ecol 85 609.1477 1478 - Cornell University1480 NICO BLU¨THGEN AND KONRAD FIEDLER Ecology, Vol. 85,...

1477

SPECIAL FEATURE

Community and Evolutionary Ecology of Nectar1

Pollination is an integral part of reproduction of most flowering plants. Worldwide, 90% ofsome 250 000 angiosperm species rely on pollination for successful reproduction, and nectar isthe primary resource provided by plants to attract mutualist visitors. Despite the ubiquity ofnectar as an ecosystem resource, our understanding of the importance of nectar has primarilybeen confined to nectar as a food reward for pollinators. However, recent findings suggest thatnectar may mediate interactions with a wide variety of plant visitors other than pollinators. Nectaris frequently consumed by organisms ranging from insect predators and parasitoids to lizards,and can confer increased plant resistance or susceptibility to antagonistic plant visitors. Thecomposition of nectar may be important in structuring insect communities, multi-trophic inter-actions, and ultimately plant communities via direct and indirect interactions. The physiologyand costs of nectar production, and the heritability of nectar traits, may limit the extent to whichselection by pollinators or other agents can shape the evolution of nectar traits. On the otherhand, the ability of floral visitors to find nectar may limit the extent to which animals can specializeon this resource.

The focus of this Special Feature is to highlight the community and evolutionary ecology ofinteractions that revolve around nectar as a resource. The last in-depth review of nectar waspublished in 1983, (The Biology of Nectaries, B. Bentley and T. Elias, Editors. Columbia Uni-versity Press, New York, New York, USA). That volume summarized advances in the under-standing of both the physiological mechanisms of nectar secretion and the ecological role ofnectar in plant–animal interactions. In the subsequent two decades, great strides have been madein these topics, and summarizing this literature is beyond the scope of this feature. Our goal isto highlight the areas that represent the most exciting recent ecological and evolutionary advances,and to draw attention to questions that are still largely unexplored in an effort to inspire futureresearch.

Traditionally, nectar has been thought of as a relatively simple food resource; however, recentchemical advances combined with field ecological investigation have shown the importance ofnectar composition to community ecology. The feature opens with strong demonstrations of theimportance of nectar composition, with three papers focusing on insect community structure(Blüthgen and Fiedler), animal foraging and nectar scent (Raguso), and plant evolution in ageographic context (Rudgers and Gardener).

In the next section, we highlight the mechanistic role that nectar plays in nonpollinatinginteractions. A considerable fraction of floral visitors do not act as pollinators, but rather consumenectar without providing pollination service. Because nectar attracts both pollinators and nectarrobbers, plants face a dilemma or possible trade-off in defending against nonpollinating floralvisitors without deterring pollinators. Irwin, Adler, and Brody offer conceptual and empiricalinsight into how plants cope with attack by nectar robbers through nectar and floral traits. Withthe defensive function of nectar in mind, Wäckers and Bonifay continue this line of reasoningby demonstrating that foliar and bracteal nectar respond in different ways to herbivory, followingpredictions of the Optimal Defense Theory. Bracteal nectar may serve a role in pollination aswell as an indirect defensive function in attracting predators and parasitoids of herbivores. How-ever, the function of foliar nectar is more likely strictly defensive. On the flip side, Adler andBronstein show that floral nectar may increase plant susceptibility to herbivore attack, especiallyin cases where pollinators act as herbivores in different life history stages. Through these inter-

1 Reprints of this 57-page Special Feature are available for $8.50 each, either as pdf files or as hard copy.Prepayment is required. Order reprints from the Ecological Society of America, Attention: Reprint Department,1707 H Street, N.W., Suite 400, Washington, DC 20006.

Spe

cial

Feat

ure

actions, nonpollinating floral visitors may impose strong selection on plants that could reinforceor conflict with selection by pollinators.

Despite the almost dogmatic view that nectar traits have evolved in response to selection bypollinators and other floral visitors, surprisingly little is known about the magnitude of additivegenetic variation for nectar traits. The feature closes with a synopsis of our current understandingof the heritability of nectar traits and constraints on their response to selection, highlighting boththe methodological advances and pitfalls to measuring heritability of nectar (Mitchell). Under-standing the degree to which nectar traits are heritable is critical to argue that the ecologicaleffects of nectar on community-level interactions, described above, can have evolutionary con-sequences.

Although this work is focused on the community and evolutionary ecology of nectar, thediversity of approaches highlighted in this Special Feature can be applied profitably to the studyof many other phenotypic traits and interactions. For example, geographic approaches to the studyof coevolution (see Rudgers and Gardener) have been successfully applied to understand howchemical traits mediate predator–prey interactions. Similarly, nectar traits that mediate communitystructure of ants (see Blüthgen and Fiedler) are functionally similar to chemical traits in plantsthat affect the community structure of grazers in terrestrial and aquatic ecosystems. The diversityof approaches and systems addressed in this feature demonstrate the fundamental importance ofnectar in mediating a wide variety of multispecies interactions. We hope this feature will inspirea future generation of biologists to integrate nectar into the conceptual framework of communityand evolutionary ecology.

—R. E. IRWINGuest EditorUniversity of Georgia

—L. S. ADLERGuest EditorVirginia Tech University

—A. A. AGRAWALSpecial Features EditorUniversity of Toronto

Key words: community ecology; evolutionary ecology; herbivory; insect community structure; multi-species interactions; mutualism; nectar; nectar robbing; nectar scent; parasitoids; plant evolution; polli-nation.

q 2004 by the Ecological Society of America

1479

Spec

ialFeatu

re

Ecology, 85(6), 2004, pp. 1479–1485q 2004 by the Ecological Society of America

COMPETITION FOR COMPOSITION: LESSONS FROM NECTAR-FEEDINGANT COMMUNITIES

NICO BLÜTHGEN1 AND KONRAD FIEDLER

Department of Animal Ecology 1, University of Bayreuth, 95440 Bayreuth, Germany

Abstract. Studies of the relationship between the composition of nectar and its con-sumers often focus on single or very few species, thus ignoring dynamics in diverse as-semblages. Conversely, most documented patterns of nectarivore communities have notbeen linked to nectar quality measures. In a study of nectar-foraging ant communities inan Australian rain forest, we found that nectar source partitioning between consumers maybe driven by two factors: (1) variation in nectar composition preferences mediated by tasteand physiological requirements, and (2) severe asymmetrical competitive interactions withinthe community. Ant communities are strongly shaped by competitive hierarchies. Whenforaging for extrafloral and floral nectar sources, wound sap, and homopteran honeydew,competitively superior weaver-ants (Oecophylla smaragdina) showed a significant pref-erence for nectar composition, whereas most other common community members werenonselective. Nectars frequently used by O. smaragdina were characterized by similar aminoacid profiles and higher sugar and amino acid concentration. We hypothesize that, for nectar–consumer relationships, as for other interactions in complex communities, the interplaybetween species-specific physiological optima and context-dependent asymmetrical com-petition is essential to explain consumers’ preferences and the dynamics of the system.

Key words: amino acids; Anonychomyrma gilberti; carbohydrates; Formicidae; interspecificcompetition; nectar composition; nectar-feeding ants; niche partitioning; Oecophylla smaragdina;tropical rain forests.

INTRODUCTION

Plant nectar is an almost ubiquitous resource in ter-restrial ecosystems and is widely utilized by a rangeof animals. Studies on nectar-mediated interactions of-ten focus on the degree of resource partitioning andspecialization among nectarivores, which may haveprofound consequences for both animals and plants.Analyses of resource partitioning are commonly basedon observed encounters, e.g., for flower visitors (Waseret al. 1996, Ollerton and Cranmer 2002), whereas themechanistic chemical basis for nectar source selectionremains largely unknown. The quality and quantity ofnectar and pollen, although representing the main fooditems collected by flower visitors, appear generally un-derstudied in regard to niche partitioning of flower vis-itors.

Thirty years ago, Baker and Baker (1973) suggestedthat variation in nectar amino acids plays a key role inflower–pollinator interactions. Nectar amino acid con-tent has been correlated with ‘‘pollinator types’’ andtheir additional diets (Baker and Baker 1975, Kevanand Baker 1983, but see Gottsberger et al. 1984). How-ever, the underlying hypotheses are still largely un-

Manuscript received 23 June 2003; revised 15 October 2003;accepted 17 October 2003; final version received 3 November2003. Corresponding Editor (ad hoc): L. S. Adler. For reprintsof this Special Feature, see footnote 1, p. 1477.

1 Present address: Department of Animal Ecology and TropicalBiology, University of Würzburg, Am Hubland, 97074 Würzburg,Germany. E-mail: [email protected]

explored (Gardener and Gillman 2002). The lack ofinformation is even more apparent for the selectionamong, and composition of, extrafloral nectar and otherplant exudates. On the other hand, phagostimulatoryeffects of different nectar characteristics have been ex-amined in great detail for a growing number of nec-tarivore species (e.g., Inouye and Waller 1984, Mar-tı́nez del Rio 1990, Lanza 1991, Erhardt and Rusterholz1998, Wäckers 1999, Wada et al. 2001), but usuallyunder isolated and controlled situations rather than incomplex natural environments. The goal of this paperis to link nectarivore community patterns to feedingpreferences and competitive abilities of individualcomponent species.

Ants are a suitable model system because they arehighly abundant on nectar sources in tropical forests.Nectar is an important part of the diet for large pro-portions of tropical ant faunas (Blüthgen et al. 2003,Davidson et al. 2003). Moreover, ant communities aredetermined by pronounced competitive hierarchies, inwhich competitively superior and territorial species(dominants) can be distinguished from submissive ants.In many ant faunas around the world, behavioral dom-inance usually coincides with numerical dominanceand is typically restricted to certain ant genera withlarge colonies (Jackson 1984, Savolainen and Vepsä-läinen 1988, Majer 1993, Andersen 1995, Davidson1997, Dejean and Corbara 2003). The ant communityinvestigated in this study (including 43 species ob-served on nectar or honeydew sources) contained two

Spe

cial

Feat

ure

1480 NICO BLÜTHGEN AND KONRAD FIEDLER Ecology, Vol. 85, No. 6

TABLE 1. Stepwise forward multiple linear regression analysis (with ridge correction, k 50.1) modeling the explanatory power of four nectar characteristics to explain foraging in-tensity of Oecophylla smaragdina ants at natural food sources.

