Interaction network and the relationships between bromeliads and hummingbirds in an area of secondary Atlantic rain forest in southern Brazil

of 9
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.
Information Report



Views: 30 | Pages: 9

Extension: PDF | Download: 0

Interaction network and the relationships between bromeliads and hummingbirds in an area of secondary Atlantic rain forest in southern Brazil
   Journal of Tropical Ecology  (2007) 23 :663–671. Copyright © 2007 Cambridge University Pressdoi:10.1017/S026646740700449X Printed in the United Kingdom Interaction network and the relationships between bromeliads andhummingbirds in an area of secondary Atlantic rain forest in southern Brazil V´ıtor de Queiroz Piacentini 1 and Isabela Galarda Varassin 2 P´os-graduac¸˜ao em Ecologia e Conservac¸˜ao, Universidade Federal do Paran´a, Curitiba, PR, Brasil (  Accepted 14 August 2007  ) Abstract:  The reciprocal importance of bromeliads and hummingbirds has been proposed for many years, evensuggesting coevolution between these two groups. Nevertheless, data are lacking that allow a better test of therelationships involved. Here we investigate the relationship between bromeliads and hummingbirds in an area of secondaryAtlanticrainforestin southernBrazil.Thestudyexaminedtheinteractionsamong12species ofbromeliadand 10 of hummingbird at Reserva Natural Salto Morato, Paran´a state. The number of flowering species of bromeliadand the species richness and abundance of hummingbirds were quantified monthly between November 2004 andOctober 2005. Focal observations on each bromeliad species were made to determine the hummingbird visitors.Neither species richness nor abundance of hummingbirds were related to bromeliad phenology. Together with themonthly variation in visit frequency by a given pollinator to a given plant, these factors indicate a generalizationin the use of bromeliads by hummingbirds and argue against tight coevolution.  Ramphodon naevius  and  Thaluraniaglaucopis  were the main pollinators in the community.  Aechmea nudicaulis  was the most generalist bromeliad species.The generalist species interacted with other generalists or with asymmetric specialists and there was no specialist– specialist interaction. This produced a strongly organized and nested matrix of interactions. This nestedness is similartootherplant-pollinatorsnetworks,supportingthehypothesisthattheevolutionaryrelationshipbetweenbromeliadsand hummingbirds is no stronger than that of other pollination networks. Key Words:  Bromeliaceae, coevolution, hummingbirds, interactions, nestedness, phenology, pollination systems INTRODUCTION Bats, butterflies and bees are known pollinators of bromeliads (Ara´ujo  et al . 2004, Machado & Semir 2006,Varassin2002),buthummingbirdsplaytheleadingrole:61% of the bromeliad species in the Bolivian Andes(Kessler & Kr¨omer 2000) and  c . 85% of the species in theAtlantic forest are pollinated primarily or exclusively byhummingbirds (Ara´ujo  et al . 2004, Varassin 2002). Onthe other hand, in some areas of the Atlantic rain forest,bromeliadspeciesrepresentmorethan30%oftheflowersused as a food resource by hummingbirds (Buzato  et al .2000). In montane forests, bromeliads are an importantsource of food, being represented by a large number of individuals (Dziedzioch  et al.  2003). 1 Corresponding author. Address: Rua Marcus Aur´elio Homem,285 88040-440, Florian´opolis – SC – Brazil. Email: 2 Email: Despite the importance of one group for the other,only in the last decade have more detailed studies aboutthe relationship between bromeliads and hummingbirdsbeen conducted (Ara´ujo  et al.  2004, Buzato  et al.  2000,Machado & Semir 2006, Sazima  et al.  1996, Varassin2002, Varassin & Sazima 2000). Nevertheless, little isknownabouttheresponseofhummingbirdcommunitiesto the flowering phenology of bromeliads, or the natureof the plant-pollinator interactions that involve bothbromeliads and hummingbirds.Bromeliads have aggregated flowering in somelocations (Kaehler  et al . 