Self and Cross Fertilization in Plants. I

8/12/2019 Self and Cross Fertilization in Plants. I

1/13

Self- and Cross-Fertilization in Plants. I. Functional DimensionsAuthor(s): David G. Lloyd and Daniel J. SchoenSource: International Journal of Plant Sciences, Vol. 153, No. 3, Part 1 (Sep., 1992), pp. 358-369Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/2995676.

Accessed: 19/10/2013 14:34

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Pressis collaborating with JSTOR to digitize, preserve and extend access to

International Journal of Plant Sciences.

http://www.jstor.org

This content downloaded from 200.24.16.109 on Sat, 19 Oct 2013 14:34:23 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/stable/2995676?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2995676?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=ucpress


2/13

Int. J. Plant Sci. 153(3):358-369. 1992.? 1992 by The Universityof Chicago.All rightsreserved.1058-5893/92/5303-0010$02.00SELF-ANDCROSS-FERTILIZATIONNPLANTS.1. FUNCTIONALIMENSIONS

DAVID G. LLOYD AND DANIEL J. SCHOENPlantand MicrobialSciences,University of Canterbury, rivate Bag,Christchurch,New Zealand;and BiologyDepartment,McGillUniversity, Montreal,QuebecH3A IBI, CanadaMany functional-ecological, morphological,and physiological-factors affect the occurrenceof self-fertilization.Six modes of self-pollination redistinguished.Thesediffer n whether hey utilizespecializedflowers,whether hey involve the transferof pollen withinor betweenflowers,whether hey are auton-omous or mediated by vectors, and their timing relative to opportunities or outcrossing.The variousmodes of selfing are subject to differentstructural onstraints.Prepotency, he preferential uccess ofcross-pollen n achieving ertilizationswhenit competes withself-pollen, nfluences he frequencyof self-fertilization n some species.The amount of self-fertilizationmay depend on environmentalconditionsand the vector species visiting each flower and may vary among the flowers of one plant. To gaininformationon the prevalenceof autonomousself-pollination,66 species for which the degreesof self-compatibilityand autofertility seed set in isolation) have been publishedweresurveyed. Partiallyself-incompatiblespecies (in which the seed set is lowerafterself-pollination han afterseparateoutcrosses)

have on average ower autofertility han self-compatiblespecies (in which self- and cross-pollinationssucceedequallywell), but some partiallyself-incompatible pecies have considerableautofertilityandsome self-compatible pecieshave none. A number of featuresof floralmorphologyand phenologyareassociatedwith high Autofertility ndices.Introduction

Thecomparison f self-andcross-fertilizationis the central opic of floralbiology.Followingthediscovery y Knight1799)andDarwin1868,1876) hatcross-fertilizations advantageouse-cause t produces uperior rogeny, herewas aperiodof intenseactivity n pollinationbiologyin the laterdecadesof the nineteenth entury.Flower tructureshatencourage ross-fertiliza-tion and reduceself-fertilization ere studiedwidely.After onsiderableebateduringhis pe-riod, t wasalso acknowledgedhat someplantsareadapted o regularelf-fertilizationnd havefloral yndromeshatcontrasttronglywith hoseassociatedwithoutcrossingsee Darwin 1876]andHenslow1879] ordivergent iewpoints ndMiuller's1883]brilliant esolution f the issue).Floralbiologybecamepopular gainnthe sec-ond halfof thetwentieth entury.nthisperiod,geneticstudieshavedominated omparisons fself- and cross-fertilization.hegeneticstudieshave providedmuch-needed mpirical nfor-mationon severalaspectsof self- andcross-fer-tilization, ncluding requencies f self-fertiliza-tion(BarrettndEckert 990), heexpressionndcauses finbreedingepressionCharlesworthndCharlesworth 987; Barrettand Charlesworth1991),andthe geneticstructuresf outcrossingand selfingpopulations Brown1990;Hamrickand Godt 1990;Ritland1990).The theoreticaleffects f themating ystemon the genetic truc-turesof regularly elfingandoutcrossing opu-lationshavealso been analyzedwidely(Brown1990;Ritland1990).Manuscript eceivedMarch1992;revisedmanuscript eceivedMay 1992.

The functionalapproach to the study of self-and cross-fertilizationcontrastswith the geneticone in emphasizing the operation of pollinationmechanismsandaspects of the naturalhistoryofflowers.Functionalstudiesof matingsystemsex-amine ecological,morphological,and physiolog-ical perspectives.They have continued from thelast centuryto the presentday with little changeand currentlyconstitutea usefulbut ratherstaticaspect of matingsystems. In recentyears, manytopics of floralecology have beenrejuvenatedbyinnovative studies of reproductivestrategiesfordeployingadaptivemechanisms,butthe newpar-adigm has had hardlyany impact on the tradi-tional topic of cross- versus self-fertilization.We believe that functionalaspects of self-fer-tilization have been underemphasizedand offerunrealizedopportunities to increase our knowl-edgeof the evolution andselectionof mating sys-tems. A number of majorfunctionalfactors,in-cludingprepotency,pollendiscounting, heeffectsof selfingon the outcrossedseed production,andreproductive assurance, remain entirely or al-most entirelyunstudied(Lloyd 1992;SchoenandLloyd 1992). Moreover,self-pollinationis not asingleunvaryingprocessthat occursin the samemanner in all species that practice any selfing(Lloyd 1979). On the contrary, self-pollinationoccurs in several fundamentally differentways,which we describe as the modes of self-polli-nation. There have been few attempts to deter-minetheirrelativefrequencies xperimentally. naddition, a number of operational factors cancause the amount of self-fertilizationto varywithin a population, including environmentalconditions and differentpollinatorspecies; thesetoo have rarelybeen examined.

