Chapter01 reprodution allocation of plants

The completion of a plant’s life cycle andthe regeneration and establishment of plantpopulations depend on the process ofreproduction: the production of physiologi-cally independent individuals. Yet there isa great deal of variation among higher plantspecies in the quantity and timing of repro-duction; individuals may concentrate alltheir reproductive output in a singleepisode, after lifespans ranging from asshort as 3 weeks to as long as severaldecades, or they may reproduce repeatedly,at regular or intermittent intervals. Thesevariations in life history are influenced byboth ecological and evolutionary factors,and a great deal of research on plant biol-ogy in the last 30 years has focused onphysiological and demographic aspects ofreproductive strategies (Harper, 1967;Silvertown and Lovett Doust, 1993).

The theory of allocation, borrowedfrom microeconomics, first introduced tobiology by MacArthur (cited in Cody, 1966)and extended to the study of plants byHarper (Harper, 1967; Harper and Ogden,1970), has provided the principal concep-tual framework for linking individual phys-iology and life-history theory. The study ofallocation in biology assumes that organ-isms have a limited supply of some criticalresource (e.g. energy, time, biomass ornutrients), which they must generallydivide between several competing func-tions, broadly defined as growth, main-tenance and reproduction. These functionsare further assumed to be mutually exclu-sive, such that allocation to one functionnecessarily leads to a decrease in the simul-taneous allocation to other functions, andas a result an optimal pattern of allocationwill exist that maximizes some output para-meter. In evolutionary ecology, this outputis a measure of fitness, an individual’s con-tribution to future generations, whichdepends on total reproductive output andthe timing and frequency of reproduction(Willson, 1983). In the study of plants, thisapproach has been widely utilized, becausethe various functions of the plant can beapproximately assigned to discrete struc-tures: carbon gain is primarily the functionof leaves, nutrient uptake of roots andreproduction of inflorescences, seeds andancillary structures (but see Bazzaz, 1993).This rough equivalence opens up numerousavenues of empirical research. Unlike theanimal ecologist’s study of the allocation ofan organism’s time, the plant ecologist’sassessment of allocation to various struc-tures can be made by harvesting plants atsingle points in the life cycle, and studiescan be conducted in both the field and thelaboratory. Proportional allocation to differ-ent structures is thus used as a measure of

1

© CAB International2000. Seeds: The Ecology of Regeneration in Plant Communities, 2nd edition(ed. M. Fenner)

2F.A. Bazzaz et al.investment in corresponding functions, fol-lowing the scheme shown in Fig. 1.1a.

The two critical assumptions of alloca-tional theory are that the resource in ques-tion is in fixed supply and that allocationamong competing functions is mutuallyexclusive, thus generating trade-offsbetween functions. In its application to thestudy of reproductive strategies, these twoassumptions are not always observed, fortwo principal reasons. First, severalprocesses, such as the photosynthesis ofreproductive parts, can lead to an increasein total resource supply associated withreproduction. And, secondly, plant struc-tures can contribute to more than one func-tion: stems can simultaneously providesupport for both leaves and fruits. If theplant reproduces vegetatively, roots, stemsand leaves may all contribute to reproduc-tion as well as to their vegetative functions.As a result, measures of allocation to struc-tures do not always reflect the plant’sinvestment in function; Fig. 1.1b shows amore complete picture of the relationshipbetween the two. The concept of reproduc-tive effort (RE), which is often equatedwith reproductive allocation (RA), shouldrefer specifically to the individual’s netinvestment of resources in reproduction,which is diverted from vegetative activity(Tuomi et al., 1983; Bazzaz and Reekie,1985). In this chapter, we discuss severalfactors that decouple RA and RE, and arguethat these two must be conceptually distin-guished in future research. We address thisdifference within ecological and evolution-ary contexts, in relation to plant size, seedproduction and life-history evolution.

As mentioned above, the process ofreproduction can be defined as the produc-tion of physiologically independent indi-viduals. In plants, this can beaccomplished either by the production ofseed, or through vegetative production ofgenetically identical offspring. The relativeimportance of these two mechanismsvaries widely in different plant communi-ties (see Grubb, 1977). Establishment fromseed dominates in early successional com-munities, following major disturbances,and also in most forests, in which regenera-

tion occurs intermittently throughout thevegetational mosaic. In various low-staturecommunities, such as savannahs, grass-lands and arctic/alpine systems, regenera-tion is primarily by clonal growth. Thestudy of allocation is considerably moredifficult when clonal reproduction is con-sidered. Clonal reproduction occurs by theproduction of vegetative tissues that con-tribute to their own growth, so it is verydifficult to assess the allocation of the par-ent plant to the growth of ramets (Ogden,1974). Numerous studies demonstrate thatthe daughter ramets are dependent on theparent for a varying length of time (seereview by Pitelka and Ashmun, 1985);because of this physiological dependence,it can be difficult to distinguish betweengrowth and reproduction. In this chapter,we shall primarily consider allocation toseed production, but with the reminderthat much of what is said may apply toclonal growth as well.

Assessment of resource allocation to

plant structures

Before embarking on a more detailed dis-cussion of the relationship between struc-ture and function and the decoupling ofRA and RE, it is necessary to clarify themeans of assessing allocation to structuresthemselves. In most studies, RA is assessedas the proportion of standing biomass orother resources found in reproductivestructures at the time of harvest. This mea-sure may be an inaccurate reflection of theallocation of total production, for severalreasons. First, plants may be harvestedbefore the end of the season, before finalreproductive output has been expressed.This problem is particularly acute in plantswith indeterminate reproductive develop-ment, which continues until frost ordrought prevents further growth. Whensuch plants are grown experimentally, theresults may be strongly influenced by thetime at which the experiment was termin-ated (e.g. Geber, 1990; Clauss and Aarssen,1994). Secondly, for perennial, iteroparousplants there is a carry-over of standing bio-

Reproductive Allocation in Plants3Fig. 1.1.Relationship between structure and function in plants. The functions represent biological processesthat most plants must accomplish to complete their life cycle. The relative importance of different functionswill vary with life history and environment. (a) In this highly simplified view, each structure corresponds to asingle function; this is the basis for the application of resource allocation theory to the study of plantfunction. (b) A more complex view considers the contribution of structures to various functions. Roots,

stems and leaves can all contribute to reproduction through various forms of clonal growth; roots contributeto support, stems and reproductive structures can be photosynthetic, and seed characteristics contribute todispersal as one component of reproduction. The relative importance of each link in this diagram will varyin different situations, leading to varying degrees of decoupling between structure and function.

4F.A. Bazzaz et al.mass from year to year in woody tissue,such that standing RA underestimates RAbased on annual production (Willson,1983). Finally, any loss of a resource dur-ing growth will lead to discrepanciesbetween the standing pattern of allocationand the total acquisition and allocation.Loss from vegetative organs will lead tooverestimates of RA, while loss from repro-ductive organs will result in underesti-mates of RA. This loss may occur by manyprocesses, and the amount of loss relativeto the standing pool will differ for differentresources. Mineral resources probably havethe lowest loss rate relative to the amountpresent in the plant under most circum-stances; losses of minerals, as well as totalbiomass and associated energy content,will be due primarily to damage, herbivory,abscission of senesced parts and dispersalof reproductive structures. These losses arefairly easily assessed in experimental situa-tions by repeated collection of abscisedparts (e.g. Harper and Ogden, 1970).

In addition to these structural losses,carbon and energy are also dissipated byrespiratory activity associated with tissueconstruction and maintenance. Respiratorycosts will only alter patterns of RA if vege-tative and reproductive tissue havemarkedly different respiratory activity. In astudy of Agropyron repens, Reekie andBazzaz (1987a) measured whole-plant res-piration associated with reproduction andfound that rates of respiration for vegeta-tive and reproductive structures wereapparently fairly similar, such that mea-sures of RA that included respiration werevery similar to those based solely on bio-mass. Laporte and Delph (1996), however,report that in the dioecious plant, Silenelatifolia, male plants have a respiratory ratethat is twice as high as that of femaleplants. In this case, ignoring the respiratorycosts associated with reproduction mayseriously underestimate resource allocationto male function (see also Goldman andWillson, 1986; Marshall et al.,1993). Of allthe resources, water has the highest lossrate, through transpiration. Measuring theplant’s allocation of water requires elabo-rate measurements of transpiration, and

has rarely been considered on a whole-plant basis. In general, for those resourceswith a large flux through the plant relativeto the amount present at one point in time,measures of standing allocation will corre-late poorly with total allocation of theresource pool.

Three different measures of resourceallocation to reproductive structures musttherefore be carefully distinguished: (i)standing RA is the proportion of a resourcecontained in reproductive structures; (ii)short-term RA is a measure of proportionalallocation of a resource over some shortinterval, relative to plant lifetime, such as agrowing season for a tree, or a period ofdays or weeks for an annual; and (iii) life-time RA is the proportion of total availableresource invested in reproductive struc-tures over the entire lifespan of an individ-ual. This last measure is of most generalinterest when comparing individuals withdifferent life-history strategies or from dif-ferent environments; many studies reportlifetime RA for annuals and other semel-parous taxa, but virtually no data exist foriteroparous perennials. Owing to the fac-tors discussed above, standing RA willreflect short-term and lifetime allocationonly under a limited set of conditions. Forcomparative purposes, measures of RAshould be based on the net production ofthe plant, including structural losses. Theinclusion of respiration is more difficultand it has been considered in so few stud-ies that no general conclusion is at presentpossible regarding its contribution to RA.

The currency of allocation

In addition to these factors, it is importantto bear in mind that the patterns ofresource allocation will differ for differentresources. In one population of Verbascumthapsus, for example, RA was 40% basedon biomass, but varied from less than 5%to almost 60% when calculated for variousmineral elements (Abrahamson andCaswell, 1982; see also Fenner, 1985a, b;Reekie and Bazzaz, 1987b; Benner andBazzaz, 1988). Which resource is the

Reproductive Allocation in Plants5appropriate currency of allocation? Themost difficult aspect of this problem arisesfrom the assumption that there is a crucialresource that limits both vegetative andreproductive functions, such that its alloca-tion will determine individual perfor-mance. It has long been recognized thatdifferent resources limit growth in differentenvironments, or even at different timesduring the plant’s lifetime. In addition, dif-ferent functions of a single plant might belimited by different resources. This prob-lem has been cast in a new light in recentyears by the observation that plant growthis not always limited by a single resource,but tends to be limited by several resourcessimultaneously (Chapin et al.,1987). Theplasticity of plant development allowsindividuals to alter the investment in theuptake of different resources – by alteringroot:shoot ratios, for example. It is oftenobserved that allocation to roots increasesin response to low nutrient and wateravailability, and allocation to leavesincreases in low light. Theoretically,growth will be maximized when invest-ment of a gram of biomass, or otherresource, in the acquisition of any singleresource from the environment leads to anequivalent increase in growth (Bloom etal., 1985). The acquisition of differentresources is thus scaled to their relativeavailability or to their need, such that theconcentrations of resources in the plant areessentially uncorrelated with their avail-ability in the environment (Abrahamsonand Caswell, 1982).

If growth and reproduction are notlimited by a single, crucial resource, then,which resource is the appropriate currencyof allocation? We suggest that the answermay be that any of the various limitingresources is appropriate. It is true thatabsolute allocation to reproduction maydiffer for various currencies. For example,the proportion of whole-plant nitrogen thatis allocated to reproductive structures isoften higher than the proportion of bio-mass, i.e. carbon (Williams and Bell, 1981;Abrahamson and Caswell, 1982; Reekieand Bazzaz, 1987b; Garbutt et al., 1990;Benech Arnold et al., 1992). However, pro-viding these resources are in fact limitinggrowth, relative rankings among species,populations, genotypes and environmentsshould be similar, regardless of the cur-rency used (Reekie, 1999). To use the aboveexample, even though the proportion ofnitrogen allocated to reproduction may behigher than that of carbon, when compar-ing the reproductive allocation of twoplants (e.g. two genotypes from contrastingpopulations), the plant with the higher bio-mass allocation to reproduction shouldalso have the higher nitrogen allocation toreproduction. If both resources are in factlimiting growth, their allocation will besubject to selection and, if selection favourshigh RA, this should be reflected in bothcurrencies. The limited data availableappear to support this suggestion. Reekieand Bazzaz (1987b) grew three genotypesof Agropyron repensin seven differentnutrient environments (i.e. various levelsof nitrogen and phosphorus availability)and calculated RA in terms of biomass,nitrogen and phosphorus. Correlationsbetween RA calculated in terms of thethree currencies across the 21 genotype ϫenvironment combinations were uniformlyhigh. Hemborg and Karlsson (1998) com-pared reproductive allocation of 13 plantpopulations representing eight species,using biomass, nitrogen and phosphorus asthe currency of allocation. Correlationsamong RA calculated in terms of the threedifferent currencies across the various pop-ulations showed that the allocation of oneresource closely reflected the allocation ofthe other resources. Ashman (1994) com-pared the resource allocation of hermaph-rodite and female plants of Sidalceaoregana,using biomass, nitrogen, phospho-rus and potassium as currencies, and foundthat conclusions regarding the relative allo-cation patterns of the plants were unaf-fected by the currency chosen.

