Self- and Cross-Fertilization in Plants. I. Functional Dimensions Author(S): David G

Home , Geitonogamy, Pollination bags

Self- and Cross-Fertilization in Plants. I. Functional Dimensions Author(s): David G. Lloyd and Daniel J. Schoen Source: International Journal of Plant Sciences, Vol. 153, No. 3, Part 1 (Sep., 1992), pp. 358- 369 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/2995676 . Accessed: 18/02/2014 15:23

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 Press is collaborating with JSTOR to digitize, preserve and extend access to International Journal of Plant Sciences.

http://www.jstor.org

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions Int. J. Plant Sci. 153(3):358-369. 1992. ? 1992 by The University of Chicago.All rightsreserved. 1058-5893/92/5303-0010$02.00

SELF-AND CROSS-FERTILIZATION IN PLANTS. 1. FUNCTIONALDIMENSIONS

DAVID G. LLOYD AND DANIEL J. SCHOEN Plant and MicrobialSciences, University of Canterbury,Private Bag, Christchurch,New Zealand; and Biology Department,McGill University, Montreal,Quebec H3A IBI, Canada

Many functional-ecological, morphological,and physiological-factors affect the occurrenceof self- fertilization.Six modes of self-pollinationare distinguished. These differin whetherthey utilize specialized flowers,whether they involve the transferof pollen within or between flowers,whether they are autonomous or mediated by vectors, and their timing relative to opportunitiesfor outcrossing.The various modes of selfing are subject to differentstructural constraints. Prepotency, the preferentialsuccess of cross-pollenin achievingfertilizations when it competes with self-pollen,influences the frequencyof self- fertilizationin some species. The amount of self-fertilizationmay depend on environmentalconditions and the vector species visiting each flower and may vary among the flowers of one plant. To gain informationon the prevalenceof autonomous self-pollination,66 species for which the degreesof self- compatibilityand autofertility(seed set in isolation) have been published were surveyed. Partiallyself- incompatiblespecies (in which the seed set is lower after self-pollinationthan after separateoutcrosses) have on average lower autofertilitythan self-compatiblespecies (in which self- and cross-pollinations succeed equally well), but some partially self-incompatiblespecies have considerableautofertility and some self-compatiblespecies have none. A number of featuresof floral morphologyand phenologyare associatedwith high AutofertilityIndices.

Introduction The functional approach to the study of self- The comparisonof self-and cross-fertilization and cross-fertilizationcontrasts with the genetic is the centraltopic of floralbiology. Following one in emphasizing the operation of pollination thediscovery by Knight(1799) and Darwin (1868, mechanisms and aspects of the naturalhistory of 1876)that cross-fertilization is advantageousbe- flowers.Functional studies of mating systems ex- causeit producessuperior progeny, there was a amine ecological, morphological,and physiolog- periodof intenseactivity in pollinationbiology ical perspectives. They have continued from the in the later decadesof the nineteenthcentury. last century to the present day with little change Flowerstructures that encouragecross-fertiliza- and currentlyconstitute a useful but ratherstatic tion and reduce self-fertilizationwere studied aspect of mating systems. In recent years, many widely.After considerable debate during this pe- topics of floral ecology have been rejuvenatedby riod, it was also acknowledgedthat some plants innovative studies of reproductive strategiesfor areadapted to regularself-fertilization and have deployingadaptive mechanisms,but the new par- floralsyndromes that contrast strongly with those adigm has had hardly any impact on the tradi- associatedwith outcrossing(see Darwin[1876] tional topic of cross- versus self-fertilization. andHenslow [1879] for divergent viewpoints and We believe that functional aspects of self-fer- Miuller's[1883] brilliant resolution of the issue). tilization have been underemphasizedand offer Floralbiology became popular again in the sec- unrealized opportunities to increase our knowl- ond halfof the twentiethcentury. In this period, edge of the evolution and selection of mating sys- geneticstudies have dominatedcomparisons of tems. A number of major functional factors, in- self- and cross-fertilization.The geneticstudies cludingprepotency, pollen discounting,the effects have provided much-neededempirical infor- of selfingon the outcrossed seed production,and mationon severalaspects of self- and cross-fer- reproductive assurance, remain entirely or al- tilization,including frequencies of self-fertiliza- most entirelyunstudied (Lloyd 1992; Schoen and tion(Barrett and Eckert 1990), the expression and Lloyd 1992). Moreover, self-pollination is not a causesof inbreedingdepression (Charlesworth and single unvaryingprocess that occurs in the same Charlesworth1987; Barrettand Charlesworth manner in all species that practice any selfing 1991),and the geneticstructures of outcrossing (Lloyd 1979). On the contrary, self-pollination and selfingpopulations (Brown 1990; Hamrick occurs in several fundamentally different ways, and Godt 1990;Ritland 1990). The theoretical which we describe as the "modes" of self-polli- effectsof the matingsystem on the geneticstruc- nation. There have been few attempts to deter- turesof regularlyselfing and outcrossingpopu- mine their relativefrequencies experimentally. In lations have also been analyzedwidely (Brown addition, a number of operational factors can 1990;Ritland 1990). cause the amount of self-fertilization to vary within a population, including environmental Manuscriptreceived March 1992; revisedmanuscript received conditions and differentpollinator species; these May 1992. too have rarelybeen examined.