2005), while flowering isdistributed throughout the year in other places (Ara´ujo et al.  2004). As birds respond to variation in resourceavailability (Poulin  et al.  1992), it is possible thatthe timing of bromeliad flowering may influence theabundanceorthepresenceofmigrantandresidentspeciesof hummingbird. Hummingbird abundance was relatedto the presence of ornithophilous bromeliads along analtitudinal gradient in the Bolivian rain forest (Kr¨omer et al.  2006).  664 V´ıTOR DE QUEIROZ PIACENTINI AND ISABELA GALARDA VARASSINThe idea of pair-wise coevolution – the extreme of spe-cialization – is well known, especially in plant-pollinatorinteractions.Notsurprisingly,coevolutionbetweenhum-mingbirds and some of their flowers, such as heliconias,hasbeensuggested(Snow&Teixeira1982,Stiles1978).Bromeliads (Benzing 1980, Reitz 1983) and humming-birds (Grantsau 1988, Sick 1997) overlap greatly intheir overall distributions and centres of speciation (inthe northern Andes) and a parallel evolution of bothgroupshasbeenproposed(Sick1997).Ontheotherhand,hummingbirdshavealifespanlongerthanthebloomingperiod of any bromeliad, so an extreme coevolution isphysiologicallyunfeasible(Waser et al .1996).Accordingto Feisinger (1983), the interactions presently observedbetween plants and hummingbirds are derived at bestfromadiffusecoevolution,i.e.agroupofpollinatorspeciesapplyingareciprocalselectiontoagroupofplantspecies.Specializationshouldnotbeviewedastheruleinplant-pollinator relationships. As an alternative, generalistpollination systems and asymmetric specialization (i.e.specializedspeciesinteractingprimarilywithgeneralists)have been presented as common situations in nature(V´azquez & Aizen 2003, 2004; Waser  et al . 1996).A lack of reciprocal specialized relationships and aneventual trend to generalization do not necessarilyimply‘disorganisation’orrandomnessinplant-pollinatorinteractions. It may reflect a ‘nested’ organisation ina plant-pollinator matrix, which can be quantified(Bascompte  et al . 2003). This nested matrix is the resultof a core of generalist pollinators interacting with plantsthat are also generalists, along with cases of asymmetricspecialization in both plants and pollinators (Bascompte et al . 2003, Jordano  et al . 2006).In this context, the relationship between bromeliadphenology and the hummingbird community in theAtlantic rain forest of southern Brazil was examined.We tested the following hypotheses: (1) both groups arestrongly linked and the number of bromeliad speciesflowering per month does relate to species richness orabundanceofhummingbirds;(2)Conversely,bothgroupsareveryweaklyinfluencedbyeachotherandthevisitingratebyhummingbirdsisaresultofrandomchoice,i.e.themost abundant hummingbirds in each month are thosewhovisitbromeliadsthemost;(3)Theguildsofpollinatorsaresimilarbetweenbromeliadspecies,withoutanytrendto specialization; (4) The hummingbird species forage onthesamebromeliadspecies,withoutanytrendtospecializ-ation;(5)Thehummingbird–bromeliadrelationshipisnotstructured and results in a random mutualistic network. STUDY SITE The Reserva Natural Salto Morato (RNSM) is locatedin Guaraquec¸aba, on the northern coast of the Figure 1.  Rainfall (bars) and monthly mean temperature (line) inGuaraquec¸aba, Paran´a state. Historical means from 1978–2005(IAPAR). state of Paran´a (25 ◦ 09   –25 ◦ 11  S, 48 ◦ 16   –48 ◦ 20  W).Recognized as a Natural Heritage Site by UNESCO in1999, the reserve has an area of 2340ha and rangesfrom 15 m to 918 m asl. Following the Koeppenclassification (McKnight & Hess 2005), the climate of the region is Af – tropical wet, no dry season – with amean annual temperature of 21 ◦ C and monthly meansrangingbetween25 ◦ Cand17 ◦ C.FromOctobertoMarchhigh temperatures and rainfall prevail at RNSM, whilefrom April to September the climate is colder and drier(Figure1). RNSM is covered by Atlantic rain forest of different successional stages.For data collection, we selected 6.35km of trails fromanalreadyestablishednetworkinRNSM.Themajorityof these trails lies in mature secondary forest ( > 60 y old)and early secondary forest (up to 35 y old), but also passthrough some more open areas of abandoned pasture inthebeginningstagesofregeneration(12yold).Elevationsin our study site