the epiphyte vegetation of the palm socratea exorrhiza

8/6/2019 The Epiphyte Vegetation of the Palm Socratea Exorrhiza

1/11

The Epiphyte Vegetation of the Palm Socratea exorrhiza: Correlations with Tree Size, TreeAge and Bryophyte CoverAuthor(s): Gerhard Zotz and Birgit VollrathSource: Journal of Tropical Ecology, Vol. 19, No. 1 (Jan., 2003), pp. 81-90Published by: Cambridge University PressStable URL: http://www.jstor.org/stable/4091827

Accessed: 24/09/2009 19:30

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=cup.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access toJournal of

Tropical Ecology.

http://www.jstor.org
http://www.jstor.org/stable/4091827?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=cuphttp://www.jstor.org/action/showPublisher?publisherCode=cuphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/4091827?origin=JSTOR-pdf


2/11

Journal f Tropical cology 2003)19:81-90.Copyright@ 2003Cambridgeniversity ressDOI: 0.1017/S0266467403003092rintedn the UnitedKingdom

Theepiphyteegetationfthepalm ocrateaxorrhizacorrelationsithree ize,treeageandbryophyteoverGerhardotz*t' ndBirgit ollrath**BotanischesnstituterUniversitatasel,Sch6nbeinstrasse, CH-4056asel, witzerland.tSmithsonianropicalesearchnstitute,pdo 072,Balboa,anama(AcceptedthJanuary002)

Abstract:We conducteda surveyof the epiphytefloragrowingon the stiltpalmSocrateaexorrhiza n a primaryowlandrain forest in Panamaby means of a canopy crane.For each palm in a 0.9-ha plot, we determineddiameterat breastheight, treeheight, per cent bryophytecover andthe number, dentityandattachment ite of all vascularepiphytes.The118 palm treeshosteda totalof 701 epiphytesandhemi-epiphytes,belongingto 66 species. Trees were estimatedto bec. 20 y old before colonization with vascularepiphytesbegan. Epiphytespecies were highly clumpedand segregatedalong the verticalaxis of the trunk.Sequentialcolonization ed to an increasednumberof species andindividualsas thetree grows. Epiphyteswere associated with bryophytepatches much more than expected by chance, but no speciesseemed to dependuponthem for establishment.The influence of tree size, age andbryophytecover on the compositionof the epiphyte communityarediscussed.Key Words:Araceae, Bromeliaceae,canopy dynamics, diversity, ferns, hemi-epiphytes,moss, Orchidaceae,vascularepiphytes

INTRODUCTIONThe diversityof vascularepiphytesin tropicalforests canbe impressive:Gentry& Dodson (1987) foundmorethan100 species of epiphytesand stranglers, ully 35% of thetotalvascularflora,in a plot of 1000 m2 in the wet forestof Rio Palenque,Ecuador.Kelly et al. (1994) reportedaneven higherpercentage n anAndean forest in Venezuela:120 of 219 species, i.e. more than 50% were true epi-phytes. A single tree may supportmore than50 differentspecies (e.g. Ingram& Nadkarni1993). However,we stillknow little of the dynamics of epiphyte communities(Benzing 1990).The studyof epiphytecommunityecology is still ham-pered by the difficultyin accessingtheupperstrataof theforest.Consequently,he numberof sampled rees is oftenlow, e.g. one (Pupulinet al. 1995) or threetrees(Freiberg1996). Researchershave frequentlyavoidedthe problemsassociated with tall trees by studying epiphyte distribu-tions in the lower strataof the forest (Kernan& Fowler1995) or in forestsof low stature Zimmerman& Olmsted1992, Zotz et al. 1999). The use of canopy cranesmaynow reduce this problem(Niederet al. 1999).Here, we reportan investigationof communitystruc-