358

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 359

In this and the following articles we examine erties of cross-fertilizationand the genetic prop- these major functional dimensions of self- and erties of self-fertilization.The distinctive nature cross-fertilizationand attempt to integratethem of geitonogamyhas led to its being the only mode with genetic approachesto the subject. This ar- of chasmogamous selfing that was distinguished ticle introduces functional aspects of self- and traditionally(Kemer 1895). A certain amount of cross-fertilizationby describingthe variousmodes geitonogamy is virtually inevitable in self-com- by which selfing occurs and the ecological, mor- patible plants that produce a number of flowers phological, and physiological factors that influ- at anthesis at the same time. Geitonogamy is ence their frequencies.It also reviews published probablythe most widespread mode of self-pol- work on spontaneousself-pollination to examine lination, but it may never achieve the predomi- functional factors that contribute to variation in nance that the autonomous modes acquirein ha- the frequency of self-fertilization. The two fol- bitually selfing species. It would be even more lowing articlespresent a phenotypic model of the important but for the fortunate and still largely selection of self-fertilization(Lloyd 1992) and de- unexplained habit shared by virtually all flower scribe and illustrateexperimental procedures for visitors of visiting only a fractionof the available estimating the selective forces and partitioning flowerson a plant before moving to the next plant the full complement of self-fertilizationinto its (Frankieet al. 1976; Kadmon and Shmida 1992; component modes (Schoen and Lloyd 1992). Robertson 1992). Charles Darwin (1859, 1876) recognized the Modesof self-pollination importanceof geitonogamywhen he arguedthat trees are likely to be self-pollinated more fre- CLEISTOGAMY quently than other plants because they display In morphologicalterms, the most distinctmode more flowersat one time. Darwin postulatedthat of selfingis cleistogamy(Kuhn 1867; Lord 1981), this could explain the higher frequencyof species which occurs in closed flowers that are structur- with separatesexes among the trees of the United ally specialized for self-fertilizationand do not Kingdom, New Zealand, and the United States outcross.Cleistogamy differs from all othermodes (but not Australia) than among other plants of of selfing in several respects. It is the only mode the same regions. A century later, Arroyo (1976) that occurs in morphologically distinct flowers. proposedthat the occurrenceof geitonogamymay Hence, it can be recognized immediately and its be an important factor in the selection of self- frequencycan be measured simply by counting incompatibility as well as separate sexes. the numbers of cleistogamously and chasmoga- A few recent studies have examined the rela- mously produced seeds. Morever, cleistogamy is tionship between the frequency of geitonogamy uniqueamong modes of selfing(Schoen and Lloyd and flower number that Darwin implied (Craw- 1984; Lively and Lloyd 1990) in possessing an ford 1984; Geber 1985; Handel 1985; Hessing advantage over outcrossing (the "cost of out- 1988; Robertson 1992). The amount of geito- crossing")that is derived from a cost of producing nogamy is influenced not only by the flower dis- males, as modeled by Maynard Smith (1971), play but also by factors that affect the extent of rather than a cost of meiosis, as described by pollen carryover(Robertson 1992). Otheraspects Williams (1971). Cleistogamousflowers cost less of geitonogamy, such as its distribution among to produce because the cost of pollen and attrac- the flowers on a plant, variation with pollinator tants is very low. Cleistogamy is also distinct in abundanceand type, changes throughouta flow- that in many species the cleistogamously and ering season, and the relative degree to which it chasmogamouslyderived seeds differin their size, displaces cross-fertilizationand the autogamous dispersal, germination, and survival character- modes of selfing, remain unexplored. All these istics (Campbellet al. 1983; Schmitt and Gamble factorsinfluence the measurementor selection of 1990). Altogether,the operation and selection of geitonogamy. cleistogamy differ widely from those of other modes of selfing(Schoen and Lloyd 1984). In this FACILITATED SELF-POLLINATION series of articles we confine our attention to the In the course of foraging for rewards, flower chasmogamous modes of selfing that occur in visitors may cause some autogamyas well as gei- flowers that can also engage in self-pollination. tonogamy (Knuth 1906-1909; Estes and Brown 1973; Hinton 1976; Schneider and Buchanan 1980; Pazy 1984). Like and com- GEITONOGAMY geitonogamy peting selfing (see below), such facilitated selfing Geitonogamy is the most distinct of the chas- (so named by Schneiderand Buchanan 1980) oc- mogamous modes of selfing because it involves curs at the same time as outcrossing.Facilitated transfer of pollen between flowers and requires selfing is primarily a by-product of adaptations the same pollination mechanism as cross-polli- for outcrossing, again resembling geitonogamy nation. Consequently,it has the ecological prop- (Lloyd 1992). In flowers that present pollen and

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions 360 INTERNATIONAL JOURNAL OF PLANT SCIENCES

stigmas during the same visits, facilitated selfing (Lloyd 1992). The original definition of compet- is almost impossible to eliminate completely un- ing selfing (Lloyd 1979) did not distinguish it less there is a mechanismthat ensuresthat stigma from facilitatedselfing and was thereforebroader contacts in a flower strictly precede pollen con- than the present definition. The two modes can tacts, as in Cypripediumspecies (van der Pijl and be separatedexperimentally in animal-pollinated Dodson 1966) or species with sensitive stigmas species (Schoenand Lloyd 1992). Competingself- (Newcombe 1922). ing is probably relatively unimportant in abiot- The amounts of facilitated selfing that animal ically pollinatedplants, which are often unisexual visitors cause in the course of their foragingac- or completely dichogamous. tivities are likely to vary enormously, depending Competingselfing occurs by a variety of mech- on the way visitors move, the time they spend anisms. In some species, such as many self-com- on each flower, and the positions of the anthers patible Brassicaceaein which the paired anthers and stigmas. There are no data available, but, surround the stigma, competing selfing results following Heine (1937), we expect that less spe- simply from the close proximity of pollen and cialized visitors that forage on promiscuously stigmas during anthesis. In other species, the ex- pollinated flowers are likely to cause more facil- act temporal relationships between selfing and itated selfing than specialized visitors with more crossing events, and thus the degree to which precise movements. competingselfing preempts outcrossing, are more complex. They depend in part on the pattern of THE THREE MODES OF AUTONOMOUS dichogamy.In incompletelyprotogynous species, SELF-POLLINATION the stigmas have an opportunity to receive out- Prior, competing, and delayed self-pollination crossingpollen first.In incompletelyprotandrous are similar in being autonomous modes of selfing species, competing selfing that occurs when stig- that occurwithout the participationof an external mas become receptivetakes place duringthe same agent. The three modes differ in their timing interval as outcrossing and is therefore likely to (Lloyd 1979). They occur before, during,and af- have a greatereffect on the amount of cross-fer- ter opportunitiesfor outcrossingin a flower, re- tilization. An intermediate situation occurs in spectively. As a result of their different timing, flowersthat open and close daily for several days they also differ in the degree to which they dis- if they undergo competing selfing only when the place cross-fertilization(Ockendon and Currah petals are moving or when they are closed (Mee- 1978) and in the conditions required for their han 1876). selection (Lloyd 1979, 1992). Delayed selfingoccurs when the movements of Numerous but brief anecdotal accounts con- flower parts at the end of anthesis lead to pollen- cerningwhen and how autonomous selfingoccurs stigmacontacts and the fertilizationof ovules that in various species are scattered through the lit- have not been previously cross-fertilized.In spe- erature. Earlier observations were collected by cies that are herkogamousduring the period when the Germanencyclopedists (Muller 1883; Kerner cross-pollinationoccurs, flower movements dur- 1895; Knuth 1906-1909). A notable modem ex- ing senescencemay cause self-pollination.In cer- ample is the work on the self-pollination of or- tain Campanulaceaeand Asteraceae,e.g., the style chids by Catling (1990), who made careful ob- armscurl aroundand touch the style wherepollen servations of floral behavior that showed exactly has been presented secondarily (Faegri and van when and how, though not precisely how much, der Pijl 1979). In some species with epipetalous autonomous self-pollination occurs in various stamens, the fall of senescent corollas may cause species. We know of no experimental attempts a portion of any remainingpollen to be brushed to determinethe relative importanceof the three against the stigmas (Hagerup 1957; Dole 1990, modes of autonomous selfing in any species. but compare Dudash and Ritland 1991). Prior selfing occurs when anthers dehisce and stigmas are receptive before anthesis and the two Factorsthat influencethe frequency pollinating surfaces are positioned and oriented of self-pollination so there is contact between them in unopened The amount of self-fertilization in a plant is buds. Some self-fertilizing species regularlyen- affectedby a number of factors that provide fur- gage in bud pollinations (Hagerup 1952). Many ther functional dimensions of self- vs. cross-fer- species may undergo an increase in prior selfing tilization. when floweropening is postponed in poor weather and herkogamyis less fully developed (again CONSTRAINTSON THE MODES there are no firm data). OF SELF-POLLINATION Competing selfing resembles facilitated selfing The morphological and phenological features in that it occursduring the same intervalas cross- of flowers impose distinct constraints on each pollination, but it differs in being achieved au- mode of self-pollination. Species that have any tonomously and, hence, it is more easily selected degree of dichogamy, either protandry or protog-