A second aspect of the currency ques-tion that has been raised in recent years isthe hypothesis that in some species growthand reproduction may be limited primarilyby the availability of meristems, ratherthan abiotic resources. In a study of popu-lation growth in the water hyacinth,

6F.A. Bazzaz et al.Eichhornia crassipes, Watson (1984) com-pared two populations growing in enclosedtanks, one of which flowered while theother remained vegetative. In this species,the rate of meristem production is fairlyconstant and every axillary meristem thatis produced is elaborated into either a veg-etative module, a new ramet, or an inflores-cence. As a result, in the population thatflowered, the rate of population growth, interms of cumulative ramet number,declined during flowering and the satura-tion population size in the tank was lowerthan in the vegetative population. A moredetailed study of potential meristem limita-tion in Polygonum arenastrum, an indeter-minate annual, was presented by Geber(1990). In P. arenastrum, like E. crassipes,every axillary meristem develops intoeither a branch or a terminal inflorescence.Geber compared growth and reproduction,based on commitment of meristems, in 26full-sib families from a natural populationgrown in a common garden experiment.Analysis of genetic and phenotypic correla-tions demonstrated that early commitmentof meristems to reproduction was nega-tively correlated with later reproductiveoutput and growth. In addition, a strongpositive genetic correlation was observedbetween the number of vegetative meris-tems and the number of reproductivemeristems that were produced late indevelopment, which was interpreted asevidence for meristem limitation – seeGeber (1990) for discussion of this interpre-tation. Both of the above studies examinedspecies whose meristems develop intoeither vegetative or reproductive shoots. Inmany species, however, meristems canhave a third fate: they can remain dormantfor prolonged periods of time and, indeed,may never give rise to an actively growingshoot. Lehtila et al. (1994) examined thequestion of meristem limitation in onesuch species, Betula pubescens. In thisspecies, meristems may develop into vege-tative shoots, generative long shoots withmale catkins or short shoots with femalecatkins or remain dormant. In spite of thecommitment of the shoot apices to catkin

production, male and female shoots hadapproximately the same bud productionrate as vegetative shoots, because new axil-lary buds developed from dormant meris-tems to compensate for the lost shootapices. This implies that this species is notlimited by meristem availability.

Studies that focus on meristem alloca-tion make an important contribution toresearch on reproductive strategies byfocusing attention on the developmentalcontext of growth and reproduction.However, it must also be kept in mind that,even in species whose meristems alwaysdevelop into either vegetative or reproduc-tive shoots, developmental programmesevolve in the context of the reproductivestrategies. Are rosette plants semelparousbecause they develop only a single apicalmeristem or do they only develop onemeristem because they are semelparous?Many of these plants, whose growth isapparently constrained by their single api-cal meristem, also grow vegetatively fromrhizomes; in one study of Agave deserti,95% of new plants were derived from rhi-zomatous production of vegetative ramets(Nobel, 1977). As is the case for differentabiotic resources, the availability of meri-stems in most plants is partially scaled tothe availability of resources. The rate ofmodule initiation by apical meristems usu-ally increases with increasing nutrient,light and, occasionally, carbon dioxideavailability and decreases with increasingdensity (Bazzaz and Harper, 1977; Harperand Sellek, 1987; Ackerly et al., 1992).Meristem activity, carbon gain and nutrientuptake are all strongly temperature-dependent as well. If meristem availabilityscales to resource availability, the completeelaboration of developed meristems doesnot necessarily indicate that meristemnumbers limit growth or reproduction.Testing the hypothesis of meristem limita-tion requires further research in two direc-tions: simultaneous comparisons ofmeristem and resource allocation, andresponses of meristem allocation to gradi-ents of resource availability.

Reproductive Allocation in Plants7Reproductive allocation (RA) versusreproductive effort (RE)RA is defined as the proportion of the totalresource supply devoted to reproductivestructures. RE, on the other hand, is usedto refer to the investment of a resource inreproduction, which results in its diversionfrom vegetative activity (Tuomi et al., 1983;Reekie and Bazzaz, 1987c; Fig. 1.2). As wehave mentioned, several factors can decouplethese two parameters in an individualReproductive allocationReproductive effortDirectRr1) TrRr+Rv2)Tr 3) Rr+Rv+Sr+ArTr+Sv+Av4)(Rr+Rv+Sr+Ar)-Pr(Tr+Sv+Av)-PrIndirect Vr-Vn5)Vn6)(Vr+Sv+Av)-(Vn+Sn+An)(Vn+Sn+An)Components of resource poolNon-reproductive plantVn =Vegetative sizeSn=Structural lossesAn=Atmospheric losses(respiration/transpiration)Reproductive plantTr=Total standing poolVr=Vegetative poolRr=Reproductive poolRv=Vegetative biomass attributable to reproductionSv=Structural losses from vegetative organsSr=Structural losses from reproductive organsAv=Atmospheric losses from vegetative organsAr=Atmospheric losses from reproductive organsPr=Enhancement of total resource supply due toreproductionFig. 1.2.Definitions of reproductive allocation (RA) and reproductive effort (RE). The resource pool of anindividual plant is divided into various components; these components can then be used to calculate variousmeasures of RA and RE. The relative importance of different components will differ in different studies, soseveral alternatives are provided for both RA and RE, depending on circumstances. Equations 1, 2, 4, 5 and 6are equivalent to the terms defined as RE2, RE3, RE5, RE6 and RE7, respectively, by Reekie and Bazzaz (1987a,c). Equations 1 and 5 are equivalent to RE/E and REsas defined by Tuomi et al.(1983). Artwork by K. Norweg.

8F.A. Bazzaz et al.plant. Three factors are considered here: (i)reproduction can lead to an increase inresource supply, either through directuptake by reproductive structures or byenhancement of uptake rates by vegetativestructures; (ii) vegetative growth may begreater in reproductive plants, such thatsome of the vegetative biomass should be

attributed to the function of reproduction;

and (iii) resources may be moved betweenvegetative and reproductive structures and,as a result, allocation of a resource to onefunction does not prevent its subsequentallocation to another.

Reproductive enhancement of the resource

supplyMany flowers and fruits are known to bephotosynthetic, and any carbon that is sup-plied in situreduces the investment inreproduction that is required from the veg-etative structures of the plant. In a study ofAmbrosia trifida, Bazzaz and Carlson(1979) reported that the contribution of insituphotosynthesis to the carbohydratedemands of male and female inflorescenceswas 41% and 57%, respectively. A compar-ative study of 15 temperate tree speciesfound that reproductive photosynthesiscontributed 2.3–64.5% of carbohydrateneeds for production of female flowers andseeds (Bazzaz et al., 1979). urik (1985)included respiratory costs associated withthe construction and maintenance of vari-ous reproductive organs in Fragaria virgini-ana, and found that in situphotosynthesiscontributed as much as 54.8% of carbohy-drate to flowers, but less than 10% toflower stalks, flower-buds and fruits. Forall organs combined, photosynthesis con-tributed 3.6–8.9% of carbon costs. The ach-enes of Ranunculus adoneus are green andmaintain a positive net assimilation ratethroughout fruit maturation (Galen et al.,1993). Shading the infructescences reducedthe weight of the mature achenes by16–18%.

Reproduction may also contributeindirectly to the plant’s total resource pool,due to enhancement of leaf photosynthetic

rates. Changes in leaf-level physiology, dueto the enhancement in sink strength duringreproduction, have been observed in fruit-trees and crop species (Neales and Incoll,1968; Deong, 1986). In the study ofAgropyron repensby Reekie and Bazzaz(1987a), changes in leaf photosynthesisranged from Ϫ33 to 64%, depending ongenotype and nutrient treatment. Leaves ofpollinated female plants in the dioeciousspecies Silene latifoliahave light-saturatedphotosynthetic rates 30% higher than thoseof unpollinated females 28 days after flow-ering (Laporte and Delph, 1996).Interestingly, male plants of S. latifoliahavehigher photosynthetic rates than pollinatedfemales, even though they have a lowerreproductive allocation and, apparently, alower sink strength (Gehring and Monson,1994; Laporte and Delph, 1996). Similarsexual differences have also been observedfor another dioecious species,Phoradendron juniperinum(Marshall etal.,1993). These apparent differences inthe effect of sink strength on photosyn-thetic rate may be related to sexual differ-ences in nutrient and water use (Gehring,1993; Gehring and Monson, 1994). In addition to direct photosynthesis byJ

reproductive structures and sink-inducedenhancement of leaf photosynthesis, repro-duction can enhance carbon gain throughchanges in canopy structure and allocationpatterns. In many plants, reproduction isassociated with stem elongation (i.e. bolt-ing). Elongation of the stem facilitates boththe receipt of pollen and the subsequentdispersal of seeds. This stem elongationcan also have beneficial effects on carbongain, in that it may reduce self-shading andmay improve the capacity of the plant tocompete for light in a closed canopy. InOenothera biennis, bolting has no effect ongrowth in an open canopy, but increasesgrowth in a closed canopy (Reekie et al.,1997). Reproduction has also been shownto enhance carbon uptake throughincreases in leaf area ratio, either throughincreases in allocation to leaves (Reekieand Bazzaz, 1992) or through changes inleaf morphology (i.e. specific leaf area)(Reekie and Reekie, 1991).

Reproductive Allocation in Plants9Although the effect of reproduction oncarbon uptake has been studied to a greaterextent, there is evidence that reproductioncan also enhance the uptake of mineralresources. By artificially manipulating lev-els of reproduction by removal of inflores-cences and comparing the total resourcepool in vegetative versus reproductiveplants, it is possible to examine the impactof reproduction on the uptake of a varietyof resources. Using this technique, Thorenet al.(1996) found that reproductive plantsof three different species of carnivorousplants in the genus Pinguiculaaccumu-lated more biomass, nitrogen and phospho-rus than equivalent vegetative plants. Themechanisms by which reproduction mayenhance nutrient uptake are not entirelyclear, but Karlsson et al. (1994) found that,in Pinguicula vulgaris, reproductive indi-viduals captured almost twice as manyprey as vegetative individuals. This couldaccount for as much as 58% of theincreased nitrogen uptake of reproductiveplants (Thoren et al.,1996). However, therewas no evidence of increased rates of preyuptake with reproduction in the other twospecies, in spite of the fact that reproduc-tion did enhance both nitrogen and phos-phorus uptake. Although reproduction mayenhance the uptake of mineral nutrients,positive effects of reproduction on carbonuptake may be more widespread. Hemborgand Karlsson (1998), using the same tech-nique as Thoren et al.(1996), examined theeffect of reproduction on biomass, nitrogenand phosphorus uptake in eight subarcticspecies. They found that reproductiveplants always accumulated more biomass(mostly carbon) than vegetative plants, butthat this was not the case for nitrogen andonly in a few cases did reproductive plantsaccumulate more phosphorus.

As a result of these direct and indirectinfluences of reproduction on the resourcebudget, the cost of reproductive organs willdiffer from that reflected in measures ofresource distribution.

Reproductive support structures

Reproduction can also lead to increases insupport structures beyond those necessaryto support leaves. This reproductive sup-port biomass may be readily apparent, suchas the stalks of rosette plants, which boltwhen they flower, or the upper stems ofmany herbaceous plants, which do not bearleaves. Thompson and Stewart (1981) havesuggested that any structures which are notfound on vegetative plants should be con-sidered part of reproductive biomass.Bazzaz and Reekie (1985) have also notedthat some existing vegetative organs willincrease in size in reproductive plants;they applied allometric relationshipsbetween leaf biomass and various supportstructures in vegetative plants in order todetermine the fraction of support biomassthat is attributable to reproduction. In con-trast to the enhanced resource uptake dueto reproduction, which decreases the car-bon cost to the vegetative plant, the consid-eration of reproductive support biomassmay significantly increase measures of RA.

In order to determine vegetative andreproductive biomass experimentally,Antonovics(1980) suggested the study ofplants in which flowering can be inducedby small variation in photoperiod, allowingthe comparison of flowering and non-flowering plants under essentially identicalresource conditions. Reekie and Bazzaz(1987a, c) followed this approach indetailed analysis of the carbon budget of A.repens, considering the influence on RE ofreproductive photosynthesis, respirationand changes in support biomass due toreproduction. Under various nutrient andlight conditions, inflorescence biomass inthis species ranged from 0 to 10% of totalbiomass. Inclusion of reproductive supportbiomass increased RA to 20–45%.Consideration of respiration had little effect,as discussed above. The inclusion of repro-ductive photosynthesis, however, decreasedRE to below 20% and, in one of the sixgenotypes, RE was negative in several treat-ments, suggesting that there was a net sup-ply of photosynthate to the vegetative plantas a result of reproduction.

