Ecology, 85(6), 2004, pp. 1503±1511 q 2004 by the Ecological Society of America THE DUAL ROLE OF FLORAL TRAITS: POLLINATOR ATTRACTION AND PLANT DEFENSE REBECCA E. IRWIN,1,2,5 LYNN S. ADLER,3 AND ALISON K. BRODY2,4 1Institute of Ecology, University of Georgia, Athens, Georgia 30602 USA 2Rocky Mountain Biological Laboratory, Crested Butte, Colorado 81224 USA 3Department of Biology, Virginia Tech, Blacksburg, Virginia 24060 USA 4Department of Biology, University of Vermont, Burlington, Vermont 05405 USA Abstract. Plants are under siege from a diversity of enemies that consume both leaf and ¯oral parts. Plants resist damage to leaves in a variety of ways, and we now have a rich literature documenting how plants defend themselves against herbivore attack. In con- trast, the mechanisms by which plants resist enemies that consume ¯oral parts or resources are much less known, even though damage to ¯oral tissue usually has tighter links to plant ®tness than damage to leaf tissue. Many plants experience nectar robbing, whereby ¯oral visitors remove nectar from ¯owers, often without pollinating. Nectar robbers can reduce plant ®tness to degrees comparable to, or even surpassing, reduction by herbivores. How- ever, because nectar attracts both pollinators and nectar robbers, plants face a dilemma in defending against nectar robbers without also deterring pollinators. Here, we extend the conceptual framework of resistance to herbivores to include resistance to nectar robbers, focusing on nectar traits. We review published data and ®nd that an array of nectar traits may deter robbers without deterring pollinators. Although resistance traits against robbers have been broadly identi®ed, the costs and bene®ts of these traits in terms of plant ®tness remain poorly understood. We present data showing that a nectar trait (dilute nectar) might Special Feature directly, as well as indirectly, bene®t plant ®tness by deterring nectar-robbing bumble bees of Ipomopsis aggregata without deterring hummingbird pollinators. However, the magni- tude of any plant ®tness bene®t will depend on the degree to which plants are pollen- vs. resource-limited in a given year. The results of our work offer both conceptual and empirical insight into how plants cope with attack by nonpollinating ¯oral visitors through a relatively unexplored trait, nectar. Key words: Bombus occidentalis; bumble bee; herbivory; Ipomopsis aggregata; nectar concen- tration; nectar robbing; plant defense; pollination; resistance. INTRODUCTION to resist enemy attack may be more dif®cult when mu- Plants are attacked by a variety of organisms that tualists and antagonists use the same resource, such as feed in different ways and on different types of tissue. nectar. Organisms that feed on nectar without providing At any given time, a plant may face damage by her- pollination service in return (henceforth referred to as bivores feeding on leaves or ¯owers, while others feed ``nectar robbers'' or ``nectar thieves'') can have sub- on roots, and still others consume sap, phloem, or xy- stantial effects on male and female plant ®tness (re- lem. The defenses deployed against such enemies can viewed in Irwin et al. 2001). A meta-analysis of ex- be constitutive or induced (Bryant et al. 1988, Karban isting studies suggests that ¯oral larceny has an overall and Meyers 1989). The defenses can be quite general detrimental effect on female plant reproduction (Irwin and effective against many organisms (Krischik et al. et al. 2001), although individual studies ®nd that nectar 1991, Omer et al. 2001), or highly speci®c and effective robbing can have a continuum of effects on plant re- against a single type or species of herbivore (Agrawal production, from positive to neutral to negative (Ma- and Karban 2000). By mounting defenses against en- loof and Inouye 2000). Nectar robbers can reduce plant emies, plants increase their resistance to further feeding ®tness to degrees comparable to, or even surpassing, and, presumably, minimize the ®tness costs of damage ®tness reduction by herbivores (Juenger and Bergelson (Karban and Myers 1989). 