358

http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/13

LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 359In this and the following articles we examinethese major functional dimensions of self- andcross-fertilizationand attempt to integrate hemwith genetic approachesto the subject. This ar-ticle introduces functional aspects of self- andcross-fertilization y describing hevariousmodes

by which selfingoccursand the ecological,mor-phological, and physiological factorsthat influ-ence their frequencies.It also reviews publishedwork on spontaneousself-pollination o examinefunctionalfactors that contributeto variation inthe frequencyof self-fertilization.The two fol-lowing articlespresenta phenotypicmodel of theselection of self-fertilizationLloyd1992) and de-scribe and illustrateexperimentalprocedures orestimating the selective forces and partitioningthe full complement of self-fertilization nto itscomponent modes (Schoenand Lloyd 1992).Modes f self-pollination

CLEISTOGAMYInmorphologicalerms,the most distinctmodeof selfing s cleistogamy(Kuhn1867;Lord1981),which occursin closed flowers that are structur-ally specializedfor self-fertilizationand do notoutcross.Cleistogamydiffers romallothermodesof selfingin severalrespects.It is the only modethat occurs in morphologicallydistinct flowers.Hence, it can be recognized mmediatelyand its

frequencycan be measuredsimply by countingthe numbers of cleistogamouslyand chasmoga-mously producedseeds. Morever,cleistogamyisuniqueamongmodesof selfing SchoenandLloyd1984; Lively and Lloyd 1990) in possessing anadvantage over outcrossing (the cost of out-crossing ) hatis derivedfrom a cost ofproducingmales, as modeled by MaynardSmith (1971),rather than a cost of meiosis, as described byWilliams(1971). Cleistogamous lowerscost lessto producebecausethe cost of pollen and attrac-tants is very low. Cleistogamyis also distinct inthat in many species the cleistogamously andchasmogamouslyderivedseeds differ n theirsize,dispersal, germination, and survival character-istics (Campbellet al. 1983;Schmitt andGamble1990).Altogether, he operationand selectionofcleistogamy differ widely from those of othermodesof selfing SchoenandLloyd 1984).In thisseries of articleswe confine our attentionto thechasmogamousmodes of selfing that occur inflowers that can also engagein self-pollination.

GEITONOGAMYGeitonogamy is the most distinct of the chas-mogamous modes of selfingbecause it involvestransfer of pollen between flowers and requiresthe same pollination mechanism as cross-polli-nation. Consequently, t has the ecologicalprop-

erties of cross-fertilizationand the genetic prop-erties of self-fertilization.The distinctive natureof geitonogamyhasled to its being theonly modeof chasmogamousselfing that was distinguishedtraditionally Kemer 1895).A certainamount ofgeitonogamy is virtually inevitable in self-com-patibleplants that producea numberof flowersat anthesis at the same time. Geitonogamy isprobablythe most widespreadmode of self-pol-lination, but it may never achieve the predomi-nancethatthe autonomousmodes acquire n ha-bitually selfing species. It would be even moreimportantbut for the fortunateand still largelyunexplained habit sharedby virtually all flowervisitors of visitingonly a fractionof the availableflowerson a plantbeforemovingto thenextplant(Frankieet al. 1976;Kadmonand Shmida 1992;Robertson 1992).

Charles Darwin (1859, 1876) recognized theimportanceof geitonogamywhen he arguedthattrees are likely to be self-pollinated more fre-quently than other plants because they displaymoreflowersat one time. Darwinpostulated hatthis could explainthe higherfrequencyof specieswith separate exesamongthe trees of the UnitedKingdom, New Zealand, and the United States(but not Australia)than among other plants ofthe same regions.A century ater,Arroyo(1976)proposed hatthe occurrence f geitonogamymaybe an important factor in the selection of self-incompatibility as well as separatesexes.A few recent studies have examined the rela-tionship between the frequencyof geitonogamyand flowernumber that Darwin implied (Craw-ford 1984; Geber 1985; Handel 1985; Hessing1988; Robertson 1992). The amount of geito-nogamyis influencednot only by the flowerdis-play but also by factorsthat affect the extent ofpollencarryover Robertson1992).Otheraspectsof geitonogamy,such as its distributionamongthe flowerson a plant, variation with pollinatorabundanceand type, changesthroughouta flow-eringseason, and the relativedegreeto which itdisplacescross-fertilizationand the autogamousmodes of selfing, remain unexplored.All thesefactors nfluencethe measurementor selection ofgeitonogamy.

FACILITATED SELF-POLLINATIONIn the course of foragingfor rewards,flowervisitorsmay causesome autogamyas well as gei-tonogamy(Knuth 1906-1909; Estes and Brown1973; Hinton 1976; Schneider and Buchanan1980; Pazy 1984). Like geitonogamyand com-

petingselfing (see below), such facilitatedselfing(so namedby SchneiderandBuchanan1980)oc-curs at the same time as outcrossing.Facilitatedselfingis primarilya by-productof adaptationsfor outcrossing, again resembling geitonogamy(Lloyd 1992). In flowersthat presentpollen and



4/13

360 INTERNATIONAL JOURNAL OF PLANT SCIENCESstigmasduringthe same visits, facilitatedselfingis almost impossible to eliminatecompletely un-less there s a mechanismthatensures hat stigmacontacts in a flowerstrictly precedepollen con-tacts,as in Cypripediumpecies(van derPijl andDodson 1966) or species with sensitive stigmas(Newcombe 1922).The amounts of facilitatedselfing that animalvisitors cause in the course of their foragingac-tivities are likely to vary enormously,dependingon the way visitors move, the time they spendon each flower, and the positions of the anthersand stigmas. There are no data available, but,following Heine (1937), we expect that less spe-cialized visitors that forage on promiscuouslypollinatedflowersare likely to cause more facil-itated selfingthan specializedvisitors with moreprecisemovements.

THE THREE MODES OF AUTONOMOUSSELF-POLLINATION

Prior,competing, and delayed self-pollinationare similar n beingautonomous modes of selfingthat occurwithout theparticipationof an externalagent. The three modes differ in their timing(Lloyd 1979). They occurbefore,during,and af-ter opportunitiesfor outcrossing n a flower,re-spectively. As a result of their differenttiming,they also differin the degreeto which they dis-place cross-fertilization(Ockendonand Currah1978) and in the conditions requiredfor theirselection (Lloyd 1979, 1992).Numerous but brief anecdotal accounts con-cerningwhen and howautonomousselfingoccursin various species are scattered through the lit-erature.Earlierobservations were collected bythe Germanencyclopedists Muller1883;Kerner1895;Knuth 1906-1909). A notablemodem ex-ample is the work on the self-pollination of or-chids by Catling(1990), who made careful ob-servations of floral behaviorthat showedexactlywhen and how, thoughnot precisely how much,autonomous self-pollination occurs in variousspecies. We know of no experimentalattemptsto determinethe relativeimportanceof the threemodes of autonomousselfingin any species.Priorselfingoccurs when anthersdehisce andstigmasarereceptivebeforeanthesis and the twopollinatingsurfaces are positioned and orientedso there is contact between them in unopenedbuds. Some self-fertilizing species regularlyen-gagein bud pollinations (Hagerup 1952). Manyspecies may undergoan increasein priorselfingwhenfloweropening s postponed n poor weath-er and herkogamy s less fully developed (againthereare no firmdata).Competingselfingresemblesfacilitatedselfingin that it occursduring he sameintervalas cross-pollination, but it differsin being achieved au-tonomouslyand,hence, it is moreeasily selected