10F.A. Bazzaz et al.A final problem is introduced by thereallocation and translocation of resourceswithin the plant. Several models andempirical studies have considered the con-sequences of carbon storage for patterns ofRA (Schaffer et al., 1982; Chiariello andRoughgarden, 1984). These models allowthe plant to store biomass for later deploy-ment to reproduction, but they still assumethat any given unit of resource may be allo-cated to only one function. In contrast,water and many mineral resources may berecycled within the plant. In the semel-parous, desert species Agave deserti, inflo-rescence development requires 18 kg ofwater, all of which can be supplied by real-locating water from the leaves (Nobel,1977). In this case, the allocation of aresource at one point in time does notreflect its functional utilization within theplant. Ashman (1994) examined the impor-tance of resorption of nitrogen and phos-phorus from reproductive structures incalculating reproductive allocation inSidalcea oregana. Two measures of RAwere calculated, one based upon the aver-age nutrient content of the reproductivestructures (‘initial’ sensuChapin, 1989) andone based upon nutrient content after anyresorption from senescing calyces, corollas,etc. had taken place (‘final’ sensuChapin,1989). The two estimates differed substan-tially. Reproductive allocation was reducedby as much as 70% by nutrient resorptionfrom senescing reproductive structures.Furthermore, measures of RA based uponresorption were more highly correlatedwith the effect of reproduction upon subse-quent performance than measures thatfailed to take resorption into consideration.Although this study provides an indicationof the possible magnitude of the errorinvolved in ignoring nutrient reallocation,ideally, reproductive allocation should alsotake into consideration the amount of timea nutrient ion is allocated to a particularfunction. Empirical determination of theextent of reallocation would require a com-plete resource budget for a plant throughoutits ontogeny. We are aware of no theoreticaltreatment that provides a framework forincluding reallocation among structures instudies of RA patterns (and consequentlyfor measures of RE), and suggest that it mayprovide a fruitful avenue for futureresearch.

There are many factors that must beincluded in evaluation of RA and RE.There is no single definition of these termsthat can both incorporate all of these fac-tors and be usable for research projects ofdifferent scale and motivation. We dosuggest, however, that each of the terms beused to refer to an exclusive set of con-cepts. RA refers to the proportion of aresource devoted to reproduction relativeto the total resource pool. Measurement isbased on the quantity of a resource inreproductive structures as a fraction of thetotal resource pool, plus consideration ofreproductive support tissues and losses ofthe resource through senescence, respiration,etc. whenever possible. The direct measure-ment of RE, on the other hand, requires cal-culation of the change in the resourcesupply due to reproduction. This quantityis removed both from the investment inreproduction and from the total quantityavailable (Bazzaz and Reekie, 1985).Indirect measurements of RE can be calcu-lated if the total size of non-reproductiveand reproductive plants in the sameresource environment are known. RE isthen defined as the proportionate change invegetative biomass resulting from repro-duction (Tuomi et al., 1983; Reekie andBazzaz, 1987c). In calculating indirect mea-sures of RE, only standing resource poolsneed be considered, excluding tissue andrespiratory losses, in order to focus on theresources available for future growth andreproduction. These definitions are formal-ized in Fig. 1.2. This distinction betweenRA and RE is critical if the concepts are toretain their usefulness both in ecologicalstudies, relating allocation to fecundity,and in the study of life-history evolution,which is based on potential trade-offsbetween reproduction and future survivaland fecundity.

Reproductive Allocation in Plants11RA and reproductive output (RO)

If two plants of equal size differ in theirallocation to reproduction, the individualwith the higher allocation would beexpected to have higher reproductive suc-cess in terms of contribution to futureregeneration of the population. However,RA is only one component of fecundityand, in order to link allocation and RO, itmust be considered in relation to plantsize. RA is a measure of the proportion ofresources devoted to reproduction, whileRO is a measure of the total quantity ofreproduction. The absolute quantity ofreproductive biomass will be the productof total plant biomass and RA. In addition,the quantity of seed biomass will dependon the seed fraction relative to other struc-tures; the number of seeds will equal thetotal seed mass divided by the mean indi-vidual seed mass, and the number of suc-cessful offspring will depend ongermination percentage and seedling sur-vivorship. The latter components of suc-cessful reproduction, germination andestablishment are treated in later chaptersof this book; we shall focus on recent stud-ies of the relative contributions of plantsize and RA to total fecundity. Variation inthese parameters can be considered at threelevels: phenotypic differences among indi-viduals resulting from environmental influ-ences; genotypic and ecotypic differencesamong individuals of a population or pop-ulations of a species; and variation amongspecies across habitat types and life-forms.

Environmental influences

Owing to the indeterminate nature ofdevelopment in most higher plants, thereproductive individuals in a populationcan vary enormously in size (Harper, 1977).Clearly, at this broad scale, variation in RAcannot compensate for such large size dif-ferences. A plant that weighs 10g anddevotes 20% of its biomass to seeds willnecessarily have a higher reproductive out-put than another individual which weighs1g, regardless of the latter’s pattern of allo-

cation. Variation in both size and RA canarise from fine-scale environmental hetero-geneity within the range of a population,and from competitive interactions, whichlead to inequalities in the distribution ofresources and result in population sizehierarchies (Weiner and Solbrig, 1984).

In annual species, the majority of thestudies seem to indicate that variation in

RO is correlated more closely with varia-tion in size than with RA. For example, in

experimental monocultures of Ambrosiaartemisiifolia, final biomass of reproduc-tive individuals varied from 1.5 to 12 g, butRA did not vary significantly with size; as aresult, most variation in reproductive out-put was due to variation in plant weight(Ackerly and Jasienski, 1990). The influ-ences of resource availability on size ver-sus allocation are also demonstrated invarious experimental studies under con-trolled conditions. In their pioneeringstudy of RA, Harper and Ogden (1970)grew Senecio vulgarisin three pot sizes. Inthe medium pots, total production wasreduced by 85% relative to plants in thelarge pots, while seed allocation wasreduced by only 21%; as a result, most ofthe reduction in RO was due to thedecrease in size. In the smallest pots, how-ever, most individuals failed to set seed, soreproduction failed completely. In a studyof several Polygonum persiariaclonesgrown along a light gradient, Sultan andBazzaz (1993a) observed changes of morethan two orders of magnitude in both totaland fruit biomass, while RA only changedfrom about 10% to 25%. Cheplick (1989)investigated allocation to above- andbelow-ground seed production in theannual grass Amphicarpum purshiigrownat two nutrient levels. Variation in fecun-dity was mostly due to variation in plantsize, but allocation to below-ground seedproduction was fairly constant at the twonutrient levels, while allocation to above-ground seeds increased. Thus, in mostcases, variation in RA is less significantthan variation in biomass in annuals, suchthat RO in many populations is stronglycorrelated with plant size (Harper, 1977;Solbrig and Solbrig, 1984). This conclusion

12F.A. Bazzaz et al.is supported by studies that utilize pathanalysis to determine the contribution ofdifferent traits to reproductive success innatural populations of annuals (Farris andLechowicz, 1990; Mitchell-Olds andBergelson, 1990; Schwaegerle and Levin,1990). These studies have all demonstratedthe importance of early size differencesand subsequent variation in growth rate tovariation in final size and fecundity.Similar conclusions emerge from studieson the role of seed size and emergence timein determining competitive hierarchies ofsize and fecundity (Stanton, 1985; Waller,1985).

The relative importance of size versusRA in determining RO in perennials is lessclear. Annual plants have only one oppor-tunity to reproduce; therefore, it is not sur-prising that variation in RA tends to berelatively low. All annuals should have auniformly high RA at maturity and it is tobe expected that most variation in RO willbe associated with size. Variation in RA inperennials, on the other hand, should begreater, given that individuals may haveopportunities to reproduce in future years.There is, indeed, a great deal of evidencethat RA in perennials varies with a varietyof environmental factors (e.g. Wankhar andTripathi, 1990; Dale and Causton, 1992;Williams, 1994; but see Benech Arnold etal., 1992). This variation can have a signifi-cant impact on total RO. For example, in astudy of the perennial grass Agropyronrepens, Reekie and Bazzaz (1987c) foundthat RO measured as the weight of theinfructescences increased over fivefold aslight and nutrient levels increased fromminimal to maximal levels. Increases inplant size could only account for 50% ofthe increase in RO; the remaining 50% wasaccounted for by an increase in RA. Therole of variation in RA in explaining life-time RO, however, has not been studied tothe same extent as it has in annuals,because of the difficulties in followingindividuals over an indeterminate lifespan(see Alvarez-Buylla and Martínez-Ramos(1992) for a detailed study of lifetime RO inan early successional tropical tree).

Given that environmentally inducedvariation in RA may be important in deter-mining RO in perennial, if not annual,species, there are relatively few studiesthat have attempted to explain this varia-tion. Instead, variation in RA has beeninterpreted, for the most part, as an in-direct consequence of the effect of the envi-ronment on plant size. RA of perennials,calculated on the basis of annual produc-tion, is frequently observed to vary withplant size (Piñero et al., 1982; Samson andWerk, 1986; Bazzaz et al., 1987; Oyama andDirzo, 1988; Weiner, 1988; see Klinkhameret al. (1990) for a discussion of statisticalproblems associated with the analysis ofsize-dependent allocations). Samson andWerk (1986) suggest that this correlationresults from underlying allometric con-straints. For example, if flowers are bornein the axils of leaves, increasing the num-ber of leaves will also increase the numberof flowers. They go on to demonstrate that,if this dependence results in a linear rela-tionship between RO and plant size, RAwill increase, decrease or remain constant,depending upon whether the x-intercept ispositive, negative or goes through the ori-gin (see Fig. 1.3). They therefore argue thatsize-dependent changes in RA are simply areflection of morphological constraints andshould be ignored when trying to deter-mine whether or not there has been directselection for phenotypic plasticity in RA.Instead, it is suggested that it is necessaryto correct for differences in plant sizeamong environments by comparing the lin-ear relationship between RO and plant sizeamong environments. Only if the slope orintercept of this relationship changesamong environments can it be concludedthat there are differences in allocation pat-terns that may be the result of selection.This approach has been widely adopted inrecent studies of RA in plants (e.g.Hartnett, 1991; Aarssen and Taylor, 1992;deRidder and Dhondt, 1992; Korpelainen,1992; Schmid and Weiner, 1993; Claussand Aarssen, 1994; Pickering, 1994). Thesestudies demonstrate that much of the varia-tion in RA among environments canindeed be correlated with differences insize. Although there are often shifts in the

Reproductive Allocation in Plants13Fig. 1.3.The implications of a linear relationship between reproductive biomass and plant size (a) on therelationship between reproductive allocation and plant size (b). Depending on whether the y-intercept ofthe relationship between reproductive biomass and plant size is zero, negative or positive, RA will not vary(A), increase (B) or decrease (C) with plant size, respectively. (From Samson and Werk, 1986.)

relationship between RO and size amongenvironments, these effects on RA are oftenless important than the size-correlatedeffects. For example, in a study of naturalpopulations of several Centauriumspecies,Schat et al. (1989) found that RA increasedmore rapidly with size in infertile sites, butthe level of allocation was higher in fertilesites because of the larger overall plantsize.

The above approach has a great deal ofmerit, in that it explicitly recognizes thatchanges in RA may not be the result ofdirect selection but simply a consequenceof morphological or allometric constraints.Unfortunately, it also has a major disadvan-tage, in that the ‘constraints’ are describedonly, in somewhat vague terms, as anysize-correlated change in RA. The idea thatsize-related changes in allocation patternsare a direct result of constraints is origi-nally derived from the zoological literature.In particular, it stems from studies of skele-tal structure, where there is a fixed rela-tionship among the sizes of various bones,which can be described by a linear rela-tionship between the logarithm of the sizeof one bone and the logarithm of another(i.e. an allometric relationship). This linearrelationship appears to hold true acrossenvironments and developmental stagesand even among species. The fact that therelationship is constant implies that it is‘constrained’ in some fashion. The con-straint in the case of skeletal structure isquite apparent. Take the bones of a skull,for example; if you increase the size of oneof the bones and do not change the size ofthe other bones in the skull in proportion,it is unlikely that the various bones willstill be able to fit together (i.e. the skullwill fall apart). Allometric relationships inplants, however, are much more variable.They vary among environments, develop-mental stages and species (Hartnett, 1991;deRidder and Dhondt, 1992; Korpelainen,1992; Clauss and Aarssen, 1994; Welhamand Setter, 1998). Given that plants have amodular structure and indeterminategrowth, as compared with the very deter-minate growth patterns of most animals,this difference should not be surprising. Touse an earlier example, RA in a particularplant species may be ‘constrained’ by thenumber of leaves in inflorescences if theflowers are produced in the axils of leaves.