were below 160 m asl.There are 28 species of epiphytic bromeliad in theRNSM, the majority of which fall into one of two genera, Vriesea  (12 species) or  Aechmea  (6 species; Gatti 2000),andthereare17hummingbirdspecies(Straube&Urben-Filho 2005). METHODSData collection We investigated bromeliad and hummingbird interac-tions during monthly field trips to the study site betweenNovember 2004 and October 2005. We quantifiedflowering bromeliad species by walking each trail andrecording all bromeliads within view that had activeinflorescences.Weidentifiedmostbromeliadsinthefield,usingpublishedguides(Reitz1983),butcollectedsamplesand/or consulted experts when necessary.  Structure of a bromeliad–hummingbird network  665We performed two censuses per monthly field trip toestimate hummingbird abundance. Each census lastedabout 4 h; during this time we observed hummingbirdactivity along our 6.35-km trail system, identifyingspecies visually or acoustically (adapted from Bibby etal . 2000). We always carried out one census in themorning, starting about 30 min after sunrise, and theother in the afternoon, starting 4–5h before sunset.Each month, we estimated the species richness of the hummingbird community based on the censuses,observations at flowering bromeliads, and opportunisticrecords. Taxonomy follows the Brazilian OrnithologicalRecords Committee ( frequency of visits of the hummingbirds to thebromeliads (number of visits per unit time) was obtainedby direct observation of bromeliads in flower, using themethod of focal observation (Dafni 1992). Floweringindividuals of each bromeliad species were observed inthe morning, starting about 30 min after sunrise, as wellas in the afternoon, ending close to sunset. Individualsof each bromeliad species were observed altogether for atleast8hpermonth,withafewexceptionsduetoadverseclimatic conditions or the absence of flowers in the daysfollowing the first observation.Weopportunisticallynoteddataonvisitstobromeliadsby other pollinators (e.g. bees), as well as hummingbirdvisits to other plants. Statistical analyses To test if monthly hummingbird abundance or spe-cies richness was influenced by bromeliad floweringphenology, we used Pearson correlations to comparethe number of flowering bromeliad species per monthto both the number of hummingbird species and overallhummingbird abundance, as well as to the monthlyabundance of the hummingbird species that wererecorded in every field trip.We used a G-test of independence to examine whethervisit frequencies of each hummingbird species to eachbromeliad species reflected the relative abundance of those species in the study site, but only when thebromeliad species were visited by at least two humming-bird species in the same month.We used the Jaccard index (Magurran 1988) to cal-culatethesimilarityofthebromeliadspeciesaccordingtotheirguildofhummingbirdvisitorsandthenwegenerateda dendrogram of similarity using the unweighted pair-group method with arithmetic averages (UPGMA) asthe grouping method and the Fitopac software package(Shepherd 1987). The same procedure was used tocalculate the similarity of the hummingbird speciesaccording to the bromeliads used by each species.Thenestedness( N  )ofthebromeliad-hummingbirdnet-work,aswellasitsconnectivity,wascalculatedusingthesoftware ANINHADO (Guimar˜aes & Guimar˜aes 2006).Wecomparedthebromeliad-hummingbirdnetworkwiththe null model ‘CE’ offered by the software: in thismodel, the probability of a cell  a ij  showing a presenceis(P i /C + P  j /R)/2,inwhichP i isthenumberofpresencesintherow i ,P  j isthenumberofpresencesinthecolumn  j , C   is the number of columns and  R  is the number of rows.Thatmeansthattheprobabilityofdrawinganinteractionis proportional to the observed number of interactions of both the animal and the plant species (Bascompte  et al .2003,Guimar˜aes etal .2007).Thisisamoreconservativeinference aboutthe significanceofnestedness thanusinganullmodelinwhicheachcellinthematrixhasthesameprobability of being occupied (Bascompte  et al . 