(Lloyd 1992). The originaldefinitionof compet-ing selfing (Lloyd 1979) did not distinguish itfromfacilitatedselfingandwasthereforebroaderthan the present definition. The two modes canbe separated xperimentallyn animal-pollinatedspecies(SchoenandLloyd1992).Competingself-ing is probablyrelativelyunimportant in abiot-icallypollinatedplants, which areoften unisexualor completelydichogamous.Competingselfingoccursby a variety of mech-anisms. In some species, such as many self-com-patibleBrassicaceae n which the paired antherssurroundthe stigma, competing selfing resultssimply from the close proximity of pollen andstigmasduringanthesis.In otherspecies, the ex-act temporal relationshipsbetween selfing andcrossing events, and thus the degree to whichcompetingselfingpreemptsoutcrossing,are morecomplex. They depend in parton the patternofdichogamy.Inincompletelyprotogynous pecies,the stigmashave an opportunityto receive out-crossingpollenfirst.In incompletelyprotandrousspecies, competingselfingthat occurs when stig-masbecomereceptive akesplaceduring hesameinterval as outcrossing and is therefore ikely tohave a greatereffect on the amount of cross-fer-tilization. An intermediate situation occurs inflowers hatopen and close daily for severaldaysif they undergocompetingselfing only when thepetals aremoving or when they are closed (Mee-han 1876).Delayedselfingoccurs whenthe movements offlowerparts at the end of anthesis leadto pollen-stigmacontactsand thefertilizationof ovulesthathave not been previouslycross-fertilized. n spe-ciesthat areherkogamousduring heperiodwhencross-pollinationoccurs,flowermovements dur-ingsenescencemaycauseself-pollination.In cer-tainCampanulaceaendAsteraceae, .g.,the stylearmscurlaroundandtouch thestylewherepollenhas been presentedsecondarily(Faegriand vander Pijl 1979). In some species with epipetalousstamens,the fall of senescentcorollasmay causea portionof any remainingpollen to be brushedagainstthe stigmas (Hagerup1957; Dole 1990,but compareDudash and Ritland 1991).

Factorshatinfluencehefrequencyof self-pollinationThe amount of self-fertilization n a plant isaffectedby a number of factors thatprovidefur-ther functionaldimensions of self- vs. cross-fer-tilization.

CONSTRAINTSON THEMODESOF SELF-POLLINATIONThe morphological and phenologicalfeaturesof flowers impose distinct constraints on eachmode of self-pollination.Species that have anydegree of dichogamy, either protandry or protog-



5/13

LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 361

yny, cannot engagein prior selfing.Conversely,delayed selfing cannot take place in species inwhich the pollen is no longerviable, or stigmasare no longerreceptive, when the opportunitiesfor cross-pollination n a flower are over. If thestructureof flowersallows pollen to be picked upduring a pollinator visit only after the stigmashave beencontacted, acilitatedselfingcannot oc-cur.Inmanyherkogamous pecies, thepollenandstigmas aretoo distantly separatedduringanthe-sis for competingselfing to be possible. The de-gree of anther-stigmaseparationaffects the fre-quency of self-pollination n some species (Rickand Dempsey 1969; Schoen 1982; Barrett andShore 1987;Holtsfordand Ellstrand1992). More-over,theamount of selfingmay dependon wheth-er the stigmas or anthersare higher(approachorreverseherkogamy) SobrevilaandArroyo 1982;Kohn and Barrett 1992).Geitonogamy s the most constrainedmode ofchasmogamousself-pollination.A considerableamount of geitonogamyis often unavoidable asa consequence of the movement of pollinatorsbetween flowersof the same plant. The amountof geitonogamy may be varied by altering thenumber or disposition of flowers or their indi-vidualattractiveness,whichalters the number ofsuccessive visits that a pollinator makes to theflowersof a plant, or by changing loralstructureso that the amountof pollen carryover s altered(Robertson 1992). All these changes, however,alternot onlytheamount ofgeitonogamybut alsothat of cross-pollination.The operation of these structuraland behav-ioral constraintsmeans that a particularspeciesof plants may be able to employ, orprevent, onlya limitedfractionof themodes of self-pollination.

THE RELATIVE COMPETITIVE ABILITIESOF SELF- AND CROSS-POLLEN

The competitive abilities of self- and cross-pollen influence the amount of self-fertilization,particularlywhen selfing occurs over the sameperiod as crossing (the geitonogamous, compet-ing, and facilitatedmodes). Darwin(1876) dem-onstrated that a number of species are highlyself-fertile when isolated and yet produce pre-dominantly outcrossed progeny when they aresurroundedby differentvarieties of the samespe-cies or different ndividuals of the same variety.The observations were made by growing plantsin close proximity and identifying outcrossedplants by their characters(when from differentvarieties)or vigor (whenfrom individuals of thesame variety).Darwin postulatedthat the pre-potency of outcrossingpollen was the most im-portantfactor in limiting the naturalfrequencyof self-fertilization.In modem times there have been a number of

observationsof the prepotencyof cross-pollen ncompetition experiments with nominally self-compatible species-those in which the successof separate self- and cross-pollinations is ap-proximately he same (Bateman 1956; Ockendonand Currah1978; Wellerand Omduff1989;Cru-zan and Barrett1992).Thephenomenon s knownas cryptic self-incompatibility.A weak self-in-compatibility reaction canalso lead to a reducedprobabilityof fruit set from self-pollinatedflow-ersthat compete with outcrossed lowers Becerraand Lloyd 1992). Not all self-compatible plantshave competitively inferior self-pollen, however(Snow and Spira 1991).Self-incompatibility s often incomplete, lead-ing to varying degreesof seed set after artificialself-pollination (pseudocompatibility).In par-tially self-incompatible species (those in whichthe seed set is lower in selfed flowers than inseparatelycrossed flowers), self-pollen performspoorlyeven in the absence of competition fromcross-pollen.We thereforeexpectthe prepotencyof outcrossedpollen to be more pronouncedincompetition experimentswith these speciesthanin cryptically self-incompatible species.A num-ber of studies of species with both gametophyticandsporophytic elf-incompatibility ystemshaveconfirmed that compatible pollen is prepotentover partially ncompatible pollen (Eenink 1982;Visser and Marcucci 1984; Bertin 1990).In pol-len competition experimentson tristylousspeciesof Pontederiaceae,Barrett and colleagues havefoundthat legitimate pollen is prepotentover il-legitimate pollen in both self-incompatible andself-compatiblepopulations (Barrettand Ander-son 1985;Cruzanand Barrett1992).We recommend that Darwin's term, prepoten-cy, be revived to cover all the abovephenomenathat cause cross-pollen to succeed in fertilizingovules more often than by chance when it com-petes with self-pollen.This definitionapplies topartially self-incompatible plants as well as tothose that exhibitcrypticself-incompatibility, utit excludespostzygotic expressionsof inbreedingdepression. When prepotency occurs, the pro-portionsof self- and cross-fertilizationneed notmatch those of self- and cross-pollination,evenin nominally self-compatiblespecies. Prepotencyis probablyanimportantdeterminantof the mat-ing systemin many speciesthat have incompleteself-incompatibilitybarriers,although t maynothave the ubiquity that Darwin postulated (Jones1928).Despite Darwin's lead, there has been noattempt to determine the degree to which pre-potency limits naturalfrequenciesof self-fertil-izationin self-compatibleorpartiallyself-incom-patible species. This would require experimentson the timingof self-pollinationas well as otherson the prepotencyof outcrossedpollen deposited