ture and dynamics of epiphytes on one host tree species,the stilt palm Socratea exorrhiza, within a 0.9-ha plot inthe reach of a construction crane on the Caribbean slopeof Panama. This particular study system was selected fora number of reasons. First, few studies can distinguish theeffects of host tree identity and tree size on the epiphytecommunity over the entire size range of a tree species:the limitation to a single host tree with a large number ofsampled individuals allowed this analysis. Second, know-ledge of average host tree growth rates made it possibleto roughly estimate tree age. Considering the almost com-plete lack of long-term observational data (but see Hietz1997), simultaneous side-by-side observations allow us toapproximate the temporal patterns of dynamic processes.Third, the simple architecture of this palm allowed thequantification of the substrate area available for epiphytecolonization, which has rarely been tried before for obvi-ous reasons (see, however, Zotz 1997). Finally, theadverse establishment conditions on a vertical trunk fea-turing a smooth bark with relatively low water-holdingcapacity made this tree suitable to study the importanceof bryophytes for the establishment of vascular epiphytes.Bryophyte mats supposedly play an important role for thedistribution of vascular epiphytes on potential host trees:they hold water, trap seeds, provide anchorage for seed-lings, and intercept nitrate from fog (Clark et al. 1998,Johansson 1974). Although these and other authors

1 Address for correspondence:Botanisches Institut der UniversititBasel, Schonbeinstrasse , CH-4056Basel, Switzerland.Email:[email protected]


3/11

82 GERHARDOTZAND BIRGITVOLLRATH

(Bennett 1987, Freiberg 1999, Hietz & Hietz-Seifert1995b) reiterate he importanceof mosses and liverwortsfor vascularepiphytes,documentation f cause andeffectis rare(Laman 1995).METHODSThis studywas carriedout in the late-1999 dry season atthe Fort ShermanCanopy Crane site, which is locatednearthe Atlantic coast of the Republicof Panama.Aver-age annual rainfall is about 3500 mm (Lerdau& Throop1999). Canopy height of this primarylowland forest isquitevariable,reachingmaxima of c. 40 m. The crane is52 m high and has a radial engthof 54 m, thus coveringc. 9000 m2 by means of a small gondola.Within the reach of the crane we located all individualsof the stilt palm Socratea exorrhiza (Mart.)H. Wendl.(syn. S. durissima (Oerst.)H. Wendl.). Data from eachpalm included diameterat breastheight (dbh;to the near-est mm; for trees > 50 mm dbh in 1999, data were sup-plied by R. Condit), height (to the nearest 0.5 m) andbryophytecover on the trunk to the nearest 5%. Onlybryophyte urfsexceedingc. 2 mm in heightwere consid-ered.Subsequently, he trunkand stilt rootswere searchedfor the presence of vascular epiphytes. Hemi-epiphyteswere also included,whether or not they had contactwiththe soil. On the other hand, vines such as the abundantPhilodendron naequilateratumLiebm.were ignored.Thegondola was positioned in the forest canopy to allowaccess to all epiphyteson Socrateaexorrhiza,but in rarecases the palmtrunkwas inspectedwith binoculars.Withfew exceptions,we were able to identify each individualto species level, even in the case of juveniles (only tinyseedlingswereignored).Inthisreport, individual' s usedsensu Sanford(1968), i.e. as 'group of stems'. For eachepiphyte,the following variableswere determined:heighton thetree,azimuth,substratequality(i.e. whetheranepi-phyte was growing in a bryophytemat or turf), size (i.e.numberandsize of stems andleaves). Voucherspecimensare depositedin the herbariumof the SmithsonianTrop-ical Research Institute, Panama (Tupper Center). Plantnamesof angiosperms ollow the flora of Panamacheckl-ist (D'Arcy 1987), while fern names areaccording o Lel-linger (1989).Nearestneighbouranalysis was conducted for all epi-phytes occurringon S. exorrhizaby means of a custom-made computerprogramwhich producednull model dis-tributionsby randomly assigning each individual to oneof the 701 attachmentsites. A nearest neighbourwasdefined as that epiphyte on a given host tree with theshortest vertical distance to a focal plant, which did notdiffer in azimuthby more than900. The programwas run240 times.By discarding he six lowest and the six highestvalues in each category we produced 95% confidenceintervals (Noreen 1989). For reasons of clarity only the

results for those six epiphyte species with > 5% of allindividualswill be presented.All other statisticalanalysiswas carried out with STATISTICA software(STATISTICA 5.1, StatSoft Inc., Tulsa, OK, USA).Wheneverpossible, we used parametricstatistics,some-times log-transformingdata before analysis (Sokal &Rohlf 1995).