Factor b† T P VIF‡

1) Amino acid profiles (NMDS)2) Total amino acid concentration (g/L)3) Sucrose/(sucrose 1 hexose)4) Total sugar concentration (g/L)

0.310.27

0.17

1.61.4

(20.2)§1.0

0.110.17

0.32

1.92.1

(1.1)1.5

Note: Summary statistics of the whole regression model, with adjusted coefficient of multipledetermination, R 5 0.26, F3,28 5 4.6, P , 0.01.2a

† Standardized regression coefficient.‡ Variance inflation factors for predictor variables, where F . 1.0.§ Factor 3 was excluded from the whole model because F , 1.0.

species that fulfill the criteria of territorial dominants:Oecophylla smaragdina (Fabricius), subfamily For-micinae (see also Hölldobler 1983, Andersen 1995),and Anonychomyrma gilberti (Forel), Dolichoderinae.Colonies of both species maintain mutually exclusiveterritories and are highly aggressive against each other,whereas the remaining species have subdominant ornondominant status. The subordinate species regularlyco-occur within the dominant’s territories and may usethe same trails and nectar sources (Blüthgen et al.2004b). Among the two dominants, only O. smarag-dina can utilize most common trees for their silk-wovennests (Blüthgen and Fiedler 2002), whereas A. gilbertinest exclusively in trunks of Syzygium erythrocalyxtrees (Monteith 1986, Blüthgen et al. 2004b). Althoughboth species have large colonies that actively forageon several trees around their nest sites, dietary choicesof the latter species could be more restricted due totheir dependence on their host tree.

Two questions are specifically addressed in thisstudy: (1) Can foraging choices by ants among naturalnectar sources be predicted by nectar composition (sug-ars, amino acids)? (2) How do dominant ants, espe-cially O. smaragdina, differ from subordinate speciesin their resource selectivity?

MATERIAL AND METHODS

The study was carried out in a tropical lowland rainforest in northern Queensland, Australia (168079S,1458279E, 80 m a.s.l.). Foraging of ants for floral nectar(FN), extrafloral nectars (EFN), and homopteran hon-eydew was assessed between 1999 and 2002 during 19months of field surveys. Surveys included the forestcanopy (0.95 ha) accessed with the aid of a tower crane(Australian canopy crane project) and haphazardly se-lected plants in the understory of the crane site and insurrounding areas along paths or scattered throughoutthe forests (within ;5 km radius of the crane site). Inorder to ensure independence of association records,only replications of the same ant species occurring onthe same plant species were considered that were sep-arated by .8 m (closer plants were often visited byants of the same colony). Plant individuals were la-beled, and repeated observations of the same individual

were pooled. In total, 43 ant species were observed tocollect FN, EFN, or wound sap from 47 plant species(432 plant individuals), and honeydew from at least 12species of honeydew producers on 25 plant species (81plant individuals) (Blüthgen and Fiedler 2002, Blüth-gen and Reifenrath 2003, Blüthgen et al. 2004b).

From a subset of these sources used by ants, car-bohydrates and amino acids were analyzed using high-performance liquid chromatography (HPLC) on 135samples collected from 92 plant individuals (Blüthgenet al. 2004a). This subset included all common nectarsources attended by ants, excluding only some raresources (representing fewer than five individual plantswhere ant visits were observed) and/or those wherenectar production was insufficient to allow quantitativesampling after ant exclusion (among which only Ichn-ocarpus fructescens R.Br. represented more than fiveindividuals with ants observed). Analyzed sources in-cluded FNs from 10 plant species and EFNs from 16plant species (48% of the plant species attended by antsfor nectar, but this species pool represented 87% of therespective plant individuals). These FNs include threesources on which ants were present but did not showany nectar consumption, although nectars were acces-sible (Blüthgen et al. 2004a), whereas nectar from nar-row flower tubes inaccessible to most ants was ex-cluded from this analysis. Moreover, wound sap exu-dates from two plant species (representing the onlyones observed to be used by ants on more than a singleplant individual), and homopteran honeydew fromthree of the most abundant honeydew producers livingon four different plant species (four honeydew sources,representing 16% of the host plant species and 30% ofthe host plant individuals) were included. Mean nectarcomposition was calculated for each of these 32 sourc-es; plant species names, sample size, and compositionaldata are presented in Appendix A. These 32 sourcesare henceforth referred to as ‘‘nectar sources.’’

For the present paper, we related the mean compo-sition of the 32 sources to their overall use by ants. Weperformed multiple regression analyses in order tomodel the contribution of four selected nectar char-acteristics describing sugar and amino acid concentra-tion and composition (Table 1) to ant species foraging.

June 2004 1481COMMUNITY AND EVOLUTIONARY NECTAR ECOLOGY

Spec

ialFeatu

re

Qualitative similarities of amino acid profiles were ex-pressed as axis scores obtained by nonmetric multi-dimensional scaling (NMDS) of all pairwise Sörensensimilarity index values (Clarke 1993). Only a singledimension was used (stress: 0.155). This axis was sig-nificantly correlated with the mean number of aminoacids present and with concentrations of several singleor grouped amino acids, e.g., hydrophilous amino acids(Appendix B). The second axis in a two-dimensionalNMDS lacked any strong relationship with amino acidparameters (Appendix B) and was therefore excludedfrom further analyses. Note that our model did not in-clude total nectar production of each plant. An indirectestimate of nectar production, namely the median num-ber of ant workers simultaneously exploiting nectarsources on each plant individual (see Dreisig 1988),showed no significant positive effect on resource choic-es by any ant species (data are not shown). Data fortotal sugar and amino acid concentration were log-transformed, and sugar composition was arcsine-trans-formed. Significant linear correlations among trans-formed predictor variables occurred between theNMDS axis and the total amino acid concentration (r5 0.55, P , 0.001), and between the latter and totalsucrose concentration (r 5 0.37, P , 0.05), whereasthe remaining variables were not intercorrelated (r #0.18, p $ 0.32). Because of collinearity, ridge correc-tion was used with k 5 0.1 (Legendre and Legendre1998), but conclusions remained stable irrespective ofchoices of k between 0.0 (no correction) and 0.5. Astepwise forward procedure included variables of effectsize F . 1.0 in the model. For each ant species, wedefined a resource use intensity (i) for each nectarsource as the dependent variable with i 5 log(numberof individuals of a plant species where the source wasconsumed 1 1).

This index (i) ranged from 0 to 3.2 and was used toaccount for variable importance of different nectarsources to the nutrition of the respective ant species(data on the number of plant individuals attended byants are shown in Appendix A; no data were availableon the total abundance of these plants).

Separate regression analyses were performed for the16 most common ant species (all were found on $10plant individuals including $5 of the 32 nectarsources). Analyses were performed using the packageStatistica (StatSoft 1999).

RESULTS

Nectar preferences among ant species.—Oecophyllasmaragdina used 19 of the 32 nectar sources analyzed,the broadest spectrum and the greatest abundance ofnectar sources observed among all ants in this study.A combination of three nectar characteristics (eachwith F .1.0) explained a significant proportion of thevariance in resource selection by this ant species: ami-no acid composition, total amino acid concentration,and total sugar concentration (Table 1). O. smaragdina

tended to prefer sources with similar amino acid pro-files (NMDS; see also Appendix B), higher total aminoacid concentration, and higher sugar concentration, asshown by significant positive individual correlations.None of these predictors alone was significant in themultiple regression model, and they were partly inter-correlated (see Methods). The significant multiple re-gression remained unaffected when honeydew sourcesor when rare sources (fewer than five plant individualsattended by any ant) were excluded from the analysis(the effect of amino acid composition became signifi-cant after excluding honeydew).

Significant (P , 0.05) or marginally significant (0.05, P , 0.06) multiple regression models were foundin only four additional ant species (see Appendix C).In these species and all other cases in which trendswere detectable (including all factors where F . 1.0,except in Leptomyrmex unicolor), ants were more fre-quent on sources with lower sugar concentration and/or lower relative sucrose content, as opposed to O.smaragdina; amino acid characteristics did not haveany effect on the foraging of these ants. Hence, onlythe competitively dominant species (O. smaragdina)demonstrated significant preferences for higher nectarquality through its foraging pattern on natural nectarand honeydew sources, whereas the rest of the com-munity was more abundant on poorer resources or non-selective. In addition, no ant species other than O.smaragdina showed a significant model when honey-dew sources were excluded from the analysis.

Monopolization of nectar sources by O. smaragdi-na.—O. smaragdina was most abundant (recorded onfive or more plant individuals) at four species of plantswith EFNs, one species with a wound sap, and twohoneydew sources (thus at seven out of 32 sources).Four of these seven cases represent the only sourcesin which at least 16 out of 17 focal amino acids werepresent (Appendix A). These four cases were: (1) hon-eydew samples collected from Sextius ‘kurandae’membracids on Entada phaseoloides or on (2) Cae-salpinia traceyi lianas (both Fabaceae sensu lato [5Leguminosae]), and (3) extrafloral nector from Fla-gellaria indica (Flagellariaceae) or from (4) Smilax cf.australis (Smilacaceae) (see Plate 1).

The former two honeydew sources represented themost frequently attended liquid food source of this antat the study site (Blüthgen and Fiedler 2002). Fur-thermore, per capita ant recruitment to homopterans onthe two leguminous liana species was significantlyhigher than to homopterans on other host plants (Blüth-gen and Fiedler 2002). The honeydew sources wereexclusively used by O. smaragdina and were alwaysattended by ants. This effective resource monopoliza-tion was probably a consequence of territorial defense(Blüthgen et al. 2004b). The two EFNs on F. indicaand S. cf. australis were also most commonly attendedby O. smaragdina and often monopolized by thesedominant ants in a similar way as honeydew sources.

Spe

cial

Feat

ure


PLATE 1. Oecophylla smaragdina typically monopolizes honeydew and nectar sources, including aggregations of Sextius‘kurandae’ membracids on Entada phaseoloides (left) and extrafloral nectaries on Smilax cf. australis leaves (middle). Othersources are often shared by different ant species, e.g. nectaries on Clerodendrum tracyanum leaves by Camponotus vitreusand Crematogaster cf. pythia (right). Photo credit: Nico Blüthgen.

Simultaneous co-occurrences of different ant speciesforaging for nectar on the same plant individual weresignificantly less common than expected on F. indicaand never observed on S. cf. australis plants; such co-occurrences between ants were frequent on many ex-trafloral and floral nectars that were poorer in aminoacids (Blüthgen et al. 2004b) (see Plate 1). Thus, hon-eydew and nectar sources that were particularly rich inamino acids were defended and monopolized by thedominant ant in a similar way.

