The effect of floral herbivory on male and female reproductive success in Isomeris arborea

Ecology, Jan, 1999 by Gary A. Krupnick, Arthur E. Weis

INTRODUCTION

Flower-feeding herbivores may consume but a small fraction of their host plant's biomass, but then can have large effects on host reproductive success (Breedlove and Ehrlich 1968, Louda 1982, Zammit and Hood 1986, Schemske and Horvitz 1988, Wallace and O'Dowd 1989, Bertness and Shumway 1992, English-Loeb and Karban 1992, Cunningham 1995). Herbivore consumption of pistils, ovaries, or ovules will have an immediate and direct effect on reproduction through female function (Zammit and Hood 1986, Schemske and Horvitz 1988, Wallace and O'Dowd 1989, Bertness and Shumway 1992, English-Loeb and Karban 1992, Pellmyr and Huth 1994). Similarly, direct anther and pollen consumption may potentially diminish male function. But floral herbivores may have additional, indirect impacts on plant reproductive potential. Attack can degrade flower appearance, reduce the quantity and quality of floral rewards, and decrease the size of the floral display. Such changes in floral signals may affect pollinator visitation rates (Karban and Strauss 1993, Cunningham 1995, Krupnick et al. 1999), which in turn can alter patterns and rates of pollen donation (male function) and pollen receipt (female function). Thus, the total negative effect of herbivore damage can be realized through several channels.

The negative impact of general herbivory on male reproductive success has only recently been explored (Allison 1990a, Quesada et al. 1995, Mutikainen and Delph 1996, Strauss et al. 1996). Floral herbivory can have particularly negative impacts due to its potential effect on pollinator response to the plant. For instance, Karban and Strauss (1993) found a drop in bee visitation to Erigeron glaucus flowers with thrips damage. Lower visitation to damaged flowers or inflorescences will lower the probability that the surviving pollen reaches receptive stigmas. For every pollen grain destroyed, several others may go to waste.

Reduced visitation can also reduce female reproductive success if seed set is pollen limited. Pollen limitation is common among plants (Burd 1994), but may vary at several spatial or temporal scales (Schemske 1977, Snow 1986, Bertness et al. 1987, Campbell 1987, Lubbers and Lechowicz 1989, Ackerman and Montalvo 1990, Johnston 1991, Dudash 1993). When pollinator service falls, fertilization rates may fall below the level supportable by the plant's resource base.

Finally, the herbivore's effect on pollinator behavior may alter plant reproductive success by changing the selfing rate, either by increasing autogamous pollen transfer or decreasing geitonogamous pollen transfer. Previous studies indicate that individuals that are more attractive to pollinators can have higher outcrossing rates (Brown and Clegg 1984, Epperson and Clegg 1987, Sun and Ganders 1990, Rausher et al. 1993) and that the lack of pollinator visits can increase autogamy in self-compatible plant species (Motten 1982, Schoen 1982, Kron et al. 1993, Rathcke and Real 1993). Hence, the reduced attractiveness of damaged plants may lead to higher selfing rates. Conversely, if an undamaged plant produces a large floral display, pollinators may visit more flowers per inflorescence per bout. This could result in greater geitonogamous pollen transfer (e.g., De Jong et al. 1992, Robertson 1992, Hodges 1995), causing outcrossing rates to drop (e.g., Barrett et al. 1994, Harder and Barrett 1995). In populations subject to inbreeding depression, such changes in selfing rate can influence offspring success.

We have explored the mechanisms through which a flower-feeding insect, Meligethes rufimanus, can affect components of male and female reproductive success in a perennial shrub, Isomeris arborea, both through direct destruction of gametes and indirect effects exerted through pollinator behavior. Previous work (Krupnick and Weis 1998) has shown that beetle damage increases flower bud abortion rate and changes sex allocation in I. arborea. A companion study of floral herbivory and pollinator behavior (Krupnick et al. 1999), finds that pollinator visitation rates decrease as a result of florivory.

This paper describes a two-year study that measured the effect of reduced pollinator services through effects on pollen export and seed production. Both naturally occurring and planted shrubs were subjected to one of two treatments: exposure to herbivore activity (most flowers damaged), and protection from herbivore activity (few flowers damaged) by application of a shortlived systemic insecticide. Various components of male and female reproductive function were analyzed by investigating the following questions: (1) Does floral herbivory indirectly reduce the export of undamaged pollen grains? (2) Do damaged plants receive fewer pollen grains than undamaged plants? (3) Does herbivory decrease fruit production? (4) Does reduced pollinator service to damaged plants result in pollen limitation? and (5) Do outcrossing rates change as a result of florivory?

METHODS

Study system

I. arborea (Capparaceae) is an endemic southern California perennial shrub. This drought-deciduous plant typically flowers from January until November (Munz 1974:330). It is self-compatible and andromonoecious; hermaphroditic and male flowers are produced sequentially on the same terminal inflorescences. Male flowers produce six stamens and a nonfunctioning, undeveloped pistil, and hermaphroditic flowers produce six stamens and a pistil with a superior ovary extended on a gynophore. The shrub is visited by flying insects, including bumblebees (Bombus spp.), non-native honeybees (Apis melifera), and hummingbirds (Calypte costae and C. anna). While it is not known which visitor is the most effective pollinator, bumblebees are a likely candidate (Grant and Grant 1967), since they contact all reproductive parts during a visit to a flower and are more common than other pollinating species (G. A. Krupnick, personal observation).