Flower trait convergence in a plant-pollinator network

LOMÁSCOLO, SILVIA B.; CHACOFF, NATACHA P.; VÁZQUEZ, DIEGO P.; CASTRO-URGAL, R

Portland

Congreso; Annual Meeting of the Ecological Society of America; 2012

Ecological Society of America

To understandbiodiversity patterns ina community, and themechanisms by whichbiodiversity is generatedand maintained, itis indispensable tostudy the patterns ofinteractions among theorganisms within thecommunity.  Interaction patterns canbe complex, but recentmethods in networkecology allow usto identify certain interactionrules that help uspredict how theorganisms in thecommunity may beaffected by theinteractions they areinvolved in. Forexample, interactionsin a communityseem to be organizedin modules, which aregroups of organismsthat tend to interactmore frequently with eachother than with thoseincluded in othermodules. It hasbeen hypothesized thatspecies within amodule will tend toshow convergence intheir traits.  We attempt tounderstand how speciesin a plant-pollinatorinteraction network affecteach other's evolutionby searching for patternsof convergence inthe traits that mediatetheir interaction.  The networkstudied consists of40 plant species, 88pollinating insect species,and  19,166 totalinteractions. The studywas done in theMonte desert of Mendoza,Argentina, between 2006and 2011. Plant traitsinclude diameter,depth, color, symmetry,and orientation ofthe corolla; height offlowers; and lengthand timing of flowering;insect traits include differentmeasures of bodysize and mouth parts.  We findthat plant species tendto converge in lengthand timing of flowering,symmetry, orientation,and flower height, amongother traits, using anEvolutionary PrincipalComponents Analysis (EPCA).Evolutionary PCA workssimilarly to regularPCA, except that theresulting ordinationcan be interpretedas an evolutionarypattern: species areordinated close toeach other if theyhave converged in thetraits included in theanalysis.  We also findthat pollination bylarge bees and hoveringflies, among others, explainthe position of plantspecies in theEPC axes, by usinga Generalized LinearModel approach.  We also findthat modules do notreflect the  convergence oftraits suggested by theEPCA.  Thesepreliminary results suggestthat pollinators affectthe evolution of planttraits and leads toconvergence of thesetraits in unrelatedplant species, although thisconvergence is notevidenced within modules.  Whether thisis also the casefor the effects ofplants on pollinatorsremains to betested.