DISCUSSION

Ant communities are strongly hierarchical. Compet-itively dominant and territorial species can be distin-guished from competitively inferior species; the dis-tribution of the latter may be partly controlled by thedominants (Savolainen and Vepsäläinen 1988, Dejeanand Corbara 2003). Dominant ants often monopolizesome nectaries and honeydew-producing homopteranaggregations, whereas subordinate ants are more op-portunistic and commonly share attended nectaries withother species simultaneously (Schemske 1982, Dejeanet al. 1997, Blüthgen et al. 2000). This pattern was alsopronounced at our study site (Blüthgen and Fielder2002, Blüthgen et al. 2004b). In the present paper, wewere able to link differences in resource monopoliza-tion with nectar production and composition. The keytraits influencing the foraging patterns of the compet-itively dominant ant Oecophylla smaragdina were ami-no acid composition and higher total concentration ofamino acids and sugars, corresponding to the impor-tance of amino acids for ant nutrition. Amino acids arethe major nitrogenous compounds in nectar and hon-eydew, and despite their high C:N ratio, these sourcescan contribute substantially to the nitrogen supply ofant colonies (Blüthgen et al. 2003). Preferences andmonopolization by the dominant ant were thus asso-ciated with high-quality resources, whereas low-qualitynectars were used more opportunistically by both dom-inant and subordinate ants. Subordinate ants, includinghighly abundant species such as Crematogaster spp.,

either showed no preference in the foraging model, orthey were more common on sources with lower sucroseor total sugar concentration, suggesting an impact ofcompetitive interactions. Even among dominant ants,nectar quality was not universally important. The sec-ond dominant ant, Anonychomyrma gilberti, did notshow any significant preference among the selectedsources, possibly due to its specialization for nest siteson its host tree Syzygium erythrocalyx, which also func-tions as a main honeydew and nectar source (Monteith1986, Blüthgen et al. 2004b).

Several alternative sources of variation may act uponant foraging that could potentially obscure stronger ef-fects of nectar composition on ant foraging. (1) Ouranalysis was based on an average nectar compositionof each source species, represented by sampling one ora few individuals at different times of the year, and theoverall visitation pattern on this source species. How-ever, these traits are variable between (and within) plantindividuals and over time, and resource selection byants is based on plant individuals rather than on averagespecies values. (2) The local pool of resources availableto each ant colony only includes a small, spatially re-stricted portion of those analyzed, and thus actualchoices are limited. (3) Niche differentiation and com-petitive avoidance among nectarivores may also occuralong niche dimensions other than nectar composition,such as seasonal or diurnal activity patterns (Hossaert-McKey et al. 2001) and differential selection of habitatpatches (Biesmeijer et al. 1999), along with spatiotem-poral heterogeneity. Despite these potential sources ofuncontrolled variation, effects of nectar compositionon foraging of O. smaragdina were significant.

In addition to observations on natural resources, weperformed cafeteria experiments offering artificial‘‘nectar’’ solutions and comparing ant foraging be-tween paired treatments (Blüthgen and Fiedler 2004).In the absence of competition, O. smaragdina and manyother ant species preferred sugar solutions containingamino acid mixtures over pure sugar solutions. Amongthe most attractive solutions to most ants was an amino


Spec

ialFeatu

re

acid mixture mimicking one of the main honeydewsources exclusively used by O. smaragdina (secretedby Sextius ‘kurandae’ on Caesalpinia traceyi; see Ap-pendix A). Most ants also preferred sucrose over mostother carbohydrates, whereas trisaccharides like me-lezitose were not particularly attractive to much of theAustralian ant community. Despite some degree of in-terspecific variation, the results indicate that, wherethey exist, gustatory preferences for sugars and aminoacid mixtures are broadly similar across the ant com-munity. In contrast, preferences for single amino acidsin sugar solutions are highly species-specific and idi-osyncratic. Interspecific variation in sugar and aminoacid preferences has also been found in other studies(Lanza and Krauss 1984, Lanza 1991, Lanza et al.1993, Koptur and Truong 1998).

These described preference patterns were found inthe absence of competitors. However, when other antspecies were present at the same bait pair in our ex-periments, discrimination between different solutionswas significantly less pronounced, i.e., the ants weremore evenly distributed between the alternative solu-tions (Blüthgen and Fiedler 2004). These findings pro-vide experimental evidence that competition stronglyaffects dietary selection (see also Nonacs and Dill1990). The high concordance of preferences amongsugars and for amino acid mixtures found in the ex-periment, in combination with the opportunistic choic-es by competitively inferior ants among natural nectarsources, suggest that the diversity of nectar-consumercommunities is more related to processes driven bycompetition and spatiotemporal variation in resourceavailability than to specializations of the componentspecies on particular resources.

Our study emphasizes the need to study interactionswithin their complex environment. For example, in iso-lated experimental situations such as in typical bio-assays and cafeteria experiments, one would concludethat ants are highly selective among different nectarcompounds. These single-species experiments are valu-able to search for potential preferences in complex en-vironments. However, when considering communityinteractions, such selectivity may only translate intoactual resource selection in competitively superior spe-cies or when competition is relaxed. In diverse tropicalinsect communities, most species are relatively rare andprobably competitively inferior. Hence we expect sub-stantial discrepancies between physiological prefer-ences for nectar types and actual resource utilizationin such communities. Such differences between real-ized and fundamental niches are likely to reflect com-petition costs.

How widespread is variation in nectar preferencesand competition in other nectarivore communities?First, gustatory preferences vary substantially betweendifferent nectarivores (Lanza and Krauss 1984, Roubikand Buchmann 1984, Alm et al. 1990, Lanza et al.1993, Roubik et al. 1995). These differences may be

linked to variable nutrient requirements across differenttaxa (Kevan and Baker 1983, Alm et al. 1990), colonies(Cassill and Tschinkel 1999), sexes (Erhardt and Rus-terholz 1998), or seasons (Sudd and Sudd 1985, Rus-terholz and Erhardt 2000). Moreover, physiologicalconstraints also may be important and may limit dietaryselection, e.g., chemoreception (Wada et al. 2001) ordigestive enzymes (Martı́nez del Rio 1990). Thus,physiological preferences for nectar composition maybe highly variable and specific, but we have little in-formation about whether and how these differencesmay translate into the structure of real nectarivore com-munities.

Secondly, if nectar represents a spatiotemporallylimited resource for nectarivores, inter- and intraspe-cific competition should be pronounced. For instance,competition between nectarivores was implied by stud-ies testing various optimal foraging models (Dreisig1995, Ohashi and Yahara 2002). Effects of interspecificcompetition between nectarivores have been demon-strated in several systems (Pyke 1982, Biesmeijer etal. 1999), including competition between unrelated tax-onomic groups (Morse 1981, Laverty and Plowright1985, Thomson 1989). Competitive hierarchies can bestrong and complex in diverse flower-visitor commu-nities (Kikuchi 1963, Nagamitsu and Inoue 1997).Choices between available flowers (such as the identityof plant species, or their positions within a plant) maydiffer between situations in which competitors werepresent or absent (competitive release) (Inouye 1978,Bowers 1985, Laverty and Plowright 1985, Thomson1989). In an experimental study of three hummingbirdspecies in Arizona, Pimm et al. (1985) showed that allspecies preferred feeding on higher sucrose concentra-tions when competition was low and increasingly usedthe poorer resources when competitors were abundant.Competition between hummingbirds was highly asym-metric in that the behaviorally dominant species af-fected the foraging of subordinate species but not viceversa.

In conclusion, gustatory preferences can vary withinand across nectarivore species in any given community.However, prevailing competition among nectarivoresin natural, diverse assemblages is expected to constrainstrongly the extent to which such preferences translateinto patterns of actual resource use and partitioning.The present study on nectar-feeding ants in combina-tion with previous work using artificial nectars (Blüth-gen and Fiedler 2004) demonstrates that competitivehierarchies may determine patterns of foraging for ami-no acids in nectars. This may be a more general phe-nomenon affecting interactions and resource partition-ing between nectarivores and nectar plants, and mayprovide a fruitful area for further studies in communitydynamics.

ACKNOWLEDGMENTSWe are grateful to Gerhard Gottsberger (University of Ulm)

for cooperation with the HPLC analysis and Hans Malchus

Spe

cial

Feat

ure


for technical assistance. We also thank the Australian CanopyCrane Company (ACC), especially Nigel Stork, and the En-vironmental Research Station in Cape Tribulation, namelyHugh Spencer, for logistical support in Australia. Plant andanimal identification was kindly supported by Mary Carver,Max Day, John Donaldson, Bruce Gray, Brian Heterick, Ber-nie Hyland, Bob Jago, Rigel Jensen, Rudy Kohout, HannaReichel, Steve Shattuck, and Andrew Small. Lynn Adler, Bri-gitte Fiala, David Inouye, Karsten Mody, and Jennifer Rudg-ers provided helpful comments and discussions on earlierversions of the manuscript. The project was funded by theDeutsche Forschungsgemeinschaft (Fi 547/9-1), a donationby the Fuchs Oil Company (Mannheim/Sydney) via the ACC,and by a doctoral fellowship of the Studienstiftung desdeutschen Volkes to N. Blüthgen.

LITERATURE CITED

Alm, J., T. E. Ohnmeiss, J. Lanza, and L. Vriesenga. 1990.Preference of cabbage white butterflies and honey bees fornectar that contains amino acids. Oecologia 84:53–57.

Andersen, A. N. 1995. A classification of Australian ant com-munities, based on functional groups which parallel plantlife-forms in relation to stress and disturbance. Journal ofBiogeography 22:15–29.

Baker, H. G., and I. Baker. 1973. Amino acids in nectar andtheir evolutionary significance. Nature 24:543–545.

Baker, H. G., and I. Baker. 1975. Studies of nectar consti-tution and pollinator–plant coevolution. Pages 100–140 inL. E. Gilbert and P. H. Raven, editors. Co-evolution ofanimals and plants. University of Texas Press, Austin, Tex-as, USA.

Biesmeijer, J. C., J. A. P. Richter, M. A. J. P. Smeets, and M.J. Sommeijer. 1999. Niche differentiation in nectar-col-lecting stingless bees: the influence of morphology, floralchoice and interference competition. Ecological Entomol-ogy 24:380–388.

Blüthgen, N., and K. Fiedler. 2002. Interactions betweenweaver ants (Oecophylla smaragdina), homopterans, treesand lianas in an Australian rainforest canopy. Journal ofAnimal Ecology 71:793–801.

Blüthgen, N., and K. Fiedler. 2004. Preferences for sugarsand amino acids and their conditionality in a diverse nectar-feeding ant community. Journal of Animal Ecology 73:155–166.

Blüthgen, N., G. Gebauer, and K. Fiedler. 2003. Disentan-gling rainforest food webs using stable isotopes: dietarydiversity in a species-rich ant community. Oecologia 137:426–435.