14F.A. Bazzaz et al.But, given that in most plants a large pro-portion of axillary buds remain dormantand never develop into inflorescences orbranches, the individual still has a greatdeal of flexibility in RA, in that it can con-trol what proportion of the population ofbuds actually develops into inflorescences(Lehtila et al., 1994). The fact that allomet-ric relationships in plants can vary impliesthat RA is not closely constrained by thesize of the plant and that it is free (i.e. vari-ation is present at a given size) to be actedupon by selection. We cannot automati-cally assume that size-dependent variationin RA is a function of allometric constraintson the basis of a principle imported fromthe zoological literature that is not applica-ble to plants. Rather, we must criticallyexamine the question of whether there areany size-related constraints based upon anunderstanding of the biology of the particu-lar species in question.

One such explanation is offered byWeiner (1988). He argues that many plantsmust reach a minimum size before repro-duction is possible. This minimum size isnecessary because the plant must haveenough resources to construct the repro-ductive support structures (e.g. the flowerstalk) before the first seed can be produced.Further, as plant size is increased beyondthe minimum size, there should be a moreor less linear increase in RO, as plant sizereflects the accumulated resources avail-able for reproduction. As a result, thereshould be a linear relationship between ROand plant size, with a positive x-interceptthat represents the minimum size requiredfor reproduction (i.e. maximum plant sizewhen RO is still zero). Given this relation-ship, RA will initially increase with sizeand eventually approach an asymptote(Fig. 1.3). With this model, variation in RAwith size is a simple consequence of thenecessity of reaching a minimum sizebefore reproduction is possible. One prob-lem with this model is that, although therelationship between RO and plant sizeoften has a positive intercept, there are alsomany cases in which the intercept is nega-tive (Samson and Werk, 1986; Shipley andDion, 1992). This implies that the mini-mum size required for reproduction is lessthan zero. Since this is clearly not possible,the simple interpretation of the x-interceptas the minimum size required for repro-duction may not be correct. Klinkhamer etal. (1992) offer a possible solution to thisdilemma by suggesting that the relation-ship between plant size and RO is, in manycases, curvilinear. This could result if therewere economies of scale such that fewerresources were required to produce a seedas more seeds are produced. Fitting a linearrelationship to such data would result inan underestimate of the x-intercept. Theyprovide an alternative model that incor-porates both a minimum size required forreproduction and the potential for a curvi-linear increase in RO with plant size. Thisis an attractive model, in that it explainssize-dependent variation in RA on thebasis of constraints that are directly relatedto plant biology. To date, however, therehas been little attempt to test the model rig-orously to see if it truly describes the rela-tionship between RO and plant size for avariety of species (e.g. see Schmid andWeiner, 1993).

Given that it is unclear to what extentsize-dependent variation in RA is a func-tion of constraints, it would be unwise tosimply dismiss the possibility that thisvariation may in fact be adaptive. There areprobably a number of reasons why it maybe adaptive for RA to vary with plant size.For example, Hara et al.(1988) have pre-sented a model of optimal allocation strate-gies which predicts that the relationshipbetween biomass and RA in herbaceousplants will depend on habitat. In species ofopen, disturbed environments, they predicta constant allocation to RA among individ-uals of a population, as observed in severalexamples above, but, for plants of closedhabitats, such as forest understorey, theypredict a decrease in RA with increasingbiomass. The difference between the twotypes of plants emerges in part from theassumptions about the relationshipbetween individual plant size and thelength of the growing season. For plants ofclosed environments, growing-seasonlength is assumed to be constant, regard-

Reproductive Allocation in Plants15less of plant size. For plants of open, dis-turbed environments, however, small sizeis assumed to be associated with a shortergrowing season, as smaller plants are oftenshaded early in growth by larger individu-als in the population. The predicted pat-terns in relation to habitat are supported bydata for several species in Japanese plantcommunities.

Genotypic variation

Relatively few studies have focused ongenotypic variation in size, RA and fecun-dity within natural populations, despite itscentral role in evolutionary dynamics.Furthermore, most of these studies havefocused on annual species, where onewould expect relatively little variation inRA. Geber (1990) studied 26 families of fullsiblings from a population of Polygonumarenastrumgrown in a common garden tominimize environmental influences. Sheobserved a twofold variation in total size,measured as either biomass or meristemnumber, and found strong genetic and envi-ronmental correlations between size andtotal reproduction. Sultan and Bazzaz(1993a, b, c) grew clones of 20 genotypesfrom two populations of P. persicariaonthree resource gradients, namely, nutrients,light and moisture. Along all three gradi-ents, fecundity was positively correlatedwith size. Genetic variation for RA wasfound in response to all three factors, butgenetic variation for total size and total fruitbiomass was present only in the moistureexperiment. Clauss and Aarssen (1994) grewthree genotypes of Arabidopsis thalianaonlight, nutrient and pot-volume gradients.Fecundity increased with plant size, butanalysis of covariance revealed that the rela-tionship between size and fecundity (i.e.RA) varied both among genotypes andamong environments. Sugiyama and Bazzaz(1997) examined the relative importance ofsize and RA in determining fecundity inAbutilon theophrastiby growing eight fami-lies of full siblings across a nutrient gradi-ent. They found that variation in RA wasmore important in determining seed output

at low nutrient levels, while variation inplant size was more important at high nutri-ent levels. Also relevant here are studies ofyield determination in crop cultivars wherethe harvested portion is a seed or fruit.These studies examine how increased yields(i.e. RO) have been achieved through cropselection. Both modern and historic culti-vars are grown in a common environmentand growth (i.e. size), harvest index (i.e. RA)and yield (i.e. RO) are compared. Thesestudies show that the substantial increasesin yield that have been achieved by cropselection are to a large extent the result ofincreases in RA rather than an increase insize (e.g. Wells et al., 1991).

Collectively, the above studies indicatethat genetic differences in both RA and sizecan be important in determining reproduc-tive output, but their relative importancevaries. Based upon the limited data, it isdifficult to make generalizations as to whensize versus RA will be more important.However, it seems logical that size shouldbe relatively more important wheneverthere is asymmetric competition thatresults in marked size hierarchies, such asat high nutrient levels and plant densities,where competition for light is the predomi-nant factor influencing plant size (e.g. athigh soil fertility in the Sugiyama andBazzaz (1997) study). On the other hand,RA should be relatively more importantwhen plant densities are low (e.g. culti-vated plants grown at spaced intervals) orwhen competition is symmetric, as is thecase when nutrients are limiting (e.g. atlow soil fertility in the Sugiyama andBazzaz (1997) study).

Variation among populations

Studies on a variety of plant speciesdemonstrate differences in RO among pop-ulations that are attributable to a combina-tion of genetic and environmental factors.In some cases, populations with greaterbiomass also have higher allocation toreproduction, and both contribute togreater reproductive biomass (e.g. Solidagospeciosa: Abrahamson and Gadgil, 1973).

16F.A. Bazzaz et al.In contrast, in five populations ofPolygonum cascadensegrowing along amoisture gradient, RA decreased while veg-etative biomass increased in successivelywetter sites; the increase in plant size wasgreater than shifts in allocation, so that ROas such increased with size. When two ofthese populations were grown under com-mon conditions, the size differences per-sisted, but allocation to reproduction wassimilar (Hickman, 1975). On the otherhand, Reekie (1991) found that differencesin RA among populations of Agropyronrepenspersisted in a common gardenexperiment, populations from disturbed

sites having a lower RA than populations

cfrom less disturbed sites. DeRidder andDhondt (1992) compared the relationshipbetween fecundity and plant size amongpopulations of Drosera intermediagrowingin disturbed versus undisturbed sites andfound that the slope of this relationshipwas steeper in more disturbed sites. InPlantago laneolata, Primack andAntonovics (1982) observed a range of RAamong eight field populations, from 17 to65%, which was attributable primarily toenvironmental influences. No data are pro-vided on relative plant size, however, inorder to evaluate the contribution of RA tofecundity. Hartnett et al. (1987) comparedtwo populations of Ambrosia trifidafromfields of different successional age. Whengrown in a common garden, the plantsfrom the older field produced 78% moreseed biomass than those from an annuallydisturbed field, but this difference was pri-marily due to a 55% increase in total sizeand secondarily to a smaller increase inRA. Scheiner (1989) compared populationsof the grass Danthoniaspicatafrom fivesites that differed in the time since the lastburning. Plants in the most recently burntsite had the highest fecundity, due to thelarger size, a higher percentage of floweringstalks and more spikelets per stalk. All ofthese factors were reduced in populationsfrom intermediate age, but in the older site,which was burned 80 years previously, aslight increase in the percentage of flower-ing stalks resulted in an increase in fecun-dity. Hartnett (1991) conducted a similar

study with the tall-grass prairie forbRatibida columnifera, but found that plantsin the most recently burned sites had thelowest fecundity. Increases in fecunditywith time from last burning were largelythe result of increases in plant size.However, there were shifts in the relation-ship between fecundity and plant sizeamong sites, such that RA increased withtime from last burning. Ostertag andMenges (1994) present a series of modelsthat predict contrasting patterns of RA withtime since last fire, depending upon firefrequency and its predictability.

In general, variation in RO amongpopulations appears to be primarily due todifferences in total size and, to a lesserextent, to changes in allocational patterns.At the same time, however, it is also clearthat there are often marked differencesamong populations in RA. Some of thesedifferences are correlated with differencesin size, but size-independent effects arealso common. Furthermore, these differ-ences in RA can persist when plants fromdifferent populations are grown in a com-mon environment. This suggests that selec-tion for differences in RA have taken place,even though it seems to have a relativelyinsignificant effect on RO when popula-tions are compared in the field. This raisesan interesting question. Why would therebe selection for differences in RA if it didnot have a significant effect on fitness? Thisapparent contradiction can be resolved ifyou consider that fitness should be mea-sured relative to competitors within agiven environment. Although there may belarge differences in size among populationsgrowing in different environments, withina given population the environment islikely to be less variable and the variationin size correspondingly less. As a result,within a given population, variation in RAmay be more significant in determining ROthan variation in size.

Variation among species and communitiesMany studies have demonstrated that RA ishigher in annuals than in perennials, and

Reproductive Allocation in Plants17higher in herbaceous plants of open habi-tats than in those of closed habitats (e.g.Gadgil and Solbrig, 1972; Pitelka, 1977;Abrahamson, 1979; see reviews byEvenson, 1983; Hancock and Pritts, 1987).Allocation of community-level productionto sexual reproduction also decreases withsuccessional age; in a recent study of old-field succession in the American Midwest,RA of community biomass declined fromabout 8% in recently abandoned fields tobelow 1% in fields abandoned for 20 ormore years (Gleeson and Tilman, 1990). Ina study of bog vegetation in England,Forrest (1971) found that total allocation ofproduction to sexual reproduction wasonly 0.5%. RA of production in forest treesis approximately 5% in both temperate andtropical forests, and can be as low as 2%(Whittaker and Marks, 1975; Kira, 1978).

Silvertown and Dodd (1996) point outthat comparisons of RA (and other traits)among species should control for phy-logeny. Otherwise, it is impossible to knowwhether the trait correlations that areobserved (e.g. differences in RA betweenannuals versus perennials) result from con-vergent evolution or simply from commondescent. They reanalysed data from the lit-erature and used phylogenetically indepen-dent contrasts to examine whether annualversus perennial and early versus late suc-cessional species differ in RA. Their analy-sis supports the conclusions of the earlierstudies, in that they found annuals andspecies of early succession have greater RAthan perennials and species of later succes-sion.

In general, comparisons of total fecun-dity of different species or life-forms aredifficult to interpret, because of variationin life-history strategies. A general discus-sion of plant reproductive capacity is pre-sented by Salisbury (1942).