RESULTSHost reecharacteristicsIn contrast to most other Arecaceae, the stilt root palmSocrateaexorrhiza ncreases n trunkdiameterwithheight(Schatz et al. 1985). In our population,the diameteratbreastheight(dbh)correlated losely with the log,,10f treeheight(log10 treeheight)= -0.25 + 0.01 dbh, r2 = 0.90, P< 0.001, n = 118). To estimatetreeage, we used dbhdatafromrepeatedmeasurements n late 1997 andearly 1999from a larger sample of palms growing in and immedi-ately adjacent o our studyarea(R. Conditet al., unpubl.data). Increments (standardizedfor 1 y) were plottedagainstthe initial dbh.We found a weak,buthighly signi-ficant negative correlationbetween dbh and subsequentincrement (r = -0.22, P < 0.001, n = 567). Mean annualincrement n dbh graduallydecreased from almost 5 mmy-1 n very small plantsto zero growthin the largestindi-viduals (Adbh = 4.58 - 0.03 dbh). The modelled growth,i.e. dbh increment,of such an 'average' tree is depictedin Figure 1. Also given are the changes in height, usingthe above regressionbetween dbh and tree height. Totalbarksurface(S), i.e. the potential targetareafor epiphytediaspores, increased exponentially with dbh (logS =-1.670 + 0.016 dbh;r2= 0.94, P < 0.001; n = 118).

TheepiphyteommunityIn the c. 9000 m2within reach of the crane, we located118 individualsof Socrateaexorrhiza.Treeheight ranged0.3-25 m, dbh5-170 mm, andS 0.01-5.7 m2. Bryophytecovervaried from0 to 30%and was positively correlatedwith dbh (r= 0.58, P < 0.001, n= 118). The remainingbark was mostly covered with crustose lichens, but wealso observed smaller mosses and liverworts (< 2 mmheight) or algae. Vascularepiphytes were found on 57trees,i.e. on 48% of all individuals.In total, we observed701 epiphytes(andhemi-epiphytes)belongingto 66 spe-cies in 15 families (Appendix 1). All but 17 individuals,which were growingon the spiny stilt roots, occurredonthe trunk.The highest number of individuals(85 speci-mens out of 12 species) was found on a large palm (dbh140 mm), an even higher diversity, 16 species, wasobserved on another large tree (dbh 157 mm; 53individuals).The three most common species were all ferns


4/11

Epiphyteson Socrateaexorrhiza 83

Socrateaexorrhiza140 - - 20120-

-15100 -80- - 1060-

40_

- tree height4 020 -

0 20 40 60 80 1000 20 40 60 80 100Age (y)

Figure1. Model growthin Socratea exorrhiza. Based on the regressionof tree size and subsequentannual ncrement n dbh an averagerelationshipof age and dbh is simulated.Tree height was then calculated from dbh as log(treeheight) [m] = -0.25 + 0.01 dbh [mm].

(Ananthacorus angustifolius, Elaphoglossum sporado-lepis, Dicranoglossumpanamense),accountingalone foralmost 30%of all individuals.Othercommonspecies re-presentingmore than 5% of the total were Scaphyglottislongicaulis (Orchidaceae),the secondaryhemi-epiphytePhilodendronschottianum Araceae)and Guzmaniasub-corymbosa Bromeliaceae).On the otherhand,almosthalfof all species were rare,i.e. occurred with just 1-3 indi-viduals on ourpopulationof palms (Appendix1).Attachmentheights, which were analysed for the 13most abundantspecies, differed significantly (one-wayANOVA, F12,497= 23.2, P < 0.001; Figure2). While sometaxa (e.g. Ananthacorus angustifolius, Elaphoglossumsporadolepis)were found at almost any possible heightalong the trunk, most species showed a rathernarrowrange.It was possible to distinguishbetweenunderstorey(e.g. Anthuriumclavigerum),midstorey (e.g. Guzmaniasubcorymbosa)and 'canopy' species (e.g. Scaphyglottisgraminifolia).Treeswithepiphytesweresignificantlyarger han hosewithout epiphytes (t-test, P < 0.001). The relationshipbetween tree size and epiphyte abundance was furtherexplored with a best-fit polynomial regression model(Figure3b):theregressionexplained34% of thevariation.Onlyon two of the 50 trees smaller hanabout80 mm dbhwere vascularepiphytesfound(two individualsof Tricho-manes spp. and one of Aspleniumserratum).Excludingthese two cases, we estimate thattreesare about 20 y oldbefore colonizationbyvascularepiphytesbegins(basedonaverage growthrates of this palm;Figure 1). Because thenumberof species and individualswereclosely correlated(log species = 0.13 + 0.58 log individuals,r2= 0.86, P


