individual treesof 16 species from January 2001 and February 2008 (seeAnderson et al. 2005 and Boesch et al. 2006 for a detailed
Anim Cogn (2013) 16:851–860description of the data collection). All trees were locatedwithin the female’s territory. Synchrony, defined as thesimultaneous production of fruit in tree individuals of thesame species within clustered time periods (fruiting seasons),was measured as the average of all Spearman rank correlationcoefficients that could be calculated for the fruiting state of allpossible pairs of trees within a species (Bjørnstad et al. 1999;Buonaccorsi et al. 2001; Koenig et al. 2003). When all treeindividuals had the same fruiting state in each month, meanrho Spearman rank correlation coefficients were equal to 1and a species was defined as having the highest synchronylevel. Low synchrony levels were calculated if, for example,not all trees carried fruit within a fruiting season or whensome trees emerged fruit in other months (e.g. Ficus sansibarica). Rho was calculated by comparing in pairs of trees the(1) absence/presence of ripe fruit (synchrony level A) and (2)amount of ripe fruit scored using the relative ranks: 0, 1, 2, 3or 4 (synchrony level B). Rank 1, 2, 3 and 4 corresponded to1–25 %, 26–50 %, 51–75 % and 76–100 % of the branchesobserved to bear fruit, respectively (see Chapman et al. 1992;Anderson et al. 2005). Since chimpanzees were suggested toalso consider the amount of fruit in trees (Normand et al.2009), we tested the effect of both synchrony values in thefinal model. Calculations of both types of synchrony levelswere conducted by Leo Polansky (unpublished data). Toestimate the density of fruit-bearing trees in the territory, wemultiplied the proportion of trees in the phenology transectthat carried fruit within the month of observation with the treedensity of each respective species in the territory. Tree density was measured by Zorro Goné Bi placing five long parallelbelt transects of 4,000 9 10 m and 3,000 9 10 m, in thenorth–south and east–west direction, respectively, within thefemales’ territory. Each transect was placed 500 m apart andcontained 200 and 150 quadrants of 10 9 20 m for north–south and east–west transects, respectively, in which thedensity of all trees with a diameter at breast height C10 cmwas recorded (Anderson et al. 2002; Goné Bi 2007).Statistical analysesWe analysed our results by running Wilcoxon matchedpairs tests and generalized linear mixed models (GLMM;Baayen 2008) in R (version 2.12.2, R Development CoreTeam 2012) using the function lmer provided by the Rpackage lme4 (Bates et al. 2011). For the Poisson model, wechecked for absence of over-dispersion and found nodeviation from the assumption that the residuals werePoisson distributed (v2 125.16,df 39, P 0.96, dispersion parameter 0.65). To check the overall significance of all predictor variables, we ran likelihood ratio testscomparing the full models with the respective null models.We only considered the effect of the individual predictors ifthe full model reached the significance (Forstmeier and853Schielzeth 2011). To create stable models, we transformedthe predictors in such a way that they resembled a roughlysymmetric distribution, prior to running the models. For thiswe log transformed the feeding duration and transformedthe number of inspected fruit-bearing trees and the estimated density of fruit-bearing trees to the third and fourthroot, respectively. After this we z-transformed all mainpredictors to establish comparable estimates. We checkedfor co-linearity by inspecting variance inflation factors(VIF) derived from a multiple regression with the randomeffects excluded (using the function ‘‘vif’’ of the R package‘‘car’’ (Fox and Weisberg 2011)). This did not indicate anyco-linearity problems with the largest VIF 1.17 in allmodels. We assessed the models’ validity by comparing theestimates derived by a model based on all data with thoseobtained from a model with data points dropped one by one,which indicated that both models were stable. In the finalmodel we included two autocorrelation terms to test fortemporal autocorrelation of the inspections. To derive theautocorrelation terms, we first calculated the residuals of thefull model and second, separately for each data point,averaged the residuals of all other data points from