Allocation of reproductive biomass to seedsThe proportion of reproductive biomassallocated to seeds varies widely. Lloyd(1988) reviews data for several species; in15 outcrossing taxa, seed and fruit alloca-

tion ranges from 34 to 83% of reproductivebiomass, while, in females of the dioeciousspecies Silene albaand in Impatiensspp.,which have cleistogamous flowers, thesevalues are greater than 90%. In a compara-tive study of numerous species with vari-ous breeding systems, Cruden and Lyon(1985) found that most outcrossing species

exhibit greater investment in male than in

Jfemale reproductive structures in the flow-ers, but following fertilization the majorityof reproductive biomass is devoted to seedsand fruits. These data suggest that the pri-mary factor influencing the evolution ofreproductive allocation is pollinator limita-tion, in the case of male function, and thenecessity of provisioning and dispersingthe offspring, in the case of female function(Antos and Allen, 1994). Because seed fill-ing is considerably more expensive thanflower production in most species, the seedweight fraction of reproductive biomasscommonly decreases when resources arelimiting. Failure to set seed may resultfrom limitation by pollinators or resources,and there is little consensus regarding theirrelative importance in natural systems(Lee, 1988; Calvo and Horvitz, 1990;Lawrence, 1993; Ramsey, 1997). Manyplants vary the relative investment in maleversus female reproductive structures as afunction of plant size; in most species, rela-tive investment in female functionincreases with size (Deong andKlinkhamer, 1989), though the oppositepattern is observed for several wind-polli-nated species: e.g. Ambrosia trifida(Abul-Fatih et al., 1979); Xanthium strumarium(Solomon,1989); A. artemisiifolia(McKoneand Tonkyn, 1986; Ackerly and Jasienski,1990). In most cases, total RO increaseswith size, even when relative allocation tomale or female shifts. However, in the caseof genotypes of Plantago majorgrown in acommon environment, it was found thatRO did not increase with size, becauselarger genotypes allocated a larger propor-tion of their resources to reproductive sup-port structures; i.e. the length of theflowering culm increased with size,decreasing the resources available for seedproduction (Reekie, 1998a). The larger

18F.A. Bazzaz et al.genotypes were isolated from habitats witha relatively tall canopy; therefore, theirlonger flowering culms may increase fit-ness by ensuring pollination (it is a wind-pollinated species) and effective seeddispersal.

Considerable attention has also beengiven to the trade-off between seed sizeand number for any quantity of seed bio-mass (Harper et al., 1970). Seed size is usu-ally small in species from more disturbedhabitats, which exhibit high levels of dis-persal in time and space. In a comparisonof nine temperate tree species, Grime andeffrey (1965) demonstrated that seedlingsurvivorship in the shade was positivelycorrelated with seed size. Foster (1986)reviews the value of large seed size fortrees of moist tropical forests. Leishmanand Westoby (1994) suggest that large seedsize may be an advantage in the shade fortwo reasons. First, large seeds, with theirgreater initial reserves, allow seedlings tosurvive longer, which may permitseedlings to survive until a gap in thecanopy is created. Secondly, large-seededspecies exhibit a greater etiolation responsethan small-seeded species, and this heightdifference will give large seeds an advan-tage where there is a steep light gradient,such as when seeds germinate below litter.Grubb and Metcalfe (1996) compared seedsize of shade-tolerant and intolerantspecies at both the intergeneric level andthe intrageneric level in the Australiantropical lowland rainforest flora. Theyfound that large seeds were associated withshade tolerance at the intergeneric levelbut not at the intrageneric level. Theyinterpreted these results to mean that seedsize is fairly low in the hierarchy of charac-teristics enabling plants to become estab-lished in shade. Seed size also increasesalong a gradient of decreasing moisture inplant communities in California, which isinterpreted in light of the requirements forrapid early growth in dry environments inorder to reach water-supplies (Baker, 1972).Armstrong and Westoby (1993) havedemonstrated that large-seeded species tol-erate defoliation better than small-seededspecies in the initial stages of growth. Inspecies of composites with larger seeds,however, a greater fraction of the seed isinvested in seed-coat, and seedlings from

larger seeds have lower initial relative

cgrowth rates (Fenner, 1983). Shipley andPeters (1990) have compiled data on 204species and found a significant negativecorrelation between seed size and seedlingrelative growth rate. Gross (1984), in astudy of seedling establishment of sixmonocarpic perennials, found that this cor-relation is environment-dependent, andwas reversed when the seeds germinatedunder a cover of Poa annua. In contrast, ina field experiment of colonization ofvarious size gaps in a grass canopy, seedsize differences among four annual speciesdid not correlate with performance in dif-ferent gap sizes (McConnaughay andBazzaz, 1987).

Seed size is frequently considered theleast plastic component of fecundity, incomparison with plant size and seed allo-cation (Harper et al., 1970). This apparentconstancy results in part from the tendencyto determine the mean weight of largenumbers of seeds, rather than the distribu-tion of individual seed weights (Fenner,1985c). Recent studies have shown thatthere is frequently considerable variationin seed size, even within individual plants;much of this variation can be attributed topositional effects during development,exemplified by the sizes of successiveseeds in pods of many legumes (Wulff,1986). Thompson (1984) reported eightfoldvariation in size within individuals ofLomatium grayi(Umbelliferae) and 15-foldvariation within a natural population.Manasse (1990) reported a range in seedsize from 0.1 to 66 g in the tropicalperennial herb Crinum erubesens(Amaryllidaceae). Sultan and Bazzaz (1993b)observed decreases in seed size ofPolygonum persicariaalong a moisture gra-dient paralleling the community-level pat-terns found by Baker, but the consequencesof this variation for germination and estab-lishment are not known. In response to ele-vated carbon dioxide levels, seed weight ofAbutilon theophrastiincreases and seednumber decreases (Garbutt and Bazzaz,

Reproductive Allocation in Plants191984). Large seed size has been shown toresult in competitive advantages in thecourse of intraspecific competition in pop-ulations of Raphanus raphanistrum(Stanton, 1985); in contrast to the patternsobserved between species, variation inseed size within a population of R.raphanistrumdid not influence initialgrowth rate (Choe et al., 1988). When rela-tive timing of emergence was controlledexperimentally, seed size also influencedthe outcome of interspecific competitionbetween two annual species grown in con-trolled conditions (Bazzaz et al., 1989).

RE and life-history evolution

As we stated above, the concept of RE isnot synonymous with reproductiveresource allocation. In order to understandthe evolution of reproductive strategies, REmust be considered not only in physiologi-cal terms, but in relation to demographiccomponents of fitness. Willson (1983)defines the three primary components ofreproductive performance as clutch size,timing of reproduction and frequency ofreproduction. RE can be defined physiolog-ically either in terms of the supply ofresources invested in reproduction whichis derived from the vegetative plant, or byexamining changes in vegetative biomassresulting from reproduction (Tuomi et al.,1983; Reekie and Bazzaz, 1987a, c). Inorder to relate these physiological mea-sures of RE to a life-history measure, therelationship between investment in vegeta-tive function and future reproduction mustbe established. Additionally, trade-offsbetween current and future reproduction(residual reproductive value (Williams,1966)) will only contribute to evolution oflife history if they have a genetic basis, butthere are still few studies that demonstratenegative genetic correlations between life-history parameters or directly assessresponses to selection (Reznick, 1985).

In demographic studies, reductions insurvival or future fecundity resulting fromreproductive activity are frequently termedthe cost of reproduction. In evolutionary

terms, however, the essential parameter istotal contribution to future generations, notsurvival of individual organisms. Con-sequently, we suggest that the negativecorrelations observed between presentreproduction and residual reproductivevalue be identified as trade-offs. The conse-quences of different reproductive sched-ules must be considered in terms of totalreproductive success, and a cost of repro-duction is incurred if investment in repro-duction at a particular time leads to adecrease in lifetime reproduction relativeto other members of the population.Geber’s (1990) study of Polygonum arenas-trumillustrates this distinction: sheobserved negative genetic correlationsbetween early and late reproduction, butno genetic correlation between early andtotal reproduction. Thus, there is a trade-off between early and late reproduction,but in this experiment there was no extracost for early reproduction.

Patterns of RA and RE are most com-monly studied in annuals; for annualplants that live in seasonal environmentsand produce only one generation each year,the only component of reproductive suc-cess is total fecundity. Because all seedsmust remain dormant until the followingyear, there is no advantage to early repro-duction due to exponential increase of pre-cocious offspring, as there is inpopulations with overlapping generations(Cole, 1954). Timing will, however, beimportant if the length of the growing sea-son is unpredictable and reproduction isindeterminate (King and Roughgarden,1982b; Geber, 1990). In determinate annuals,a physiological measure of RE can beobtained from a careful study of the indi-vidual’s resource budget. However, theoptimal value of RE will not depend on theinteraction of demographic parameters, buton the developmental problem of maximiz-ing the product of plant size and RA. Anumber of models exist that predict theoptimal pattern of allocation for annualplants. By assuming that the structure andfunction of vegetative and reproductiveorgans are strictly distinguished, the mod-els can derive constraint functions for

20F.A. Bazzaz et al.growth and fecundity, and then apply opti-mization techniques (Fox, 1992). In envi-ronments with a fixed growing season, theoptimal allocation strategy involves a sin-gle switch from purely vegetative to strictlyreproductive growth (Cohen, 1971). If theloss of vegetative and reproductive biomassduring the growing season is considered,multiple switches between vegetative andreproductive activity may be optimal (Kingand Roughgarden, 1982a). In contrast, ifthe length of the growing season variesunpredictably, a gradual transition fromvegetative to reproductive allocation is pre-dicted (King and Roughgarden, 1982b;Kozlowski, 1992) and, if seasonality influ-ences growth and reproduction differen-tially, plants should initially store resourcesand subsequently allocate them to repro-duction (Schaffer et al., 1982; Chiarielloand Roughgarden, 1984). Several of thesemodels have also explored the factors thatmay control the time of the switch fromvegetative to reproductive growth. Cohen(1976) demonstrated that early reproduc-tion would be advantageous if relativegrowth rate decreases with size, if the prob-ability of mortality increases with time or ifthe reproductive structures are photosyn-thetic. High rates of vegetative tissue losswill also favour an early switch to repro-duction (Kozlowski, 1992). On the otherhand, late reproduction will be favoured ifa large size confers particular advantages,such as improving the ability of the plantto compete for light (Schaffer, 1977) orimproving access to pollinators (Cohen,1976). Late reproduction will also befavoured if flowering uses reserves storedduring the vegetative phase (Chiariello andRoughgarden, 1984; Kozlowski andWiegert, 1986).

Trade-offs between current reproduc-tion and residual reproductive value havebeen demonstrated for several taxa. Forexample, Sarukhan, Piñero and their co-workers have conducted detailed studies ofresource budgets and demographic parame-ters in the tropical palm Astrocaryum mex-icanum(Martínez-Ramos et al., 1988).Using matrix methods, they have estimatedlifetime reproductive schedules anddemonstrated a strong negative correlationbetween fecundity at various ages andresidual reproductive value (Piñero Jet al.,1982; see also Oyama and Dirzo, 1988).Primack and Hall (1990) manipulated levelof reproduction in Cypripedium acaulebycontrolling pollination and examined theimpact of reproduction on subsequent sur-vival, growth and reproduction over a 4-year period. Reproduction decreasedgrowth and reduced the probability offlowering in subsequent years. In trees,negative correlations are frequentlyobserved between reproductive activityand vegetative growth within a season(Kozlowski, 1971). Newell (1991) foundthat branches of Aesculus californicathatfruited in one season had greatly reducedleaf-area development the following sea-son, but it was not possible to examineinteractions at the whole-plant level.Cipollini and Whigham (1994) experimen-tally manipulated reproduction in thewoody shrub Lindera benzoin.Fruit thin-ning enhanced fruit production the follow-ing year.

Although there are many studies thatdemonstrate negative correlations betweencurrent reproduction and residual repro-ductive value, there are also many studiesthat fail to find any relationship. For ex-ample, Horvitz and Schemske (1988)directly manipulated reproductive outputby removing flower-buds in a population ofthe perennial, tropical herb Calathea ovan-densis. In contrast to the studies above,reproduction did not result in any signifi-cant changes in survival, growth or repro-duction in the following season. Similarstudies have been conducted with a widevariety of taxa, including Pinguicula alpina(Karlsson et al., 1990), Viscaria vulgaris(ennersten, 1991), Alaria nana(Pfister,1992) and Blandfordia grandiflora(Ramsey, 1997), with identical results. Thisapparent lack of any trade-off between cur-rent reproduction and residual reproduc-tive value is puzzling. True, there areknown physiological mechanisms that mayreduce the impact of reproduction ongrowth (see literature discussed above),but, if current reproduction really has no

Reproductive Allocation in Plants21effect on residual reproductive value, thenwhy has there been no selection to increaseRA to the point where there is a trade-off?Some other factor must be constraining orlimiting RA. A partial answer to this ques-tion may be pollen limitation (Calvo andHorvitz, 1990). Reproduction in some ofthese species is known to be limited bypollen availability (e.g. see ennersten,1991; Ramsey, 1997), but this cannotexplain the lack of trade-off in all of thesespecies. It is also possible that RA is con-strained by extreme events, rather than‘normal’ or ‘average’ conditions. Severalstudies have shown that the trade-offbetween current reproduction and residualreproductive value varies with environ-ment (Zimmerman, 1991; Syrjanen andLehtila, 1993; Agren and Willson, 1994;Ramadan et al., 1994; Saikkonen et al.,1998). For example, Primack et al. (1994)found that, in Cypripedium acaule, plantshad to be placed under severe physiologi-cal stress by extensive defoliation beforeexperimentally increasing level of repro-duction by hand pollination had any nega-Jtive effects on growth or flowering the

following year. If RA is limited by theseextreme events, it is not surprising that itmay be difficult to detect trade-offs undermany circumstances. It will also be diffi-cult to detect trade-offs when reproductionis largely dependent upon stored reserves.Several studies have shown that storedreserves canprovide a buffer that will maskeffects of reproduction on subsequentgrowth and reproduction in the short term(Westley, 1993; Cunningham, 1997; Geber etal., 1997).