5/11

84 GERHARD ZOTZ AND BIRGIT VOLLRATH

e - a Niphidiumrassifoliumde I--r-- n I Scaphyglottisraminifolia

d I Scaphyglottisongicauliscd - H I l I IElaphoglossumporadolepis

be ISobralia fragransC Dicranoglossumanamense

bc Guzmaniaubcorymbosab - I - -I Ananthacorusngustifolius

a i Tillandsiancepsa - A [] Aspleniumerratum

a F DichaeaanamensisI 5-ange a,-F5%--H PhilodendronchottianumS25- 75%o Median a F-F Anthuriumlavigerum

20 15 10 5 0Heightof attachmentm)

Figure2. Vertical distributionsof the 13 most abundantepiphyte species on Socratea exorrhiza.Different letters beside each box plot indicatesignificantdifferencesin a post-hocLSD test (P < 0.05). For reasons of clarityoutliersare not shown.

Scaphyglottis longicaulis or Philodendron schottianum.The latterfindingis not surprising,however,consideringthe verticaldistributionsof the threespecies (Figure2).The temporalpatternof the colonizationof Socrateabybryophyteswas not distinguishable rom that of vascularepiphytes(Figure3). The spatialassociationof these twoplant groupswas analysedby comparingthe numberofvascular epiphytes occurringdirectly on bark to thosefoundrooted n patchesof bryophytes.First,we estimatedthe barkareacoveredby bryophytes.We obtained7.9 m2,i.e. 5% of the total bark surface (155 m2) of all of thepalmsin the studyarea.The numberof epiphytesgrowingin patchesof mosses andliverworts,however,was consid-

erably higher than expected from this surface area, i.e.267 or 38% of all 701 individuals(X2= 1509;P < 0.001).This association with bryophytes was highly variable(Table4). While some species, for exampleAnanthacorusangustifolius or Dicranoglossum panamense almostalways anchored in patches of mosses and liverworts,others like Scaphyglottis longicaulis were rarely associ-ated with them.

DISCUSSIONEven architecturally imple trees like Socratea exorrhizamay host an impressive number of vascular epiphytes


6/11


3530-25- *20 - *1510- * * * **5 -o

S I, I I I I80- 0

60 - o40- S0 02 - 0 8

16- 0C o012-o oo o8 0- 0o 00 0 04- oo o o00

oo .0

0 20-5 (D0 1 O 00 25 50 75 100 125 150 175

dbh(mm)0 c.5 c.10 c.20 c.40 c.60 c.80

Estimatedree age (y)Figure3. Relationshipsof the relativebryophytecover (a), the numberof epiphyteindividuals(b) and the numberof epiphyte species (c) andhost tree size (dbh;in mm), respectively.The estimated tree age (y) isalso given (compareFigure1). Eachsymbolis a different ree(n = 118).The lines arebest-fitpolynomialregressions:bryophyte over(%)= 0.52-0.03 dbh + 0.004 dbh2 + 8.7 x 10-7 dbh3, r2= 0.22; individuals= 1.23-0.03 dbh - 0.0006 dbh2+ 1.2 x 10-5 dbh3, r2=0.34; and speciesnumber= 0.21 - 0.01 dbh+ 3.6 x 10-6dbh2 + 2.6 x 10-6dbh3,r2= 0.47.