Blüthgen, N., G. Gottsberger, and K. Fiedler. 2004a. Sugarand amino acid composition of ant-attended nectar and hon-eydew sources from an Australian rainforest. Austral Ecol-ogy 29, in press.

Blüthgen, N., and K. Reifenrath. 2003. Extrafloral nectariesin an Australian rainforest—structure and distribution.Australian Journal of Botany 51:515–527.

Blüthgen, N., N. E. Stork, and K. Fiedler. 2004b. Bottom-up control and co-occurence in complex communities: Hon-eydew and nectar determine a rainforest ant mosaic. Oikos105, in press.

Blüthgen, N., M. Verhaagh, W. Goitı́a, K. Jaffé, W. Morawetz,and W. Barthlott. 2000. How plants shape the ant com-munity in the Amazonian rainforest canopy: the key roleof extrafloral nectaries and homopteran honeydew. Oec-ologia 125:229–240.

Bowers, M. A. 1985. Experimental analyses of competitionbetween two species of bumble bees (Hymenoptera: Api-dae). Oecologia 67:224–230.

Cassill, D. L., and W. R. Tschinkel. 1999. Regulation of dietin the fire ant, Solenopsis invicta. Journal of Insect Behavior12:307–328.

Clarke, K. R. 1993. Non-parametric multivariate analyses ofchanges in community structure. Australian Journal ofEcology 18:117–143.

Davidson, D. W. 1997. The role of resource imbalances inthe evolutionary ecology of tropical arboreal ants. Biolog-ical Journal of the Linnean Society 61:153–181.

Davidson, D. W., S. C. Cook, R. R. Snelling, and T. H. Chua.2003. Explaining the abundance of ants in lowland tropicalrainforest canopies. Science 300:969–972.

Dejean, A., T. Bourgoin, and M. Gibernau. 1997. Ant speciesthat protect figs against other ants: result of territorialityinduced by a mutualistic homopteran. Ecoscience 4:446–453.

Dejean, A., and B. Corbara. 2003. A review of mosaics ofdominant ants in rainforests and plantations. Pages 341–347 in Y. Basset, V. Novotny, S. E. Miller, and R. L. Kitch-ing, editors. Arthropods of tropical forests: spatio-temporaldynamics and resource use in the canopy. Cambridge Uni-versity Press, Cambridge, UK.

Dreisig, H. 1988. Foraging rate of ants collecting honeydewor extrafloral nectar, and some possible constraints. Eco-logical Entomology 13:143–154.

Dreisig, H. 1995. Ideal free distributions of nectar foragingbumblebees. Oikos 72:161–172.

Erhardt, A., and H. P. Rusterholz. 1998. Do peacock butter-flies (Inachis io L.) detect and prefer nectar amino acidsand other nitrogenous compounds? Oecologia 117:536–542.

Gardener, M. C., and M. P. Gillman. 2002. The taste ofnectar—a neglected area of pollination ecology. Oikos 98:552–557.

Gottsberger, G., J. Schrauwen, and H. F. Linskens. 1984.Amino acids and sugars in nectar, and their putative evo-lutionary significance. Plant Systematics and Evolution145:55–77.

Hölldobler, B. 1983. Territorial behavior in the green tree ant(Oecophylla smaragdina). Biotropica 15:241–250.

Hossaert-McKey, M., J. Orivel, E. Labeyrie, L. Pascal, J. H.C. Delabie, and A. Dejean. 2001. Differential associationswith ants of three co-occurring extrafloral nectary-bearingplants. Ecoscience 8:325–335.

Inouye, D. W. 1978. Resource partitioning in bumblebees:experimental studies of foraging behavior. Ecology 59:672–678.

Inouye, D. W., and G. D. Waller. 1984. Responses of honeybees (Apis mellifera) to amino acid solutions mimickingfloral nectars. Ecology 65:618–625.

Jackson, D. 1984. Competition in the tropics: ants on trees.Antenna 8:19–22.

Kevan, P., and H. B. Baker. 1983. Insects as flower visitorsand pollinators. Annual Review of Entomology 28:407–454.

Kikuchi, T. 1963. Studies on the coaction among insects vis-iting flowers. III. Dominance relationships among flower-visiting flies, bees and butterflies. Science Reports of theTôhoku University Series IV (Biology) 29:1–8.

Koptur, S., and N. Truong. 1998. Facultative ant–plant in-teractions: nectar sugar preferences of introduced pest antspecies in south Florida. Biotropica 30:179–189.

Lanza, J. 1991. Response of fire ants (Formicidae: Solenopsisinvicta and S. geminata) to artificial nectars with aminoacids. Ecological Entomology 16:203–210.

Lanza, J., and B. R. Krauss. 1984. Detection of amino acidsin artificial nectars by two tropical ants, Leptothorax andMonomorium. Oecologia 63:423–425.

Lanza, J., E. L. Vargo, S. Pulim, and Y. Z. Chang. 1993.Preferences of the fire ants Solenopsis invicta and S. gem-inata (Hymenoptera: Formicidae) for amino acid and sugarcomponents of extrafloral nectars. Environmental Ento-mology 22:411–417.


Spec

ialFeatu

re

Laverty, T. M., and R. C. Plowright. 1985. Competition be-tween hummingbirds and bumble bees for nectar in flowersof Impatiens biflora. Oecologia 66:25–32.

Legendre, P., and L. Legendre. 1998. Numerical ecology.Second edition. Elsevier, Amsterdam, The Netherlands.

Majer, J. D. 1993. Comparison of the arboreal ant mosaic inGhana, Brazil, Papua New Guinea and Australia—its struc-ture and influence on arthropod diversity. Pages 115–141in J. LaSalle and I. D. Gauld, editors. Hymenoptera andbiodiversity. CAB International, Wallingford, UK.

Martı́nez del Rio, C. 1990. Dietary, phylogenetic, and eco-logical correlates of intestinal sucrase and maltase activityin birds. Physiological Zoology 63:987–1011.

Monteith, G. B. 1986. Some curious insect–plant associationsin Queensland. Queensland Naturalist 26:105–114.

Morse, D. H. 1981. Interactions among syrphid flies andbumblebees on flowers. Ecology 62:81–88.

Nagamitsu, T., and T. Inoue. 1997. Aggressive foraging ofsocial bees as a mechanism of floral resource partitioningin an Asian tropical rainforest. Oecologia 110:432–439.

Nonacs, P., and L. M. Dill. 1990. Mortality risk vs. foodquality trade-offs in a common currency: ant patch pref-erences. Ecology 71:1886–1892.

Ohashi, K., and T. Yahara. 2002. Visit larger displays butprobe proportionally fewer flowers: counterintuitive be-haviour of nectar-collecting bumble bees achieves an idealfree distribution. Functional Ecology 16:492–503.

Ollerton, J., and L. Cranmer. 2002. Latitudinal trends inplant–pollinator interactions: are tropical plants more spe-cialised? Oikos 98:340–350.

Pimm, S. L., M. L. Rosenzweig, and W. Mitchell. 1985. Com-petition and food selection: field tests of a theory. Ecology66:798–807.

Pyke, G. H. 1982. Local geographic distributions of bum-blebees near Crested Butte, Colorado: competition andcommunity structure. Ecology 63:555–573.

Roubik, D. W., and S. L. Buchmann. 1984. Nectar selectionby Melipona and Apis mellifera (Hymenoptera: Apidae) and

the ecology of nectar intake by bee colonies in a tropicalforest. Oecologia 61:1–10.

Roubik, D. W., D. Yanega, M. Aluja, S. L. Buchmann, andD. W. Inouye. 1995. On optimal nectar foraging by sometropical bees (Hymenoptera, Apidae). Apidologie 26:197–211.

Rusterholz, H. P., and A. Erhardt. 2000. Can nectar propertiesexplain sex-specific flower preferences in the Adonis Bluebutterfly Lysandra bellargus? Ecological Entomology 25:81–90

Savolainen, R., and K. Vepsäläinen. 1988. A competitionhierarchy among boreal ants: impact on resource partition-ing and community structure. Oikos 51:135–155.

Schemske, D. W. 1982. Ecological correlates of a neotropicalmutualism: ant assemblages at Costus extrafloral nectaries.Ecology 63:932–941.

StatSoft. 1999. Statistica for Windows 5. 5. StatSoft, Tulsa,Oklahoma, USA.

Sudd, J. H., and M. E. Sudd. 1985. Seasonal changes in theresponse of wood-ants (Formica lugubris) to sucrose baits.Ecological Entomology 10:89–97.

Thomson, J. D. 1989. Reversal of apparent feeding prefer-ences of bumble bees by aggression from Vespula wasps.Canadian Journal of Zoology 67:2588–2591.

Wäckers, F. L. 1999. Gustatory response by the hymenop-teran parasitoid Cotesia glomerata to a range of nectar andhoneydew sugars. Journal of Chemical Ecology 25:2863–2877.

Wada, A., Y. Isobe, S. Yamaguchi, R. Yamaoka, and M.Ozaki. 2001. Taste-enhancing effects of glycine on thesweetness of glucose: a gustatory aspect of symbiosis be-tween the ant, Camponotus japonicus, and the larvae of thelycaenid butterfly, Niphanda fusca. Chemical Senses 26:983–992.

Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, andJ. Ollerton. 1996. Generalization in pollination systems,and why it matters. Ecology 77:1043–1060.

APPENDIX A

A table showing sugar and amino acid composition of nectar and honeydew sources is available in ESA’s Electronic DataArchive: Ecological Archives E085-038-A1.

APPENDIX B

A table of NMDS (nonmetric multidimensional scaling) ordination of amino acid profiles is available in ESA’s ElectronicData Archive: Ecological Archives E085-038-A2.

APPENDIX C

Multiple linear regression models for foraging patterns of 16 ant species are available in ESA’s Electronic Data Archive:Ecological Archives E085-038-A3.

1486

Spe

cial

Feat

ure


WHY ARE SOME FLORAL NECTARS SCENTED?

ROBERT A. RAGUSO1

Department of Biological Sciences, University of South Carolina, Columbia, South Carolina 29208 USA

Abstract. Despite recent interest in the non-sugar components of floral nectar, nearlynothing is known about the ecological importance and phylogenetic distribution of scentednectar. If present, the scent of nectar would provide an honest signal to nectar-feedinganimals. Nectar odors may directly impact plant reproductive fitness, through pollinatorattraction or deterrence of nectar robbers and florivores. In addition, nectar odors mayindirectly impact plant fitness through antimicrobial activity, pleiotropic interactions withplant defense, and communication with predators and parasitoids. The literature providesonly circumstantial evidence for scented nectar, through the study of bee honey odors. HereI confirm the presence of scent in the nectar of four out of seven angiosperm speciessampled with solid-phase micro-extraction and gas chromatography–mass spectrometry. InAbelia x grandiflora and Hedychium coronarium, nectar odors are a hydrophilic subset ofthe compounds emitted by surrounding floral tissues, suggesting passive absorption by thenectar standing crop. Sucrose solution applied to the petals of a nectarless flower, Magnoliagrandiflora, absorbed a hydrophilic subset of scent compounds after one hour, lendingsupport to this hypothesis. Nectar from Oenothera primiveris and Agave palmeri containedunique scent compounds compared to the floral tissues. The presence of fermentation vol-atiles in A. palmeri nectar suggests a dynamic role for yeast in its floral biology. Thesedata highlight the need for systematic studies on the distribution and mechanistic importanceof scented floral nectar to plant–animal interactions.

