HANDBOOK OFEXPERIMENTALPOLLINATION

BIOLOGYEdited BY

C. Eugene JonesDepartrn€nt of Biological Science

California Slale Univ€rsity, FullertonFullerton. Califo.nia

R. John LittleRancho Santa Ana Botanic Garden

Claremont, California

It6N

l^<17

0 - + \ z - 2 ' 16 - l - t

OA./f;\ SCIENTIFIC AND ACADEMIC EDITIONSDZI-/ 1,.; "3;^ "';,i:^';^"., :;;3,1; ";&;;";:i;,":,.

COMPONENT ANALYSIS OFCOMMUNITY.LEVEL

INTERACTIONS IN POLLINATIO]SYSTEMS

James D. ThomsonEcologf aml Evolution Depattnenl

Slate Uniwnit, oI New YotkStonf Rrcok, Ne|| York

ABSTRACT

In studies ot communiti€s, ore almost always asks br@d quenions and is usually compelled topursue these questions rather superlicially. I suSgest that dissectin8 pollination systems inlocomponenls pbose irteractiors can be modeled is the best insurance aSainst making e(o6 ofinlerpretalion in studies ar this level. I give eradpl€s of ene cha.acteristics thar mate polli'nalion syslems .esislant tocoarse analysis, and try lodcmonslrale rhal coDponent analysis doesmore lhan simply to provide insiShr into the mechnnisds ofcommunity levelinreraclions: wbenpu$ued as a syslematic res€arch program, i1 reduccs the probability of nisinrerpretation.

KEY wORDST Con6unily, competilion, component analysis, pollen carryover, bees, fo.aging,consrancy, irflorescerc. architectur€, nectar, rnimic.y, time lag, scalc, presentation.

CONTENTS

Hrporheucal and Real b\amples of Componenl Inrencl ion"

22

Majo. Cooponerrs ol PlanlPollinator Systems

451452

45345ESumminS Up459459

FOREWORD

This chapter was originally to be coauthoredby Dr. R. C. Plowright of thc Dcpartm€nt ofZoology, University of Toronlo Much of theoriginal work described was done by him and

his sludents, and the editorial viewpoint heis intendcd to represent his views as welmine. Unfortunately, at the last minute heunable to f'articipate in the preparationof

452 VII / SYSTEMS ECOI,OG! IN POLLINAIION BIOLOGY

chapter, and I have hastily comPiled it fromrough outlines that we had prePared jointlv.

Therefore. while Dr. Plowright deservos muchof the crcdit for its contents, he cannot be heldresponsible for any misrePr$entalrcns orfaults of expo6ition, which r€main my respon-sibility.

INTRODUCTTON

At least since lhe work of Robertson (1895)'pollination ecologists have been inlercsted incommunitylev€l intemctions, i.e., ecologicalafld evolulionary connections among sP€ciesthat extend beyond the dir€ct relationship ofaplant and its pollimtor. Especially within thelast decade. Dollination community ecologyhas gained mom€nlum as pollination ecologistshave worked th€ir way uP from sp€cific to gen_

eral questions, and as communily ecologistshav€ adopted pollination systems as highlyquantifiable arenas for study. The recognitionthat few plants and few pollinators are obliSatespecialists with regard to pollination compelsthe conclusion that "pollination webs," likefood webs, will often be complex and cross'connected. The success of a pa icular sPeciesh likely to d€pend in part on its position in thismatrix. In this chaDtcr I wish to outline aframework for investigating such phenomena'and to Resh out the framework with a hetero-geneous collection of small, but exemplary'caso studies. I am concerned only with animalpollination, and my examples will reveal a biastoward bumblebees.

One can identify two extreme strategies forcommunity studies, the "toPiown" and "bot-

tom-up" approaches. To illustrate the contrast,consider the case of competition b€tween twoplant species for pollinator servic€s, reviewedby Waser (this book). At the broadest, "totr

down" level, one may either compule simplemeasures, such as blooming'time ov€rlap6'which may have something to do with comPe'tition. or conduct a large-scale experiment in'volvinS removals or additioff of the potentialcomFJetitors, as Was€r (1978a) did Althoughthe oxp€riment is likely to yield an unambig-uously intorpretable result (and the measure'ment of overlap is likely not to), neither aP

proach gives information on the mechanism ofthe intemction. Unraveling mechanisms de-pends on a more detailed, "bottom_up" analy-iis of the cozponenls of the interaction in the

style of C. s. Holling By specifying the be'

havior of small system subunlts, one can pro-

ceed to model system behavior, either in a log-ical, veibal way or explicitly mathematicaUy'in a systems model such as those of [,€vin and

Anderson (1970), Was€r (1978b), or Le z-

man and Cass (this book)Lertzman and Gass discuss the general

philosophical rationale for constructing this

soft of model, and I will therefore not pursue

this point. What I wish to stress is that polli'

nalion systems contain a numb€r ofquirk that

seriously impede "totrdown' analysh and

that dissection of componenti may often be

necessary, not only to see the details of how the

sysiem works, but to avoid erroneous conclu-

sions about the Seneral nature of the intem€-tion under study.

MAJOR COMPONENTS OT PLANT-POLLINATOR SYSTEMS

I first will separate the variable3 of interest

into three major groups. The first of lhese, and

rhe largesl. is lhe colleclion ol all planl attri'

but€s that afrect pollination success l lump all

thes€ characteristics under the general head-

iB plant prcsentation {Fig 22-l) Some of

these are apparenl to pollinatols whereas olh-

ers are cryPiic. A Partial listing of the formerwould include: densitl sftatial pattem; timingof bloomi flower morPhology, color, and sc€nt;

amount, type, and timing of floral rewards;plant height; inflorescence architectur€ (rn-

iludinq flowers per infiorescence and infloresc'

enos pet ptant, as well as rhe lype of infla

resc€nc€); and variation in gender. AsP€cts

lhat arc nol apParcnt to the Pollinators but slill

determin€ repmductive success include such

things as the clonal structure' if any; the ge_

netic neiShborhood sizei the typ€ of incomPal-

ibility syst€m, ifpresent. and the number olin_

ompatibility alleles or mating typ€s; mepollen production and ovule numtter; the func_

tional response of seed set to varying amountl

of pollen; and the se€d_maturing resources

FiBurc 221.