2003).Since the bromeliad–hummingbird network belongs to awider network that encompasses the whole communityof plants and their pollinators, we also calculated, forcomparisons, the nestedness of a more comprehensivematrix that included all the other plants besides thebromeliads that were recorded to receive visits of hummingbirds.Forallanalyses,weconsideredresultssignificantifP ≤ 0.05. RESULTSHummingbird community composition and bromeliadphenology We found 13 bromeliad species flowering at our studysiteduringthestudyperiodandquantifiedhummingbirdvisitationfor12ofthem(Table1).Thereareatleastsevenotherbromeliadspeciesthatoccurinthestudyarea(pers.obs.).  Vriesea erythrodactylum  was not included due toobserver difficulty in viewing flowers.Bromeliads in the RNSM flower throughout the year.In each month there were at least two bromeliad speciesin flower, except for October 2005, when only onespecies bloomed (Table 1). The months with the greatestnumber of blooming species were January (five species)and December, February and April (four species).We found nine hummingbird species in our study siteduring the study period (Table 2), although a tenthspecies ( Calliphlox amethystina ) was observed during afield trip in September 2004.  Ramphodon naevius  and Thalurania glaucopis  were recorded at least once duringeach month of our study and were considered residentspecies.  Phaethornis squalidus  was also considered aresident, even though it was not recorded in four of themonths. Amaziliaversicolor  , Anthracothoraxnigricollis , Florisuga fusca, Aphantochroa cirrochloris  and  Lophornischalybeus  were considered summer migrants, i.e. they  666 V´ıTOR DE QUEIROZ PIACENTINI AND ISABELA GALARDA VARASSIN Table 1.  Monthly flowering of bromeliad species at Reserva Natural Salto Morato between November 2004 and October 2005. Species are listedaccording to their flowering sequence.N D J F M A M J J A S O Aechmea nudicaulis  (L.) Grisebach  x Nidularium innocentii  Lemaire  x x x Vriesea erythrodactylon  Morren ex Mez  x x x x Aechmea pectinata  Baker  x x x Nidularium procerum  Lindman  x x x Vriesea ensiformis  (Vellozo) Beer  x x x x Vriesea incurvata  Gaudichaud  x x x x Vriesea rodigasiana  Morren  x x x Aechmea ornata  Baker  x x Vriesea carinata  Wawra  x x x Aechmea organensis  Wawra  x x Vriesea  cf  . friburgensis  Mez  x x Ananas bracteatus  (Lindley) Schultes f.  x Number of species 3 4 5 4 3 4 3 2 2 2 2 1 werepresentatRNSMonlyduringthe(austral)summer.Despite being recorded once in the surroundings of thereserve in May,  Amazilia fimbriata  was recorded insideRNSM only in September and October.We also observed the bromeliad  Aechmea nudicaulis being visited by three species of bee and a bird, thebananaquit  Coereba flaveola  (Linnaeus), and  Vriesearodigasiana ,  Aechmea ornata  and  Ananas bracteatus  beingvisited by bees, the latter bromeliad also visited by twospecies of Lepidoptera. Hummingbirds were seen visitingplants of various other families, including:  Ramphodonnaevius  and  Florisuga fusca  visiting  Costus spiralis  Rosc.; R. naevius  and  Thalurania glaucopis  visiting  Psychotrianuda  Wawra,  Erythrina speciosa  Andrews and  Heliconiavelloziana Emygdio; R.naevius , T.glaucopis , Anthracothoraxnigricollis  and  Aphantochroa cirrochloris  visiting  Musarosacea  Jacq.;  T. glaucopis  and  F. fusca  visiting  Hedychiumcoronarium Koenig;and T. glaucopis visiting Rubus sp.andan unidentified Marantaceae (aff.  Calathea ). Note that Musa rosacea  and  Hedychium coronarium  are not nativespecies. Bromeliad–hummingbird interactions Aechmeanudicaulis wasvisitedmostfrequentlyandbythelargest number of species (Table 3).  