Direct tests of the principles of life-history theory require that a genetic basisexists for the trade-offs between current andresidual reproduction found within individ-uals of a population. Few studies of plantsare available that consider genetic correla-tions among life-history traits in iteroparousspecies. Those studies that are available sug-gest that, as in the case of phenotypic corre-lations, the evidence for a negativecorrelation between current and residualreproduction is variable. Law (1979) studiedvariation in fecundity in two seasons amonggenetic families from two populations of Poaannua, in which most individuals actuallyreproduce over 2 years. Using a populationmodel to evaluate the contribution of eachreproductive episode, he found a strong neg-ative genetic correlation between reproduc-tive value in successive years. Genetictrade-offs between current reproduction andfuture reproduction have also been found inGladiolus(Rameau and Gouyon, 1991) andPseudotsuga menziesii(El-Kassaby andBarclay, 1992). However, there are also stud-ies that have failed to find any evidence of agenetic trade-off. Galen (1993) examined theeffect of early reproduction on subsequentcc

growth and survival in the alpine perennialPolemonium viscosumJ

. Measurements weremade on 34 maternal half-sib families over aperiod of 4 years. Approximately half (62%)of the families did not exhibit early flower-ing, but there was little evidence of a genetictrade-off between early reproduction andsubsequent survival. ackson and Dewald(1994) examined the effect of a mutation thatincreased reproductive allocation in theperennial grass Tripsaum datyloides.Although the mutation substantiallyincreased RA (plants produce only pistillateflowers), there were no effects on vegetativegrowth. As with phenotypic trade-offs, thereis evidence that genetic trade-offs vary withenvironment. Reekie (1998b) grew 15 mater-nal half-sib families of Plantago majorinmown versus unmown grassland sites.Variation in RA among families had no effecton growth the following year in the mownsites, but there was a negative correlationbetween RA and growth in the unmownsites.

It must be remembered that failure todetect a significant effect in an experimentdoes not necessarily mean that there is noeffect; it only means that the effect, if itexists, is too small to be detected, given thelevel of variance within the study.Therefore, the failure of a number of studiesto detect significant trade-offs between cur-rent reproduction and residual reproduc-tive value does not mean that there is notrade-off. However, the wide variation,both among and within studies, in whethera trade-off is detected does indicate that

22F.A. Bazzaz et al.the trade-off is highly variable and dependsupon environment. Variation in the magni-tude of the trade-off is likely to select fordifferent patterns of RA. For example, inthe case of P. major, the lower genetic trade-off in mown sites as compared withunmown sites (Reekie, 1998b) may explainwhy genotypes isolated from mown siteshave a higher RA than genotypes fromunmown sites (Warwick and Briggs, 1980).Variation in the phenotypic trade-offbetween current and residual reproduction,on the other hand, will select for pheno-typic plasticity in RA. Individuals shouldincrease RA in those environments wherethe phenotypic trade-off is minimal andreduce RA in those environments where thetrade-off function is steeper. To date, therehas been little attempt to test this predictionin the field. However, several studies dopresent evidence that suggests the timing ofreproduction in some species coincideswith environmental conditions that mini-mize the trade-off (Brewer, 1995; Reekie etal., 1997; Shitaka and Hirose, 1998).

Evolutionary models incorporate thenegative correlations between life-historycomponents as physiological constraintsunderlying life-history evolution (e.g.Gadgil and Bossert, 1970; Schaffer, 1974). Itmust be remembered, however, that thegenetic variation observed in natural popu-lations may arise from the selectiveneutrality of genotypes resulting from neg-atively correlated life-history components(Falconer, 1952; Lande, 1982) or from theinteraction of phenotypic plasticity andenvironmental heterogeneity (Bazzaz andSultan, 1987; Sultan, 1987). The variationrevealed by quantitative genetic analysiscan only predict evolutionary trajectoriesin the immediate future. The most interest-ing evolutionary change may result notfrom evolution within these physiologicalconstraints, but from the evolution of theconstraints themselves.

Summary

The theory of resource allocation has beenwidely used to examine the physiological

basis for variation in RO and the evolu-tion of life-history strategies in plants.Application of this theory depends on theassumption that the resource supply isfixed and that allocation of resources tovarious plant structures is equivalent toallocation to corresponding functions.Neither of these assumptions is valid formany (or possibly any) plants, and vari-ous physiological processes (such as pho-tosynthesis of reproductive parts andinternal reallocation of resources) decou-ple structure and function. As a result,measures of RA are not adequate forassessing RE, and these two conceptsmust be distinguished. In an ecologicalcontext, RO can be considered as theproduct of plant size and RA. The relativeimportance of these two terms in deter-mining RO appears to vary with speciesand environmental conditions; plant sizeis more important in annual than peren-nial species, particularly when there isasymmetric competition for resources,resulting in marked size hierarchies. Inperennial species, RA exhibits a great dealof phenotypic plasticity as well as genetic differentiation among popula-tions. Studies that examine the evolutionof different patterns of RA have focusedon genetic differentiation among popula-tions and have largely ignored the impor-tance of phenotypic plasticity. In thestudy of life-history evolution, particu-larly for iteroparous species, RE needs tobe related to demographic parameters,such as changes in future survival andRO. Studies where this has been done sug-gest that both the phenotypic and genetictrade-off between current reproductionand future performance is highly variable.Models of life-history evolution have usu-ally considered trade-offs between fitnesscomponents as consequences of physio-logical constraints, but little is knownabout the evolution of the constraintsthemselves and how these constraintsmay be modified by the environment. Toexplain both phenotypic plasticity in RAand genetic differentiation in RA amongpopulations, it will be necessary toaddress this gap in our knowledge.

Reproductive Allocation in Plants23References

Aarssen, L.W. and Taylor, D.R. (1992) Fecundity allocation in herbaceous plants. Oikos65, 225–232.Abrahamson, W.G. (1979) Patterns of resource allocation in wildflower populations of fields and

woods.American Journal of Botany66, 71–79.

Abrahamson, W.G. and Caswell, H. (1982) On the comparative allocation of biomass, energy, and

nutrients in plants. Ecology63, 982–991.

Abrahamson, W.G. and Gadgil, M. (1973) Growth form and reproductive effort in Goldenrods

(Solidago, Compositae). American Naturalist107, 651–661.

Abul-Fatih, H.A., Bazzaz, F.A. and Hunt, R. (1979) The biology of Ambrosia trifidaL. III. Growth and

biomass allocation. New Phytologist83, 829–838.

Ackerly, D.D. and Jasienski, M. (1990) Size-dependent variation of gender in high density stands of

the monoecious annual, Ambrosia artemisiifolia(Asteraceae). Oecologia82, 474–477.

Ackerly, D.D., Coleman, J.S., Morse, S.R. and Bazzaz, F.A. (1992) CO2and temperature effects on leaf

area production in two annual plant species. Ecology73, 1260–1269.

Agren, J. and Willson, M.F. (1994) Cost of seed production in the perennial herbs Geranium macula-tumand G. sylvaticum: an experimental field study. Oikos70, 35–42.

Alvarez-Buylla, E.R. and Martínez-Ramos, M. (1992) Demography and allometry of Cecropia obtusi-folia, a neotropical pioneer tree – an evaluation of the climax–pioneer paradigm for tropicalrainforests. Journal of Ecology80, 275–290.

Antonovics, J. (1980) Concepts of resource allocation and partitioning in plants. In: Staddon, J.E.R.

(ed.) Limits to Action: the Allocation of Individual Behavior. Academic Press, New York,pp.1–25.

Antos, J.A. and Allen, G.A. (1994) Biomass allocation among reproductive structures in the dioecious

shrub Oemleria cerasiformis: a functional interpretation. Journal of Ecology82, 21–29.

Armstrong, D.P. and Westoby, M. (1993) Seedlings from large seeds tolerate defoliation better: a test

using phylogenetically independent contrasts. Ecology74, 1092–1100.

Ashman, T.L. (1994) Reproductive allocation in hermaphrodite and female plants of Sidalcea ore-ganassp. spicata(Malvaceae) using four currencies. American Journal of Botany81, 21–29.

Baker, H.G. (1972) Seed weight in relation to environmental conditions in California. Ecology 53,

997–1010.

Bazzaz, F.A. (1993). Use of plant growth analysis in global change studies: modules, individuals and

populations. In: Schulze, E.D. and Mooney, H.A. (eds) Design and Execution of Experiments onCO2Enrichment. Commission of the European Communities, Luxembourg, pp.53–71.

Bazzaz, F.A. and Carlson, R.W. (1979) Photosynthetic contribution of flowers and seeds to reproduc-tive effort of an annual colonizer. New Phytologist82, 223–232.

Bazzaz, F.A. and Harper, J.L. (1977) Demographic analysis of the growth of Linum usitatissimum.

New Phytologist78, 193–208.

Bazzaz, F.A. and Reekie, E.G. (1985) The meaning and measurement of reproductive effort in plants.

In: White, J. (ed.) Studies on Plant Demography: a Festschrift for John L. Harper. AcademicPress, London, pp.373–387.

Bazzaz, F.A. and Sultan, S. (1987) Ecological variation and the maintenance of plant diversity. In:

Urbanska, K. (ed.) Differentiation in Higher Plants. Academic Press, London, pp.69–93.

Bazzaz, F.A., Carlson, R.W. and Harper, J.L. (1979) Contribution to reproductive effort by photosyn-thesis of flowers and fruits. Nature279, 554–555.

Bazzaz, F.A., Chiariello, N.R., Coley, P.D. and Pitelka, L.F. (1987) Allocation resources to reproduc-tion and defense. Bioscience37, 58–67.

Bazzaz, F.A., Garbutt, K., Reekie, E.G. and Williams, W.E. (1989) Using growth analysis to interpret

competition between a C3and a C4annual under ambient and elevated CO2. Oecologia79,223–235.

Benech Arnold, R.L., Fenner, M. and Edwards, P.J. (1992) Mineral allocation to reproduction in

Sorgum bicolorand Sorgum halpensein relation to parental nutrient supply. Oecologia92,138–144.

Benner, B.L. and Bazzaz, F.A. (1988) Carbon and mineral element accumulation and allocation in

two annual plant species in response to timing of nutrient addition. Journal of Ecology76, 19–40.Bloom, A.J., Chapin, F.S., III and Mooney, H.A. (1985) Resource limitation in plants – an economic

analogy. Annual Review of Ecology and Systematics16, 363–392.

Brewer, J.S. (1995) The relationship between soil fertility and fire-stimulated floral induction in two

populations of grass-leaved golden aster, Pityopsis graminifolia.Oikos74, 45–54.

Calvo, R.N. and Horvitz, C.C. (1990) Pollinator limitation, cost of reproduction, and fitness in plants:

a transition matrix demographic approach. American Naturalist136, 499–516.

Chapin, F.S., III (1989) The cost of tundra plant structures: evaluation of concepts and currencies.

American Naturalist133, 1–19.

24F.A. Bazzaz et al.Chapin, F.S., III, Bloom, A.J., Field, C.B. and Waring, R.H. (1987) Plant responses to multiple envi-ronmental factors. Bioscience37, 49–57.

Cheplick, G.P. (1989) Nutrient availability, dimorphic seed production, and reproductive allocation

in the annual grass Amphicarpum purshii. Canadian Journal of Botany67, 2514–2521.

Chiariello, N. and Roughgarden, J. (1984) Storage allocation in seasonal races of an annual plant:

optimal versus actual allocation. Ecology65, 1290–1301.

Choe, H.S., Chu, C., Koch, G., Gorham, J. and Mooney, H.A. (1988) Seed weight and seed resources in

relation to plant growth rate. Oecologia76, 158–159.

Cipollini, M.L. and Whigham, D.F. (1994) Sexual dimorphism and cost of reproduction in the

dioecious shrub Lindera benzoin(Lauraceae). American Journal of Botany81, 65–75.