6/13

362 INTERNATIONAL JOURNAL OF PLANT SCIENCES

at various times relative to thedepositionof self-pollen (Schoenand Lloyd 1992).THE ENVIRONMENTAL CONDITIONS

OF POLLINATIONThe environmental conditions of pollination

may cause variationin the amount of any of themodes of selfing. It has been shown in a consid-erable number of species that the frequencyofself-fertilizationvaries with seasonal or weatherconditions or even on different parts of a plant(Glendinning1962; Rust and Clement 1977;An-tonovics and Levin 1980; Stephenson 1982).These studies have identified ecological factorsthat influence the amounts of self-fertilization,such as population densityor size (StephensandFinkner 1953; Bateman 1956; Ganders 1975; Va-quero et al. 1989), but they have not attemptedto explore the floralevents that alter the depo-sition of self- and cross-pollen.Relevant factorsinclude the temperature, ight,andhumidity,thetime of the floweringseason when a flower isproduced, and the age of a flower. In general,unfavorablepollination conditions are likely toincrease the amount of autogamy, both becauseself-pollenthen competes less with cross-pollenand because the degree of temporal and spatialseparationof thepollenandstigmasmaydecreasein flowersthatdevelop underpoorconditions orremainunopenedforlong periods. By comparingthe frequencyof self-fertilizationamongthe fruitof individual plants, Schoen and Brown (1991)have provided evidence that selfing may be in-duced in environmental conditions associatedwith poor pollination.In some species the environmental variationamong flowers in opportunities for cross-polli-nation may be so extremethat some flowershaveno opportunities or outcrossingas seed or pollenparentsand can produce only self-fertilized off-spring.Muller(1883, p. 18) stated that in someaquatic plants flowers remain closed and polli-nate themselves if the water level is unusuallyhigh. Such flowersmay merge nto cleistogamousflowers, dependingon when the arrestin devel-opmentoccurs.Some intermediate onditionsarereviewed n Sculthorpe 1967).In some terrestrialspecies, such as Australasiansun orchids (Thel-ymitra species [Jones 1988]) and Gentiana li-neata (Webb 1984), there may also be a dichot-omous switch between conditions that allowflowers o openand thosethatdo not. Previouslywe described his extreme of conditional self-fer-tilization as inducedselfing Schoen and Lloyd1984).It is likely, however, that there are all de-greesof environmentalvariation n the frequencyof self-fertilization,and now we preferto distin-guish environmentaleffects as a general dimen-sion of self- vs. cross-fertilizationrather than aspecialcircumstance.

THEBEHAVIORF DIFFERENTECTORSThe amount of selfingmay also dependon theanimal species that visit a particularflower. Inpromiscuouslypollinatedspecies with readilyac-cessiblerewards,such as many ApiaceaeandAs-teraceae, hediversevisitorspotentiallymaycausewidely differingfrequencies of self- and cross-pollination. There have been, however, almostno studies of variation in the frequencyof self-pollinationcausedbydifferentvisitors(Andersonand Symon 1988).An extreme example of variationin self-pol-linationcausedbydifferent isitorsinvolves long-staying squatters, ncludingaphids,thrips, spi-der mites, and nitulid beetles. The squattersarepredators hat eat pollen or suck plantjuices anduse the flowers as a protectedhaven. In movingaround flowers, heymay cause varyingamountsof self-pollination,dependingin part on the po-sitionsand orientationsof thepollenandstigmas.Bakerand Cruden 1991) demonstratedhat thripsand/or aphidscausea significantamountof self-ing in the course of wandering over flowers ofRanunculus scleratus and Potentilla rivalis.Squat-tersflyrarelyor not at all,and thus cause virtuallyno cross-pollination. Experimental proceduresthat exclude mobile short-stay pollinators,suchas the use of pollination bags or cages, do notremove squatters.In such experiments,any self-pollinationcausedby squatterswill be classed as

autonomous, togetherwith thetrulyautonomousactivitiesof theplantsthemselvesandthe actionsof physical factors such as wind and rain. Thequasi-autonomouseffectsof squatterscan be ex-aminedby intensive pesticidetreatments(Bakerand Cruden 1991).VARIATION AMONG FLOWERS ON ONE PLANTIn multiovulate ovaries,any proportionof theovules in a single flower can be self-fertilized.Variationamongflowers n the proportionof self-fertilized ovules may arise because of variation

in the timing of self-pollinationor from the be-havior of pollinators.It is useful to distinguishbetweenwhole-flower and part-flower elf-polli-nation. A failureto recognizeinterflowervaria-tion can lead to a bias in estimates of the fre-quency of self-fertilization(Schoen and Brown1991).Asurvey f the literature n the extentof autonomouself-fertilization

Whether the deposition of self-pollen occursautonomously (prior, competing, and delayedselfing)or is mediated by a pollen vector(geito-nogamy and facilitated selfing)affects the con-ditions under which self-fertilizationoccurs and,hence, how readily it is selected (Lloyd 1992).The traditionalexplanation of self-pollinationas



7/13

D o _ m > > o -i s _ = =m;~~~~~~~~~~~~~~~~~~~~~~1-9 =9sX g (< r o s) = = o S -~~~~ _ e~oo N r- .: o, cJ -- _ coooooFooo Oa - -E; o . R N ~~~~~O-N mOmO?m o o m oe u~~~~~~~~~~~~~~~~~~~~~~~~no4 , o- o- rl V- C) C eoooNoeooS~~~~~~~~' o- -o 00-4>< e(U: 0~~~~~~~~~'

b 2S g> 3g 2' a2 4 i

. . . . . . . . . . . . . .. _R ~ ~ :::::::::::::::::::