withina small areaof lowlandforest:we observeda max-imum of 16 species on a single large tree (Figure 3c),and distinguisheda total of 66 taxa on the 57 trees withepiphytes.This representsabouthalf of all epiphytetaxain the 0.9-ha plot (Zotz & Schultz, unpubl.data).Specieswere clearly segregatedvertically(Figure2). Severalspe-cies occupieda rather imitedrange,but others ike Anan-thacorus angustifolius could be found over almost theentire trunk.Competitionseems an unlikely explanationfor thispattern,considering he low densitiesof epiphytes(less than1 individualm-2).Alternatively,we assumethatdifferences in light and humidity requirements mayexplainmuch of the observedsegregation.Unfortunately,there is little informationon the ecophysiologyof most ofthe component species to test this suggestion.Although overall diversity of epiphytes on Socrateaexorrhizawas high, colonizationwas probablyvery slow.The youngest palms with vascularepiphytes(two speciesof ferns;Figures 3b,c) were estimatedto be about 20 yold. Dudgeon (1923) describeda very similarpictureinhis classic study in a Himalayanforest: it took vascularepiphytes about two decades to get established onQuercus incana trees. A somewhat faster developmentwas reported rom a montane orestin the BolivianAndes(Ibisch 1996). There, 10-y-old Alnus acuminata treeshosted a largernumberof epiphytespecies, amounting oabout half of the maximumdiversityfound on 'mature'trees. Finally, Catling et al. (1986) found an averagenumberof three species per tree in a 13-y-old grapefruitorchardin Belize, while 30-y-old plantationsaveragedeight species per tree.Thus,colonizationof treesby vas-cular epiphytes seems to be rather slow in general, butmostly faster than in the palm treeof this study.At the time of the initial colonizationby vascularepi-phytes, Socratea trees had grown to about 5 m height(Figure 1), i.e. probablystill did not offer suitablesub-strate for those species restrictedto higher strata (e.g.

Table 1. Correlationbetween plant size, estimatedas length of longest leaf (L) or shoot (S), and host tree dbh for the 13 most common epiphytespecies. Shownare the resultsof separateSpearman ankordercorrelationanalyseswith samplesize (n), correlation oefficient(SpearmanR), t-valueand P level. Significance s indicatedby asterisks.Species Size Abbreviation n SpearmanR t(n-2) PAnanthacorusangustifolius L Aa 76 -0.16 -1.4 0.17Anthurium lavigerum L Ac 21 0.52 2.63 0.017*Aspleniumserratum L As 18 0.05 0.2 0.84Dichaea panamensis S Dp 16 0.45 1.91 0.07Dicranoglossumpanamense L Dcp 56 -0.13 -0.93 0.36Elaphoglossumsporadolepis L Es 70 0.15 1.27 0.21Guzmaniasubcorymbosa L Gs 40 0.48 3.34 0.002*Niphidiumcrassifolium L Nc 24 0.67 4.15 0.000*Philodendronschottianum L Ps 49 -0.05 -0.33 0.74Scaphyglottisgraminifolia S Sg 34 0.17 0.97 0.34Scaphyglottis ongicaulis S S1 51 -0.15 -1.05 0.30Sobralia ragrans S Sf 31 -0.04 -0.19 0.85Tillandsiaanceps L Ta 23 0.03 0.13 0.9


7/11


Table 2. Correlationbetween the percent occurrenceof the 13 most commonepiphytespecies andthe totalnumberof taxaper tree. Foreach specieswe give the % occurrence n the 57 epiphyte assemblages,correlationcoefficient (SpearmanR), t-value and P level. Significanceis indicatedbyasterisks.Species %Occurrence SpearmanR t(n-2 PAnthurium lavigerum 26 0.75 2.75 0.03*Ananthacorusangustifolius 32 0.85 3.90 0.01*Aspleniumserratum 14 0.29 0.76 0.48Dichaea panamensis 5 0.37 0.98 0.37Dicranoglossumpanamense 21 0.92 5.84 0.001*Elaphoglossumsporadolepis 26 0.85 4.02 0.01*Guzmaniasubcorymbosa 9 0.87 4.38 0.004*Niphidiumcrassifolium 12 0.87 4.24 0.01*Philodendronschottianum 42 -0.22 -0.54 0.61Scaphyglottisgraminifolia 11 0.63 1.97 0.10Scaphyglottis ongicaulis 12 0.80 3.32 0.02*Sobraliafragrans 16 0.85 3.95 0.007*Tillandsiaanceps 23 0.68 2.29 0.06

2.753.90

7060-50 - o40 - ++30 - 30 - Dcp Nc +

0-

1 2 3 4 5 6-9 10 >10Numberof species per palm

Figure4. The per cent occurrenceof the most commonepiphyte species in communitiesdifferingin species numberfrom 1 to 16. The lines weredrawnby hand.Dcp = Dicranoglossumpanamense (open circles), Nc = Niphidiumcrassifolium(closed squares),Ps = Philodendronschottianum(crosses).