availabl€comp€uuaamong fn

pollinalo6

Planr prev

choice offomgingplants (diringl floveplant; d€gforaging hof unfari

fectivenesand applypollen calcass elser

Pollinalstrict rcf(pollinatonknowlcdS(oftheirFwill deFr

ception rtions. inclnonlineariand thc lildenslly mmighr diflson, l98lquire viril

foragoN iers and thing, wher

3as) serr.ds oal 5ql uaealeq palredxs eq IEurl)]r^r3s uorlPUrllod roJ uorl!l3duor rul].dsJrlurliql srstr88ns uorls.nb eqr Jo Surserqd rqJ

..ilsn rolPurlloour puP uroolq lo eurn ur { Eou"lsqns dEI-.'!^o lrql suooDlndod sarrads ruEld o^\l ur3,r{l'iq uo't.[latur aql Jo.rnlEu aqt sr l?qr\A,, iuorlsrnb r€lnjrlred a 01 uorlBlrr ur Luoql lLrasrrdllra I rilllr8ot slusr'rit. .srql .r1 oI suorl€c-rtoluor Jo tsrutrrqJ pelElouu? uE s? reloertr.qt prrSar ,{Ert] rapsar aqt suolrErelur at?rs-enrt ol relruEr{r Surpurl ur eq ol lno urnlirur strqsp ..tt'ru"q..tu,. parpnlsun {t!&"1.Luos luelodlur ̂{oq 8u!^\oqs Jo l?o3 .ql qtra'suals,{s Jo ,{1.u?^ € |!orJ s.rpnts luiuoduoi-alSurslo uorrrrlss E ssncsrp Ilra I '�aole8 lsrYesirpnls qrns ou purls srellEu sv selpnls ural'3uot ,{q ..pa^tos,. {t38rrt eq um saqrunuluol.ld!t'rs sruos lnq1 l.eJ I puE pl59 ,LIl u' ,{ttsec{1.^rt"l

prUrtu€nb.q uecsLu.ls,{s uorlpurtlodJo stJldse Iupur 'rl"llsuourp Itpelesd-ar looq srql ur srrldEqtr eql sE'l'^eaoH Ilru-nuuor slorta E lnq rcleurlod rtSurs pur sarr-.ds lutld lt;urs r ftuo lou raprsuor rsnu euourqa {ltel]ods! Is?l p?lerrlduor trt?urpprqroJ? sr 5^oq€ prlD stuouodluor .ql P SurpnlcurtePour suals,ts Ie'!roJ " BurtqLUassE'IIr"elJ

SNOIJJYdIINIrNlNOdI^tO) tO S3'IdhtVXS.IV3d CNV'IY]IJSHJOdIH

t-zz 3t!ur uaoqs ar? sluluoduroc rof€ur rseql 3uoru"sdrqsuo|llar eql tuptd eqt Jo uon?luasard.ql srilt€,{t.lr?rp.ruur qlrqa'srolrsr^ raaou.(q rDtrru lo uoDeldrp aql sr etdu"xa snor^qouv slrnqPaal ,{rEuortnlo^. !ur?l-8uol ot uotl-rpp" ur {ltrarrp uorl?tussard lu?td lr.llP oq?Iiu lsuods?r rol€urllod lorluoat cnrueB repun.ilr liql uorlPrurs.rd lu?ld Jo sllrdsE ssoqluo s.rnssrrd a^nJales lexe urnl ur PlnoqssuonEluesrrd luar4rp qlra stu"td Jo sserrnsa^'Fnporosr tuouarauo slurto aqt Joss?.trns.^rtrnpord .ql aioJ.rlql pu?.roheqcq 3ur3EioJ riqt rurul'trp suortEtuas!.rd .slueld eql

lo suonda ed .srolEur od eql 'IlalEtul]In

1086r'{ue^'rt :9t61 'qrDuraH) 3urpr"^\!r {tq3'qs" s.r.i o0 .Lurs .qt e^rlrrrd srnbruqcsr riql

p3lr.Jr.d !.\rq leql saeq r.qto serr'qa '8ur-prraiiuou se lucql r^re ed eDJaraqr pue sre-aou isrqt rlpu?q ol aoq ur"el re^eu srrSPmJ.uros 'seeqelqunq ur lsErl tP :uonr?llxr prE^\-er roJ senbruqJ'l redord uJ?!tol sJolrsh ernb-?r sr?^\og xeldluor {llernlcnrls (186t 'uos-uroqJ) s3l]rds roleurltod 8uo!u€ r.Jrp lqSruqrrqa'.lP.rs lErlEds euros c^lo^ur lsnu ,{lrsuePreaou Jo sloeussessE .srolEurlod slrr eql PUs'ezrs eruers ouur ol esuo<jser ur s.rluEaurluou's3?l lrorl seDrlu.pr u.IEls[u 8u'pnl]u' rsuorl-rolsrp Jo ,(leu?^ e ss?duorus {?u uollda.-I!d s!ql pup .suo!lrlulsrrd

,slu?td !q1r^recradsrol€urlod q.rLta ur ,{EA lqr uo pu5d5p raesuodsar'peslsut slu"td pooJ lpluerod rreql Jos]lls .lrtrrEqr lue^at srtl Jo ltE jo !3p?l^lou)lat.louror .^Erl r3^. Jr uropPs ltra srol?urllooasnm.q uort?lu.s.Id lu?td Jo uorlcaud lcols?.q lc^raorl lou l|a rsuodser rolEurtlod

llooq srql u' eret{hesla ss"cpu" uEruzlre-t Iq prlE t re^oIrrEr u. odJo tuelya aql puE sELuSDs ol tr 3ul{lddr pu?sraqlu" ruo{ uellod 8ur^otuer lE ssous^rl.aJ-Jo :,{u?uol|llal :SururtdErl llu.rul!nrf, or iopnl.t-ur slr$se r0q1o al|srraou r?rtrurEJun JoSurtdlu?s Jo lu:lx: :.{ru?tsuoc lqBrrq 3ur3"rolpue IruBlsuor serreds reaog Jo aer8ep llu"tdrad puE .cuacserouu' rad perrsr^ sr.^\og lSur-tro^lJo perds :(?ruErsrp puE uoDup) slueldu33al3q lu3ru3^oru :Ilrtepu ltrs BurSPJoJ:.iPId JO iCroqi pu? l$r^ ol sorceds Jo lcloql:Surpntcur ',{roaql 3u!8?roJ leurlldo Jo surrluor erE !s:ql Jo rsqunu v uon?luaserd luEld

Jo eredsE luerEddE aql ot puodser srol"urlodqrrq^{ ur sIE^\ rql Suru"au { ErUt]eds arvodt-ar rotrut od sr ,{ro8?lE rotuu puoirs v

'lueld tql urqtra slrn{ SuotuPro lrn{ € urqlra saln^o SuourP uonDaduorroJ terluslod eql puE luEld aql ol etqruE^"

rr.l.'lr ur p.trEr.p.rs!o3rr?tr q.E. urqrrn slu.uoduoJ suorrlr.lur rorDur od-rurd lo s.uoiJtJ ru.uoduo rofrE rql I'za !n8'l

-?3 oqt I

lIls lnq r

-ouur Jo

-u0 ornl

Surrurl :u

Jo suros'pr3rt tErIIE du,nl-|lllE tu?