8/13

I-- 1- I--,I-,-,00 00 ~~~~ N (7 DN ON -S') b c '- -- CD oNr- ON r- ON zz-, q, (7N -, 5, 0~~~o -- (7 o 00 ONA~~ ~~~O ON ON 'O ON (00 _0 ;3N (O_N (ON_ _ -_N ;I~~~~~~~~~~~~~~~1-w _0 k3- 3k0 3 3 3 keu ; ? __ ? < t Ch ' .,_ . > uN ~ ~~~~~~~~~~~~~~~P3Z~~~~~~~~~~~~~~~~~0S C,8t8844UOo _ooslvoo>o-O4-4)

Cd 1% en ' N 0ONN C N C oo N No, r- M o N as

0 a,-O 90s O ,O r- o CD N t r- ?, un N ?, 00 r- -t r- c O r 0

Tl

S~~~~~~~~~~~~~~~~o oo) *ON_ _ N 1,D 00 _ 00_ t- _, en en C) t-o 0L t Q e ^ Q~~~| ^< Noi I 0(t Uc~~~~~~~~~~~~~L cn

| :s S t E 'i < 2 w 34 fi t U e , . g 4 * . 2: . . . . . . . . . . . . . . . . . 1 Qu : : * : : : : : : *^~~~~~~~~~~~~~~~~~~~~~~

: . . . . . . . : Y . . . . . . : .s E E E B 8 S g o ? _ o~-C



9/13


a means ofreproductiveassurance Darwin1859;Muller 1883)emphasizes ts autonomous nature.We still know little about the frequencyof theautonomous modes of self-pollination,however.Here we examine informationpertainingto theoccurrenceof autonomous selfing in floweringplants. Wecannot estimatethe actualfrequenciesof autonomous self-fertilization from the pub-lished literature, since we do not know of anymeasurements of the natural frequency of self-fertilization hathave separated he autonomousand vector-mediated modes. Nevertheless, it ispossible to gainindirect nformation romstudiesthat have measuredseed set or fruit set in plantsthat have been isolated fromvisitors-henceforthreferred o as their autofertility(Drayner 1959).An autofertility evel greaterthan zero is a nec-essaryprerequisite or autonomousself-fertiliza-tion, but it does not guarantee hat self-fertiliza-tion will occur under natural conditions. Theexperimentalproceduresthat are used to deter-mine autofertility preclude competition withcross-pollen,butin naturecross-pollenmay growthroughthe style more rapidly than self-pollenand reduce the amount of self-fertilization.Thenaturalfrequencyof self-fertilizationcan be ob-tained only from a knowledgeof the genotypesof open-pollinated eedparentsand theirprogenyat one or more marker loci (Clegg 1980). Theautofertilityof isolated plants reflectsthe poten-tial, rather than the actual, rate of autonomousself-fertilization.We surveyed reports of plant reproductivebi-ology publishedbetween 1975 and 1991 in sev-eralmajorbotanical,ecological,andevolutionaryjournals. The summarized results are from 66cosexual,predominantlyhermaphrodite pecies.We consideredonly those studies reportingdataon autofertility togetherwith data on seed setand/or fruit setfollowingartificial elf- and cross-pollinationsof separate lowers(tables 1, 2). Thedata from artificialpollinationsare necessarytodeterminewhethera given species is self-incom-patible and to verify that seed set occurs in theconditions underwhichplantsaregrown.We ex-cluded a few speciesthat set almost no seeds fol-lowing cross-pollination and a few others thatexhibitedsignificant evels of seed set after com-binedemasculationandbagging,an indicationofapomixis.Inaddition, speciesthat arecompletelyself-incompatibleor produced only occasionalseeds after selfingwere not considered becauseeven if autonomousself-pollinationoccurred nthem, there would be little or no autofertility.

Wefirstcalculated heSelf-compatibility ndexfor each species-the average seed or fruit setafterself-pollinationdivided by the seed or fruitset after cross-pollination (Becerraand Lloyd1992).The index correctsthe success of self-pol-linations forvariability n seedproductioncaused

by variationsin plant vigor, physiologicallimi-tations of seed production,or the techniquesorconditions of pollination.The Self-compatibilityIndices for the 66 spe-cies range from a little above zero to more thanone, with no conspicuousgaps. Thus there is nononarbitraryboundarybetweenpartially self-in-compatible and self-compatibleplants. Instead,self-incompatibilitymust be regardedas a quan-titative phenomenon(Becerraand Lloyd 1992).The values above one arepresumablythe resultof experimentalerror,since there are few biolog-ical groundsfor obtaininga greaterseed set afterself-pollination than after cross-pollinationbe-tween members of the same population. Somevalues below one alsorepresent andomvariationin populationsin which separateself- and cross-pollinations areequallysuccessful.

To compare major segments of the self-com-patibilitycontinuum,we divided therangeof Self-compatibilityIndices into two groups(tables 1,2). The 37 samples with indicesgreater han 0.75are describedas self-compatible.The boundarywas chosen at 0.75 because its reciprocal (1.33)is close to the highest index obtained(1.39); thisassumes thatin plantsin whichseparate elf-andcross-pollinationssucceed equallywell, randomvariationsin the actualcounts of small samplesareequallylikelyto show less frequentsuccess ofself- or cross-pollinations.The geometric mean(appropriate or dividends varying around 1.0)forthe Self-compatibility ndices of the self-com-patible plantsis 1.02. This indicatesthat, on av-erage, separateself-pollinationsdojust as well ascross-pollinations-although cross-pollen mightstill be prepotent n competitionwith self-pollen.There were 29 species with Self-compatibility n-dices between zero and 0.75, which we describeas self-incompatible.On average,selfs succeeded40% as frequentlyas crosses in this group.For each sample we also calculatedan Auto-fertilityIndex, the seed (or fruit) set of isolatedplants divided by that of artificialcross-pollina-tions. Again the denominator removes some ofthe effectsof variablepollinatingconditions. TheAutofertility Indices ranged continuously fromzero to above one (tables 1, 2), indicatingthat alldegreesof autonomousself-pollinationoccur.Low levels of autofertilitydo not necessarilyimply lack of self-pollinationunder naturalcon-ditions.Infact,estimationsof themating systemsof severalspeciesin the tableswith low Autofer-tility Indices reveal that they have significantamountsof natural self-fertilization Mitchell-Olds and Waller 1985; Johnston 1990; SchoenandLloyd 1992).In these species, it is likely thatone of the mediated modes of self-pollination soccurring,but biparental inbreedingcannot beruledout as an alternativeexplanation.To see how autofertility s related to self-com-



