Plant Ecology 169: 161–170, 2003. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.
161
Root herbivores, pathogenic fungi, and competition between Centaurea maculosa and Festuca idahoensis Wendy L. Ridenour and Ragan M. Callaway* Division of Biological Sciences, University of Montana, Missoula, Montana 59812, USA; *Author for correspondence (e-mail:
[email protected]; phone: 406–243-5077; fax: 406–243-4184) Received 9 October 2001; accepted in revised form 21 June 2002
Key words: Agapeta zoegana, Allelopathy, Biological control, Centaurea maculosa, Compensatory growth, Competition, Festuca idahoensis, Indirect interactions, Interference, Sclerotinia sclerotiorum Abstract We used a common garden experiment to evaluate the isolated and combined effects of a biocontrol agent, the insect (Agapeta zoegana, Lepidoptera), and a native North American fungal pathogen (Sclerotinia sclerotiorum) on competition between the noxious weed Centaurea maculosa and the native Festuca idahoensis. In 0.5-m 2 plots with 24 plants per plot, competition between Centaurea and Festuca was highly asymmetrical, with Centaurea strongly reducing the final biomass and reproduction of Festuca, and Festuca having no effect, or possibly a positive effect, on Centaurea. The direct effects of the biological control agents differed entirely. All Centaurea individuals died in plots receiving Sclerotinia, but Agapeta did not significantly reduce the growth of Centaurea, and apparently stimulated a compensatory reproductive response in the weed. Individual Centaurea plants that had been damaged by Agapeta produced more flowerheads, and the number of Centaurea plants with Agapeta root damage in a plot was positively correlated with total Centaurea biomass. These differences in the direct effects of the consumers were reflected in their indirect effects. In plots where Sclerotinia killed Centaurea (strong direct effects) Festuca growth and reproduction was equal to that in Festuca plots without Centaurea and the reproductive output of Festuca increased substantially (strong indirect effects). However, in the absence of Sclerotinia, the application of Agapeta did not significantly decrease Centaurea biomass (weak direct effects) and actually stimulated small, but significant decreases in Festuca reproduction and trends towards lower Festuca biomass (weak and opposite indirect effects – Agapeta does not eat Festuca). If the direct effects of biocontrol agents on Centaurea are weak, as suggested by our results, natives are unlikely to be released from the competitive effects of Centaurea, and natives may suffer from Centaurea’s compensatory response to herbivory. Introduction A large body of research has demonstrated that herbivory can alter interactions among plants (Louda et al. 1990; Clay 1990; Crawley 1992). Plants experiencing herbivory are typically at a competitive disadvantage, an observation which provides the theoretical rationale for importing large numbers of exotic insect herbivores to control exotic, highly competitive invasive plants. However, there have been relatively few controlled experimental studies of the effects of consumers on competition between exotic and native plants in general (Clay 1990; Clay et al. 1993; Calla-
way et al. 1999), and most studies of the effects of biological control agents on the growth of their target species or inter-plant interactions have focused on a single biocontrol species. In natural and biocontrolweed systems, however, the additive effects of multiple consumers can be much greater than their isolated effects (Charudattan 1986; Berlow 1999), and the assumption that multiple biocontrol agents can do what one agent cannot is widespread (Harris 1984; Goeden and Andrés 1999). However, the effects of multiple species cannot always be predicted from paired experiments (Adler and Morris 1994; Kareiva 1994; Wootton 1994). Van der Putten et al. (2001)
162 have emphasized the importance of linking multitrophic interactions of plants, herbivores, and pathogens in order to understand community process. Intermountain grassland of western Montana is a good system in which to investigate the direct and indirect effects of biocontrols. A diverse native flora exists, but intermountain grassland has been extensively invaded by exotic perennials, including Centaurea maculosa Lam. which was introduced into North America from Eurasia (Müller-Schärer and Schroeder 1993; Sheley et al. 1998). Centaurea maculosa is among the most widespread and destructive grassland invaders in the Western United States and Canada (Tyser and Key 1989; Griffith and Lucey 1991; Sheley and Jacobs 1997). The negative effects of Centaurea species on native plants are well documented (Muir and Majak 1983; Rice et al. 1992; Lesica and Shelley 1996). Near our experimental site C. maculosa appears to have reduced the diversity of native grassland species by more than 90% (Ridenour and Callaway 2001). Centaurea maculosa and the closely related C. diffusa may suppress natives via allelopathic effects (Muir and Majak 1983; Callaway and Aschehoug 2000; Ridenour and Callaway 2001), and competition for resources (Callaway and Aschehoug 2000). Despite more than 50 years of chemical control efforts and biological control introductions starting in 1970 (the seed attacking Urophora affınis) C. maculosa had spread onto > 2.5 million hectares in the Northwest by the mid-1980’s (Chicoine et al. 1985) and the number of counties in the region reporting C. maculosa increased from 39 in 1974 to 133 in 1994 (Rice 1994). At least 12 insect species have now been introduced to North America from Eurasia in an effort to increase “cumulative stress” (i.e. Harris (1984)) and reduce the competitive effects of C. maculosa on North American natives (Sheley et al. 1998). Additionally, a fungus species native to the intermountain prairie, Sclerotinia sclerotiorum, has been shown to damage C. maculosa and benefit native Pseudoroegneria spicata grasses (Jacobs et al. 1996). Therefore, the intermountain prairie has native competitors, native pathogens that damage Centaurea, and introduced insect herbivores that may interact with Centaurea in complex ways. Studies of complex interactions among C. maculosa, competitors, and consumers in its native range in Europe have provided valuable insight into the natural ecology of this species (Müller 1989; Müller-Schärer 1991; Müller and Steinger 1990; Steinger and Müller-Schärer 1992;
Weiner et al. 1997). However, these studies provide little evidence that C. maculosa is strongly limited by herbivory in its natural range. Quantitative studies of interactions in invaded North American communities are crucial to formulating realistic evaluations and expectations for controlling Centaurea species in intermountain prairie. Here we report on the results from a common garden experiment (see also Callaway et al. (1999)) designed to study: 1) the direct effects of C. maculosa on F. idahoensis, 2) the isolated and combined indirect effects of Sclerotinia and Agapeta on F. idahoensis, and 3) the isolated and combined direct effects of two consumers (the fungus Sclerotinia sclerotiorum and Agapeta zoegana, Lepidoptera) and one native competitor (Festuca idahoensis) on C. maculosa.