Clauss, M.J. and Aarssen, L.W. (1994) Patterns of reproductive effort inArabidopsis thaliana: con-founding effects of size and developmental stage.Ecoscience1, 153–159.Cody, M.L. (1966) A general theory of clutch size. Evolution20, 174–184.

Cohen, D. (1971) Maximizing final yield when growth is limited by time or by limiting resources.

Journal of Theoretical Biology33, 299–307.

Cohen, D. (1976) The optimal timing of reproduction. American Naturalist110, 801–807.

Cole, L.C. (1954) The population consequences of life history phenomena. Quarterly Review of

Biology29, 103–137.

Cruden, R.W. and Lyon, D.L. (1985) Patterns of biomass allocation to male and female functions in

plants with different mating systems. Oecologia66, 299–306.

Cunningham, S.A. (1997) The effect of light environment, leaf area, and stored carbohydrates on

inflorescence production by a rainforest understory palm. Oecologia111, 36–44.

Dale, M.P. and Causton, D.R. (1992) The ecophysiology of Veronica chamaedrys, Veronica montana

and Veronica officinalis. III. Effects of shading on the phenology of biomass allocations: a fieldexperiment. Journal of Ecology 80, 505–515.

DeJong, T.J. (1986) Effects of reproductive and vegetative sink activity on leaf conductance and water

potential in Prunus persicacultivar Fantasia. Scientific Horticulture29, 131–138.

DeJong, T.J. and Klinkhamer, P.G.L. (1989) Size-dependency of sex-allocation in hermaphoditic,

monocarpic plants. Functional Ecology3, 201–206.

deRidder, F. and Dhondt, A.A. (1992) The reproductive behaviour of a clonal herbaceous plant, the

long-leaved sundew Drosera intermedia, in different heathland habitats. Ecography15,144–153.

El-Kassaby, Y.A. and Barclay, H.J. (1992) Cost of reproduction in Douglas-fir. Canadian Journal of

Botany70, 1429–1432.

Evenson, W.E. (1983) Experimental studies of reproductive energy allocation in plants. In: Jones, C.E.

and Little, R.J. (eds) Handbook of Experimental Pollination Biology.Van Nostrand Reinhold,New York, pp. 249–274.

Falconer, D.S. (1952) The problem of environment and selection. American Naturalist86, 293–298.Farris, M.A. and Lechowicz, M.J. (1990) Functional interactions among traits that determine repro-ductive success in a native annual plant. Ecology71, 548–557.

Fenner, M. (1983) Relationships between seed weight, ash content and seedling growth in twenty-four species of Compositae. New Phytologist95, 697–706.

Fenner, M. (1985a) The allocation of minerals to seeds in Senecio vulgarisplants subjected to nutri-ent shortage. Journal of Ecology74, 385–392.

Fenner, M. (1985b) A bioassay to determine the limiting minerals for seeds from nutrient-deprived

Senecio vulgarisplants. Journal of Ecology74, 497–506.Fenner, M. (1985c) Seed Ecology. Chapman and Hall, London.

Forrest, G.I. (1971) Structure and production of North Pennine blanket bog vegetation.Journal of

Ecology59, 453–479.

Foster, S.A. (1986) On the adaptive value of large seeds for tropical moist forest trees, a review and

synthesis. Botanical Review53, 260–299.

Fox, G.A. (1992) Annual plant life histories and the paradigm of resource allocation. Evolutionary

Ecology6, 482–499.

Gadgil, M. and Bossert, W.H. (1970) Life historical consequences of natural selection.American

Naturalist104, 1–24.

Gadgil, M. and Solbrig, O.T. (1972) The concept of r- and K-selection: evidence from wild flowers

and some theoretical considerations. American Naturalist106, 14–31.

Galen, C. (1993) Cost of reproduction in Polemonium viscosum: phenotypic and genetic approaches.

Evolution47, 1073–1079.

Galen, C., Dawson, T.E. and Stanton, M.L. (1993) Carpels as leaves: meeting the carbon cost of repro-duction in an alpine buttercup. Oecologia95, 187–193.

Garbutt, K. and Bazzaz, F.A. (1984) The effects of elevated CO2on plants. III. Flower, fruit and seed

production and abortion. New Phytologist98, 433–446.

Reproductive Allocation in Plants25Garbutt, K., Williams, W.E. and Bazzaz, F.A. (1990) Analysis of differential response of five annuals

to elevated CO2during growth. Ecology71, 1185–1194.

Geber, M.A. (1990) The cost of meristem limitation in Polygonum arenastrum: negative genetic corre-lations between fecundity and growth. Evolution44, 799–819.

Geber, M.A., De Kroon, H. and Watson, M.A. (1997) Organ preformation in mayapple as a mechanism

for historical effects on demography. Journal of Ecology85, 211–223.

Gehring, J.L. (1993) Temporal patterns in sexual dimorphisms in Silene latifolia(Caryophylaceae).

Bulletin of the Torrey Botanical Club120, 405–416.

Gehring, J.L. and Monson, R.K. (1994) Sexual differences in gas exchange and response to environ-mental stress in dioecious Silene latifolia(Caryophyllaceae). American Journal of Botany81,166–174.

Gleeson, S.K. and Tilman, D. (1990) Allocation and the transient dynamics of succession on poor

soils. Ecology71, 1144–1155.

Goldman, D.A. and Willson, M.F. (1986) Sex allocation in functionally hermaphroditic plants: a

review and critique. Botanical Review52, 157–194.

Grime, J.P. and Jeffrey, D.W. (1965) Seedling establishment in vertical gradients of sunlight. Journal

of Ecology53, 621–642.

Gross, K. (1984) Effects of seed size and growth form on seedling establishment of six monocarpic

perennial plants. Journal of Ecology72, 369–387.

Grubb, P.J. (1977) The maintenance of species richness in plant communities: the importance of the

regeneration niche. Biological Reviews 52, 107–145.

Grubb, P.J. and Metcalfe, D.J. (1996) Adaptation and inertia in the Australian tropical lowland rain-forest flora: contradictory trends in intergeneric and intrageneric comparisons of seed size inrelation to light demand. Functional Ecology 10, 512–520.

Hancock, J.F. and Pritts, M.P. (1987) Does reproductive effort vary across different life forms and seral

environments? A review of the literature. Bulletin of the Torrey Botanical Club114, 53–59.

Hara, T., Kawano, S. and Nagai, Y. (1988) Optimal reproductive strategy of plants, with special refer-ence to the modes of reproductive resource allocation. Plant Species Biology3, 43–59.Harper, J.L. (1967) A Darwinian approach to plant ecology. Journal of Ecology55, 247–270. Harper, J.L. (1977) The Population Biology of Plants. Academic Press, London.

Harper, J.L. and Ogden, J. (1970) The reproductive strategy of higher plants. I. The concept of strategy

with special reference to Senecio vulgarisL. Journal of Ecology58, 681–698.

Harper, J.L. and Sellek, C. (1987) The effects of severe mineral nutrient deficiencies on the demogra-phy of leaves. Proceedings of the Royal Society of London, Series B232, 137–157.

Harper, J.L., Lovell, P. and Moore, K. (1970) The shapes and sizes of seeds. Annual Review of Ecology

and Systematics1, 327–356.

Hartnett, D.C. (1991) Effects of fire in tallgrass prairie on growth and reproduction of prairie cone-flower (Ratibida columnifera: Asteraceae). American Journal of Botany78, 429–435.

Hartnett, D., Hartnett, B. and Bazzaz, F.A. (1987) Persistence of Ambrosia trifidapopulations in old

fields and responses to successional changes. American Journal of Botany74, 1239–1248.

Hemborg, A.M. and Karlsson, P.S. (1998) Somatic costs of reproduction in eight subarctic plant

species. Oikos82, 149–157.

Hickman, J.C. (1975). Environmental unpredictability and plastic energy allocation strategies in the

annual Polygonum cascadense. Journal of Ecology63, 689–701.

Horvitz, C.C. and Schemske, D.W. (1988) Demographic cost of reproduction in a neotropical herb, an

experimental approach. Ecology69, 1741–1745.

Jackson, L.L. and Dewald, C.L. (1994) Predicting evolutionary consequences of greater reproductive

effort in Tripsacum dactyloidesa perennial grass. Ecology75, 627–641.

Jennersten, O. (1991) Cost of reproduction in Viscaria vulgaris(Caryophyllaceae): a field experiment.

Oikos61, 197–204.

Jurik, T.W. (1985) Differential costs of sexual and vegetative reproduction in wild strawberry popula-tions. Oecologia66, 394–403.

Karlsson, P.S., Svensson, B.M., Carlsson, B.A. and Nordell, K.O. (1990) Resource investment in

reproduction and its consequences in three Pinguiculaspecies. Oikos59, 393–398.

Karlsson, P.S., Thoren, L.M. and Haslin, H.M. (1994) Prey capture by threePinguiculaspecies in a

subarctic environment. Oecologia99, 188–193.

King, D. and Roughgarden, J. (1982a) Multiple switches between vegetative and reproductive growth

in annual plants. Theoretical Population Biology21, 194–204.

King, D. and Roughgarden, J. (1982b) Graded allocation between vegetative and reproductive growth

for annual plants in growing season of random length.Theoretical Population Biology22, 1–16.

Kira, T. (1978) Community architecture and organic matter dynamics in tropical lowland rainforests

of Southeast Asia with special reference to Pasoh Forest, West Malaysia. In: Tomlinson, P.B. and

26F.A. Bazzaz et al.Zimmermann, M.H. (eds) Tropical Trees as Living Systems. Cambridge University Press,Cambridge, pp.561–590.

Klinkhamer, P.G.L., de Jong, T.J. and Meelis, E. (1990) How to test for proportionality in the repro-ductive effort of plants. American Naturalist 135, 291–300.

Klinkhamer, P.G.L., Meelis, E., de Jong, T.J. and Weiner, J. (1992) On the analysis of size-dependent

reproductive output in plants. Functional Ecology6, 308–316.

Korpelainen, H. (1992) Patterns of resource allocation in male and female plants ofRumex acetosa

and Rumex acetosella. Oecologia89, 133–139.

Kozlowski, J. (1992) Optimal allocation of resources to growth and reproduction: implications for age

and size at maturity. Trends in Ecology and Evolution7, 15–19.

Kozlowski, J. and Wiegert, R.G. (1986) Optimal allocation of energy to growth and reproduction.

Theoretical Population Biology29, 16–37.

Kozlowski, T.T. (1971) Growth and Development of Trees, Vol. 2. Academic Press, New York.Lande, R. (1982) A quantitative theory of life history evolution. Ecology63, 607–615.

Laporte, M.M. and Delph, L.F. (1996) Sex-specific physiology and source–sink relations in the

dioecious plant Silene latifolia. Oecologia106, 63–72.

Law, R. (1979) The cost of reproduction in annual meadow grass. American Naturalist 113, 3–16.Lawrence, W.S. (1993) Resource and pollen limitation: plant size-dependent reproductive patterns in

Physalis longifolia. American Naturalist141, 296–313.

Lee, T.D. (1988) Patterns of fruit and seed production. In: Lovett Doust, J. and Jovett Doust, L. (eds)

Plant Reproductive Ecology: Patterns and Strategies. Oxford University Press, New York,pp.179–202.

Lehtila, K., Tumori, J. and Sulkinoja, M. (1994) Bud demography of the mountain birch Betula pubes-censssp. tortuosanear the tree line. Ecology75, 945–955.

Leishman, M.R. and Westoby, M. (1994) The role of large seed size in shaded conditions: experimen-tal evidence. Functional Ecology8, 205–214.

Lloyd, D.G. (1988) Benefits and costs of biparental and uniparental reproduction in plants. In:

Michod, R.E. and Levin, B.R. (eds) The Evolution of Sex. Sinauer Associates, Sunderland,Massachusetts, pp.233–252.

McConnaughay, K.D.M. and Bazzaz, F.A. (1987) The relationship between gap size and performance

of several colonizing annuals. Ecology 68, 411–416.

McKone, M.. and Tonkyn, D.W. (1986) Intrapopulation gender variation in common ragweed

(Asteraceae, Ambrosia artemisiifoliaL.), a monoecious, annual herb.Oecologia70, 63–67.

Manasse, R.S. (1990) Seed size and habitat effects on the water-dispersed perennial, Crinum

erubescens(Amaryllidaceae). American Journal of Botany77, 1336–1342.

Marshall, J.D., Dawson, T.E. and Ehleringer, J.R. (1993). Gender-related differences in gas exchange

are not related to host quality in xylem-tapping mistletoe, Phoradendron juniperinum(Viscaceae). American Journal of Botany80, 641–645.