Vriesea rodigasiana was visited by the second largest number of speciesand  Ananas bracteatus  had the second highest frequencyof hummingbird visits. Four bromeliad species werevisited by two or more hummingbird species in the samemonth. When we determined the relative abundance of hummingbird species using the frequency of their visitsto bromeliads, it was not equal to relative abundancesderived from census data ( Aechmea nudicaulis : G = 56.2,df  = 4,P < 0.05; Aechmea pectinata :G = 17.4,df  = 3,P < 0.05;  Vriesea rodigasiana : G = 18.2, df  = 2, P  <  0.05; Aechmea ornata : G = 24.9, df  = 1, P < 0.05).The frequency of visits by hummingbird species variedfrom month to month. With the same observation efforton  Vriesea ensiformis ,  Ramphodon naevius  made 16 visitsin January, one in February and six in March. Similarly,in May  Ramphodon naevius  visited  Aechmea ornata  eighttimes, whereas in the following month it visited only Table2. MonthlyoccurrenceofhummingbirdspeciesatReservaNaturalSaltoMorato(RNSM)betweenNovember2004andOctober2005.Numbersrefertotheabundanceobservedduringthecensuses.An‘x’meansthatthespecieswasobservedinRNSMthatmonth,butwasnotrecordedduringthe censuses.N D J F M A M J J A S O Ramphodon naevius  (Dumont) 12 4 19 18 15 15 15 17 15 29 28 9 Thalurania glaucopis  (Gmelin) 3 6 7 2 7 x 5 6 9 6 6 3 Phaethornis squalidus  (Temminck) 1 2 2 1 x x x x Florisuga fusca  (Vieillot) 10 5 2 5 3 Anthracothorax nigricollis (Vieillot) 2 1 2 x 1 x Amazilia versicolor   (Vieillot) 3 12 5 1 2 x Lophornis chalybeus  (Vieillot) x 1 1 Aphantochroa cirrochloris  (Vieillot) 1 x x Amazilia fimbriata  (Gmelin) 1 xNumber of species 7 7 7 7 5 3 3 3 2 3 4 6  Structure of a bromeliad–hummingbird network  667 Table 3.  Total observation effort, rate of hummingbird visits and number of visiting species to each bromeliad species at Reserva Natural SaltoMorato between November 2004 and October 2005. Rn = Ramphodon naevius , Tg = Thalurania glaucopis , Ps = Phaethornis squalidus , Ff  = Florisuga fusca , Av = Amazilia versicolo r, Af  = Amazilia fimbriata , Lc = Lophornis chalybeus  and Ca = Callyphlox amethystina  (Boddaert).Hummingbird visits (records h − 1 )Observ. effort (h)  Rn Tg Ps Ff Av Af Lc Ca  Numberofspecies Aechmea nudicaulis  9 0.11 3.22 0.11 1.22 0.55 5 Nidularium innocentii  13.8 0.51 1 Nidularium procerum  10 0.7 1 Aechmea pectinata  14 0.86 0.57 2 Vriesea ensiformis  26 0.84 1 Vriesea incurvata  33 0.33 1 Vriesea rodigasiana  16 0.31 0.5 0.12  ∗  4 Aechmea ornata  16 0.62 1.5 2 Vriesea carinata  18 0.17 1 Aechmea organensis  12 0.83 0.08 2 Vriesea  cf.  friburgensis  12 0.25 0.08 2 Ananas bracteatus  8 0.12 2.75 2 ∗ A single opportunistic record from September 2004. twice,and Thalurania glaucopis ,absentinMay,visitedthesame bromeliad 24 times.The number of flowering bromeliad species in eachmonthwasnotrelatedtoeithermonthlyspeciesrichness(Pearson, r = 0.47; P = 0.12; n = 12) or abundance of hummingbirds(r = 0.42; P = 0.17; n = 12). Nor did wefind a relationship between the number of floweringbromeliad species and the abundance of the twohummingbird species that were recorded every month inour study site ( Ramphodon naevius : r =− 0.18; P = 0.59;n = 12; Thaluraniaglaucopis :r =  –0.19;P = 0.56;n = 12).The dendrogram of similarity grouped the bromeliadsaccording to the visitors/pollinators of each species,ranging from the most specialized species (those visitedonly by  Ramphodon naevius ) to the more generalist,such as  Aechmea nudicaulis , which was visited by fivehummingbird species (Figure 2). Figure 2.  Dendrogram of similarity of bromeliad species at ReservaNatural Salto Morato in relation to their hummingbird visitors(cophenetic correlation = 0.96). The analysis of similarity on the use of bromeliadsby the hummingbirds grouped  Ramphodon naevius  and Thalurania glaucopis  as the two most differentiated in thecommunity. Not surprisingly, these were also the mostgeneralisthummingbirds.Theremaininghummingbirdsvisitedonlyonebromeliadspeciesandthusweregroupedaccording to the species visited (Figure 3).The matrix of interactions between bromeliad andhummingbird species had a nestedness value of  N  = 0.922(P = 0.01;Figure4a),with25%connectivity.Amplifyingthematrixtoincludetheotherplantsobservedtobefoodresourcesforthehummingbirds,thenestednessvalueremainedsimilar, N  = 0.918(P = 0.01;Figure4b),but with a slightly lower connectivity: 19%, as expectedfrom the well-known negative relationship betweenconnectivity and number of species. Figure 3.  Dendrogram of similarity of hummingbird species at ReservaNatural Salto Morato according to the bromeliads used by each species(cophenetic correlation = 0.99).
View more...
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