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 361 yny, cannot engage in prior selfing. Conversely, observationsof the prepotencyof cross-pollen in delayed selfing cannot take place in species in competition experiments with nominally self- which the pollen is no longer viable, or stigmas compatible species-those in which the success are no longer receptive, when the opportunities of separate self- and cross-pollinations is ap- for cross-pollination in a flower are over. If the proximatelythe same (Bateman 1956; Ockendon structureof flowersallows pollen to be picked up and Currah1978; Wellerand Omduff 1989; Cru- during a pollinator visit only after the stigmas zan and Barrett1992). The phenomenonis known have been contacted,facilitated selfing cannot oc- as cryptic self-incompatibility. A weak self-in- cur. In many herkogamousspecies, the pollen and compatibility reaction can also lead to a reduced stigmas are too distantly separatedduring anthe- probabilityof fruit set from self-pollinatedflow- sis for competing selfing to be possible. The de- ers that compete with outcrossedflowers (Becerra gree of anther-stigmaseparation affects the fre- and Lloyd 1992). Not all self-compatible plants quency of self-pollination in some species (Rick have competitively inferior self-pollen, however and Dempsey 1969; Schoen 1982; Barrett and (Snow and Spira 1991). Shore 1987; Holtsfordand Ellstrand1992). More- Self-incompatibilityis often incomplete, lead- over, the amount of selfingmay dependon wheth- ing to varying degrees of seed set after artificial er the stigmas or anthersare higher (approachor self-pollination (pseudocompatibility). In par- reverse herkogamy)(Sobrevila and Arroyo 1982; tially self-incompatible species (those in which Kohn and Barrett 1992). the seed set is lower in selfed flowers than in Geitonogamyis the most constrainedmode of separatelycrossed flowers), self-pollen performs chasmogamous self-pollination. A considerable poorly even in the absence of competition from amount of geitonogamy is often unavoidable as cross-pollen.We thereforeexpect the prepotency a consequence of the movement of pollinators of outcrossed pollen to be more pronounced in between flowers of the same plant. The amount competition experimentswith these species than of geitonogamy may be varied by altering the in cryptically self-incompatible species. A num- number or disposition of flowers or their indi- ber of studies of species with both gametophytic vidual attractiveness,which alters the number of and sporophyticself-incompatibility systems have successive visits that a pollinator makes to the confirmed that compatible pollen is prepotent flowersof a plant, or by changingfloral structure over partiallyincompatible pollen (Eenink 1982; so that the amount of pollen carryoveris altered Visser and Marcucci 1984; Bertin 1990). In pol- (Robertson 1992). All these changes, however, len competition experimentson tristylousspecies alternot only the amount of geitonogamybut also of Pontederiaceae, Barrett and colleagues have that of cross-pollination. found that legitimate pollen is prepotent over il- The operation of these structuraland behav- legitimate pollen in both self-incompatible and ioral constraints means that a particularspecies self-compatiblepopulations (Barrettand Ander- of plants may be able to employ, or prevent, only son 1985; Cruzanand Barrett 1992). a limited fractionof the modes of self-pollination. We recommend that Darwin's term, prepotency, be revived to cover all the above phenomena that cause cross-pollen to succeed in fertilizing THE RELATIVE COMPETITIVE ABILITIES ovules more often than by chance when it com- OF SELF- AND CROSS-POLLEN petes with self-pollen. This definition applies to The competitive abilities of self- and cross- partially self-incompatible plants as well as to pollen influence the amount of self-fertilization, those that exhibit crypticself-incompatibility, but particularlywhen selfing occurs over the same it excludes postzygotic expressions of inbreeding period as crossing (the geitonogamous, compet- depression. When prepotency occurs, the pro- ing, and facilitated modes). Darwin (1876) dem- portions of self- and cross-fertilizationneed not onstrated that a number of species are highly match those of self- and cross-pollination, even self-fertile when isolated and yet produce pre- in nominally self-compatiblespecies. Prepotency dominantly outcrossed progeny when they are is probablyan importantdeterminant of the mat- surroundedby differentvarieties of the same spe- ing system in many species that have incomplete cies or differentindividuals of the same variety. self-incompatibilitybarriers, although it may not The observations were made by growing plants have the ubiquity that Darwin postulated (Jones in close proximity and identifying outcrossed 1928). plants by their characters(when from different Despite Darwin's lead, there has been no varieties) or vigor (when from individuals of the attempt to determine the degree to which pre- same variety). Darwin postulated that the "prepotency limits natural frequencies of self-fertil- potency" of outcrossingpollen was the most im- ization in self-compatibleor partiallyself-incom- portant factor in limiting the natural frequency patible species. This would require experiments of self-fertilization. on the timing of self-pollinationas well as others In modem times there have been a number of on the prepotencyof outcrossedpollen deposited