10/13


11/13


ConclusionWehavepostulated hat avariety ofecological,morphological, and physiological factors affecthowmuchself-fertilization akes place in a flowerand how it occurs. Some of these factors have

not beenexaminedseriously n even a single spe-cies. The effects on natural frequencies of self-fertilizationof, e.g., prepotency,structuralcon-straints, short-termenvironmental fluctuations,and the mode of selfing are virtuallyunknown.The survey of the literatureconfirms that theamounts of autonomous self-fertilizationvarywidelyamongspecies and areinfluencedby boththe degree of self-compatibility and featuresoffloralmorphologyand phenology.The functional dimensions of self- and cross-fertilizationdeserve to be consideredalongwiththe currentlymore popular geneticfactors, suchas the degreeof overdominanceat loci causing

inbreeding depression (which is poorly knownafterthree-quarters f a centuryof effort)and theamount of recombinationbetween the viabilityloci and the (equallypoorly known)mating sys-tem loci. We will not understandthe evolutionof self-fertilizationproperlyuntil we know moreabout its functional dimensions as well as thegenetic aspects.For this purpose,it is necessaryto analyzethe effects of these factorson the se-lection of self- andcross-pollinationand to mea-sure their operationin natural populations.Weexamine these aspectsin the followingtwo arti-cles in this series.

AcknowledgmentsWe are grateful to Spencer Barrett, LyndaDelph, and KentHolsingerfor theirhelpful com-ments on an earlier draft of this article.

LiteratureitedAnderson, G. J., and D. Symon. 1988. Insect foragersonSolanum flowersin Australia. Ann. Mo. Bot. Gard. 75:842-852.Andersson,S. 1989. The evolution of self-fertilityn Crepistectorum Asteraceae).Plant Syst. Evol. 168:227-236.Antonovics, J., and D. A. Levin. 1980. The ecologicalandgeneticalconsequencesof density-dependent egulation nplants. Ann. Rev. Ecol. Syst. 11:411-452.Arroyo,M. T. K. 1976. Geitonogamy n animal pollinated

tropicalangiosperms: stimulus for the evolution of self-incompatibility.Taxon 25:543-548.Baker, J. D., and R. W. Cruden. 1991. Thrips-mediatedself-pollinationof two facultativelyxenogamouswetlandspecies.Am. J. Bot. 78:959-963.Barrett,S. C. H., and J. M. Anderson. 1985. Variation nexpression n trimorphic ncompatibilityn Pontederia or-data L. (Pontederiaceae). heor.Appl. Genet. 70:355-362.Barrett,S. C. H., and D. Charlesworth. 1991. Effectsof achange n thelevel of inbreeding n thegenetic oad. Nature352:522-524.Barrett,S. C. H., and C. G. Eckert. 1990. Variation andevolution of mating systemsin seed plants.In S. Kawano,ed.Biologicalapproaches nd evolutionary rends nplants.Academic Press, London.Barrett,S. C. H., and K. Helenurm. 1987. The reproductivebiology of borealforestherbs. I. Breeding ystems and pol-lination. Can. J. Bot. 65:2036-2046.Barrett,S. C. H., and J. S. Shore. 1987. Variationand evo-lutionof breeding ystems nthe Turnera lmifolia omplex(Turneraceae). volution 41:340-354.Bateman,A. J. 1956. Cryptic self-incompatibility n thewallflower:Chieranthus hieri L. Heredity10:257-261.Becerra, ., and D. G. Lloyd. 1992. Competition-dependentabscission of self-pollinatedflowersof Phormiumtenax(Agavaceae): second action of self-incompatibility t thewhole-flower evel. Evolution 46:458-469.Bertin,R. I. 1990. Self-sterilityand crypticself-fertility nCampsisradicans Bignoniaceae).Bot. Gaz. 150:397-403.Bowers,K. A. W. 1975. The pollinationecologyof Solanumrostratum Solanaceae).Am. J. Bot. 62:633-638.Brown, A. H. D. 1990. Genetic characterization f plantmating systems. Pages 43-63 in A. H. D. Brown,M. T.Clegg,A. L. Kahler,and B. S. Weir,eds. Plantpopulationgenetics, breedingand geneticresources.Sinauer,Sunder-land, Mass.

Campbell,C. S., J. A. Quinn,G. P. Cheplick,and T. J. Bell.1983. Cleistogamy n grasses. Ann. Rev. Ecol. Syst. 14:411-441.Catling, P. M. 1990. A review of spontaneousself-polli-nation n theOrchidaceae. n J. Arditti,ed. Orchidbiology.Vol. 4. Comstock,Ithaca,N.Y.Charlesworth,D., and B. Charlesworth. 1987. Inbreedingdepressionand its evolutionaryconsequences.Ann. Rev.Ecol.Syst. 18:237-268.Clegg,M. T. 1980. Measuringplant mating systems. Bio-science30:814-818.Crawford,T. J. 1984. What s a population?Pages 135-173in B. Shorrocks, d. Evolutionary cology. Blackwell,Ox-ford.Crome,F. H. J., and A. K. Irvine. 1986. Two bob eachway : hepollinationandbreeding ystemof theAustralianmountainforest tree Syzygiumcormiflorum Myrtaceae).Biotropica18:115-125.Cruzan,M. B., and S. C. H. Barrett. 1992. Contribution fcryptic ncompatibility o the matingsystemof Eichhorniapaniculata Pontederiaceae).Evolution (in press).Darwin,,C. 1859. On the originof species.Murray,London.1868. The variation of animals and plants underdomestication.2 vols. Murray,London.1876. The effects of cross and self fertilisation nthe vegetablekingdom.Murray,London.Dickinson, T. A., and J. B. Phipps. 1986. Studies in Cra-taegus (Rosaceae:Maloideae).XIV. The breedingsystemof Crataegus rus-gallisensu lato in Ontario.Am. J. Bot.73:116-130.Dole, J. A. 1990. Role of corollaabscission n delayedself-pollinationof Mimulusguttatus Scrophulariaceae). m. J.Bot. 77:1505-1507.Drayner,J. M. 1959. Self- and cross-fertility n fieldbeans(Viciafaba L.). J. Agric. Sci. 53:387-403.Dudash, M. R., and K. Ritland. 1991. Multiplepaternityand self-fertilizationn relation to floralage in Mimulusguttatus Scrophulariaceae). m. J. Bot. 78:1746-1753.Eenink,A. H. 1982. Compatibilityand incompatibility nwitloof-chicory Cichorium ntybus L.). 3. Gametic com-petition after mixed pollination and double pollinations.Euphytica31:773-786.Estes,J.R., and L. S. Brown. 1973. Entomophilous ntraflo-ralpollination n Phyla incisa. Am. J. Bot. 60:228-230.