IlNvr

pu€ \rs,ilBql si{rn-!llod leW

srqt SulclEr.uoS t

-zlre'I .ropuP ullo.I',{llmn?ur-3ol E ur r:

-aq eqt R!.ql ur uoll-,{tEuP ..dl'?p sursrul

tsr SNo|IJY)J3JNI NO|MIttod r3^1-r-,{rrNnn'\]OJlO SISIIVNV INI:IOdNOJ

454 \'rl / SYSTEMS ECOll)Cl IN POLLINATION

was€r. this book). Bu1 lhe easy cquat ion oicc-ological ovcrhp and competilion has beenunder a11ack for some rime (c 8 , Colwcll andFutuyma, l97l i Connel l . 1975. 1980: andmany olhers). 3nd certxinly cannot bc consid'ered to be univcrsally truc. With specilic ref-erence lo pollination systems. sevcral aulhorsnos hale suggested that s imi lat i l i€s in bloomlime and use of pollinators may rn some cNesindi€ate mulual ism .r lher !han compcrr l ronfor pollimlion, espccially whcn flowe|s appearsimilar. In thesc cases. lhe similarily of thellowers has been explarncd as convergenccdriven by a sort ol N'liillctian mimic.y in whicha number ofspccies ma) all re€eive better pof

lination by rcsembling each olhcr (see Cranl1966r Macio.. 19?lr watt e1al. . 197'1: B.ownand Kodric-Brown. 1979r Parrish and Bazzaz.1979; Thomson, 1980. l98l : Schemskc. l98l iLi t l le, this booki Powel land Jones. this book)Thus the.e have bcen two outcomes suggest€dfor plant-planl inleractrons, and thcy are oFposite in elTect: competition and mutualism l

am awa.e of no cfilical tesls of a mimicry hy-pothesis. but w€ can identify certain compGnents that are necessary for such a relationshipto work. Il these componenl inleractions arenot pres€nt in nature. lhe mirnicry hypoth€sisis weakened. Proc€eding through a series ofsreps, I $ill dcmonstratc the slyle of thc

A hypothesis of Mitllerian mimicry firsl ofal l dcpends on a dcnsi ly 'dcpendent or l re 'qucncydepcndenl advantagc. su€h that adcnse populalion receivcs disproporliona!clymore of b€l ler pol l inat ion shen i t compnses alrrger fraclion of the locrl floru' Do pollinatorsbehavc this w,yl Such relat ionships havc beendcscribcd trom .eal sysrems Levin,rnd Andervrn (1970) revicw rhe cvidenc€ for "Arnel ls

dominating flowcr phenomenon. one exprcs-sion ol lrequ€ncy-dependenl advanugc How_evcr. rhis phcnomenon properly describcs rn

incrersc in tu$e. .r)rrllx{t' $iih incrcasinglrequcnc). Morc !o lhc speci l lc point js a dcm_onslrr l ion thl t v is i la i ion ratc on certarn rnsect 'pol l inated spccics not onl t is highcr whercthcir densi ly is higher. but incrcases lur lhc.{hcre simil,rr flowcrs ol ditre.cnl spc€ics co-exist (Thomson. 1980, 1981. 1982a) Thus. atlcr:r .un,( nul l ;nrr"r ' . in( ludrnt bumblebcc'

EIOLOCY

do react posiliyely 1o ffowe. density, and' 'sum the densi l ic: ot \ imi lsr Bo$er specie! in

rher perceplron oi density This behavior saF

i lhes onc or the pf lmar) condit ion\ for mim-

rcry. Al(hough nece\saty. Ihis condit ion is nol'u l ic icnr. o$ing to "minor derr i ls of pol l ina

tor forJging and 'patrdl dispersiun ot lheplanl ' . t t i i nol certain lhal increared visi tal ioncauce\ increa\ed hlne\s. especial l ) in Iherc cir-

uum\rdnlcs, $hcre di trcrent spccies arc in-volved. lf pollinators .€spond to mixed'speciesarrJ) ' ol s imi lar l lo$ef\ b) moving incon'

slantly from onc plant species to another' lhehetero\pecihc trrnsler ol pol lcn md) sclual l )reducc seed \el by \sr ious mechanisms d(-s€ribed by waser (1978a,b. this book) Besidesthe $ashge ol pollen and the loss of stigmalic,ur lace emphd\rzed in Watis ( lq78a)modeh. "allelopathic" poll€n has been found

rn tuo Senerd of ( omposrtae I Pa henium Sn'

kada and Jayachandra. 1980i Hietuciumthom,on er al . , lq82br. The\e pol lenc rnhibi lsuccessful fruiling in c€rhin olh€r specieswhen appl ied In mi)\ lures si lh conspecihc pol-

len gra'ns.The degree of constancy displayed bv lle

pol l inaror. lan arpect of pol l inaror re\pon'e)rhu' r ' . lear l) rmporlanr to planl-planl Inte _

actions (Levin and Anderson. I9?0: Straw'19721 Bobisud and Neuhaus, 1975) l l musl becmphdsi/€d rhdr mrny $i ld pol l inalo's mav bewillins to foragc inconstantly. and that theconcept of pollinator constancy has perhapsbeen undLl) inf luenced by the wcl l -srudicd butunusually high constancy ol honevbee workers(Free. 1970i Crant, 1950i Faegri and van derPi l l . lq7e. p. 50r Indrvrdudl bumblebee.. lore\amplc. communl! v is ir lwo or mote pldnl