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions 362 INTERNATIONAL JOURNAL OF PLANT SCIENCES at various times relative to the deposition of self- THE BEHAVIOROF DIFFERENTVECTORS pollen (Schoen and Lloyd 1992). The amount of selfing may also depend on the THE ENVIRONMENTAL CONDITIONS animal species that visit a particularflower. In OF POLLINATION promiscuouslypollinated species with readilyac- cessible rewards,such as many Apiaceae and As- The environmental conditions of pollination teraceae,the diversevisitors potentiallymay cause may cause variation in the amount of any of the widely differing frequencies of self- and cross- modes of selfing. It has been shown in a consid- pollination. There have been, however, almost erable number of species that the frequency of no studies of variation in the frequency of self- self-fertilizationvaries with seasonal or weather pollinationcaused by differentvisitors (Anderson conditions or even on different parts of a plant and Symon 1988). (Glendinning 1962; Rust and Clement 1977; An- An extreme example of variation in self-pol- tonovics and Levin 1980; Stephenson 1982). linationcaused by differentvisitors involves long- These studies have identified ecological factors staying"squatters," including aphids, thrips, spi- that influence the amounts of self-fertilization, der mites, and nitulid beetles. The squattersare such as population density or size (Stephensand predatorsthat eat pollen or suck plant juices and Finkner 1953; Bateman 1956; Ganders 1975; Va- use the flowers as a protected haven. In moving quero et al. 1989), but they have not attempted around flowers,they may cause varyingamounts to explore the floral events that alter the depo- of self-pollination, depending in part on the position of self- and cross-pollen. Relevant factors sitions and orientationsof the pollen and stigmas. include the temperature,light, and humidity, the Bakerand Cruden(1991) demonstratedthat thrips time of the flowering season when a flower is and/or aphids cause a significantamount of self- produced, and the age of a flower. In general, ing in the course of wandering over flowers of unfavorable pollination conditions are likely to Ranunculus scleratus and Potentilla rivalis. Squat- increase the amount of autogamy, both because ters fly rarelyor not at all, and thus cause virtually self-pollen then competes less with cross-pollen no cross-pollination. Experimental procedures and because the degree of temporal and spatial that exclude mobile short-stay pollinators, such separationof the pollen and stigmasmay decrease as the use of pollination bags or cages, do not in flowersthat develop under poor conditions or remove squatters.In such experiments, any self- remainunopened for long periods. By comparing pollination caused by squatterswill be classed as the frequencyof self-fertilizationamong the fruit autonomous, togetherwith the truly autonomous of individual plants, Schoen and Brown (1991) activities of the plants themselves and the actions have provided evidence that selfing may be in- of physical factors such as wind and rain. The duced in environmental conditions associated quasi-autonomouseffects of squatterscan be ex- with poor pollination. amined by intensive pesticide treatments (Baker In some species the environmental variation and Cruden 1991). among flowers in opportunities for cross-pollination may be so extremethat some flowershave VARIATION AMONG FLOWERS ON ONE PLANT no opportunitiesfor outcrossingas seed or pollen In multiovulate ovaries, any proportionof the parents and can produce only self-fertilized off- ovules in a single flower can be self-fertilized. spring. Muller (1883, p. 18) stated that in some Variationamong flowersin the proportionof self- aquatic plants flowers remain closed and polli- fertilized ovules may arise because of variation nate themselves if the water level is unusually in the timing of self-pollination or from the be- high. Such flowersmay mergeinto cleistogamous havior of pollinators. It is useful to distinguish flowers, depending on when the arrest in devel- between whole-flower and part-flowerself-polli- opment occurs.Some intermediateconditions are nation. A failure to recognize interflowervaria- reviewedin Sculthorpe(1967). In some terrestrial tion can lead to a bias in estimates of the fre- species, such as Australasiansun orchids (Thel- quency of self-fertilization (Schoen and Brown ymitra species [Jones 1988]) and Gentiana li- 1991). neata (Webb 1984), there may also be a dichot- omous switch between conditions that allow A surveyof the literatureon the extent flowersto open and those that do not. Previously we describedthis extreme of conditional self-fer- of autonomousself-fertilization tilization as "inducedselfing" (Schoen and Lloyd Whether the deposition of self-pollen occurs 1984). It is likely, however, that there are all de- autonomously (prior, competing, and delayed greesof environmentalvariation in the frequency selfing) or is mediated by a pollen vector (geito- of self-fertilization,and now we prefer to distin- nogamy and facilitated selfing) affects the con- guish environmental effects as a general dimen- ditions under which self-fertilizationoccurs and, sion of self- vs. cross-fertilizationrather than a hence, how readily it is selected (Lloyd 1992). special circumstance. The traditionalexplanation of self-pollinationas

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions D o _ m > > o -i s _ = =m =~~~~~~~~~~~~~~~~~~~~~~1s;>U 1 -9 =9s

X g (< r o s) = = o S <28 = = =Y~~~~~~~~~~~~~~cr Y c

0 N~~~~~~~~~~~~~~~~~~~~~~~0

z~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l0 01

g~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~(r 00 X- --

< Q dw o ?-oo O O Q 9~~~~~~00 08-0 d OCJod OOCZ - wOda X Q >-~~~~r _ ~~ooe oN r-o o: o,. cJ -- _ coooooFooo Oa - -E

; o . R N ~~~~~Ob -N mOmO?m

9~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Z

N > m k~~~~~~~~~~~~~~~~~~Cce 3 o o oe ct CZ C CZ E u~~~~~~~~~~~~~~~~~~~~~~~~nvo4C, o- o- rl V- C) C eoooNoeoo S~~~~~~~~' o- -o 00- -4> < e

(U : 0~~~~~~~~~'

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

3 ...... _ R ~ ~ :::::::::::::::::::

...... e...... ^ O ._ ;2E en::C:::::C:0C v::::: ~ r-C,::::::::~ CD C=: E. ?' t

c ON8n - ,E CD = tr enX ON-t CD 88X CDv) cn 00

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions I-- 1- I--,I-,-, N 00 00 ~~~~ (7 DN ON -S c ') b c '- - - CD oN r- ON r- 00 ON zz-, q, (7N -, 5, (7 ON ON ON 'O ON ( 0~~~o(O- -- (ON o A~~ ~~~O 00 _0 ;3N _N _ _ -_N ; I~~~~~~~~~~~~~~~1-0w _0 k3- 3k0 3 3 3 k eu ; ? __ ? < t Cho' t< 00 t'?r r r ,,- > .,_ . >

ut~~~~~~~0C0=C Er CP 0 R 0r dC

o~~~~~~~~tZS S Su> N ~ ~~~~~~~~~~~~~~~CP3 I Z~~~~~~~~~~~~~~~~~0ZS 8t8844U< C,

W~~~~~~~~~~~~~~~~~Woa m m . ) OCS g~~~~~~~~~~~~~~~~P*U) w*4 < >nvn>Oo< _ooslvoo>o-O< < <- < < < < <

4-4)

Cd 1%o en ' N 0ONN C N "C oo N 0 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

Wl T

S~~~~~~~~~~~~~~~~oo) *ON_ O _ _DN oo _0 _ 1, 00_0 t- _, en en C) t- o 0

L t Q e ^ Q~~~

| <3 -o< mo ooo o oN Oo oooo> ^< No i I 0(t .c~~~~~~~~~~~~~LU cn

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

: ...... 1 Q u : : * : : : : : e~~~~~~~~~~~~~~~~~~~~~~~~z: *^ : . . . . Z. . . : Y ...... : . s ES E E B 8 S g c~-Co ? _ o