12/13

368 INTERNATIONAL JOURNAL OF PLANT SCIENCES

Faegri, K., and L. van der Pijl. 1979. The principles ofpollinationecology.3d ed. PergamonPress, Oxford.Firmage,D. H., and R. R. Cole. 1988. Reproductive uccessand inflorescencesize of Calopogontuberosus Orchida-ceae).Am. J. Bot. 75:1371-1377.Frankie,G. W., P. A. Opler,and K. S. Bawa. 1976. Foragingbehaviourof solitarybees:implicationsfor outcrossingofa neotropical oresttreespecies. J. Ecol. 64:1049-1057.Ganders,F. R. 1975. Mating patternsin self-compatibledistylouspopulationsof Amsinckia Boraginaceae).Can.J.Bot. 53:773-779.Geber, M. A. 1985. The relationshipof plant size to self-pollination n Mertensiaciliata.Ecology66:762-772.Glendinning,D. R. 1962. Naturalpollinationof cocoa.Na-ture 193:1305.Haber, W. A., and G. W. Frankie. 1982. Pollination ofLuehea Tiliaceae) n Costa Rica deciduous orest.Ecology63:1740-1750.Hagerup, . 1952. Budautogamyn some northern rchids.Phytomorphology :51-60.. 1957. Wind autogamy n Arbutus.Bull. Jard. Bot.Etat Brux.27:41-47.

Hamrick, . L., and M. J. W. Godt. 1990. Ailozymediversityin plant species. Pages 43-63 in A. H. D. Brown,M. T.Clegg,A. L. Kahlen,and B. S. Weir,eds. Plantpopulationgenetics, breedingand genetic resources.Sinauer,Sunder-land, Mass.Handel,S. N. 1985. The intrusionof clonal growthpatternsof plant breedingsystems.Am. Nat. 125:367-384.Harding,J., C. B. Mankinen,and M. H. Elliott. 1974. Ge-netics of Lupinus.VII.Outcrossing, utofertility, nd vari-ability in naturalpopulationsof the nanusgroup.Taxon23:729-738.Heine, E. M. 1937. Observations n the pollinationof NewZealand loweringplants. Trans.R. Soc. N.Z. 67:133-148.Henslow,G. 1879. On the self-fertilization f plants.Trans.Linn.Soc., 2d ser., Bot. 1:317-398.Hessing, M. B. 1988. Geitonogamouspollination and itsconsequences n Geraniumcaespitosum.Am. J. Bot. 75:1324-1333.Hinton, W. F. 1976. The evolution of insect-mediated elf-pollination from an outcrossing system in Calyptridium(Portulacaceae).Am. J. Bot. 63:979-986.Holtsford, T. P., and N. C. Ellstrand. 1992. Genetic andenvironmental ariation nfloral raitsaffecting utcrossingrate in Clarkia tembloriensis Onagraceae).Evolution 46:216-225.Janzen,D. H., P. Devries, T).E. Gladstone.M. L. Higgins,and M. Lewinsohn. 1980. Self- and cross-pollinationofEncyclia cordigera Orchidaceae)n Santa Rosa NationalPark,CostaRica. Biotropica12:72-74.Johnston, M. 0. 1990. Natural selection, inbreedingde-pressionand self-fertilizationn two species of Lobeliawithdifferentpollinators.Ph.D. diss. University of Chicago,Chicago.Jones,D. F. 1928. Selectivefertilization.Universityof Chi-cago Press, Chicago.Jones,D. L. 1988. Nativeorchidsof Australia.ReedBooks,FrenchsForest,New SouthWales,Australia.Kadmon, R., and A. Shmida. 1992. Departurerules usedby bees foraging or nectar:a field test. Evol. Ecol. 6:142-151.Karron,J. D. 1989. Breeding ystemsand levels of inbreed-ing depression n geographically estrictedandwidespreadspeciesof Astragalus Fabaceae).Am. J. Bot. 76:331-340.Kerner,A. 1895. Thenaturalhistoryofplants.Vol.2. Black-ie, London.Kimata,M. 1978. Comparative tudieson the reproductivesystemsof Mazus aponicusand M. miquelii Scrophular-iaceae).PlantSyst. Evol. 129:243-253.Knight,T. A. 1799. An account of some experimentsonthe fecundationof vegetables.Phil. Trans. R. Soc. Lond.1799:195-204.

Knuth, P. 1906-1909. Handbook of flower pollination. 3vols. Oxford University Press, Oxford.Kohn,J. R., andS. C. H. Barrett. 1992. Experimentaltudieson the functionalsignificanceof heterostyly.Evolution(inpress).Kuhn, M. 1867. Einige Bemerkungen uberVandelliaundden Bliltenpolymorphismus. ot. Z. 25:65-67.Lively, C., and D. G. Lloyd. 1990. The cost of biparentalsex under individual selection.Am. Nat. 135:489-500.Lloyd, D. G. 1979. Some reproductive actorsaffecting heselectionof self-fertilizationn plants. Am. Nat. 113:67-79. 1992. Self- and cross-fertilizationn plants. II. Theselectionof self-fertilization. nt.J.PlantSci. 153:370-380.Lord,E. M. 1981. Cleistogamy: tool for the studyof floralmorphogenesis, unctionandevolution. Bot. Rev. 47:421449.Lyons, E. E., and J. Antonovics. 1991. Breeding systemevolutionin Leavenworthia: reeding ystemvariationandreproductive uccess n naturalpopulationsof Leavenwor-thia crassa(Cruciferae).Am. J. Bot. 78:270-287.McMullen,C. K. 1987. Breedingsystems of selected Ga-lapagosIslands angiosperms.Am. J. Bot. 74:1694-1705.MaynardSmith, J. 1971. The origin and maintenanceofsex. Pp. 163-175 in G. C. Williams,ed. Groupselection.AldineAtherton,Chicago.Meehan,T. 1876. The sleepof plants as an agentof self-fertilization.Proc. Acad. Nat. Sci. Phila. 1876:84-85.Melampy, M. N., and A. M. Hayworth. 1980. Seed pro-ductionand pollenvectors n severalnectarlessplants.Evo-lution 34:1144-1154.Mitchell-Olds,T., and D. M. Waller. 1985. Relative per-formance of selfed and outcrossed progenyin Impatienscapensis.Evolution39:533-544.Motten,A. F. 1986. Pollinationecology of the springwild-flowercommunityof a temperatedeciduous forest. Ecol.Monogr. 56:21-42.