specics on a single fomging trip Such in€on'sranc) md) ha\e J samphng luncl ion in l rack'ing changcs in rcsources (Heinr ich l9?6.la7a. Oncr and Hernr ich lq '?6): in oth€r(dse!, be(ause Indir id ' ral bces con\ i : renlb 'c-

lurn to thc same individual plants in a mixed-sDecies traplinc le.g.. Anlia his da ^ndnrbus sp.. Ttromson. Maddison, and PIGwright, unpub. ms.), the bebavior scems l€ssinle.prelable as sampling and probably repre_sents simple hatvesting. Indeed' in the hvpcthet ical sit uation under consideration-similarflowers, inlermingled to some cxtent-where

wards, optil

stanl sfnletWh€n flo

ble animalsit may PanaagSregatlonstancy," Thpecially eviings. If iw

visns willbrdisplay noason. l9?0;\

involv€d(clsiz€ increas

the two plaage (Thom

The argtthat the ovplanB h n

effectsi i.o.

Thomson,the interpl

di t ion ofdinteminglthorougn Ir€al planl.

spac€" dcll982ai loc

tionshipssively by Iempirical

COMPONf,NT ANALISIS OF COMMUNITY-LEVIL POLLINATION INTERACTIONS 455

lensity, and

dition is not" oi pollina-

ed visitation

)a! acfually

ok). Besidesolstigmatic's (1978a)

\eniun, Sn-

llens inhibil

+ecilic pol-

yed by rhe

planl Inter 'r70i Stra*.

d rhal rheas perhapssrudi€d bur

ich, I976.

and PIG

ably repre-lhe hypc

in most cases the plants yield roughly equal re-wards, optimal foraging models that includetravel costs usually would predict an incon-slant strategy.

When flowertonstancy does occur in flexi-ble animals like bumblebeesor hummingbirds,it may partially or wholly reflect simple spatialaggregalion of flowef species ("passive con-stancy," Thomson, 1982a). which will b€ es-pecially evident in monocultural crop plant-ings. If two planl species intermingle in amosaic of monospeci6c patches, heterosFr€cificvisits *ill be comparatively rar€ evcn if visitorsdisplay no active constancy (L€vin and Ander-son, l9?0; Waser, 1978a), especially ifthe vis-itors mainlain reslricled foraging areas thatare smallrelative 1() the patchsizeofthe plantsinvolved(cf. Solbrig, 1975). However, as patchsize increases, any "mimelic" advantages pre-sumably will be dissipated because the polli'nators no longer will sum the contributions ofth€ two plant species in deciding where to for_ase (Tbomson, I98la).

The argument so far should suffice to showlhat the overall effect of pollinator sbaring bvplants is not easily predicted wirhout payingcarefu I altention to th€ 6ne structure of th€ in_teraction. A numb€r of components are in_volv€d, and different componenis haveoppositeelTects; i.e., som€ tend to produce mutualismand some comp€tition. Figure 22-2 (fromThomson, 1982a) is an att€mpt lo summarizethe interplay among spatial pattern. visitationrate, and reproduclive success, with the €on-dition of decreasing conslancy with increasingintermingling. No existing study has beenthorough enough to determine whether anyreat plant-plant interaction actually 6ts thecurves of Fig. 22-2, bul experimental studieshav€ shown that different natural syslernsoccur in different regions of the "success

space" d€picted (Locations I and 2, Thomsofl.l982ai location 3, Waser, l9?8b)

Another phenomenon affecting plant rcla-tionships is pollen carryover, treated exten-sively by Lertzman and Gass in this book. Allempirical tr€atmenls of carryove. to date haveconsideredone speciesonly, but i$ importancein helerospecific visit sequen€es is readily aPpareni. Whcn a poll;nator visits a fiower ofspecies A, then B. then A, how rnuch pollen

Highly Inr..minll.dinr.rminal.d .lumP.

Fisrre 22.2 A sunmary ofthc r4vs in rhich spalialin_

r.rminglinB nay ailct the inl$action bel*ecn l*oPlanr

speci.slhal shar. pollinatos Whcn plants dcli 1n sP

arat.. discreic Palches. tbe patchesma! compet for vn

no6. As th. speciesi.lermingl. nore, visilation rate nav

rise as rhc pollinato6 rcspond lo th. grcalcr denmv.

Ho4ver, seed s.r nay droPorat crtrcncs ofior.rmin'

8li.s o*inB to lhc incr€asin8 lraction oi herer6pcci6c

will the second A receive from tbe 6rst? Al-though data ar€ lacking, assumptions of pub-lished compelition models range from Stmw s(1972) propodal that carryover is essentiallycomplete, to Levin and Andersont (1970)view that it is completely lacking. The Levin-Ande$on model reflected th€ then-pr€vailingview that carryove. within a sp€cies is gener'ally low. with most pollen being placed on lhenext ffower visited (Levin and Kerster. 1967,1969). Rec€nl experimental work suggeststhat carryover is often more extensive thanrhis, ln Erythroniun ome.icanun, E. grundi

forun. Clintonia borcalk, ^nd Dienilla loni.e.a (all flowers wirh open corollas, plentifulexposed pollen, and imprecise channeling ofvisirors over the rcproduclive parts), lhe pollenput on a nectar-feeding bumblebee by on€flower declines roughly as an exponentraldecay with a "grain half-life" of between oneand two flower visiis. However, the lail of thefall-off curve usually is longer tban predictedby exponential decayi i.e.. a small fraction ofgmins trav€l considerable distances The.e isgreat flower-t}flo$er variability in deposition'

456 V|I / SYSTEMS LCOLOCY l\ POI-LINATIO\ BlOl.O(;\

much of which depends on chrnce posi t ion ing

of the bccon the 8ower. a l though Eowcrs wi lh

exlra nectar .eceivc nore grains (Thomson

and Plowr ight . 1980i Thomson. unpub ) . Pollen loss from the bee's bod) is due moslly logrooming, which is donc n st ly in f l i8h l ln