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 365 a means of reproductiveassurance (Darwin 1859; by variations in plant vigor, physiological limi- Muller 1883) emphasizes its autonomous nature. tations of seed production, or the techniques or We still know little about the frequencyof the conditions of pollination. autonomous modes of self-pollination, however. The Self-compatibilityIndices for the 66 spe- Here we examine information pertainingto the cies range from a little above zero to more than occurrence of autonomous selfing in flowering one, with no conspicuous gaps. Thus there is no plants. We cannot estimate the actual frequencies nonarbitraryboundary between partially self-in- of autonomous self-fertilization from the pub- compatible and self-compatible plants. Instead, lished literature, since we do not know of any self-incompatibilitymust be regardedas a quan- measurements of the natural frequency of self- titative phenomenon (Becerraand Lloyd 1992). fertilizationthat have separatedthe autonomous The values above one are presumablythe result and vector-mediated modes. Nevertheless, it is of experimentalerror, since there are few biolog- possible to gain indirectinformation from studies ical grounds for obtaining a greaterseed set after that have measuredseed set or fruit set in plants self-pollination than after cross-pollination be- that have been isolated from visitors-henceforth tween members of the same population. Some referredto as their autofertility(Drayner 1959). values below one also representrandom variation An autofertilitylevel greaterthan zero is a nec- in populations in which separateself- and cross- essary prerequisitefor autonomous self-fertiliza- pollinations are equally successful. tion, but it does not guaranteethat self-fertiliza- To compare major segments of the self-com- tion will occur under natural conditions. The patibilitycontinuum, we divided the rangeof Self- experimentalprocedures that are used to deter- compatibility Indices into two groups (tables 1, mine autofertility preclude competition with 2). The 37 samples with indices greaterthan 0.75 cross-pollen,but in naturecross-pollen may grow are described as self-compatible. The boundary through the style more rapidly than self-pollen was chosen at 0.75 because its reciprocal (1.33) and reduce the amount of self-fertilization.The is close to the highest index obtained (1.39); this natural frequencyof self-fertilizationcan be ob- assumes that in plants in which separateself- and tained only from a knowledge of the genotypes cross-pollinations succeed equally well, random of open-pollinatedseed parentsand their progeny variations in the actual counts of small samples at one or more marker loci (Clegg 1980). The are equally likely to show less frequentsuccess of autofertilityof isolated plants reflects the poten- self- or cross-pollinations. The geometric mean tial, rather than the actual, rate of autonomous (appropriatefor dividends varying around 1.0) self-fertilization. for the Self-compatibilityIndices of the self-com- We surveyed reports of plant reproductivebi- patible plants is 1.02. This indicates that, on av- ology published between 1975 and 1991 in sev- erage, separateself-pollinations do just as well as eralmajor botanical, ecological, and evolutionary cross-pollinations-although cross-pollen might journals. The summarized results are from 66 still be prepotentin competition with self-pollen. cosexual, predominantlyhermaphrodite species. There were 29 species with Self-compatibilityIn- We considered only those studies reportingdata dices between zero and 0.75, which we describe on autofertility together with data on seed set as self-incompatible.On average,selfs succeeded and/or fruit set following artificialself- and cross- 40% as frequentlyas crosses in this group. pollinations of separateflowers (tables 1, 2). The For each sample we also calculated an Auto- data from artificialpollinations are necessary to fertility Index, the seed (or fruit) set of isolated determinewhether a given species is self-incom- plants divided by that of artificialcross-pollina- patible and to verify that seed set occurs in the tions. Again the denominator removes some of conditions underwhich plants are grown. We ex- the effectsof variable pollinatingconditions. The cluded a few species that set almost no seeds fol- Autofertility Indices ranged continuously from lowing cross-pollination and a few others that zero to above one (tables 1, 2), indicatingthat all exhibited significantlevels of seed set after com- degrees of autonomous self-pollination occur. bined emasculationand bagging,an indication of Low levels of autofertility do not necessarily apomixis. In addition, species that are completely imply lack of self-pollinationunder naturalcon- self-incompatible or produced only occasional ditions. In fact, estimations of the mating systems seeds after selfing were not considered because of several species in the tables with low Autofer- even if autonomous self-pollination occurredin tility Indices reveal that they have significant them, there would be little or no autofertility. amounts of natural"self-fertilization" (Mitchell- We firstcalculated the Self-compatibilityIndex Olds and Waller 1985; Johnston 1990; Schoen for each species-the average seed or fruit set and Lloyd 1992). In these species, it is likely that after self-pollination divided by the seed or fruit one of the mediated modes of self-pollination is set after cross-pollination (Becerra and Lloyd occurring, but biparental inbreeding cannot be 1992). The index correctsthe success of self-pol- ruled out as an alternative explanation. linations for variabilityin seed productioncaused To see how autofertilityis related to self-com-

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions 366 INTERNATIONAL JOURNAL OF PLANT SCIENCES

16 -v 100 14 A 0~~~~~~~~~~~~~go0 H 12 a 80 I 0 n 7070 60 0 ?0 4 50 Of 6 (40 0 uJ E 3 0 0 0. 2 310 0 0 o ?0 0 0 0 I~ 0 K 0 10 20 30 40 50 60 70 80 90 100 16 %Fruit (seed) set following artificial self-pollination 114 B Fig.2 The relationshipbetween the frequenciesof fruit or 6 12 seed set after self-pollinationand in isolated flowersamong 10 the samples of self-compatibleand self-incompatiblespecies in tables 1 and 2. 0 6 5- 4 What is surprisingis that this association is not D Z 2 stronger, as is seen in the wide ranges of auto- 0 fertility in self-incompatibleand self-compatible 00. 2 0.4 0.6 0.8 1.0 1.2 1.4 1.6C species and in the scatterin figure2. In some self- AUTOFERTILITYINDEX incompatible species, the evolution of even a re- stricted Fig.1 The distributionsof AutofertilityIndices for (A) self- ability of self-pollen to succeed in fertil- compatiblespecies and (B) self-incompatiblespecies. ization is sufficient to allow some autonomous selfingdespite the partialincompatibility barrier. On the other hand, the range of autofertility in self-compatiblespecies shows that complete self- patibility, we comparedfirst the AutofertilityIn- compatibility, at least as evident in separateself- dices for the self-compatible and self-incompat- and cross-pollinations, does not guarantee that ible groups.Among the self-incompatiblesamples, much, or even any, autonomous self-pollination the AutofertilityIndices rangefrom zero (six spe- will be possible. The degree of self-compatibility cies) to 0.76 (fig. 1B); the geometricmean of non- apparentlyis not the only factor that influences zero values is 18%of the potential seed set. The the extent of autonomous pollination. AutofertilityIndices for the self-compatibleplants To identify morphological and phenological range even more widely, from zero (11 samples) factors that affect the ease of autonomous polli- to 1.52 (fig. 1A);five samples have values of one nation, we have examined a number of features or more, and the geometricmean of nonzero val- of the floralbiology of the species reviewed (tables ues is 47%. The results indicate a significant re- 1, 2). The species lacking autofertilityaltogether lationship between self-compatibility and auto- or having low Autofertility Indices tend to be fertility, as one would expect. The association concentrated in a few families, particularlythe between self-compatibility and autofertility is Balsaminaceae, Fabaceae, Lobeliaceae, Orchi- confirmedwhen the average frequenciesof seed daceae,and Scrophulariaceae.These familieshave set in isolated flowers and after self-pollination specialized pollination mechanisms and charac- are graphedagainst each other (fig. 2; r = +.32, teristically have bilaterally symmetrical flowers, n = 65, P = .01 when both frequencies are trans- fused petals, and, in most species, relatively few formed to angles to improve the approximation stamens per flower. In most species there is a to a bivariate normal distribution). The corre- strongherkogamy mechanism separatingthe pol- lation between the two indices cannot be tested len and stigmas. The three species with unisexual directly because both indices include the cross- flowers, Caperoniapalustris, Dieffenbachia lon- pollination success in their computations. Fur- gispatha, and Croton hirtus, also have low Au- thermore, several detailed studies have shown tofertilityIndices. Moreover, all species reported that autofertilityis correlatedwith the estimated as dichogamous have low autofertility,e.g., spe- selfingrate among populations of a species (Har- cies of Impatiens, Lobelia, and Isopyrum. ding et al. 1974; Schoen 1982; Lyons and An- At the other extreme, the species with more tonovics 1991). In these species, at least some of than 50%autofertility usually have radiallysym- the self-fertilization appears to result from the metrical flowers, separatepetals, and more than operation of one or more of the autonomous five stamensper flower(tables 1, 2). These species modes of selfing. are likely to have less precise placement of floral The association between the degrees of self- parts, and this may allow less herkogamy and compatibilityand autofertilityis not unexpected. greaterlevels of autonomous selfing.

