The consequences of floral herbivory for pollinator service to Isomeris arborea

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

INTRODUCTION

Herbivores that feed upon floral structures reduce the reproductive success of many plant species (Breedlove and Ehrlich 1968, Inouye 1982, Zammit and Hood 1986, Schemske and Horvitz 1988, Evans et al. 1989, Wallace and O'Dowd 1989, Pettersson 1991, English-Loeb and Karban 1992, Muenchow and Delesalle 1992, Cunningham 1995). It is reasonable to believe that these reductions are primarily due to the direct loss of gametes. Floral herbivory, however, may hinder plant reproduction indirectly by reducing pollinator service. When herbivores attack flowers, their damage can (1) degrade the advertising properties of each flower, (2) reduce the number of flowers per display, and (3) lower pollinator rewards. Pollinators are known to respond to variation in floral morphology: Studies utilizing either natural variation in flower size (Galen and Newport 1988, Stanton and Preston 1988, Galen 1989, 1996, Young and Stanton 1990, Campbell et al. 1991, Mitchell 1992) or experimental manipulation of petal length (Johnson et al. 1995) have shown that pollinators prefer larger flowers. Intraspecific variation in color (Waser and Price 1981) and scent (Galen 1985) also influences visitation. Pollinators not only respond to attributes of the flowers themselves, but also to the number of flowers, with larger inflorescences usually being the more attractive (although in many of these cases the proportion of flowers visited actually declines as flower number increases [reviewed in Snow et al. 1996]). Flowers rich in nectar (Zimmerman 1983, 1988, Galen and Plowright 1985, Thomson 1988, Campbell et al. 1991, Real and Rathcke 1991, Mitchell 1993, 1994, Hodges 1995) and other pollen rewards (Harder 1990, Cresswell and Robertson 1994) also can receive more visits than flowers with lower levels. Given this sensitivity in pollinators to floral state, external agents that alter display or reward can alter the plant's chance of being serviced (Strauss 1997). Herbivores can be such agents (Karban and Strauss 1993, Strauss et al. 1996, Lehtila and Strauss 1997).

We have examined the effects of flower predation on pollinator service and reproductive success in Isomeris arborea (Capparaceae), a shrub whose flowers are attacked by a pollen beetle, Meligethes rufimanus (Coleoptera: Nitidulidae). This beetle alters floral advertisement: Petals and anthers wilt and the corolla does not completely open. In this paper, we quantify damage to display and rewards, determine if damage reduces service, and ask at what scale (e.g., flower, plant, or neighborhood) pollinators respond to damage. Reduced pollinator service can in turn lead to reduced pollen export and receipt, which reduces reproductive success through male and female function, respectively. A companion paper (Krupnick and Weis 1999) examines the differential consequences of reduced pollinator activity for male and female components of reproductive success.

METHODS

Study system

I. arborea is a Southern California drought-deciduous perennial shrub (Munz 1974:330) that can flower from January until November if water is available. The species has a disjunct range, with both Mojave desert populations and coastal populations ranging from Santa Barbara, California, to Northern Baja California, Mexico. This self-compatible species is andromonoecious, with hermaphroditic and male flowers produced on the same inflorescence. Bisexual flowers produce six stamens and a superior ovary extended on a gynophore; male flowers produce six stamens and a nonfunctioning, undeveloped pistil. Visitors to the yellow flowers include bumblebees (Bombus spp.), nonnative 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 (G. A. Krupnick, personal observation).

I. arborea is host to the pollen beetle, M. rufimanus, which feeds on capers and crucifers in northern temperate regions. In coastal California, adult beetles feed on pollen from developing and mature I. arborea flowers between January and June. Females oviposit in developing flower buds. The resulting larvae mature within the buds, where they consume developing anthers. Most damaged buds abort (Krupnick 1996), but some flowers succeed in blooming, with a reduced number of functional anthers and possible damage to the ovary wall. At the end of the third instar, the beetle larvae fall to the ground to pupate.

Study plots and damage manipulation

We evaluated beetle damage and pollinator response in both natural and planted arrays of I. arborea within the Ecological Reserve at the University of California, Irvine. The natural population consists of over 200 I. arborea shrubs growing on a hill ([approximately]75 m in diameter) with other coastal sage scrub species (e.g., Artemisia californica, Eriogonum fasciculatum, and Encelia californica). We manipulated damage levels through insecticide applications. During 1992, 60 plants within this population received one of three treatments: protection from natural levels of herbivory ("protected"), which results in low levels of damage; exposure to herbivory with a water application ("exposed"), which results in high levels of damage and is a sham-control for spraying; or exposure to herbivory without a water application ("exposed without spray"), which results in high levels of damage. Protected plants were sprayed with a systemic insecticide (Dimethoate CA267, obtained from Platte Chemical Company, Fremont, Nebraska) dissolved in water (29.5 mL/L); exposed plants were sprayed with water only. Both insecticide spray and water were applied at the rate of 40 mL per plant, once every two weeks, between January and June from 1993 to 1995. M. rufimanus is the only herbivore that feeds upon I. arborea during this time period (G. A. Krupnick, personal observation), and thus it is the only herbivore species being removed by the insecticide treatment. The insecticide treatment has no effect on inflorescence or fruit production of I. arborea in a herbivore-free environment (Krupnick 1996).