Et\lhtuniun grundiforun, pollen dcpositionfalls oll f.tsler when bees fir- belwccn llowers

than when they valk . l l is c lear thdl most he1

crospe€ilic lransfe.s of pollcn will involvcf i ights. and tha! pol len loss wi l l o f ten be sub

slant ia l . Whcther rh is loss wi l l have.rnr- !p-preciablc imprct on thc sced sel of thc rccipts

ents dcpends on lhc nunber o l Sra ins actual ly

carr ied, which wi l l v , r ry s i lh thc pol l inator i onthe numbcr of ovules to bc fcrtilizedi and onlhe funclional response of seed set to pollinr_

lion intensity. Despitc bcing one of thc norc

impor ianl componcnls o l anv quantr tat rve

model of pollination success. this funcrional re

sponse xpparently has been describcd lor vety

lew plan! spccies (oendlrcn Jnni&sa. Silan'

der rnd Primack. 1978i Pll$iloru rti./olia ffidGelaniu' na&lotun, A. Sno$. pcrs. comm ).Extensivc hand-pollinations in (li onia bGrcalis and Medeold rils,,iana have r-ieldcdonly ,t very weak posilivc rel,rrionship betwccnthe number ol outcr6s grains applicd and thenumber or proportion ol ovules devclopcd (Plo-

wright and Thonlson, unpub. dau). The weak-ness ofrhe pollination/seed set rclalionship lbrthese spccics is due largely !o cxtensive l .u j labortion, which may represcnr incompaljblc orscnicompat ib le pol l ina l ions (c f . Ber l in , I 98 I ) ,or possib ly denciencicsol technique In! iewolthe general lack of data, lhe cxtcnt 10 Fhichthe seed set .esponse is couplcd to pollination

should become an aclivc rcscarch topic: {ith

oul such dala, the significance of pollinationr,t1e and oi pollen car.yover is very dillicult 10

The role ol grooning in pollen carrr-over isanolhcr poo.ly unde.stood rclaltunship {ithinrporianl. somelimes clabor,rle .amifi citions.Hart l inS and Plowr ight ( l979a,br Plowr ighland Ha.1lin8. unpub. ms ) found that ! bum-blcbcc could pollinale as man! as 50 norelsaf tcr i ts , r r iva l on a hcad of .ed c lover / I / i

.lalium pntense), a much hrge. number than

supposed prcviousl]. I ,a.arcrr. is sclLinconrpat ib le. and r r had bccn though! that only a

few florcr\ could bc lcrtilized on I new plant

bcfue sel l po l lcn over lhelmcd rhe oulcrosspol len on thc bee's body (Free and Eut lcr .

1959i Michcner, 197,1) . The outcross pol len

thar ; dcposited on clover florcts .rppirrcntl)-!

comcs from rhe proboscidial loss! (Furgala er

a l . . 1960i Spencer-Booth, 1965). which nrust

nor becomc ovcrhid by inconpatiblc sclf pol-

lcn unt i l thc bee grooms. af tcr lcrv ing the

llower head. Be€s seenr unable 10 gtuom the

fossa c lcan, and thcy sweep pol lcn in to i t dur

ing othcr groo,n ing movemcnts. whi le nor

firmly- demonstratcd. this mcch.tnism rs thc

most l ike ly cxplannt ion tor the cx lcnivewi lh in-head cNrryover . Bccs aho rcguldtc the

number of llorels visilcd on a head according

to lhe an unt o i ncctar in lhc lbrcts (Hrr

t l ing, 1979i Ha. l l ing and Plowr ish l . 1979a) l l

the.c is hye.ing of pollen iro succcssivc

hcads in the proboscidial fossa. wrth thc depth

of cach heads st r r tum pruport iooal lo lhe

numberof lhrets v is i lcdon rhe he: ld , rnd l l lhc

. ]moun! o i pol lcn rcmovcd i ronr thc supcr l ic ia l

s t .ata du. ing a v is i l to . t head a lso dcpends on

the number of florers lisited. interestinS con-

scqucnces follow. Since ncclrr-rich plants rre

thc onlJ" ones whose sligmas will cu1 down lo

the deepcsl l,ryers and the onl) oncs whose pol

len wi l l subsequent ly bc swept in tu those lay

ers, thefc can be ssso.trtive maling lbr neclar

levels wirhout any d iscr iminal ion by thc pol l i -

naloA. Neclar r ich p laols could a lso hate I

largcr neighborhood sizc in thrt somc of rheir

dceply dcposited grains m.rr- lravcl considera

ble dislanccs bcfore bcing brouShl to lighl (scc

Leruman and Grss. this book, on thc cllccts

of pol len la ler ingi .The conncctions belwccn neclrf sccrction

and pol l ina lor lbraSinS hale been made explicil in r systems modcl lor pollination in rcd

clovcr xssembled by Plowright .rnd Harlling(unpub. s . ) us ing p! . rmeler v l lues f rom l ic ld

expcrimenls rnd the |lssurnplion ofexponcntirldccry oi pollcn. A flow chrrl for lhis model isg iven in F ig. 22-3. Al though th is form ol thc

nrodel does nol incorpor.rte thc pollcnlrlcnng

elTccts discussed above. il docs !llo{ onc 1o ex-

t

I

,I58 VII / SYSTEMS ECOLOGY IN POLLINATION BIOLOCY

amin€ the efects of changing the rewardschedule of the nowers, the number of incom_patibility alleles in the clover population, andthe rul€s of movement of the bees. Analyzingbee movement in th€ cont€xt of plant-pollina-tor interaction was. in fact, thc original ratiGnale for buitding th€ model. Plorrright andHartling used it to show that the flexiblc for-aging behavio. actually displayed by bunble-bees also resuh€d in higher seed sets thanbehavior representing previous casual descritrtions of be€ behavior ( Michener. I 974, p. 3 27 ).The model also points out the inadequacy ofestimating seed set in a stand of clover fromsimply determining the frequency of interplanttransitions. Behavior that maximizes inter-plant moves does not maximize seed set.