Key words: constancy; fragrance; headspace; honest signals; nectar; plant defense; plant vol-atiles; pollination; solid-phase micro-extraction; SPME.

INTRODUCTION

In the generation since Herbert and Irene Baker’spioneering studies of nectar composition (Baker andBaker 1973, 1975), pollination biologists have em-braced the idea that floral nectars are more than sugarsolutions (reviewed by Bentley and Elias 1983, Adler2000, Gardener and Gillman 2002). Studies of nectarchemistry now explore the connections between nectaramino acids and pollinator guilds (Gottsberger et al.1984, Forcone et al. 1997, Erhardt and Rusterholz1998) and between nectar phenolic compounds andplant defense (Guerrant and Fiedler 1981, Hagler andBuchmann 1993). Despite nascent interest in non-sugarnectar chemistry, there have been no systematic anal-yses of scent compounds in floral nectar. Flowers com-monly are scented, which can have both positive andnegative implications for plant–animal interactions(Raguso 2001). Similarly, scented nectar has the po-tential to impact plant reproductive fitness directly, asa sensory signal to which both beneficial and harmfulfloral visitors respond (Galen 1999). Scent compoundsin nectar may also function indirectly, through anti-microbial activity, defense physiology, or signaling topredators and parasitoids (Pichersky and Gershenzon

Manuscript received 12 June 2003; revised 15 October 2003;accepted 16 October 2003; final version received 3 November2003. Corresponding Editor (ad hoc): R. E. Irwin. For reprintsof this Special Feature, see footnote 1, p. 1477.

1 E-mail: [email protected]

2002). Here I outline several potential functions ofscented nectar, review the fragmentary evidence for itspresence in flowering plants, and present new dataquantifying scented nectar in a small sample of angio-sperms.

BACKGROUND: SCENT IN NECTAR AND HONESTFLORAL SIGNALS

Most floral fragrances are chemically heterogeneousblends (e.g., terpenoids, aromatics, fatty acid- and ami-no acid-derived compounds) emitted from floral tissuesand vegetation (Raguso and Pichersky 1999). Odor-mediated pollinator attraction is considered a derivedcondition because the original functions of floral odorsprobably were defensive (Pellmyr and Thien 1986,Armbruster 1997). Alone, or with visual cues, floralscent communicates the location, quality, and abun-dance of nectar and pollen to flower-visiting animals(Dobson 1994). Diverse groups of nectar- and pollen-foraging insects utilize odor contrast within flowers toreduce handling times (Brantjes 1976, Lunau 1992),and changes in such odors to discriminate against un-rewarding flowers (Lex 1954). Nectar that possessesits own odor provides an honest signal, facilitating re-mote detection by nectar-foraging animals (Heinrich1979), especially in cases in which the scent of nectaris qualitatively or quantitatively distinct from overallfloral scent.

Floral ‘‘truth in advertising’’ holds potential benefitsfor plants if it increases pollen donation and receipt


Spec

ialFeatu

re

and for pollinators if it promotes efficient learning,flower handling, and profitability (Dobson and Bergs-tröm 2000). Conversely, one potential constraint on theevolution of nectar odor is its effect on pollinator move-ment within vs. between conspecific plants (Weather-wax 1986). For example, if visitation frequency amongflowers of the same plant is high, geitonogamy (DeJong et al. 1993), pollen discounting (Johnson and Nils-son 1999), and the dislodging of pollen from stigmaticsurfaces (Young 1988) could outweigh the benefits ofhonest signals. Similarly, nectar scent would not benefitplants that manipulate pollinator movement via within-plant heterogeneity in nectar rewards (Soberón-Mai-nero and Martı́nez del Rı́o 1985, Real and Rathcke1988).

With these costs and benefits of scented nectar inmind, the effectiveness of scented nectar for plant–pollinator interactions hinges upon a pollinator’s abilityto detect nectar or its absence (Bell 1986, Thakar et al.2003). Heinrich (1979) established that Bombus vagansworkers distinguish nectar-rich Trifolium repens (Fa-gaceae) flowers from depleted ones by the intensity of‘‘clover scent’’ and reject empty flowers without land-ing. Marden (1984) elegantly confirmed the olfactorybasis of this behavior using honey- and fingerprint-scented artificial flowers. Similarly, many butterfliesand moths feed more readily from artificial flowers withscented honey water instead of sugar solutions (Weiss2001). In addition, honeybees and bumble bees showgreater constancy (consecutive visits to conspecificflowers) when flowers differ in multiple sensory sig-nals, e.g., color and odor (Free 1970, Gegear and Lav-erty 2001). However, some of the patterns previouslydescribed could reflect gustatory instead of (or in ad-dition to) olfactory conditioning, as proposed by Gar-dener and Gillman (2002) for amino acids. Assays ofthe behavioral responses of olfaction-impaired insectsto nectar with vs. without scent would distinguish be-tween olfactory and gustatory responses. Alternatively,one could score visitor preference or constancy amongartificial flowers with free nectar odors (e.g., linaloolor 2-phenyl ethanol) vs. those with nonvolatile con-jugates of the same compounds (e.g., linalyl-b-D-gly-cosides), which are common in floral tissues (Watanabeet al. 1993). Finally, electrophysiological recordings ofchemoreceptor responses to odors (vs. sugars, aminoacids) are needed to determine whether gustatory re-sponses to scent compounds in nectar occur in a givenfloral visitor.

BEYOND POLLINATION: ALTERNATIVE HYPOTHESESON THE ROLE(S) OF SCENT IN NECTAR

Honest signaling is relevant to plant–pollinator com-munication, but scent in nectar may also serve defen-sive functions. Protecting nectar crops from harmfulvisitors, particularly ants, has been an important themein the post-Baker era (Janzen 1977a, Feinsinger andSwarm 1978). Calyx odors deter ants from robbing

Polemonium viscosum (Polemoniaceae) flowers (Galen1983), and scent in nectar might serve a similar func-tion. Hummingbird-pollinated flowers are vulnerableto nectar larceny by bees (Waser 1979, Irwin and Brody1999), birds (McDade and Kinsman 1980), and flowermites (Colwell 1995). The costs vs. benefits of nectarrobbing are highly debated (Irwin et al. 2001). If thenet impact of robbers on plant fitness is positive, thenselection might favor nectar odors that attract ratherthan repel robbers. For example, Lara and Ornelas(2002) showed that reduced nectar crops in Moussoniadeppeana (Gesneriaceae), due to mite infestation, werecorrelated with enhanced seed set through increasedhummingbird probing per flower. Indeed, Heyneman etal. (1991) have shown that flower mites use odor todiscriminate between nectars of host and non-hostplants (and sucrose solution) in binary choice assays.Conversely, if the net impact of robbers on plant fitnessis negative, then the opposite pattern might hold, al-though there also may be costs of deterring pollinators.Clearly, the magnitude and direction of selection onodors that attract and repel nectar robbers will dependon the net costs and benefits of robbing to the plant(Maloof and Inouye 2000).

Yet another potential defensive function is to protectnectar resources from microbial infestation. Yeaststhrive in floral nectar, producing fermentation productsthat repel pollinators or impair their effectiveness (Jan-zen 1977b, Kevan et al. 1988). Indeed, the presence ofshort-chain alcohol and ketone odors in nectar mightresult from yeast infection (Williams et al. 1981). Theessential oils common to floral scents have well-doc-umented antimicrobial properties (Lokvam and Brad-dock 1999), particularly the water-soluble terpene al-cohols and phenolic compounds (Knobloch et al. 1989).Lawton et al. (1993) observed inhibition of yeastgrowth by petal extracts and single compounds (cin-namic alcohol) from Guettarda poasana (Rubiaceae).However, the critical experiments in which yeast andbacterial cultures are plated with natural concentrationsof scented nectar have yet to be performed.

Nectar odor may also arise through pleiotropic in-teractions with plant defense. For example, male but-terflies and moths are attracted from a distance by thevolatile derivatives of pyrrolizidine alkaloids presentin Eupatorium (Asteraceae) floral nectar (Pliske 1975),which they collect and utilize in predator defense andcourtship (Brown 1984, Boppré 1990). In addition, re-cent studies support the role of floral and extrafloralnectars in mediating interactions between plants andthe third trophic level (Wäckers and Bonifay 2004, inthis Special Feature). Agrawal and Rutter (1998) haveshown that Azteca ants respond to olfactory signals inplants whose defenses have been induced through ac-tual or simulated herbivore wounding. Scented extra-floral nectar may enhance the recruitment or condition-ing of beneficial insects as an inducible defense, justas scented floral nectar is proposed to enhance polli-

Spe

cial

Feat

ure

1488 ROBERT A. RAGUSO Ecology, Vol. 85, No. 6

nator constancy (Robacker et al. 1988). Scented floralnectar could be an artifact in such plants if floral andextrafloral nectaries are served by common phloemducts (Fahn 1988). Thus, nectar odors, like petal pig-ments, may arise through pleiotropic interactions withplant defense, rather than pollinator adaptation (e.g.,Armbruster 2002).

ARE FLORAL NECTARS SCENTED? INDIRECTEVIDENCE FROM THE LITERATURE

What is known about the distribution and frequencyof scented nectars? Unfortunately, little can be gleanedfrom traditional sources on floral biology and polli-nation. Faegri and van der Pijl (1971), Kevan and Baker(1983), Barth (1991), Kearns and Inouye (1993), andProctor et al. (1996) all omit scent from discussions ofnectar composition. A review of the chemical ecolog-ical literature from 1993 to 2003 shows 157 articles onplant volatiles, 127 on nectar chemistry, and 78 onfloral scent or fragrance (Appendix A). Only one ofthese studies (Ecroyd et al. 1995) involved the directanalysis of scent in nectar, from spadices of Dactylan-thus taylorii (Balanophoraceae), a rare parasitic plantendemic to New Zealand.