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 367

Conclusion inbreeding depression (which is poorly known We have postulatedthat a variety of ecological, afterthree-quarters of a centuryof effort)and the morphological, and physiological factors affect amount of recombination between the viability how much self-fertilizationtakes place in a flower loci and the (equally poorly known) mating sys- will and how it occurs. Some of these factors have tem loci. We not understand the evolution not been examined seriouslyin even a single spe- of self-fertilizationproperly until we know more cies. The effects on natural frequencies of self- about its functional dimensions as well as the fertilization of, e.g., prepotency, structuralcon- genetic aspects. For this purpose, it is necessary straints, short-term environmental fluctuations, to analyze the effects of these factors on the selection of self- and cross-pollinationand to mea- and the mode of selfing are virtually unknown. The survey of the literature confirms that the sure their operation in natural populations. We in two arti- amounts of autonomous self-fertilization vary examine these aspects the following widely among species and are influencedby both cles in this series. the degree of self-compatibility and features of Acknowledgments floral morphology and phenology. The functional dimensions of self- and cross- We are grateful to Spencer Barrett, Lynda Delph, and Kent Holsingerfor their helpful com- fertilizationdeserve to be considered along with draft of this article. the currentlymore popular genetic factors, such ments on an earlier as the degree of overdominance at loci causing

Literaturecited Anderson, G. J., and D. Symon. 1988. Insect foragerson Campbell,C. S., J. A. Quinn, G. P. Cheplick,and T. J. Bell. Solanum flowers in Australia. Ann. Mo. Bot. Gard. 75: 1983. Cleistogamyin grasses. Ann. Rev. Ecol. Syst. 14: 842-852. 411-441. Andersson,S. 1989. The evolution of self-fertilityin Crepis Catling, P. M. 1990. A review of spontaneous self-polli- tectorum(Asteraceae). Plant Syst. Evol. 168:227-236. nation in the Orchidaceae.In J. Arditti,ed. Orchidbiology. Antonovics, J., and D. A. Levin. 1980. The ecologicaland Vol. 4. Comstock, Ithaca,N.Y. geneticalconsequences of density-dependentregulation in Charlesworth,D., and B. Charlesworth. 1987. Inbreeding plants. Ann. Rev. Ecol. Syst. 11:411-452. depressionand its evolutionaryconsequences. Ann. Rev. Arroyo,M. T. K. 1976. Geitonogamyin animal pollinated Ecol. Syst. 18:237-268. tropicalangiosperms: a stimulus for the evolution of self- Clegg, M. T. 1980. Measuringplant mating systems. Bio- incompatibility.Taxon 25:543-548. science 30:814-818. Baker, J. D., and R. W. Cruden. 1991. Thrips-mediated Crawford,T. J. 1984. Whatis a population?Pages 135-173 self-pollinationof two facultativelyxenogamous wetland in B. Shorrocks,ed. Evolutionaryecology. Blackwell,Ox- species. Am. J. Bot. 78:959-963. ford. Barrett,S. C. H., and J. M. Anderson. 1985. Variationin Crome, F. H. J., and A. K. Irvine. 1986. "Two bob each expressionin trimorphicincompatibility in Pontederiacor- way":the pollinationand breedingsystem of the Australian data L. (Pontederiaceae).Theor. Appl. Genet. 70:355-362. mountain forest tree Syzygium cormiflorum(Myrtaceae). Barrett,S. C. H., and D. Charlesworth. 1991. Effectsof a Biotropica 18:115-125. changein the level of inbreedingon the geneticload. Nature Cruzan,M. B., and S. C. H. Barrett. 1992. Contributionof 352:522-524. crypticincompatibility to the mating system of Eichhornia Barrett,S. C. H., and C. G. Eckert. 1990. Variation and paniculata(Pontederiaceae). Evolution (in press). evolution of mating systems in seed plants. In S. Kawano, Darwin,,C. 1859. On the originof species.Murray, London. ed. Biologicalapproaches and evolutionarytrends in plants. 1868. The variation of animals and plants under Academic Press, London. domestication.2 vols. Murray,London. Barrett,S. C. H., and K. Helenurm. 1987. The reproductive 1876. The effects of cross and self fertilisationin biology of borealforest herbs. I. Breedingsystems and pol- the vegetablekingdom. Murray,London. lination. Can. J. Bot. 65:2036-2046. Dickinson, T. A., and J. B. Phipps. 1986. Studies in Cra- Barrett,S. C. H., and J. S. Shore. 1987. Variationand evo- taegus (Rosaceae:Maloideae). XIV. The breedingsystem lutionof breedingsystems in the Turneraulmifolia complex of Crataeguscrus-galli sensu lato in Ontario. Am. J. Bot. (Turneraceae).Evolution 41:340-354. 73:116-130. Bateman, A. J. 1956. Cryptic self-incompatibilityin the wallflower:Chieranthus chieri L. Heredity 10:257-261. Dole, J. A. 1990. Role of corolla abscissionin delayed self- Becerra,J., and D. G. Lloyd. 1992. Competition-dependent pollinationof Mimulusguttatus (Scrophulariaceae). Am. J. abscission of self-pollinatedflowers of Phormium tenax Bot. 77:1505-1507. (Agavaceae):a second action of self-incompatibilityat the Drayner,J. M. 1959. Self- and cross-fertilityin field beans whole-flowerlevel. Evolution 46:458-469. (Viciafaba L.). J. Agric. Sci. 53:387-403. Bertin, R. I. 1990. Self-sterilityand cryptic self-fertilityin Dudash, M. R., and K. Ritland. 1991. Multiple paternity Campsisradicans (Bignoniaceae). Bot. Gaz. 150:397-403. and self-fertilizationin relation to floral age in Mimulus Bowers,K. A. W. 1975. The pollinationecology of Solanum guttatus(Scrophulariaceae). Am. J. Bot. 78:1746-1753. rostratum(Solanaceae). Am. J. Bot. 62:633-638. Eenink,A. H. 1982. Compatibilityand incompatibilityin Brown, A. H. D. 1990. Genetic characterizationof plant witloof-chicory(Cichorium intybus L.). 3. Gametic com- mating systems. Pages 43-63 in A. H. D. Brown, M. T. petition after mixed pollination and double pollinations. Clegg,A. L. Kahler,and B. S. Weir, eds. Plant population Euphytica31:773-786. genetics, breedingand genetic resources.Sinauer, Sunder- Estes,J. R., and L. S. Brown. 1973. Entomophilousintraflo- land, Mass. ral pollination in Phyla incisa. Am. J. Bot. 60:228-230.