Manipulations of models of this sort canshed lighton other pollinaiion interactions. Forexample, it is possibl€ to determine the nectarsecretion rate thal yields optimum seed set. Inobligate outcrossers like red clover, this opti-mum clearly will often involve a "compro-

mise" nectar production. rich enouSh to pr€-vent visitors from visiting few flowers per plantbut not so rich as to make them stay so longthat their outcross polten becomes €xhaustedor overlaid (see Heinrich and Rav€n, 1972).The residence dme of pollinaton on an inflo-rescenc€ may also vary with the architectureofthe inflorescence and with the pattern of nec-tar secrction. For instance. Pyke (1978) hasproposed that many planls with vertical inflo_rescences present more nectar in lhe lowerflowers, and that this presentatior prompts thevisitors to move upward and to depan when re'wards diminish. Again. analysis of the com-ponents of presentation. pcrception. and re-sponse €conomically reveah the workings of asystem subunit and suggests a co€volutjonaryinierpretation. However. one must be verycautious in interpreting inflorescence design asa means of optimizing outcrossing by balanc-ing vhitalion and carryover as above. sincemany plants have multiple infforescenc€s. Insuch cases, a spike or head that is optimallysuited for receiving a pollinator's fint visit toihe plant will be progressivety less suiled forsuccessive visits made within the same plant.Theories of inffor€scenc€ evolution (e-9.. will_

son and Rathcke, l9?4; Willson and Pricc,1977) must reflect this unpredictabilily as wellas the modular nature of most infloresccnces(Harpe., 1977).

SUMMING UP

As I warned in the introduction, this is pri-marily a didactic essay. I concede that it alsomay seem. depending on the tolemnce of th€reader, €ither fragmentary atd incomplote orsimply pr€mature. My main point is to advccate a systems aPproach, yet I have presentedonly one real systems model (th€ clovermodel), and €ven that only includes a smallsubset of the componenls that I hav€ identifiedas important. My explanation of this embar'rassing situation, as might be cxpeclod, hasseveral subunils.

First, I fe€l ihat the advantages of a com-ponent-analysis ̂pprcach in rcvealing detailare evideni enough in subunit mod€ls (such asthe red clover model cit€d here and Lertzmanand Oass's carryover models) that one neednot wail for a complete community'level treat_ment to proclaim them. The wait would belong because the intensity ofstudy required toyi€ld the neccssaty detail means that onlysmall portions of systems usually can be suc'cessfully analyzed in the couNe of a disserta_tion project. Unless long-term coopomtivestudies b€come moae common than they arcnow, our knowledge of communitylevel inter-aciions will continue to consist of a number ofisolated glimpses of various parts of varioussystems. Of course, this is the prevailing situ-ation in all of community ecology, and if pol_lination ecologists are forccd to draw too manyinfcrences from too few studies-if bumblc-bees, and Delphiniun and lpomopsis andErythrcnium, reach the talisman status ofintertidal invertebrates and W€sl Indian an-oles-w€ are no wors€ or bttcr o{T than ourcolloagues. This stale of affairs provides oneanswer to thequestion. "Why is so much detailneeded?" Because we will p.obably contioueto be heavily dep€ndent on a small number ofstudies, and continue to "solve" other systemsby analogy and extension, it is critical to real_ize that small differences in structure may

ently. BrcaU

difcrenc€s.

in most aPD[is to menipulmaximlza r Isccd sct. lt i!the rclcventin arriving alimum of tril

Ono still tfon,ard quc

P€rimcnts IHowcvar, tl

i'l undctccllConsider ewhethcr orcompct€ fo

to scttlo thiconceivrbLthc flo\fcrt(see ThoN'tail" that tGshum. ibpollinatots,1949: Srng1956; Hcirritory holdrto me $flspons. of !

may ba to

of compctiHowevcr,I

speci€s hacthem (sa.

ExamPlm€ that fiircms usu!amount oldetail lnulsmall-scali

t only a

Butler.

ParenllYrgala el

self pol-4ng ihe

hil€ not

late tbe)codlngs (HaFt79^). t r

re depth

nd if thepe cial

ng con-

ose lay-

he poll''

of their

ght (s€e

tartlingom 6eld

nodel is

ayering

COMPONENT ANALYSTS OF COMMUNITY-LEVEI' POLLINATION INTERACTTONS 45'

Fignre 22-1. A flo* chart oI rh' Aombus Ttifoli'n nciLer discu$cd in lh' texr Addnionaldetails mav be lound in

Plowrisbl and Hartling (unplb ms )

;;i.tr" ,r.1.r., ...d.'lv (r.€quar r r .qu . .c l . . ) to e rar '; ' i . H , I r - . " r L x 1 " s

i--.rc* b- d.."r.v! oo re .phs : cooP ' t e "o o

P . " " . G d f d . * h b . h * . " - t y p "

ttil-al'x,-.: ;.:::'i::"''* "

; - ry * P . "b 'b l l t v rP , " r

ampue * . co@d ib i l l t c h " * . ' ! , ^ ' " 3 1 r . * .

Price,

|s prF

of lh€

:nli6ed

,d, has

il treat-)uld b€

ti only

ng situ-I if pol-

h detail

COMPONENT ANALYSIS Of COMMUNITY'LEVEL POLLINAIION INTERACTIONS 459

make otherwise rimrlar svsrems behave difrer' esting or mor€ imponant lhan l^iCt;t::f^.::

"nii". g.""ting interactions into components ological and evolutionary Processes Howev€r'

l#ir#l^"". it" importance of rhese see-mtngly rriviat derails of plant pr€se nratron

dinerences "J$.Tllilj":iyff:"."ilH11",:il"::f:A second, more practical need for detarl rs

* l*, ,ppiill*iri* "here ihe soal usLrallv comperitive or mutualistic and .r€lecr

ror or

r. - -,"i.tfrt. lf'" poffination syst€m 50 as lo against certain elolulronary drrectrons l nev

."r'.ii.'" p*r."rit q"anlitv such as fruit or cannot be safelv ignored

seed sel. lt is obvious that an understandrng-oI

rhe relevant components \rill often be helplul ACI(\OWLEDCMENT

ilil:i'.:\'JJi:::::,::i'.'S,li:l i"Til' :::ii::*tffll"J ,'l};l$;it,'J;:lTl;,:ri,One nill might oblect that ce arn slmrgnr

io_rr'c r"rro"i"e ..nt*s "l rh. Plosnsht kb ror rh'tr

forward questions can be settled by slnSle ex- di"(u$ion or 'h.i. idea. L.{er Hanlins La*t'nm

oerimenti withour inv€nigalinS Inlricacies Hddcr' 'nd r*o( lNdt) M'n) (mkrvt'n' rr'"1 f " " . . " " t . , f t . r .

are easi ly- imagined c i rcum_ * 'nr ' * ' r * nu 'crou: b l isr h ' lPedshap' rhrch 'p ier '