Muller,H. 1883. The fertilisationof flowers. Macmillan,London.Newcombe, F. C. 1922. Significanceof the behaviour ofsensitivestigmas. Am. J. Bot. 9:99-122.Ockendon,D. J., and L. Currah. 1978. Time of cross- andself-pollinationaffects he amounts of self-seed set by par-tially self-incompatibleplantsof Brassicaoleracea.Theor.Appl.Genet. 52:233-237.Opler,P. A., H. G. Baker, and G. W. Frankie. 1975. Re-productive biology of some Costa Rican Cordia species(Boraginaceae). iotropica7:234-247.Osbom, M. M., P. G. Kevan, and M. L. Lane. 1988. Pol-lination biology of Opuntia polyacantha and Opuntiaphaeacantha Cactaceae)n southem Colorado.PlantSyst.Evol. 159:85-94.Patt, J. M., M. W. Merchant,D. R. E.Williams,and B. J. D.Meeuse. 1989. Pollinatorbiologyof Planthera tricta Or-chidaceae) n OlympicNational Park,Washington.Am. J.Bot. 76:1097-1106.Pazy, B. 1984. Insect induced self-pollination.Plant Syst.Evol. 144:315-320.Ramirez, N., and Y. Brito. 1990. Reproductivebiology ofa tropicalplant swampcommunity n the VenezuelanLla-nos. Am. J. Bot. 77:1260-1271.Rick, C. M., and W. H. Dempsey. 1969. Position of thestigma n relationto fruitsettingof the tomato. Bot. Gaz.130:180-186.Ritland,K. 1990. Gene dentityand thegeneticdemographyof plant populations. Pages 181-199 in A. H. D. Brown,M. T. Clegg,A. L. Kahler,and B. S. Weir,eds. Plantpop-ulation genetics, breedingand geneticresources.Sinauer,Sunderland,Mass.Robertson,A. W. 1992. The relationshipsbetween floraldisplaysize, pollen carryover ndgeitonogamyn Myosotiscolensoi Kirk)Macbride Boraginaceae). iol.J. Linn.Soc.(in press).



13/13

LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 369Robertson,J. L., and R. Wyatt. 1990. Reproductivebiologyof the yellow-fringed rchid Plantheraciliaris.Am. J. Bot.77:388-398.Rust, R. W., and S. L. Clement. 1977. Entomophilouspol-lination of the self-compatible pecies CollinsiaparvifloraFisher and Meyer.J. Kans. Entomol. Soc. 50:37-48.Schemske,D. W. 1978. Evolution of reproductive harac-

teristicsin Impatiens(Balsaminaceae):he significanceofcleistogamyand chasmogamy.Ecology59:596-613.Schlessman,M. A. 1985. FloralbiologyofAmericanginseng(Panax quinquefolium).Bull. Torrey Bot. Club 112:129-133.Schlising,R. A. 1976. Reproductiveproficiencyn Paeoniacalifornica Paeoniaceae).Am. J. Bot. 63:1095-1103.Schmitt, J., and S. E. Gamble. 1990. The effectof distancefrom the parental site on offspringperformanceand in-breeding xpressionn Impatiens apensis: test of the localadaptationhypothesis.Evolution44:2022-2030.Schneider,E. L.,andJ. D. Buchanan. 1980. Morphologicalstudies of the Nymphaeaceae.XI. The floral biology ofNelumbopentapetala.Am. J. Bot. 67:182-193.Schneider,E. L., and D. M. Nichols. 1984. Floral biologyof Argemoneaurantiaca Papaveraceae).Bull.TorreyBot.Club 111:1-7.Schoen,D. J. 1982. The breedingsystem of Giliaachillei-folia: variation n floralcharacteristics ndoutcrossing ate.Evolution 36:352-360.Schoen, D. J., and A. H. D. Brown. 1991. Whole- andwithin-flowerself-pollination n Glycine argyreaand G.clandestinaand the evolution of autogamy.Evolution45:1651-1665.Schoen, D. J., and D. G. Lloyd. 1984. The selection ofcleistogamyand heteromorphicdiaspores.Biol. J. Linn.Soc. 23:303-322.1992. Self- andcross-fertilizationn plants.III. Es-timation of selectiveforces and modes of self-pollination.Int. J. Plant Sci. 153:381-393.

Sculthorpe,C. D. 1967. The biology of aquatic vascularplants.Arnold, London.Simpson,B. B., J. L. Neff, and G. Dieringer. 1986. Repro-ductivebiologyof Tinantia nomala Commelinaceae).ull.TorreyBot. Club 113:149-158.

Snow, A. A., and T. P. Spira. 1991. Differentialpollen-tubegrowthratesand non-random ertilizationn Hibiscusmos-cheutos Malvaceae).Am. J. Bot. 78:1419-1426.Sobrevila,C., and M. T. K. Arroyo. 1982. Breeding ystemsin a montane ropicalcloudforest n Venezuela.Plant Syst.Evol. 140:19-37.Stephens,S. G., and M. D. Finkner. 1953. Naturalcrossingin cotton. Econ. Bot. 7:257-269.Stephenson,A. G. 1982. When does outcrossingoccur n amass-flowering lant?Evolution36:762-767.Threadgill,P. F., J. M. Baskin,andC. C. Baskin. 1981. Thefloralecologyof FraseracaroliniensisGentianaceae).Bull.TorreyBot. Club 108:1-23.van der Pijl, L., and C. H. Dodson. 1966. Orchidflowers:theirpollinationand evolution.Universityof MiamiPress,Miami.Vaquero,F., F. J. Vences, P. Garcia, L. Ramirez, and M.Perezde la Vega. 1989. Matingsystemin rye:variabilityin relation to the populationand plant density. Heredity62:17-26.Visser,T., and C. Marcucci. 1984. The interactionbetweencompatibleand self-incompatiblepollen of apple and pearas influencedby the ratio in the pollen cloud. Euphytica33:699-704.Webb, C. J. 1984. Constraintson the evolution of plantbreeding ystemsand theirrelevance o systematics.Pages249-270 in W. F. Grant, ed. Plant biosystematics.Aca-demic Press,Toronto.Weller, S. G., and R. Ornduff. 1989. Incompatibility nAmsinckiagrandiflora Boraginaceae): istributionof cal-loseplugsand pollentubes following nter-andintramorphcrosses. Am. J. Bot. 76:277-282.Williams, G. C. 1971. Introduction.Pages 1-15 in G. C.Williams, ed. Groupselection. Aldine Atherton,Chicago.Young,H. J. 1986. Beetle pollinationof Dieffenbachiaon-gispatha(Araceae).Am. J. Bot. 73:931-944.Zapata,T. R., and M. T. K. Arroyo. 1978. Plantreproduc-tive ecology of a secondarydeciduous tropical forest inVenezuela.Biotropica10:221-230.

Documents

Self and Cross Fertilization in Plants. I