Table3. The identityof nearestneighboursof the six most common epiphytes.For each species the observedoccurrence of species pairs and thelimits of 95% confidence ntervalsof randomdistributions in parentheses)aregiven. Bold numbers ndicatespecies pairs,which were observed morecommonlythanexpected;numbers n italics are combinations ess common thanexpected.Note thatthe matrixis asymmetricbecause the fact thatindividualA is nearestneighbour o individual B does not imply that B is nearestneighbour o A.

Aa Es Dcp S1 Ps GsAnanthacorusangustifolius(Aa) 28 3 20 2 2 1(2;16) (3;13) (2;10) (2;10) (1;9) (1;8)Elaphoglossumsporadolepis(Es) 3 38 2 3 0 2(3;13) (2;12) (2;9) (1;9) (1;9) (1;8)Dicranoglossumpanamense(Dcp) 12 5 26 0 0 0(2;11) (1;10) (1;9) (1;8) (1;7) (1;6)Scaphyglottis ongicaulis (S1) 1 3 0 36 0 0(0;10) (1;9) (0;8) (0;9) (1;7) (0;6)Philodendronschottianum Ps) 2 1 0 1 23 2

(2;9) (1;9) (1;7) (1;7) (0;8) (0;6)Guzmaniasubcorymbosa Gs) 1 2 1 0 0 25(1;8) (1;8) (0;7) (0;6) (0;6) (0;7)


8/11


Table4. Association of 13 species of vascularepiphyteswith bryophytemats.Given are the observedand the expectedfrequencies(assumingsimilarproportions or all species) for individualsgrowing in bryophytes (bryophyte+) or on bare bark (bryophyte ). Bold numbersindicatesignificantdifferencesbetween observedand expected frequencies (x2, P < 0.05). For species abbreviations ee Table 1.Species abbreviation

Aa Ac As Dp Dcp Es Gs Nc Ps Sg Sl Sf Tan 76 21 18 16 56 70 40 24 49 34 51 31 23ObservedfrequenciesBryophyte+ 57 1 3 3 46 45 6 8 1 16 2 13 4Bryophyte- 19 20 15 13 10 25 34 16 48 18 49 18 19ExpectedfrequenciesBryophyte+ 30 8 7 6 22 28 16 10 20 14 20 12 9Bryophyte 46 13 11 10 34 42 24 14 30 20 31 19 14

Scaphyglottis pp., cf. Figure2). Further rowthcoincidedwith a strongincreasein tree height andbryophytecover(Figures 1, 3a). Both processes along with possibleweatheringof the bark (cf. Catling et al. 1986) shouldlead to habitatdiversificationandareexpectedto facilitatethe colonization of palmtreesby an increasingnumberofepiphytes. This expectation was only partly fulfilled.While we observeda substantiallyhigher density, speciesnumbersper unit surfacearea did not change with plantsize.

Overall,the correlationbetweentree size andepiphyteload was quiteweak (Figure3), indicating hat local idio-syncrasies such as proximity to sources of propagules,variation n microclimate(due to, for example, local dif-ferences in forest canopy height) or possible accidentscausedby foragingarborealanimals(Perry1978), fallingpalm fronds,or simply chance have a strong nfluence onthe compositionof the epiphytecommunityof a particulartree. The observationthatmany of the largertrees weredevoid of epiphytes(Figure3) in spite of their size andage (Figure1) is not exceptional.Forexample,Johansson(1974) found that only 50% of all larger trees in anAfrican rain forestcarriedepiphytes.However,in spiteofthe unpredictability f the residentepiphyteflora of anyone tree and the dismissal of Went's (1940) concept ofspecies specificityas a rareexception(Benzing 1990), wereiteratethe suggestion of Zotz et al. (1999): each treespecies in a given areaof forestmay host a specific spec-trum of epiphyte taxa. Each tree species offers a uniqueset of architectural,morphological, hemical andphenolo-gical traits,which should give rise to a similarly uniquesubset of epiphytesfrom the local species pool, both interms of species composition and, in particular,relativeabundances.It was alreadypointed out by Went (1940)that accumulations f humusmaymitigatepossibleeffectsof a host tree species. Hence, it is conceivable thatdiffer-ences in the epiphytespectrabetweenhost trees are minoror absent in forests, where branchescarrya dense coverof bryophytes, lichens and dead organic material, forexamplemany montane forests.Although vascular epiphytes were more commonlyassociatedwith patches of bryophytesthan expected by