The apicultural literature provides a contrastingview, in which floral nectar is assumed to be a majorsource of flavors and fragrances that characterize com-mercial honeys (Bicchi et al. 1983, Gil et al. 1995).The characteristic aromas of honeys from single plantsources are chemically screened by the apicultural in-dustry as a means of authentication (Guyot et al. 1998).However, pollen and beeswax derived from plant resinsalso contribute odors to honey (Tomás-Barberán et al.1993). The decomposition of pollen amino acids, fla-vonoids, and phenolics, along with enzymatic modi-fication of nectar odors during honey processing, mayproduce additional volatiles (Aparna and Rajalakshmi1999). Thus, nectar odors are not solely responsiblefor the full aroma spectrum in bee honeys.

The remaining evidence consists of anecdotal ac-counts (e.g., Towner 1977) that lack direct analyticalmeasurements. It is now possible to detect tissue-spe-cific scent patterns within living or recently dissectedflowers (MacTavish et al. 2000, Irwin and Dorsett2002) using solid-phase micro-extraction (SPME),which traps plant odors by exposing an adsorbent fiberwithin a chamber in which plant odors equilibrate (Ar-thur and Pawliszyn 1990). This approach, combinedwith gas chromatography–mass spectrometry (GC–MS), is sensitive enough to identify plant odors fromspecific tissues (Bartak et al. 2003). I will present directevidence for scent in floral nectar using these analyticalmethods.

DIRECT EVIDENCE FOR SCENT IN NECTAR THROUGHSPME ANALYSES

SPME–GC–MS was used to sample scent com-pounds from flowers and nectar of seven species from

six angiosperm families (see Appendix B). Samplingwas constrained by time and plant availability, butbracketed a range of pollination strategies and phylo-genetic affinities. Agave nectar samples were collectednear Portal, Cochise County, Arizona, USA and werestored frozen (2208C) until analysis. The remainingtaxa were studied at the University of South Carolina,Columbia, Richland County, South Carolina, USA.Vouchers were deposited at the A. C. Moore Herbarium(University of South Carolina Herbarium).

For each species, 1–2 g (fresh mass) of newlyopened, excised flowers were sealed within 8 3 8 cmoven bags, in which odors equilibrated for 30 min.Floral nectar (5–10 mL) was drawn using glass capil-laries and blotted onto 1 3 1 cm squares of Whatman# 2 filter paper. Filter papers were sealed within 10-mL glass beakers and equilibrated for 30 min. SPMEfibers (65-m-PDMS/DVB, polydimethylsiloxane/divi-nylbenzene; Supelco, Bellefonte, Pennsylvania, USA)were exposed to equilibrated headspace air for 15 min,then injected into a Shimadzu QP5000 GC–MS with aDB5 capillary column (for details, see Raguso et al.2003). Control samples were collected from water-moistened filter paper in glass beakers, analyzed as justdescribed, and compared with nectar samples. Chro-matogram peaks were tentatively identified using massspectral libraries and co-injection with standard com-pounds (Levin et al. 2001), and peak areas were in-tegrated using Shimadzu software and expressed as per-centage of total emissions per sample.

In total, 130 scent compounds were detected in theseven species sampled (Appendix B). Campsis radi-cans (Bignoniaceae) and Lonicera sempervirens (Ca-prifoliaceae) produce red, tubular flowers that are pol-linated by hummingbirds and bees in southeasternNorth America (James 1948, Bertin 1982); they arenearly scentless to the human nose. These flowers pro-duced relatively small amounts of scent compounds(See Appendix B), of which none was detected in 10-mL nectar samples. Foeniculum vulgare (Apiaceae)umbels produced a distinctive, spicy fragrance, but fil-ter paper blotted directly on the glistening nectaries ofmany tens of flowers did not yield detectable odor com-pounds.

The remaining taxa emitted complex fragrances fromfloral and nectar samples (Appendix B; Fig. 1). Somenectar odors were similar to those emitted by floraltissues of the same species. This pattern was observedfor non-native Abelia x grandiflora (Caprifoliaceae)flowers, which attract noctuid moths (Haynes et al.1991), butterflies, and diverse bees in South Carolina(R. A. Raguso, personal observation). Nectar from A.x grandiflora contained 2-phenylethanol and pheny-lacetaldehyde, which account for ,10% of total floralemissions (Fig. 1A). These compounds function as sa-lient conditioning stimuli and feeding attractants for arange of butterfly, moth, and bee species (Raguso2003). Similarly, only 12 of the 71 compounds emitted


Spec

ialFeatu

re

FIG. 1. Superimposed GC–MS chromatograms from floral (upper trace) and nectar (lower trace) odor samples from (A)Abelia x grandiflora, (B) Hedychium coronarium, and (C) Agave palmeri. Boxed peaks in panels (A) and (B) are a series ofmonoterpene hydrocarbons and benzaldehyde (1); note their absence in nectar samples. Arrows indicate shared scent com-pounds between floral tissues and nectar; asterisks denote odors unique to nectar. Linalool is the large peak at 11.5 min inpanel (B). Sesquiterpene compounds with retention times greater than 13 min are not shown (see Appendix B). Unlabeledpeaks in lower traces represent ambient contaminants.

by night-blooming Hedychium coronarium (Zingiber-aceae) flowers were detected in nectar samples (Fig.1B). The dominant nectar odorants (linalool and severalaldoximes, nitriles, and nitro-compounds) are dispro-portionately represented with respect to their abun-dance in floral samples (Appendix B), but are char-

acteristic of a majority of hawkmoth-pollinated plantsworldwide (Kaiser 1993, Knudsen and Tollsten 1993).

In contrast, the remaining two species produced nec-tars with unique scent compounds not found in floraltissue. The nectar of night-blooming Oenothera pri-miveris (Onagraceae; see Plate 1) had a sharp, pungent

Spe

cial

Feat

ure


FIG. 2. Evidence for passive extraction of Magnolia grandiflora volatiles by sucrose solution. (A) SPME–GC–MS chro-matograms of headspace from M. grandiflora petals (top trace), sucrose solution after 1-h incubation on petals (middle trace),and sucrose control (bottom trace). Box 1 encloses several large monoterpene peaks; note their absence in sucrose samples.Boxes 2 and 3 indicate geraniol and cis-jasmone, respectively, present in the incubated sucrose (small arrows) but absent inthe control trace. Remaining peaks in the sucrose samples are ambient contaminants. Panels (B) and (C) show the massspectra of geraniol from the floral and sucrose samples, respectively (SI unit conversion: 1 dalton 3 (1.66 3 1018) 5 1 mg).

odor consisting of methyl benzoate and an unidentifiedcompound not emitted by corolla tissues or pollen, andlacked the blend of aldoximes and terpenoids emittedby petals (Appendix B). The nectar of Agave palmeri(Agaveceae) has a chemically complex odor redolentof rotting fruit (Schaffer and Schaffer 1977). SPMEanalysis identified 10 scent compounds also emitted byfloral tissues (fruity esters, dimethyl disulfide, terpe-noids, and sorbic acid esters), along with seven com-pounds unique to nectar. The presence of specific short-chain alcohols and ketones in the latter category sug-gests that fermentation occurred in the nectar (Williamset al. 1981, Nout and Bartelt 1998) (Appendix B; Fig.1C).

These results demonstrate that nectar odor can beanalyzed directly using SPME–GC–MS. The samplingmethodology was suitable for strongly scented flowerswith tubular corollas or spurs and large nectar volumes.Alternative methods will be needed to study umbels or

capitula of minute flowers with negligible nectar vol-umes. For example, I could not verify whether Foen-iculum nectar is scented by using the filter paper-SPMEapproach. However, Patt et al. (1999) demonstrated thateulophid parasitoid wasps learn to find and feed fromartificial flowers more effectively when they includehoney water or Foeniculum nectar than when sugarsolutions are used, providing behavioral data sugges-tive of nectar scent in this system.

HOW DO NECTARS BECOME SCENTED?

The simplest hypothesis for the presence of scent infour of seven nectar samples studied is that volatilesreleased by adjacent floral tissues are partially solublein the aqueous medium of nectar (Weidenhamer et al.1993). The partial extraction of corolla odors by nectarshould occur in strongly fragrant flowers when nectaraccumulates within odor-emitting tissues. I tested thishypothesis using flowers of Magnolia grandiflora


Spec

ialFeatu

re

PLATE 1. Oenothera primiveris (Onagraceae) flowers atdawn, Mohawk Dunes, Arizona, USA. Photo credit: RobertA. Raguso.

(Magnoliaceae), which emit geraniol, aliphatic methylesters, and cis-jasmone (Azuma et al. 1997), but pro-duce no nectar. Two 100-mL aliquots of 20% mass/mass sucrose solution were pipetted onto the abaxialsurface of an excised Magnolia petal, incubated atroom temperature for 1 h, and then transferred to filterpaper for SPME analysis as described, with fresh su-crose solution on filter paper as a control.

The results (Fig. 2) indicate that geraniol and cis-jasmone from Magnolia petals were taken up by theartificial nectar in one hour and were volatilized fromthe filter paper. No such odors were emitted by controlfilter paper with sucrose solution (Fig. 2). Interestingly,b-myrcene and trans-b-ocimene, which were nearly asabundant as geraniol in M. grandiflora floral samples,were not detected in the sucrose solutions (Fig. 2). Asimilar pattern is evident in comparisons between He-dychium coronarium floral and nectar odors (AppendixB; Fig. 1B), in which linalool dominated the nectarsample (80% of total scent), but trans-b-ocimene wasnearly absent. The 10-fold greater solubility of alcohols(e.g., linalool and geraniol) over hydrocarbons (e.g.,b-myrcene, trans-b-ocimene) is the simplest explana-tion for this pattern (Weidenhamer et al. 1993). Passiveacquisition of nectar odors from scented corollas orpollen contamination (grains dislodged into the nectar;Erhardt and Baker 1990) may be a widespread phe-nomenon and deserves further study. The larger ques-tion of whether such odors can be perceived as signalsdistinct from the scent of the surrounding corolla mustbe addressed using behavioral and electrophysiologicalassays.

The presence of methyl benzoate and an unknowncompound in Oenothera primiveris represents a dif-

ferent pattern, one of contrast between nectar and co-rolla odors. These compounds must be actively secretedwithin the hypanthium (through an unknown mecha-nism), because they are absent from the headspace ofdissected petals, styles, and stamens (data not shown).The disparity in scent composition between Agave nec-tar and floral samples is even more complex. The emis-sion of sorbic esters from Agave nectar, if verified, hasno precedent in studies of floral scent (Knudsen et al.1993). Sorbic acid is widely utilized as a food preser-vative, due to its antimicrobial activity against yeasts(Stratford and Anslow 1998). The most abundant esterin our Agave samples, ethyl sorbate, is a known by-product of bacterial spoilage of wines treated with sor-bic acid (Chisholm and Samuels 1992). Agave palmeriflowers produce large nectar pools, remain open for 4–6 days, and are visited by a broad spectrum of bats,hawkmoths, and bees during the warmest, most humidspan of the Sonoran Desert year (Slauson 2000). Inshort, they are excellent candidates for microbial in-festation, as suggested by the presence of fermentationodors in the nectar sample (Appendix B).