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions 368 INTERNATIONAL JOURNAL OF PLANT SCIENCES

Faegri, K., and L. van der Pijl. 1979. The principles of Knuth, P. 1906-1909. Handbook of flower pollination. 3 pollinationecology. 3d ed. PergamonPress, Oxford. vols. Oxford University Press, Oxford. Firmage,D. H., and R. R. Cole. 1988. Reproductivesuccess Kohn, J. R., and S. C. H. Barrett. 1992. Experimentalstudies and inflorescencesize of Calopogon tuberosus(Orchida- on the functionalsignificance of heterostyly.Evolution (in ceae). Am. J. Bot. 75:1371-1377. press). Frankie,G. W., P. A. Opler,and K. S. Bawa. 1976. Foraging Kuhn, M. 1867. Einige Bemerkungeniuber Vandellia und behaviourof solitarybees: implicationsfor outcrossingof den Bliltenpolymorphismus.Bot. Z. 25:65-67. a neotropicalforest tree species. J. Ecol. 64:1049-1057. Lively, C., and D. G. Lloyd. 1990. The cost of biparental Ganders, F. R. 1975. Mating patterns in self-compatible sex under individual selection. Am. Nat. 135:489-500. distylouspopulations of Amsinckia(Boraginaceae). Can. J. Lloyd, D. G. 1979. Some reproductivefactors affecting the Bot. 53:773-779. selection of self-fertilizationin plants. Am. Nat. 113:67- Geber, M. A. 1985. The relationshipof plant size to self- 79. pollination in Mertensiaciliata. Ecology66:762-772. 1992. Self- and cross-fertilizationin plants. II. The Glendinning,D. R. 1962. Naturalpollination of cocoa. Na- selectionof self-fertilization.Int. J. PlantSci. 153:370-380. ture 193:1305. Lord,E. M. 1981. Cleistogamy:a tool for the study of floral Haber, W. A., and G. W. Frankie. 1982. Pollination of morphogenesis,function and evolution. Bot. Rev. 47:421- Luehea(Tiliaceae) in Costa Rica deciduousforest. Ecology 449. 63:1740-1750. Lyons, E. E., and J. Antonovics. 1991. Breeding system Hagerup,0. 1952. Bud autogamyin some northernorchids. evolution in Leavenworthia:breeding system variationand Phytomorphology2:51-60. reproductivesuccess in naturalpopulations of Leavenwor- . 1957. Wind autogamyin Arbutus.Bull. Jard. Bot. thia crassa (Cruciferae).Am. J. Bot. 78:270-287. Etat Brux. 27:41-47. McMullen, C. K. 1987. Breedingsystems of selected Ga- Hamrick,J. L., and M. J. W. Godt. 1990. Ailozymediversity lapagosIslands angiosperms.Am. J. Bot. 74:1694-1705. in plant species. Pages 43-63 in A. H. D. Brown, M. T. MaynardSmith, J. 1971. The origin and maintenanceof Clegg,A. L. Kahlen,and B. S. Weir, eds. Plant population sex. Pp. 163-175 in G. C. Williams, ed. Group selection. genetics, breedingand genetic resources.Sinauer, Sunder- Aldine Atherton,Chicago. land, Mass. Meehan,T. 1876. The "sleep of plants"as an agent of self- Handel,S. N. 1985. The intrusionof clonal growthpatterns fertilization.Proc. Acad. Nat. Sci. Phila. 1876:84-85. of plant breedingsystems. Am. Nat. 125:367-384. Melampy, M. N., and A. M. Hayworth. 1980. Seed pro- Harding,J., C. B. Mankinen,and M. H. Elliott. 1974. Ge- ductionand pollen vectorsin severalnectarless plants. Evo- netics of Lupinus.VII. Outcrossing,autofertility, and vari- lution 34:1144-1154. ability in naturalpopulations of the nanus group. Taxon Mitchell-Olds,T., and D. M. Waller. 1985. Relative per- 23:729-738. formance of selfed and outcrossed progeny in Impatiens Heine, E. M. 1937. Observationson the pollinationof New capensis.Evolution 39:533-544. Zealandflowering plants. Trans.R. Soc. N.Z. 67:133-148. Motten, A. F. 1986. Pollinationecology of the springwild- Henslow,G. 1879. On the self-fertilizationof plants.Trans. flower community of a temperatedeciduous forest. Ecol. Linn. Soc., 2d ser., Bot. 1:317-398. Monogr. 56:21-42. Hessing, M. B. 1988. Geitonogamouspollination and its Muller, H. 1883. The fertilisationof flowers. Macmillan, consequencesin Geraniumcaespitosum. Am. J. Bot. 75: London. 1324-1333. Newcombe, F. C. 1922. Significanceof the behaviour of Hinton, W. F. 1976. The evolution of insect-mediatedself- sensitive stigmas. Am. J. Bot. 9:99-122. pollination from an outcrossing system in Calyptridium Ockendon,D. J., and L. Currah. 1978. Time of cross- and (Portulacaceae).Am. J. Bot. 63:979-986. self-pollinationaffects the amounts of self-seed set by par- Holtsford, T. P., and N. C. Ellstrand. 1992. Genetic and tially self-incompatibleplants of Brassicaoleracea. Theor. environmentalvariation in floraltraits affecting outcrossing Appl. Genet. 52:233-237. rate in Clarkia tembloriensis(Onagraceae). Evolution 46: Opler, P. A., H. G. Baker, and G. W. Frankie. 1975. Re- 216-225. productive biology of some Costa Rican Cordia species Janzen, D. H., P. Devries, T). E. Gladstone. M. L. Higgins, (Boraginaceae).Biotropica 7:234-247. and M. Lewinsohn. 1980. Self- and cross-pollinationof Osbom, M. M., P. G. Kevan, and M. L. Lane. 1988. Pol- Encyclia cordigera(Orchidaceae) in Santa Rosa National lination biology of Opuntia polyacantha and Opuntia Park, Costa Rica. Biotropica 12:72-74. phaeacantha(Cactaceae) in southem Colorado.Plant Syst. Johnston, M. 0. 1990. Natural selection, inbreedingde- Evol. 159:85-94. pressionand self-fertilizationin two species of Lobeliawith Patt, J. M., M. W. Merchant,D. R. E. Williams,and B. J. D. different pollinators. Ph.D. diss. University of Chicago, Meeuse. 1989. Pollinatorbiology of Plantherastricta (Or- Chicago. chidaceae)in OlympicNational Park,Washington. Am. J. Jones, D. F. 1928. Selectivefertilization. University of Chi- Bot. 76:1097-1106. cago Press, Chicago. Pazy, B. 1984. Insect induced self-pollination.Plant Syst. Jones, D. L. 1988. Native orchidsof Australia.Reed Books, Evol. 144:315-320. FrenchsForest, New South Wales, Australia. Ramirez, N., and Y. Brito. 1990. Reproductivebiology of Kadmon, R., and A. Shmida. 1992. Departurerules used a tropicalplant swamp communityin the VenezuelanLla- by bees foragingfor nectar:a field test. Evol. Ecol. 6:142- nos. Am. J. Bot. 77:1260-1271. 151. Rick, C. M., and W. H. Dempsey. 1969. Position of the Karron,J. D. 1989. Breedingsystems and levels of inbreed- stigma in relation to fruit setting of the tomato. Bot. Gaz. ing depressionin geographicallyrestricted and widespread 130:180-186. species of Astragalus(Fabaceae). Am. J. Bot. 76:331-340. Ritland,K. 1990. Gene identityand the geneticdemography Kerner,A. 1895. The naturalhistory of plants.Vol.2. Black- of plant populations. Pages 181-199 in A. H. D. Brown, ie, London. M. T. Clegg,A. L. Kahler,and B. S. Weir, eds. Plant pop- Kimata,M. 1978. Comparativestudies on the reproductive ulation genetics, breedingand genetic resources.Sinauer, systems of Mazusjaponicus and M. miquelii(Scrophular- Sunderland,Mass. iaceae). Plant Syst. Evol. 129:243-253. Robertson, A. W. 1992. The relationshipsbetween floral Knight, T. A. 1799. An account of some experimentson displaysize, pollen carryoverand geitonogamyin Myosotis the fecundationof vegetables.Phil. Trans. R. Soc. Lond. colensoi(Kirk) Macbride (Boraginaceae). Biol. J. Linn. Soc. 1799:195-204. (in press).