1997, Irwin and Brody 2000). Damage to ¯oral tissue Resistance is the ability of plants to decrease the can have stronger links to plant ®tness than does dam- frequency of attack (Painter 1958). The ability of plants age to leaf tissue (reviewed in Strauss et al. 2003); nectar robbing can reduce female ®tness by 50% (e.g., Manuscript received 9 June 2003; revised 2 September 2003; Irwin and Brody 2000, Lara and Ornelas 2001) and accepted 8 September 2003. Corresponding Editor: A. A. Agra- estimates of male ®tness by up to 80% (Irwin and Bro- wal. For reprints of this Special Feature, see footnote 1, p. 1477. dy 1999). Yet, the mechanisms by which plants resist 5 Address for correspondence: Institute of Ecology, Ecol- ogy Building, University of Georgia, Athens, Georgia 30602 nectar robbers are less well known than mechanisms USA. E-mail: [email protected] associated with resistance to herbivores. This lack of 1503 1504 REBECCA E. IRWIN ET AL. Ecology, Vol. 85, No. 6 understanding is surprising for a number of reasons. sexual organs of ¯owers (Inouye 1980). For simplicity, First, nectar robbing is extremely common among ¯ow- we include all ¯oral larcenists (both nectar robbers and ering plants (Irwin and Maloof 2002). Most plant spe- thieves) under the term ``nectar robber'' except in cases cies with tubular corollas or ¯owers with nectar spurs in which we describe scenarios only applicable to nec- experience some form of ¯oral larceny. Moreover, the tar thieves. mechanisms by which nectar robbers affect plant ®tness Although one might ascribe adaptive function to are analogous to those of herbivores. These include traits that deter nectar robbers, it is likely that many, direct mechanisms, such as damage to ¯oral reproduc- if not most, of the resistance traits against robbers that tive organs (Galen 1983, Adler et al. 2001), as well as we will describe are exaptations (Armbruster 1997) and indirect mechanisms, such as changes in pollinator be- thus did not evolve in response to selection by robbers, havior (Strauss et al. 1996, Irwin and Brody 2000). per se. Traits that confer resistance to robbers most Concepts of resistance have been applied to other an- likely represent traits involved in one set of interactions tagonistic interactions besides herbivory, such as pre- (interactions between plants and pollinators or inter- dation, providing a unifying theory across taxa (e.g., actions between plants and herbivores) that have been Agrawal et al. 1999, BroÈnmark et al. 1999). Therefore, co-opted into other, very different interactions. Many it seems likely that the concept of resistance can be of these traits may also re¯ect pleiotropic effects be- applied to understand how plants cope with nectar rob- tween traits that confer resistance to herbivores and bers in cases in which robbers negatively affect plant resistance to nectar robbers. ®tness. Resistance to nectar robbing, however, presents Toxins and chemical deterrents plants with a conundrum not experienced in preventing Plant compounds associated with herbivore resis- attack by leaf-feeding herbivores. The nature of this tance (secondary compounds or toxins) are surprisingly quandary lies in providing nectar as a reward for mu- common in ¯oral nectar (Baker 1977, Adler 2000). Al- tualist pollinators. Pollinator attraction is critical to the though the bene®ts of ``toxic nectar'' are debated (Ad- reproductive success of the majority of ¯owering plants ler 2000), a few examples suggest that toxic nectar may (reviewed in Kearns et al. 1998). However, the payment reduce the frequency of nectar robbing. In Catalpa spe- for attracting pollinators can be costly, both in the ciosa, nectar containing iridoid glycosides deterred amount of resources invested (Pyke 1991) and in at- nectar-thieving ants but did not deter legitimate bee tracting nectar robbers. Thus, how might plants attract pollinators (Stephenson 1981, 1982). In Gelsemium pollinators while avoiding nectar robbers? sempervirens, experimental manipulation of alkaloids The aim of this contribution is to extend the existing in nectar deterred nectar-robbing carpenter bees, but at conceptual framework of resistance to herbivores to a cost of reduced visitation by legitimate bee polli- include nectar robbers. We review the literature on re- nators (L. S. Adler and R. E. Irwin, unpublished man- Special Feature sistance to robbers, highlighting the subtle but impor- uscript). Petal tissue can also contain secondary com- tant similarities between resistance to herbivores and pounds (Euler and Baldwin 1996). Guerrant and Fiedler robbers. We then present new data and suggest future (1981) suggest that corolla-tube damage by ants on a areas of research in this ®eld. Although this work is variety of tropical species can release secondary com- focused on resistance to nectar robbers, the predictions pounds from petal tissue into the nectar. Secondary and results presented are applicable and functionally compounds thus released may serve to protect the similar to other plant±animal interactions in which plants from nectar-robbing or nectar-thieving ants, al- plants face trade-offs in attracting mutualists and an- though this has never been explicitly tested. These stud- tagonists, such as plant interactions with seed dispers- ies demonstrate that secondary compounds may deter ers and seed predators (Herrera 1982, Cipollini and robbers; however, the bene®ts of toxic nectar will