CONCLUSIONS

I have outlined several hypotheses for the potentialorigins and functions of scent in nectar and have pre-sented new data confirming that scented nectar is foundin diverse families of angiosperms. The need for well-conceived systematic studies of this phenomenon, us-ing standardized, noninvasive analytical methodolo-gies, cannot be overstated. Nearly all questions con-cerning this topic remain open to exploration. Is nectarodor correlated with specific pollination strategies, in-florescence architectures, habitats, or phylogenetic lin-eages? Why do the hydrophilic scent compounds mostlikely to be absorbed passively by nectar also appearto show the greatest antimicrobial activity? Why aresome floral nectars scented and others not? Weatherwax(1986) proposed that odor in nectar would be mostappropriate in plants whose male or female fitnesscould be satisfied by one pollinator visit, or in casesin which floral competition for pollinators was limiting.Marden (1984) predicted that selection would opposescent in nectar if it elicited pollinator behaviors thatdisrupted pollen flow, out-crossing efficiency, or sexualfunction. I have identified additional circumstances inwhich processes, independent of pollination, may se-lect for odoriferous nectar. However, these hypotheseswill remain purely speculative until chemical analysesand experimental manipulations of scent in nectar be-come integrated into studies of plant reproductive ecol-ogy.

ACKNOWLEDGMENTS

Many of the ideas presented here were inspired by thegroundbreaking studies of Judie Bronstein and Heidi Dobson.I am grateful to Beverly Rathcke and Leslie Real for earlyinstruction in nectar ecology, and to Lynn Adler, AnuragAgrawal, Laurel Hester, Becky Irwin, and three anonymous

Spe

cial

Feat

ure


reviewers for constructive editorial suggestions. Thanks toLaura Kakuk for translating Lex’s paper and to Glenn Svens-son and Michael Hickman for risking life and limb to collectAgave nectar. This work was supported by NSF grantsDEB9806840 and DEB0317217.

LITERATURE CITED

Adler, L. S. 2000. The ecological significance of toxic nectar.Oikos 91:409–420.

Agrawal, A. A., and M. T. Rutter. 1998. Dynamic anti-her-bivore defense in ant-plants: the role of induced responses.Oikos 83:227–236.

Aparna, A. R., and D. Rajalakshmi. 1999. Honey—its char-acteristics, sensory aspects and applications. Food ReviewsInternational 15:455–471.

Armbruster, W. S. 1997. Exaptations link evolution of plant–herbivore and plant–pollinator interactions: a phylogeneticinquiry. Ecology 78:1661–1672.

Armbruster, W. S. 2002. Can indirect selection and geneticcontext contribute to trait diversification? A transition-probability study of blossom-colour evolution in two gen-era. Journal of Evolutionary Biology 15:468–486.

Arthur, C. L., and J. Pawliszyn. 1990. Solid phase microex-traction with thermal desorption using fused silica opticalfibers. Analytical Chemistry 62:2145–2148.

Azuma, H., M. Toyota, Y. Asakawa, R. Yamaoka, J. G.Garcia-Franco, G. Dieringer, L. B. Thien, and S. Kawano.1997. Chemical divergence in floral scents of Magnoliaand allied genera (Magnoliaceae). Plant Species Biology12:69–83.

Baker, H. G., and I. Baker. 1973. Amino-acids in nectar andtheir evolutionary significance. Nature 241:542–545.

Baker, H. G., and I. Baker. 1975. Studies of nectar consti-tution and plant–pollinator coevolution. Pages 100–140 inL. E. Gilbert and P. H. Raven, editors. Coevolution of an-imals and plants. University of Texas Press, Austin, Texas,USA.

Bartak, P., P. Bednar, L. Cap, L. Ondrakova, and Z. Stransky.2003. SPME—a valuable tool for investigation of flowerscent. Journal of Separation Science 26:715–721.

Barth, F. G. 1991. Insects and flowers. The biology of apartnership. Princeton University Press, Princeton, NewJersey, USA.

Bell, G. 1986. The evolution of empty flowers. Journal ofTheoretical Biology 118:253–258.

Bentley, B., and T. Elias. 1983. The biology of nectaries.Columbia University Press, New York, New York, USA.

Bertin, R. I. 1982. Floral biology, hummingbird pollinationand fruit production of trumpet creeper (Campsis radicans,Bignoniaceae). American Journal of Botany 69:122–134.

Bicchi, C., F. Belliardo, and C. Frattini. 1983. Identificationof the volatile components of some Piedmont honeys. Jour-nal of Apicultural Research 122:130–136.

Boppré, M. 1990. Lepidoptera and pyrrolizidine alkaloids.Exemplification of complexity in chemical ecology. Journalof Chemical Ecology 16:165–185.

Brantjes, N. B. M. 1976. Senses involved in the visiting offlowers by Cucullia umbratica (Noctuidae, Lepidoptera).Entomologia Experimentalis et Applicata 20:1–7.

Brown, K. S., Jr. 1984. Adult-obtained pyrrolizidine alka-loids defend ithomiine butterflies against a spider predator.Nature 309:707–709.

Chisholm, M. G., and J. M. Samuels. 1992. Determinationof the impact of the metabolites of sorbic acid on the odorof a spoiled red wine. Journal of Agricultural and FoodChemistry 40:630–633.

Colwell, R. K. 1995. Effects of nectar consumption by thehummingbird flower mite Proctolaelaps kirmsei on nectaravailability in Hamelia patens. Biotropica 27:206–217.

De Jong, T. J., N. M. Waser, and P. G. L. Klinkhamer. 1993.Geitonogamy: the neglected side of selfing. Trends in Ecol-ogy and Evolution 8:321–325.

Dobson, H. E. M. 1994. Floral volatiles in insect biology.Pages 47–81 in E. Bernays, editor. Insect–plant interac-tions. Volume 5. CRC Press, Boca Raton, Florida, USA.

Dobson, H. E. M., and G. Bergström. 2000. The ecology andevolution of pollen odors. Plant Systematics and Evolution222:63–87.

Ecroyd, C. E., R. A. Franich, H. W. Kroese, and D. Steward.1995. Volatile constituents of Dactylanthus taylorii flowernectar in relation to flower pollination and browsing byanimals. Phytochemistry 40:1387–1389.

Erhardt, A., and I. Baker. 1990. Pollen amino acids—an ad-ditional diet for a nectar feeding butterfly. Plant System-atics and Evolution 169:111–121.

Erhardt, A., and H.-P. Rusterholz. 1998. Do peacock butter-flies (Inachis io L) detect and prefer nectar amino acidsand other nitrogenous compounds? Oecologia 117:536–542.

Faegri, K., and L. van der Pijl. 1971. The principles of pol-lination ecology. Second edition. Pergamon Press, Oxford,UK.

Fahn, A. 1988. Secretory tissues in vascular plants. NewPhytologist 108:229–257.

Feinsinger, P., and L. A. Swarm. 1978. How common areant-repellant nectars? Biotropica 10:238–239.

Forcone, A., L. Galetto, and L. Bernardello. 1997. Floralnectar chemical composition of some species from Pata-gonia. Biochemical Systematics and Ecology 25:395–402.

Free, J. B. 1970. Effect of flower shapes and nectar guideson the behaviour of foraging honeybees. Behaviour 37:269–285.

Galen, C. 1983. The effects of nectar thieving ants on seed-set in floral scent morphs of Polemonium viscosum. Oikos41:245–249.

Galen, C. 1999. Flowers and enemies: predation by nectarthieving ants in relation to variation in floral form of analpine wildflower, Polemonium viscosum. Oikos 85:426–434.

Gardener, M. C., and M. P. Gillman. 2002. The taste ofnectar—a neglected area of pollination ecology. Oikos 98:552–557.

Gegear, R. J., and T. M. Laverty. 2001. The effect of variationamong floral traits on the flower constancy of pollinators.Pages 1–20 in L. Chittka and J. D. Thomson, editors. Cog-nitive ecology of pollination; animal behavior and floralevolution. Cambridge University Press, Cambridge, UK.

Gil, M. I., F. Ferreres, A. Ortiz, E. Subra, and F. A. Tomás-Barberán. 1995. Plant phenolic metabolites and floral or-igin of rosemary honey. Journal of Agricultural and FoodChemistry 43:2833–2838.

Gottsberger, G., J. Schrauwen, and H. F. Linskens. 1984.Amino acids and sugars in nectar, and their putative evo-lutionary significance. Plant Systematics and Evolution145:55–77.

Guerrant, E. O., Jr., and P. L. Fiedler. 1981. Flower defensesagainst nectar-pilferage by ants. Biotropica 13(Supple-ment):25–33.

Guyot, C., A. Bouseta, V. Scheirman, and S. Collin. 1998.Floral origin markers of chestnut and lime tree honeys.Journal of Agricultural and Food Chemistry 46:625–633.

Hagler, J. R., and S. L. Buchmann. 1993. Honeybee foragingresponses to phenolic-rich nectars. Journal of the KansasEntomological Society 66:223–230.

Haynes, K. F., J. Z. Zhao, and A. Latif. 1991. Identificationof floral compounds from Abelia grandiflora that stimulateupwind flight in cabbage looper moths. Journal of ChemicalEcology 17:637–646.


Spec

ialFeatu

re

Heinrich, B. 1979. Resource heterogeneity and patterns ofmovement in foraging bumblebees. Oecologia 40:235–245.

Heyneman, A. J., R. K. Colwell, S. Naeem, D. S. Dobkin,and B. Hallet. 1991. Host plant discrimination: experi-ments with hummingbird flower mites. Pages 455–485 inP. W. Price, T. M. Lewinsohn, G. W. Fernandes, and W. W.Benson, editors. Plant–animal interactions: evolutionaryecology in tropical and temperate regions. John Wiley, NewYork, New York, USA.

Irwin, R. E., and A. K. Brody. 1999. Nectar-robbing bumblebees reduce the fitness of Ipomopsis aggregata (Polemon-iaceae). Ecology 80:1703–1712.

Irwin, R. E., A. K. Brody, and N. M. Waser. 2001. The impactof floral larceny on individuals, populations and commu-nities. Oecologia 129:161–168.