Martínez-Ramos, M., Sarukhan, J. and Piñero, D. (1988) The demography of tropical trees in the con-text of forest gap dynamics: the case of Astrocaryum mexicanumat Los Tuxtlas tropical rainforest.In: Davy, A.J., Hutchings, M.J. and Watkinson, A.R. (eds) Plant Population Ecology. BlackwellScientific Publications, Oxford, pp.315–342.

Mitchell-Olds, T. and Bergelson, J. (1990) Statistical genetics of an annual plant,Impatiens capensis.

II. Natural selection. Genetics124, 417–421.

Neales, T.F. and Incoll, L.O. (1968) The control of leaf photosynthesis rate by the level of assimilate

concentration in the leaf: a review of the hypothesis. Annals of Botany30, 349–363.

Newell, E.A. (1991) Direct and delayed costs of reproduction in Aesculus californica. Journal of

Ecology 79, 365–378.

Nobel, P.S. (1977) Water relations of flowering of Agave deserti. Botanical Gazette138, 1–6.

Ogden, J. (1974) The reproductive strategy of higher plants. II. The reproductive strategy of Tussilago

farfaraL. Journal of Ecology 62, 291–324.

Ostertag, R. and Menges, E.S. (1994) Patterns of reproductive effort with time since last fire in Florida

scrub plants. Journal of Vegetation Science5, 303–310.

Oyama, K. and Dirzo, R. (1988) Biomass allocation in the dioecious tropical palm Chamaedorea

tepejiloteand its life history consequences. Plant Species Biology3, 27–33.

Pfister, C.A. (1992) Costs of reproduction in an intertidal kelp: patterns of allocation and life history

consequences. Ecology73, 1586–1596.

Pickering, C.M. (1994) Size-dependent reproduction in Australian alpine Ranunculus. Australian

Journal of Ecology19, 336–344.

Piñero, D., Sarukhan, J. and Alberdi, P. (1982) The costs of reproduction in a tropical palm,

Astrocaryum mexicanum. Journal of Ecology 70, 473–481.

Pitelka, L.F. (1977) Energy allocation in annual and perennial lupines (Lupinus, Leguminosae).

Ecology58, 1055–1065.

Reproductive Allocation in Plants27Pitelka, L.F. and Ashmun, J. (1985) Physiology and integration of ramets in clonal plants. In: Jackson,

J.B.C., Buss, L.W. and Cook, R.E. (eds) Population Biology and Evolution of Clonal Organisms.Yale University Press, New Haven, pp.399–436.

Primack, R.B. and Antonovics, . (1982) Experimental ecological genetics in Plantago. VII.

Reproductive effort in populations of P. lanceolataL. Evolution36, 742–752.

Primack, R.B. and Hall, P. (1990) Costs of reproduction in pink lady’s slipper orchid: a four year

experimental study. American Naturalist136, 638–656.

Primack, R.B., Miao, S.L. and Becker, K.R. (1994) Costs of reproduction in the pink lady’s slipper

orchid (Cypripedium acaule): defoliation, increased fruit production and fire. American Journalof Botany81, 1083–1090.

Ramadan, A.A., El-Keblawy, A., Shaltout, K.H. and Lovett-Doust, J. (1994) Sexual polymorphism,

growth, and reproductive effort in Egyptian Thymelaea hirsuta(Thymelaeceae). AmericanJournal of Botany81, 847–857.

Rameau, C. and Gouyon, P.H. (1991) Resource allocation to growth, reproduction and survival in

Gladiolus: the cost of male function. Journal of Evolutionary Biology4, 291–308.

Ramsey, M. (1997) No evidence of demographic costs of seed production in the pollen-limited peren-nial herb Blandfordia grandiflora(Liliaceae). International Journal of Plant Sciences158,785–793.

Reekie, E.G. (1991) Cost of seed versus rhizome poduction in Agropyron repens.Canadian Journal of

Botany69, 2678–2683.

Reekie, E.G. (1998a) An explanation for size-dependent reproductive allocation inPlantago major.

Canadian Journal of Botany76, 43–50.

Reekie, E.G. (1998b) An experimental field study of the cost of reproduction in Plantago majorL.

Ecoscience5, 200–206.

Reekie, E.G. (1999) Resource allocation, trade-offs, and reproductive effort in plants. In: Vuorisalo,

T.O. and Mutikainen, P.K. (eds) Life History Evolution in Plants. Kluwer Academic Publishers,The Netherlands, pp.173–194.

Reekie, E.G. and Bazzaz, F.A. (1987a) Reproductive effort in plants. 1. Carbon allocation to reproduc-tion. American Naturalist129, 876–896.

Reekie, E.G. and Bazzaz, F.A. (1987b) Reproductive effort in plants. 2. Does carbon reflect the alloca-tion of other resources? American Naturalist129, 897–906.

Reekie, E.G. and Bazzaz, F.A. (1987c) Reproductive effort in plants. 3. Effect of reproduction on vege-tative activity. American Naturalist129, 907–919.

Reekie, E.G. and Bazzaz, F.A. (1992) Cost of reproduction in genotypes of two congeneric plant

species with contrasting life histories. Oecologia90, 21–26.

JReekie, E.G. and Reekie, J.Y.C. (1991) An experimental investigation of the effect of reproduction on

canopy structure, allocation and growth in Oenothera biennis.Journal of Ecology79,1061–1071.

Reekie, E., Parmiter, D., Zebian, K. and Reekie, J. (1997) Trade-offs between reproduction and growth

influence time of reproduction in Oenothera biennis.Canadian Journal of Botany75,1897–1902.

Reznick, D. (1985) Costs of reproduction, an evaluation of the empirical evidence. Oikos 44,

257–267.

Saikkonen, K., Koivunen, S., Vuorisalo, T. and Mutikainen, P. (1998) Interactive effects of pollination

and heavy metals on resource allocation in Potentilla anserinaL.Ecology79, 1620–1629.Salisbury, E.J. (1942) The Reproductive Capacity of Plants. Bell, London.

Samson, D.A. and Werk, K.S. (1986) Size-dependent effects in the analysis of reproductive effort in

plants. American Naturalist 124, 667–680.

Schaffer, W.M. (1974) Optimal reproductive effort in fluctuating environments. American Naturalist

108, 783–790.

Schaffer, W.M. (1977) Some observations on the evolution of reproductive rate and competitive abil-ity in flowering plants. Theoretical Population Biology11, 90–104.

Schaffer, W.M., Inouye, R.S. and Whittam, T.S. (1982) Energy allocation by an annual plant when the

effects of seasonality on growth and reproduction are decoupled.American Naturalist120,787–815.

Schat, H., Ouborg, J. and DeWit, R. (1989) Life history and plant architecture: size-dependent repro-ductive allocation in annual and biennial Centauriumspecies. Acta Botanica Neerlandica 38,183–201.

Scheiner, S.M. (1989) Variable selection along a successional gradient. Evolution43, 548–562.

Schmid, B. and Weiner, J. (1993) Plastic relationships between reproductive and vegetative mass in

Solidago altissima. Evolution47, 61–74.

Schwaegerle, K.E. and Levin, D.A. (1990) Environmental effects on growth and fruit production in

Phlox drummondii. Journal of Ecology78, 15–26.

28F.A. Bazzaz et al.Shipley, B. and Dion, J. (1992) The allometry of seed production in herbaceous angiosperms.

American Naturalist139, 467–483.

Shipley, B. and Peters, R.H. (1990) The allometry of seed weight and seedling relative growth rate.

Functional Ecology 4, 523–529.

Shitaka, Y. and Hirose, T. (1998) Effects of shift in flowering time on the reproductive output of

Xanthium canadensein a seasonal environment. Oecologia114, 361–367.

Silvertown, . and Dodd, M. (1996) Comparing plants and connecting traits.Philosophical

Transactions of the Royal Society of London B – Biological Sciences351, 1233–1239.

Silvertown, J.W. and Lovett Doust, J. (1993) Introduction to Plant Population Biology. Blackwell

Scientific Publications, Oxford.

Solbrig, O.T. and Solbrig, D.J. (1984) Size inequalities and fitness in plant populations.Oxford

Survey in Evolutionary Biology1, 141–159.

Solomon, B.P. (1989) Size-dependent sex ratios in the monoecious, wind-pollinated annual,

Xanthium strumarium. American Midland Naturalist 121, 209–218.

Stanton, M. (1985) Seed size and emergence time within a stand of wild radish (Raphanus

raphanistrumL.): the establishment of a fitness hierarchy. Oecologia67, 524–531.

Sugiyama, S. and Bazzaz, F.A. (1997) Plasticity of seed output in response to soil nutrients and den-sity in Abutilon theophrasti: implications for maintenance of genetic variation. Oecologia112,35–41.

Sultan, S. (1987) Evolutionary implications of phenotypic plasticity in plants.Evolutionary Biology

21, 127–178.

Sultan, S. and Bazzaz, F.A. (1993a) Phenotypic plasticity in Polygonum persicaria. I. Diversity and

uniformity in genotypic norms of reaction to light intensity. Evolution47, 1009–1031.

Sultan, S. and Bazzaz, F.A. (1993b) Phenotypic plasticity in Polygonum persicaria. II. Norms of reac-tion to soil moisture and the maintenance of genetic diversity.Evolution47, 1032–1049.

Sultan, S. and Bazzaz, F.A. (1993c) Phenotypic plasticity in Polygonum persicaria. III. The evolution

of ecological breadth for nutrient environment. Evolution47, 1050–1071.

Syrjanen, K. and Lehtila, K. (1993) The cost of reproduction in Primula veris: differences between

two adjacent populations. Oikos67, 465–472.

Thompson, J.N. (1984) Variation among individual seed masses in Lomatium grayi(Umbelliferae)

under controlled conditions: magnitude and partitioning of variance. Ecology65, 626–631.

Thompson, K. and Stewart, A.J.A. (1981) The measurement and meaning of reproductive effort in

plants. American Naturalist117, 205–211.

JThoren, L.M., Karlsson, P.S. and Tuomi, J. (1996) Somatic cost of reproduction in three carnivorous

Pinguiculaspecies. Oikos76, 427–434.

Tuomi, J., Hakala, T. and Haukioja, E. (1983) Alternative concepts of reproductive effort, costs of

reproduction, and selection in life-history evolution. American Zoologist 23, 25–34.

Waller, D.M. (1985) The genesis of size hierarchies in seedling populations of Impatiens capensis.

New Phytologist100, 243–260.

Wankhar, B. and Tripathi, R.S. (1990) Growth and reproductive allocation of Centella asiaticaraised

from stem cuttings of different sizes in relation to light regimes, soil texture and soil moisture.Acta Oeclogica11, 683–692.

Warwick, S.I. and Briggs, D. (1980) The genecology of lawn weeds. V. The adaptive significance of

different growth habits in lawn and roadside populations of Plantago majorL. New Phytologist85, 289–300.

Watson, M.A. (1984) Developmental constraints, effect of population growth and patterns of resource

allocation in a clonal plant. American Naturalist123, 411–426.

Weiner, J. (1988) The influence of competition on plant reproduction. In: Lovett-Doust, J. and Lovett-Doust, L. (eds) Plant Reproductive Ecology. Oxford University Press, Oxford, pp.228–245.

Weiner, J. and Solbrig, O.T. (1984) The meaning and measurement of size hierarchies.Oecologia61,

334–336.

Welham, C.V.J. and Setter, R.A. (1998) Comparison of size-dependent reproductive effort in two dan-delion (Taraxacum officinale) populations. Canadian Journal of Botany76, 166–173.

Wells, R., Anderson, W.F. and Wynne, J.C. (1991) Peanut yield as a result of fifty years of breeding.

Agronomy Journal83, 957–961.

Westley, L.C. (1993) The effect of inflorescence bud removal on tuber production inHelianthus

tuberosusL. (Asteraceae). Ecology74, 2136–2144.

Whittaker, R.H. and Marks, P.L. (1975) Methods of assessing terrestrial productivity. In: Lieth, H. and

Whittaker, R.H. (eds) Primary Productivity of the Biosphere. Springer-Verlag, New York,pp.55–118.

Williams, G.C. (1966) Natural selection, the costs of reproduction, and a refinement of Lack’s princi-ple. American Naturalist100, 687–690.

Williams, M.. (1994) Reproductive resource allocation in rhizoma peanut. Crop Science34,

477–482.

Reproductive Allocation in Plants29Williams, R.B. and Bell, K.L. (1981) Nitrogen allocation in Mojave desert winter annuals. Oecologia

48, 145–150.

Willson, M.F. (1983) Plant Reproductive Ecology. Wiley, New York.

Wulff, R.D. (1986) Seed size variation in Desmodium paniculatum. I. Factors affecting seeds size.

Journal of Ecology74, 87–97.

Zimmerman, J.K. (1991) Ecological correlates of labile sex expression in the orchid Catasetum viridi-flavum. Ecology72, 597–608.