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions LLOYD & SCHOEN-FUNCTIONAL DIMENSIONS OF SELF- AND CROSS-FERTILIZATION 369

Robertson,J. L., and R. Wyatt. 1990. Reproductivebiology Snow, A. A., and T. P. Spira. 1991. Differentialpollen-tube of the yellow-fringedorchid Plantheraciliaris. Am. J. Bot. growthrates and non-randomfertilization in Hibiscusmos- 77:388-398. cheutos(Malvaceae). Am. J. Bot. 78:1419-1426. Rust, R. W., and S. L. Clement. 1977. Entomophilouspol- Sobrevila,C., and M. T. K. Arroyo. 1982. Breedingsystems lination of the self-compatiblespecies Collinsiaparviflora in a montanetropical cloud forestin Venezuela.Plant Syst. Fisher and Meyer. J. Kans. Entomol. Soc. 50:37-48. Evol. 140:19-37. Schemske,D. W. 1978. Evolution of reproductivecharac- Stephens,S. G., and M. D. Finkner. 1953. Naturalcrossing teristics in Impatiens (Balsaminaceae):the significanceof in cotton. Econ. Bot. 7:257-269. cleistogamyand chasmogamy.Ecology 59:596-613. Stephenson,A. G. 1982. When does outcrossingoccur in a Schlessman,M. A. 1985. Floralbiology of Americanginseng mass-floweringplant? Evolution 36:762-767. (Panax quinquefolium).Bull. Torrey Bot. Club 112:129- Threadgill,P. F., J. M. Baskin,and C. C. Baskin. 1981. The 133. floralecology of Fraseracaroliniensis (Gentianaceae). Bull. Schlising,R. A. 1976. Reproductiveproficiency in Paeonia TorreyBot. Club 108:1-23. californica(Paeoniaceae). Am. J. Bot. 63:1095-1103. van der Pijl, L., and C. H. Dodson. 1966. Orchid flowers: Schmitt, J., and S. E. Gamble. 1990. The effect of distance theirpollination and evolution. Universityof Miami Press, from the parental site on offspringperformance and in- Miami. breedingexpression in Impatienscapensis: a test of the local Vaquero, F., F. J. Vences, P. Garcia, L. Ramirez, and M. adaptationhypothesis. Evolution 44:2022-2030. Perezde la Vega. 1989. Matingsystem in rye:variability Schneider,E. L., and J. D. Buchanan. 1980. Morphological in relation to the population and plant density. Heredity studies of the Nymphaeaceae. XI. The floral biology of 62:17-26. Nelumbopentapetala. Am. J. Bot. 67:182-193. Visser,T., and C. Marcucci. 1984. The interactionbetween Schneider,E. L., and D. M. Nichols. 1984. Floral biology compatibleand self-incompatiblepollen of apple and pear of Argemoneaurantiaca (Papaveraceae). Bull. TorreyBot. as influencedby the ratio in the pollen cloud. Euphytica Club 111:1-7. 33:699-704. Schoen, D. J. 1982. The breedingsystem of Gilia achillei- Webb, C. J. 1984. Constraintson the evolution of plant folia: variationin floralcharacteristics and outcrossingrate. breedingsystems and their relevanceto systematics.Pages Evolution 36:352-360. 249-270 in W. F. Grant, ed. Plant biosystematics.Aca- Schoen, D. J., and A. H. D. Brown. 1991. Whole- and demic Press, Toronto. within-flowerself-pollination in Glycine argyrea and G. Weller, S. G., and R. Ornduff. 1989. Incompatibilityin clandestinaand the evolution of autogamy.Evolution 45: Amsinckiagrandiflora (Boraginaceae): distribution of cal- 1651-1665. lose plugsand pollen tubes followinginter- and intramorph Schoen, D. J., and D. G. Lloyd. 1984. The selection of crosses. Am. J. Bot. 76:277-282. cleistogamy and heteromorphicdiaspores. Biol. J. Linn. Williams, G. C. 1971. Introduction.Pages 1-15 in G. C. Soc. 23:303-322. Williams, ed. Group selection. Aldine Atherton,Chicago. 1992. Self- and cross-fertilizationin plants. III. Es- Young, H. J. 1986. Beetle pollinationof Dieffenbachialon- timation of selective forces and modes of self-pollination. gispatha(Araceae). Am. J. Bot. 73:931-944. Int. J. Plant Sci. 153:381-393. Zapata,T. R., and M. T. K. Arroyo. 1978. Plant reproduc- Sculthorpe,C. D. 1967. The biology of aquatic vascular tive ecology of a secondary deciduous tropical forest in plants. Arnold, London. Venezuela.Biotropica 10:221-230. Simpson, B. B., J. L. Neff, and G. Dieringer. 1986. Repro- ductivebiology of Tinantiaanomala (Commelinaceae). Bull. TorreyBot. Club 113:149-158.

This content downloaded from 131.212.205.195 on Tue, 18 Feb 2014 15:23:21 PM All use subject to JSTOR Terms and Conditions