Irwin, R. E., and R. Dorsett. 2002. Volatile production bybuds and corollas of two sympatric, confamilial plants,Ipomopsis aggregata and Polemonium foliosissimum. Jour-nal of Chemical Ecology 28:565–578.

James, R. L. 1948. Some hummingbird flowers east of theMississippi. Castanea 13:97–109.

Janzen, D. H. 1977a. Why don’t ants visit flowers? Biotropica9:1252.

Janzen, D. H. 1977b. Why fruits rot, seeds mold and meatspoils. American Naturalist 111:691–713.

Johnson, S. D., and L. A. Nilsson. 1999. Pollen carryover,geitonogamy, and the evolution of deceptive pollinationsystems in orchids. Ecology 80:2607–2619.

Kaiser, R. 1993. The scent of orchids. Elsevier, EditionesRoche, Basel, Switzerland.

Kearns, C. A., and D. W. Inouye. 1993. Techniques for pol-lination biologists. University Press of Colorado, Boulder,Colorado, USA.

Kevan, P. G., and H. G. Baker. 1983. Insects as floral visitorsand pollinators. Annual Review of Entomology 28:407–453.

Kevan, P. G., D. Eisikowitch, S. Fowle, and K. Thomas. 1988.Yeast-contaminated nectar and its effects on bee foraging.Journal of Apicultural Research 27:26–29.

Knobloch, K., A. Pauli, B. Iberl, H. Weigand, and N. Weis.1989. Antibacterial and antifungal properties of essentialoil components. Journal of Essential Oil Research 1:119–128.

Knudsen, J. T., and L. Tollsten. 1993. Trends in floral scentchemistry in pollination syndromes: floral scent composi-tion in moth-pollinated taxa. Botanical Journal of the Lin-nean Society 113:263–284.

Knudsen, J. T., L. Tollsten, and G. Bergström. 1993. Floralscents—a checklist of volatile compounds isolated by head-space techniques. Phytochemistry 33:253–280.

Lara, C., and J. F. Ornelas. 2002. Effects of nectar theft byflower mites on hummingbird behavior and the reproduc-tive success of their host plant, Moussonia deppeana (Ges-neriaceae). Oikos 96:470–480.

Lawton, R. O., L. D. Alexander, W. N. Setzer, and K. G.Byler. 1993. Floral essential oil of Guettarda poasana in-hibits yeast growth. Biotropica 25:483–486.

Levin, R. A., R. A. Raguso, and L. A. McDade. 2001. Fra-grance chemistry and pollinator affinities in Nyctaginaceae.Phytochemistry 58:429–440.

Lex, T. 1954. Duftmale an Blüten. Zeitschrift für verglei-chende Physiologie 36:212–234.

Lokvam, J., and J. F. Braddock. 1999. Anti-bacterial functionin the sexually dimorphic pollinator rewards of Clusiagrandiflora (Clusiaceae). Oecologia 119:534–540.

Lunau, K. 1992. Innate recognition of flowers by bumblebees: orientation of antennae to visual stamen signals. Ca-nadian Journal of Zoology 70:2139–2144.

MacTavish, H. S., N. W. Davies, and R. C. Menary. 2000.Emission of volatiles from brown Boronia flowers: somecomparative observations. Annals of Botany 86:347–354.

Maloof, J. E., and D. W. Inouye. 2000. Are nectar robberscheaters or mutualists? Ecology 81:2651–2661.

Marden, J. H. 1984. Remote perception of floral nectar bybumblebees. Oecologia 64:232–240.

McDade, L. A., and S. Kinsman. 1980. The impact of floralparasitism in two neotropical hummingbird-pollinatedplant species. Evolution 34:944–958.

Nout, M. J. R., and R. J. Bartelt. 1998. Attraction of a flyingnitidulid beetle (Carpophilus humeralis) to volatiles pro-duced by yeasts grown on sweet corn and a corn-basedmedium. Journal of Chemical Ecology 24:1217–1239.

Patt, J. M., G. C. Hamilton, and J. H. Lashcomb. 1999. Re-sponses of two parasitoid wasps to nectar odors as a func-tion of experience. Entomologia Experimentalis et Appli-cata 90:1–8.

Pellmyr, O., and L. B. Thien. 1986. Insect reproduction andfloral fragrances: keys to the evolution of the angiosperms?Taxon 35:76–85.

Pichersky, E., and J. Gershenzon. 2002. The formation andfunction of plant volatiles: perfumes for pollinator at-traction and defense. Current Opinions in Plant Biology5:237–243.

Pliske, T. E. 1975. Pollination of pyrrolizidine alkaloid-con-taining plants by male Lepidoptera. Environmental Ento-mology 4:475–479.

Proctor, M., P. Yeo, and A. Lack. 1996. The natural historyof pollination. Timber Press, Portland, Oregon, USA.

Raguso, R. A. 2001. Floral scent, olfaction, and scent-drivenforaging behavior. Pages 83–105 in L. Chittka and J. D.Thomson, editors. Cognitive ecology of pollination; animalbehavior and floral evolution. Cambridge University Press,Cambridge, UK.

Raguso, R. A. 2003. Olfactory landscapes and deceptive pol-lination: signal, noise and convergent evolution in floralscent. Pages 631–650 in G. J. Blomquist and R. Vogt, ed-itors. Insect pheromone biochemistry and molecular biol-ogy. Academic Press, New York, New York, USA.

Raguso, R. A., R. A. Levin, S. E. Foose, M. W. Holmberg,and L. A. McDade. 2003. Fragrance chemistry, nocturnalrhythms and pollination ‘‘syndromes’’ in Nicotiana. Phy-tochemistry 63:265–284.

Raguso, R. A., and E. Pichersky. 1999. A day in the life ofa linalool molecule: chemical communication in a plant–pollinator system. Part 1: Linalool biosynthesis in floweringplants. Plant Species Biology 14:95–120.

Real, L., and B. J. Rathcke. 1988. Patterns of individualvariability in floral resources. Ecology 69:728–735.

Robacker, D. C., B. J. D. Meeuse, and E. H. Erickson. 1988.Floral aroma: how far will plants go to attract pollinators?BioScience 38:390–398.

Schaffer, W. M., and M. V. Schaffer. 1977. The reproductivebiology of Agavaceae: I. Pollen and nectar production infour Arizona agaves. Southwestern Naturalist 22:157–168.

Slauson, L. A. 2000. Pollination biology of two chiropter-ophilous agaves in Arizona. American Journal of Botany87:825–836.

Soberón-Mainero, J., and C. Martı́nez del Rı́o. 1985. Cheat-ing and taking advantage in mutualistic associations. Pages192–216 in D. H. Boucher, editor. The biology of mutu-alism. Croom Helm, London, UK.

Stratford, M., and P. A. Anslow. 1998. Evidence that sorbicacid does not inhibit yeast as a classic ‘‘weak acid preser-vative.’’ Letters in Applied Microbiology 27:203–206.

Thakar, J. D., K. Kunte, A. K. Chauhan, A. V. Watve, andM. G. Watve. 2003. Nectarless flowers: ecological corre-lates and evolutionary stability. Oecologia 136:565–570.

Tomás-Barberán, F. A., F. Ferreres, F. Tomas-Lorente, and A.Ortiz. 1993. Flavonoids from Apis mellifera beeswax. Zeit-schrift für Naturforschung C 48:68–72.

Spe

cial

Feat

ure


Towner, H. F. 1977. The biosystematics of Calylophus (On-agraceae). Annals of the Missouri Botanical Garden 64:48–120.

Wäckers, F. L., and C. Bonifay. 2004. How to be sweet?extrafloral nectar allocation by Gossypium hirsutum fits op-timal defense theory predictions. Ecology 85:1512–1518.

Waser, N. M. 1979. Pollinator availability as a determinantof flowering time in ocotillo (Fouquieria splendens). Oec-ologia 39:107–121.

Watanabe, N., S. Watanabe, R. Nakajima, J.-H. Moon, K. Shi-mokihara, J. Inagaki, H. Etoh, T. Asay, K. Sakata, and K.Ina. 1993. Formation of flower fragrance compounds fromtheir precursors by enzymic action during flower opening.Bioscience, Biotechnology and Biochemistry 57:1101–1106.

Weatherwax, P. B. 1986. Why do honeybees reject certainflowers? Oecologia 69:567–570.

Weidenhamer, J. D., F. A. Macias, N. H. Fischer, and G. B.Williamson. 1993. Just how insoluble are monoterpenes?Journal of Chemical Ecology 19:1799–1807.

Weiss, M. R. 2001. Vision and learning in some neglectedpollinators: beetles, flies, moths and butterflies. Pages 171–190 in L. Chittka and J. D. Thomson, editors. Cognitiveecology of pollination; animal behavior and floral evolu-tion. Cambridge University Press, Cambridge, UK.

Williams, A. A., T. A. Hollands, and O. G. Tucknott. 1981.The gas chromatographic–mass spectrometric examinationof the volatiles produced by the fermentation of a sucrosesolution. Zeitschrift für Lebensmitteluntersuchung undForschung 172:377–381.

Young, H. J. 1988. Differential importance of beetle speciespollinating Dieffenbachia longispatha (Araceae). Ecology69:832–844.

APPENDIX A

A table summarizing literature search results for papers relevant to scent compounds in nature is available in ESA’sElectronic Data Archive: Ecological Archives E085-039-A1.

APPENDIX B

A table showing scent compounds identified from floral and nectar headspace using SPME-GC-MS is available in ESA’sElectronic Data Archive: Ecological Archives E085-039-A2.

1495

Spec

ialFeatu

re


EXTRAFLORAL NECTAR AS A RESOURCE MEDIATINGMULTISPECIES INTERACTIONS

JENNIFER A. RUDGERS1,3 AND MARK C. GARDENER2,4

1Center for Population Biology, One Shields Avenue, University of California, Davis, California 95616 USA2Ecology and Evolution Research Group, Department of Biological Sciences, The Open University, Walton Hall,

Milton Keynes, MK7 6AA UK

Abstract. Extrafloral (EF) nectar resources can affect the dynamics of species inter-actions at the community scale. Furthermore, selection acting on EF nectary traits mayextend beyond simple mutualisms between plants and the enemies of herbivores to involveother community members that use EF nectar. We examine how EF nectaries influence andare influenced by interactions with multiple species, highlighting our review with originaldata from the association between ant

ecol 85 609.1477 1478 - Cornell University1480 NICO BLU¨THGEN AND KONRAD FIEDLER Ecology, Vol. 85,...

Documents

Transcript of ecol 85 609.1477 1478 - Cornell University1480 NICO BLU¨THGEN AND KONRAD FIEDLER Ecology, Vol. 85,...