:i;;;; ;;;"; ;""'"".e oraetitrnav re'utr ;1h":l$::,ff'::jl:,Y,"3,,';ffff::ii"il:i'in undelecred f fa*s in experimenlal 0." ,9n: i . i -" i i . "" i e. ,o.myors. icnc.sandsomcblthcN, 'Conrider again the "simple quer(ion ol ; ai sLi.ncn and tns,nccnns R6'arch (oun('l or

whether or not two co€xisting plant sP€cres catada

comDete for pollination Elsewhere rn thts

booi, waser argues cogendv for erpenments LITERATURE CITEDh qettle lhis question Hov€ver' the simplesl

concei"uUt.e*pe''rnent-remouing or bagging B'ain R leSl Pat'rnrv and frul ProdErion 'h trLm'

ii"i'"".., "il,* O *. p.enrial comp€rirors pet ct..p.t tdnpi, ndian!' autt E.ot sd 44

i:. ' i i", "".ls8..iterar-isnoresthe de- "#,ll l, ' i":lf 'I. ', \euh.u" rq'5 porind,o,,o..

rait rhar lh€ pollinarot communitv do.s *t "],*y ""a Jf -.""r .i dr. :@cics o..o/o8,,

rerhume i lset l randomly ever) day lmporlan( 2l 2hr '7?2

""iii*i"", i*f"a'"g honeybees {Rrbbands' arorn J H andA Kodric'Brcwn re?e coNr3enc!'

iil,j,"iJ,r*l;,o,lu"tbl*:.,,y*lilq: :iTfrilil,1ff#,1:'"ft;.X:::f;:J%"];li'l950: Heinr ich, l916l. and bi tds lar least ler

l ' , ir"r i"ri ' , . t- '*r" tq7lr. rend to return , "13'.u. ^. * -o o r Fu,uymd ralr on,r'. m.d-io th. sa.e foruging a"as each day The *_ ""..*"' or nrhe brcadth and os'ndp l'"/o'r

sDonse of such anrmals to the sudden dr 'aP <) (hr '116

":";;;;i; iormerlv fa"oreo flower tvpe connerr'r H ro75 som' m(hani'n5,prod-u!'n8 rruc'

fr;;t;"t'";;;,i.r specier in the area ',"*';.:':Hi;"JH;'S 'ii#;:"Ti ?"il::which *oufd be rnlerpreted ar a connrmallon 6ond \.ds.t. F,otagl ond E\oluuonot Connuni'e.

of comoelition for visitors between the plants Hailai! u.iv.sitv Pr.s. canbridse, M'$ Pp a6o-

Howev;r, ir is possible thal the animah would 4ao

l ' . , . ' ." l . , l 'L '* * tn. u,"u 11t1' '" lemoved - ra8o Dn't lrvandt

sDecies had not been present initialty to attract rros or lh' ghct of comp'tuon Ptsl oild r5: l :r l -

,i",n'i"*lir"','*". tst8). lf so' the exp€ri- ,"llli * -o ,. *. *. "t . tsie rhe Ptincit)tes olment reveats nothing aboul competilion be- P; iatian Ftatoe\' P.cdnon P''s o\ro'd

,*"" ift" "p""i." "nait "atLrralcircumslances FtL ! B tslo t6e't Potlination of ctops Ac d'm"

txamples lite thore given above conlince Prcs' London

.ii"ji'f,"ir"* tJ i"ir';bout potlination svs' Fre' J B andc c Bur'r' rq5e '"udl'b'€s corrios'

r.r.'".",irv irrr requrre disseition and "o'e f"T;|,:"i., K w luckcr' dnd r c Hordasav rabo

amount of mod€l building My em?hdsis^on PlLn" 'n rhe prob*B ross' or honc-jba: ro'Pin8

alr" i f .* t "o, be taken to imply lhat | f ind ccna,n r .suma B" .wov at: t tq-r t t

- ' - . . - ."_."ff-Ji In""ftu"i"uf processes more inter- crait K' A 1966 A htpolh6is concrn'nB rb' P!'va_

J 6 | \ l t \ \ \ l l \ t \ | ( o l ( x \ l \ f o l I l \ \ l l { ) \ l J t o l ( x \

. n ! . , i r , . L l r , , r , t , r I ( . L i , r r . r ' r r n x t [ , L

l l , l . h r , r \ r I r | s r r \

( : . , r \ l r ) ; L r r l l i r . r . , i n r r r \ , t , i r i r r ' r / n

l l , : f r f I r ' ' l r , r i r r l i / r r , r , r ' r r ' r . i r r ! \ ! I

t . L , . t t s . \ r r l . ' l

1 1 . , , 1 , ! . 1 | \ ! r ' )

r f LB i r r b t r n l . l . l , ! i l , , r . , f r r i " l , r l r . r \ r L r ! l L | \ '

, . . : i , r . ! i ! 1 , \ . r ' L r f r r r f r r f 1 i I L \ r \ . r r i i , ! h

\ l \ r Lh r . . . L r L t r r r ! , r . 1n rL f ) l . l l

l l . n . f r I K L r l R ( l ' 1 , ' \ r ! l L l

l , t ] ] l ] l I l . r . L n d

l , \ L ! r , h ! r . r ' , i t L l ! , \ . r n 11 , \ 1 , . ! L , r r : . r . r :

i f . L l \ r r / t , , . / r , / r / d \ 1 , 1 / i ' r / ' , i r l i r r , ' r \ l . l

r L l \ r , 1 . l \ t \ L . t \ : . . \ l r . I ' r , ) \ , J '

e ' r ! l , . , ! r n t r r \ h L r r r \ r . \ , r i ! 1 . \ r l

, f ! . L r l , r \ f l . L . Lb , , r f r r \ r L , | i . r r / , / I i

l l r . f . , . l l ' r _ r r \ . r , r . r i ! r ' . . r . 1 r r , . , 1 , 1

, L l r , r l { l [ f b . . . 1 i ] t n , r l r 1 : s

l q q \ l . L \ , , t L r l l , t r , f , , r t [ , ] . ' i r !

l . LL , L rh , r \ . / r ' ' , r , i

. . r , , , i , r ' r 6 1 r , 1 i l i i

l l L , . . 1 l l r i , l ! ) l l R . , ! f f r _ l I n r r L i ' , r , 1 )