eitherthe same respective individual (term individual) or the samerespective species (term species). The contribution of theresiduals to these averages was weighted by the time lagbetween the particular data point and the others. We modelled the weight functions as a Gaussian distribution with amean of zero (maximum weight at time lag 0). Its standard deviation was obtained by maximizing the likelihoodof the full model with both autocorrelation terms included.Since there is some uncertainty about the validity of P values of fixed effects in the framework of GLMMs (Bolkeret al. 2008), we additionally tested the effect of (1) thenumber of fruit-bearing inspected trees on the number ofempty inspected trees, controlling for feeding duration and(2) the synchrony level on the inspection probability, controlling for the estimated density of fruit-bearing trees in theterritory, using a partial rank correlation permutation test,programmed by R. Mundry in Visual Basic. We controlledfor multiple testing (three likelihood tests and two Wilcoxonmatched paired tests) by using the Fisher’s Omnibus test(Haccou and Meelis 1994) which revealed an overall significant P value (v2 111.32, df 10 and P \0.0001).All tests were two tailed.ResultsIs fruit discovery followed by increased inspection?To investigate whether chimpanzees use fruiting synchrony, we first recorded the frequencies of the females’feeding and inspection behaviour and tested whether the123
854Ficus sansibaricaFicus umbellataFicus saussureanaFicus ottoniifoliaFicus politaFicus luteaFicus kamerunensisFicus elasticoidesEribroma oblongumLandolphia foretianaIrvingia grandifoliaMyrianthus arboreusDaniellia oblongaGrewia malacocarpaDuguetia staudtiiGarcinia kolaMusanga cecropioidesScottelia klaineanaScytopetalum tieghemiiNauclea diderrichiiErythroxylum mannii6050cumulative number of new inspected treeFig. 1 Chimpanzee femalesincreased inspection of trees(full or empty) after feeding ontrees of the same species. Dayzero represents the day of firstobserved feeding on each fruitspecies. To make data fromeach species visible, wesummarized the cumulativenumber of inspections perspecies and averaged thenumber of new inspectionsperformed by the differentfemales when the days after orbefore first feeding overlapped.To show at which point in timeeach fruit observation periodstarted and ended, we extendedthe Y-axis below zero. However,all Y-values lower than zeroshould be considered equal tozeroAnim Cogn (2013) 16:851–860403020100-10-40-30-20-10010203040days before and afer first observed feedingdiscovery of edible fruit at the start of a season was followed by increased inspection of trees of that same species.The target females fed on an average of 7.14 trees(SD 3.89, range: 1–21) and 4.03 species (SD 1.66,range: 1–9) each day, of which 4.63 trees (SD 2.88,range: 0–16) and 0.57 species (SD 1.15, range: 0–7)were not revisited and were new within each observationperiod. They inspected 4.97 trees (SD 4.75, range: 0–24)and 3.31 species (SD 2.65, range: 0–14) per day, ofwhich 4.42 trees (SD 3.97, range: 0–22) and 0.47 species (SD 0.94, range: 0–5) were new. Although targetfemales already started inspecting trees before we sawthem eat fruit of that same species, they increasedinspection after the first observed moment of feeding(Fig. 1). When we excluded the use of sensory cues byonly considering inspections of empty trees, femalesinspected trees significantly more after than before they123were first observed to feed on fruit of the same species(Wilcoxon paired signed rank (exact): T? 140,P 0.0013, Nb of species inspected 17 (4 ties)). Tomake a valid comparison, we only considered fruit speciesfor which the first feeding observation took place at least1 week before the end and at least 1 week after the start ofthe observation period (N 17). To verify whether thefemales indeed already started inspecting trees before theyhad been feeding on fruit from the same species, and to seewhat could have triggered this, we separately analysedfemale follows consisting of 16–44 consecutive days.Within these unique consecutive follows, the observerswere able to follow the females without interruption andwe were thus certain that the females had not yet fed on theconcerned fruit species. We confirmed that females wereindeed inspecting the fruiting state of trees days before theyfed on fruit of that same species (Fig. 2). In addition, the