Methods We conducted the common garden experiment at The University of Montana Diettert Experimental Gardens. These gardens occupy land once covered by intermountain grassland, and are near natural intermountain grasslands that have been heavily invaded by C. maculosa. Centaurea maculosa and F. idahoensis plants were started from seed in March 1994 in a greenhouse and transplanted into the garden plots in May 1994. The experimental design consisted of six treatments, each of which was replicated ten times for a total of 60 0.25-m 2 plots that were randomly located with respect to each other in the experimental area (see Callaway et al. (1999)). Treatments were: 1) Centaurea alone, no consumers, 2) Festuca alone, no consumers 3) Centaurea with Festuca, no consumers 4) Centaurea with Festuca and Agapeta, 5) Centaurea with Festuca and Sclerotinia, and 6) Centaurea with Festuca, Agapeta, and Sclerotinia. Each plot was 0.50 m from any other plot, and within each plot, individual plants were located 10 cm from all neighbors. Twenty-four individuals were grown in each plot. In “control” (1 & 2) plots, all 24 individuals were either C. maculosa or F. idahoensis. In “treatment” (3,4,5, & 6) plots, 12 Centaurea and 12 Festuca were planted such that individual plants alternated by species in rows and columns of a grid. This design permitted two scales at which we could measure treatment effects; that of the individual plant, and that of the combined 12 target plants in the entire 0.25-m 2 plot (reported as plot biomass). In other words, because the two “control” monoculture plots
163 had 24 conspecific individuals we only weighed the combined biomass of the 12 individuals that occurred in the same locations within the plots as their 12 conspecifics in the plots with Centaurea and Festuca planted together. This standardized all measured plants so that they had the same number of neighbors. Biomass in the 0.25 m 2 plots was converted to g/m 2. To investigate the effects of treatments on soil water availability, PVC monitoring tubes were installed 30-cm deep in the center of each of the 60 garden plots. Soil moisture was measured using Frequency Domain Reflectometry (Troxler, Sentry 200-AP) once per week at 15-cm and 30-cm depths from April to late October 1994 and April to late October 1995. Two biocontrol treatments were applied, both singly and in combination. These biocontrols were the lepidopteran root-mining herbivore Agapeta zoegana and the fungal pathogen Sclerotinia sclerotiorum. Agapeta was introduced into the United States from Eurasia, where both Agapeta and C. maculosa are native, for the biocontrol of Centaurea species. Agapeta is highly host specific to several closely related species of Centaurea. Adult insects deposit their eggs at the root-shoot interface of Centaurea plants, and the larvae mine Centaurea taproots. We acquired Agapeta from the Montana State University Agricultural Experimental Station in Corvallis, Montana. In August 1994, twenty plots with mixtures of Centaurea and Festuca were treated with Agapeta to determine whether or not root herbivory on Centaurea would shift the balance of competition in favor of Festuca. After a period of plant establishment and growth, fine-mesh cages were placed over all plots and three adult Agapeta were introduced into each of the ten Agapeta treatment plots and the ten Agapeta-Sclerotinia plots. Cages were placed over the other 40 plots that were not treated with Agapeta to control for the effects of caging. After ten days, the time allowed for Agapeta egg-laying, cages were removed from all plots. Sclerotinia is a soil-borne fungus native to the Northern Rocky Mountains and was acquired from David Sands, Montana State University. In August, two days after applying Agapeta, ten of the plots that were planted with mixtures of Centaurea and Festuca and treated with Agapeta were also infected with Sclerotinia by applying 2 cm 3 of Sclerotinia-infected grain to the base of each Centaurea stem, and ten other plots were treated with Sclerotinia without Agapeta. Several hours before applying Sclerotinia, plots were watered manually to promote establish-
ment, and plots were watered for several days afterwards. In September of 1995, volumetric measurements (maximum height, basal diameter, and crown diameter) and number of green leaves (fall green-up) of the four Festuca plants nearest the center in each plot were collected. At the scale of individuals, we nondestructively measured the growth of the four interior F. idahoensis and C. maculosa individuals, thus controlling for the number of interspecific neighbors, edge effect, and intraspecific density. At the wholeplot scale, all plots were harvested at the end of the growing season in September 1995, and aboveground biomass was dried at 60 °C and weighed. Although we did not measure belowground biomass because of difficulty inherent to collecting the fine fragile roots of Centaurea, we pulled out the taproots of Centaurea, measured their diameters and searched them for signs of Agapeta damage. We also counted the number of total Festuca florets on the 12 target individuals per plot (see first paragraph of methods) and Centaurea flowerheads per individual and total on the 12 target individuals per plot to quantify reproductive output.