chance,themajorityof individuals, .e. c. 60%,had colon-ized naked bark or crustose lichens. While there wereobvious species-specificdifferencesin the degree of theassociation with bryophytes Table4), none seemedto bedependenton bryophytesfor establishmenteven on thesmoothbarkof a palm. This findingcontrastswith otherstudies, in which the establishment of bryophyteswasdepictedas a necessarysuccessionalstep in the develop-ment of epiphytecommunities(Dudgeon 1923, Van Oye1924). Ourresultsare,however,consistent with the notionthatmosses andliverwortsstrongly acilitatethe establish-ment of seedlings of vascular epiphytes. Presently,ourevidence is correlational: nly experimentssuch as thoseby Laman(1995) canreveal the natureof this association.Althoughwe emphasizethepositiveeffects of cryptogamson vascularepiphytes,this interactionmay have a nega-tive outcome as well for seedlings.For example,Zotz &Andrade 2002) report he repeatedobservationof folioselichen thalli over-growing and presumablykilling theseedlingsof epiphyticorchids.The question whetherthere is true succession amongvascular epiphytes is debated (Benzing 1990). Whilesome studies (Catlinget al. 1986, Catling & Lefkovitch1989, Johansson 1974) describe the replacement anddecline of early seral stages by laterones, othersfind noindication of such processes (Yeaton& Gladstone 1982,Zotz et al. 1999). Similar o these latterstudies,most spe-cies in the presentinvestigationincreasedin numbersinmore complex epiphyte assemblageson older trees,nonedecreased n occurrence Table 2). Figure4 suggeststhreedifferentgroupsof colonists. The hemi-epiphyticPhilo-dendron schottianum and the epiphytic Aspleniumserratum were frequentcomponents of palms with epi-phytes, irrespectiveof species numbers, .e. maybe calledpersistentpioneers.The majorityof the remainingspeciesshowed a constant increase in frequencieswith speciesnumberper tree,while a last grouponly occurred n morecomplex assemblages. Notably, the latter species, e.g.Niphidiumcrassifoliumor Scaphyglottisspp., were thoserestricted to the upper part of the trunk (Figure 2, cf.Zotz & Winter 1994).The observation that the number of individuals per


9/11


10/11


SOKAL,R. R. & ROHLF,F. J. 1995.Biometry.FreemanandCo., NewYork. 887 pp.

VAN OYE, M. P. 1924. Epiphytesdes troncesd'arbresauCongo Belge.Revue Generalede Botanique36:481-498.

WENT, F. W. 1940. Soziologie der Epiphyteneines tropischenRegen-waldes.Annalesdu JardinBotaniquede Buitenzorg50:1-98.

YEATON,R. I. & GLADSTONE,D. E. 1982. The patternof coloniza-tion of epiphytes on Calabash trees (Crescentia alata HBK.) inGuanacasteProvince,CostaRica. Biotropica 14:137-140.

ZIMMERMAN, . K. & OLMSTED, . C. 1992. Host treeutilizationbyvascularepiphytes n a seasonally nundated orest(Tintal) n Mexico.Biotropica24:402-407.

ZOTZ,G. 1997. Substrateuse of threeepiphyticbromeliads.Ecography20:264-270.

ZOTZ,G. & ANDRADE,J.-L. 2002. Ecologiade epifitasy hemiepifitas.Pp. 271-296 in Guariguata,M. & Kattan,G. (eds). Ecologia de Bos-ques iluviososNeotropicales.IICA, San Jos6,Costa Rica.