Anim Cogn (2013) 16:851–860855Table 1 Effect of the number of inspected fruit-bearing trees andfeeding duration on the number of empty trees that the lueP valueIntercept-2.410.30-8.14 \0.0001# Inspected fruit-bearingtrees per day0.540.153.93 \0.0001Feeding duration per day0.210.111.810.0704of fruit-bearing trees but no feeding within the consecutivefollowing periods (Nb 8).Fig. 2 Chimpanzee females inspected (empty) trees before feeding.Day zero represents the day of first observed inspection of a fruitbearing tree. The species-specific differences in the number of the lastanalyzed day are determined by either the day at which the femalesstarted feeding or the day at which observers lost contact with thetarget femalesFig. 3 Chimpanzee females also inspected more trees after thanbefore they were first observed to see fruit. Each circle represents themean number of new empty trees inspected per day of a fruit species.The lines in between represent the differences between the means ofbefore and after the first time the females were observed to see(inspect) the fruit belonging to the same fruit speciesnumber of new empty trees inspected per day was significantly higher after than before they were first observed toinspect fruit-bearing trees (i.e. to see fruit of the samespecies), suggesting that it is not only the taste of fruit thattriggered the chimpanzee’s expectations but also theobservation of the fruit itself (T? 35, P 0.015, Nb ofinsp
Received: 30 August 2012/Revised: 28 December 2012/Accepted: 18 February 2013/Published online: 11 April 2013 Springer-Verlag Berlin Heidelberg 2013 . (Normand et al. 2009), we tested the effect of both synchrony values in the fi
between these two species. Since humans have a more extended ankle compared to chimpanzees at midstance during walking (bipedal and quadrupedal for humans and chimpanzees respec-tively), we postulate that PTO would differ significantly between humans and chimpanzees.
10 Best Botanical Gardens In The World Source: Internet 4. Berlin-Dahlem Botanical Garden and Botanical Museum, Berlin, Germany - Opened to the public in 1910, this botanical garden has an area of 43 hectares and 22,000 plant hespecies. The garden is part of the Free University of Berlin and t Botanical Museum is attached to the garden
HISTORIC The New York Botanical Gardens AND/OR COMMON The New York Botanical Gardens LOCATION CITY. TOWN -VICINITY OF New York COUNTY Bronx STATE New York PHOTO REFERENCE PHOTO CREDIT NEGATIVE FILED AT The New York Botanical Gardens The New York Botanical Gardens DATE OF PHOTO circa 1962; confirmed 1975 IDENTIFICATION DESCRIBE VIEW. DIRECTION. ETC.
4 Evolution! 4 billion years of earth!3.5 billion years of life!650 million years of multi-celled organisms!600 million years of nervous system! 80 million years of mammals! 60 million years of primates! 6 million years ago: last common ancestor with chimpanzees, our closest relative among the “great apes” (gorillas, orangutans, chimpanzees, bonobos, humans)
African Primates 10: 1-12 (2015)/ 1 Characterization of Nest Sites of Chimpanzees (Pan troglodytes schweinfurthii) in Kibira National Park, Burundi Dismas Hakizimana1,2, Alain Hambuckers1, Fany Brotcorne1, and Marie-Claude Huynen1 1Primatology Research Group, Behavioral Biology Unit, University of Liège, Liège, Belgi
She proposed that some of them were cultural in origin. The most conspicu-ous one was nut cracking, which is absent in the Gombe chimpanzees, in spite of the presence of oil-palm nuts. Observations of this behavior were first reported in the 1840s in Liberian chimpanzees.23
the evolution of cooperation. In fact, chimpanzees show important di-vergences from humans in some aspects of their social behavior. Al-though chimpanzees have relatively sophisticated perspective-taking abilities (Call & Tomasello, 2008) and are capable of recognizing cues of need in others (Melis & Tomasello, 2013; Melis et al., 2011;
Botanical name of Ceylon cinnamon Botanical names for plants were developed to identify individual plants scientifically because there were many different common names given to one particular plant. It is also reliable to identify a plant by its botanical name, rather than by its common name. Sometimes there can be