, i r . f j t . l

L . L \ r r I \ 1 l r i r I l ' r l , s . r . L r L [ . ] r I

n r f h ! r r t \ t , t r : t ! , n , t t . \ r \ . r I

l r , r l ) \ L , J \ r \ \ \ , , d r , - I r ) ' I ( , n f . r L i I

t , l r . r r \ b (1 r l r t r \ : , ! L r r !

r 4 \ , r r t t r i 1 6

| ! ! , , , ) \ . L t r d l l \ k .F r r ' t ,

: , n t r r l t r ! 1 . \ . n r , r r J , , t l , r i r \ / I l r r r r ' l | !

l ! r ' ) l ) tD \ 1 \ r l r f f n , l t l L t . r . J . r . r r / r l

L L i r , , r . \ I l ! _ 1 l r f i [ ! . , 1 . ] , L l h ! r . , \ L . L L l f r . f n

, , , r ! , r i \ . , L l . f r r r i h r r n r r r r r b n : , r , t . L r

/ / , i r , ' , , " i r , \ . r l , ' i : i

\ 1 , ! , . . 1 \ \ ! '

\ 1 . r , , , ! \ ! i 6 \ , { , , ! $ l r ! r \ , L l r r , r . ! . r h t I \

i , , I L ! f h t r r . \ 1 , , / , , , r , , i r ' ) [ 1 l

\ l . l , . i ! , . ( l ) r ) ' l / r , \ , ! , , 1 | n , r ' r r ' , i I [ ] n i !

I n r f . L l l ' , s i l l . L r . , r ( l t I \ . r LL \ r ' . n ( r

( ) , L r : t r r r L l l l | . r r . . r , _ r \ \ t r \ L l , , l . u I u r \ . i

r , , , , ' \ n , . L L l r r i ! . r r r ' L r r r J ! / . , i t / , n l f ' '

l ' , , n ' 1 , \ l ) n r r l I \ U i / / L / i ! L ) I i r r r . r l

t , , . , r f r r t l L ! , \ LL r i 1 \ i r L l \

r r r 1 t \ L . L i 1 , , { n r , t r t r r r r \ / , r ' i , . 1 . n , r l l l

l \ l r . ( r l e -N ( ) f r { L J ; . L t t i !

1 , n l f L , L r , 1 $ i r , r i . : I r i r r , 1 , . , , r 1 r ' r s l r ' l

l r L h h . L t r L l ' ( l 1 l r ' l ) l l r . r , , f , ! r t r , . r \ , ! L I L r ! l L r L l L I

h , , ) . \ h ! t ! . / l L , , / i , , 1 ! J r r

R , , l r . n ! r , ( I t r l l i . i h , \ i r ! , : l r . _ a . \ , ' r '

r r : L l ! , r , i , ! t r t ! L L , r . , r r i t r L i ] | r r r '

l \ , , . 1 . , , r ' r . .

s ! ! n . r r l r \ ! \ l l r , . r , r r " r ' ! ! I I I

r . { i l r . , , , ! r L r l . . . l r . r L r L i r ,

\ ] . , . J . r . I \ , n L l t l r l j f l l I \ r r ' r , t I l

r r N L r , i , 1 . . - . r r

, . r 1 , , , , r i r r r / . l i \ r r u l r L , I i : '

\ r f , . s t i L j l l r l . , r ' i , r i t . ,' r ' r ! r . . L l . l U I ' r r ! i l r ! ' r / \ r \ r I \ \

s f , r , , L . . ( ) I ! ' ( ) , L l : . , 1 , r ! 1 , . r . ! t . . r

. , 1 . i I r . i r / L L . , , ,

s l , r n ! r , r l , L , \ r f ' l l i ! ! . 1

f , ' r . . . , , L l r . f 1 . 1 , L . L | . , , 1 r .

f . , , . , . ! , 1 , . . , . / r f , A , ! I ! i t L l

! , . , $ r t \ 1 i r \ \ l . L , l , \ r L l r ' , , r ' I r r . r l

: 1 . , r \ . , r l ! , r ' l l L : , r r L i \ i J I r ' q ' I l L l

\ t r r ! l . L | , l L L r , . 1 . , l ! r P . . L | , l i I r r

l t N l , l i ! t r , : r f r \ ' . l i ' L ' , r \ l _ i ' l r

l r t r L l r , . \ r l t \ l r r l , I r , , r . . , r .i t L , , . l t ) ( / ' i r , . , / , , ,

. r A , , , \ , , . / , r l \ , , i r ; n r I

l l f , 1 l ) ! \ l l ( L ' r ' L r L l . i l , i . : i , I l

L ! i | ! . r i , ! L r ! i | / r i n r r r r r

r 4 , 1 , r l l \ . i 1 . r r r , l r r ' l

i i r f t ! . L L , r i 1 L l , r r , . r ' . i ! , r , L l r . l

r r \ \ t . r 1 . L r l l, . r t r \ ( . . r ! l L r { . 1 . . / , , r r i

t r , l . L l l r r . r ) i , r i l l ' i 1

l r r r ! r . I l ) l l I \ i L l ' ! r ' r r l l i I I ' r \ l l r

! s l I I r r ! l ! . 1' , , r r / , , r r r i r l { , , r \ 1 ( r r l , , ! . r \ (

l , , , ! , r ) L , J R ( I ' i i f f I ) \ r r . r . r r \

\ r r r . . | , : f r r I L l . . ! l r , , f , : r . . . f r r L r . ' .

. . r , t : . f t r . . r , i / ) r , r , r r i r r ' ! , 1 . , r r r , i , r r , ' - r f \

\ \ . t r | l \ \ 1 /

, r t f r r r r n I f h i : l

r r r L l \ r ! t L , . r r , l . r \ r r i f f r \ ' ( r r . ! r l r i l \

f h f , i r t , L l r ' r r - t { J

\ r ' \ r l I ( l l ! r L , ! l \ { , \ l i . l / , \ r . r . L r

, i r , r , r L I r f t r r . i ) , . , r , , , r r I , . ,

\ \ l L . , i \ l | . , f J l ' \ \ j i , . , l q ' _ r r , : 1 L L ' ,

r r l . t r r \ r . r i / . , ,

\ \ L - { r \ l I r J l l I i i l , ! | . ! I \ L L r : r r l r . !

, r ' 1 . 1 , f . , L l : . r r r , , r / , i ! ! L , . n J l r 4 i r l r l i

FLO

l ' , r 1 .

\ i . rt h