ZOTZ, G., BERMEJO,P. & DIETZ,H. 1999. The epiphyte vegetationof Annona glabra on Barro Colorado Island, Panama. Journal ofBiogeography26:761-776.

ZOTZ,G. & WINTER,K. 1994. Annual carbonbalance and nitrogenuse efficiency in tropicalC3 and CAM epiphytes. New Phytologist126:481-492.


11/11


Appendix1. Vascularepiphytesandhemi-epiphytesoccurringon Socratea exhorriza.Species are listed by decreasingabundance.Species names offloweringplantsfollow the Flora of PanamaChecklistandIndex (D'Arcy 1987), andLellinger(1989). Hemi-epiphytesaremarkedwith an asterisk.Species Family IndividualsAnanthacorusangustifolius Vittariaceae 76Elaphoglossumsporadolepis Lomariopsidaceae 70Dicranoglossumpanamense Polypodiaceae 56Scaphyglottis ongicaulis Orchidaceae 51Philodendronschottianum* Araceae 50Guzmaniasubcorymbosa Bromeliaceae 40Scaphyglottisgraminifolia Orchidaceae 34Sobraliafragans Orchidaceae 31Niphidiumcrassifolium Polypodiaceae 24Tillandsiaanceps Bromeliaceae 23Anthurium lavigerum* Araceae 21Aspleniumserratum Aspleniaceae 18Dichaea panamensis Orchidaceae 16Vrieseagladioliflora Bromeliaceae 14Epidendrumnocturnum Orchidaceae 13Epidendrumdifforme Orchidaceae 11Polybotryavillosula Dryopteridaceae 11Anthuriumriedrichsthalii Araceae 11Columneabillbergiana Gesneriaceae 8Campyloneurumhyllitidis Polypodiaceae 7Codonanthemacradenia Gesneriaceae 7Syngoniumpodophyllum* Araceae 7Campyloneurumccultum Polypodiaceae 6Trichomanesangustifrons Hymenophyllaceae 6unidentified uveniles Polypodiaceae 5Peperomiarotundifolia Piperaceae 5Anthurium f. acutangulum Araceae 5unidentified uveniles Araceae 5Vittaria ineata Vittariaceae 4Peperomiaebingeri Piperaceae 4Scaphyglottisprolifera Orchidaceae 4Topobeapraecox * Melastomataceae 4Trichomanes kmanii Hymenophyllaceae 4Clusia cf. uvitana* Clusiaceae 4Tillandsia bulbosa Bromeliaceae 4Microgramma ycopodioides Polypodiaceae 3Pleopeltis percussa Polypodiaceae 3Encyclia chimborazoensis Orchidaceae 3Pleurothallisorbicularis Orchidaceae 3Trichomanesovale Hymenophyllaceae 3Anthurium f. salviniae Araceae 3Pleopeltis cf. panamensis Polypodiaceae 2Polypodiumtriseriale Polypodiaceae 2Polystachya oliosa Orchidaceae 2Sobraliapanamensis Orchidaceae 2Guzmaniamosaica Bromeliaceae 2Vrieseasanguinolenta Bromeliaceae 2Philodendron radiatum* Araceae 2Philodendronsp. Araceae 2Anetiumcitrifolium Vittariaceae 1Ravnia triflora* Rubiaceae 1Polypodiumcostaricense Polypodiaceae 1Peperomiamacrostachya Piperaceae 1Dimerandraemarginata Orchidaceae 1Elleanthuslongibracteatus Orchidaceae 1Ornithocephalus f. bicornis Orchidaceae 1Pleurothallis verecunda Orchidaceae 1unidentified uvenile Orchidaceae 1Lomariopsisvestita Lomariopsidaceae 1Trichomanesgodmanii Hymenophyllaceae 1Drymoniaserrulata Gesneriaceae 1Epiphyllumphyllanthus Cactaceae 1Aechmeatillandsioidesvar. kienastii Bromeliaceae 1Monsteradilacerata Araceae 1Philodendronsagittifolium* Araceae 1Stenospermation ngustifolium Araceae 1

the epiphyte vegetation of the palm socratea exorrhiza

Documents