Results The total aboveground biomass of the 12 target individuals in the Festuca control plots was more than two times greater than conspecifics in the Centaurea × Festuca treatment plots (One-way ANOVA, F treatment = 4.32, df = 5,59, P = 0.003, post-ANOVA Tukey HSD, P < 0.01), (Figure 1). Festuca floret production was 10–20 times greater in Festuca control plots than in Centaurea × Festuca treatment plots (One-way ANOVA, F treatment = 8.34, df = 5,59, P < 0.001 postANOVA Tukey HSD, P < 0.001; Figure 2). The interfering effect of Centaurea on Festuca, however, did not appear to be caused by decreased soil water (see Callaway et al. (1999)). Soil moisture content at 15 and 30 cm depths did not vary significantly between treatments throughout the summer in either year (Repeated-measures ANOVA, F treatment × time at 15 cm = 0.93, df = 5,59, P = 0.606; F treatment × time at 30 cm = 1.04, df = 5,59, P = 0.412). Likewise, the interfering effect of Centaurea on Festuca was not associated with reduced levels of soil nutrients. Available soil nitrate, ammonium and phosphorous did not vary significantly between treatments during the first year of the study (One-way ANOVA, NO 3: F treatment
164
Figure 1. Above-ground biomass of Festuca idahoensis in a common garden experiment either in a matrix of conspecifics, a matrix of Centaurea maculosa, and for the interspecific matrix with the biocontrols Agapeta zoegana and Sclerotinia sclerotiorum applied alone or in combination. For the conspecific matrix we only included the biomass of the 12 individuals that were in the same locations occupied by F. idahoensis in the interspecific matrices. Error bars represent one standard error, and letters designate means that were significantly different in post-ANOVA Tukey HSD tests, P < 0.05).
= 0.511, df = 5,59, P = 0.767; NH 4: F treatment = 0.446, df = 5,58, P = 0.814; P: F treatment = 1.20, df = 5,59, P = 0.321). In contrast to the strong negative effect of Centaurea on Festuca, Centaurea plants grown with Festuca competitors were almost 4 times larger than Centaurea’s grown with conspecific competitors (Figure 3). The direct effects of the two biocontrol agents differed markedly (Figure 3). All of the Centaurea plants in both Sclerotinia treatments (Centaurea with Festuca with Sclerotinia and Centaurea with Festuca with Agapeta and Sclerotinia) died within two weeks of application. However, Centaurea biomass in Centaurea x Festuca with Agapeta treatments was not less than when the biocontrol larvae were absent (Figure 3). Examination of the taproots of the 12 focal Centaurea plants in the ten plots with Centaurea, Festuca, and Agapeta found that a mean of 4.0 ± 0.7 (1 s.e.) Centaurea individuals (30%) had signs of root damage. Elimination of Centaurea by Sclerotinia had strong positive effects on the growth of Festuca (Figures 1 and 2). By October 1994, only two months after application of the fungal pathogen, individual Festuca plants in the Sclerotinia treatment plots had similar basal diameters and heights as those in the Festuca controls (data not shown). By October 1995, aboveground Festuca biomass in the Sclerotinia treat-
Figure 2. Floret production of Festuca idahoensis in a common garden experiment either in a matrix of conspecifics, a matrix of Centaurea maculosa, and for the interspecific matrix with the biocontrols Agapeta zoegana and Sclerotinia sclerotiorum applied alone or in combination. For the conspecific matrix we only included the biomass of the 12 individuals that were in the same locations occupied by F. idahoensis in the interspecific matrices. Error bars represent one standard error, and letters designate means that were significantly different in post-ANOVA Tukey HSD tests, P < 0.05).
Figure 3. Above-ground biomass of Centaurea maculosa in a common garden experiment either in a matrix of conspecifics, a matrix of Festuca idahoensis, and for the interspecific matrix with the biocontrols Agapeta zoegana and Sclerotinia sclerotiorum applied alone or in combination. For the conspecific matrices we only included the biomass of the 12 individuals that were in the same locations occupied by C. maculosa in the interspecific matrices. Error bars represent one standard error, and letters designate means that were significantly different in post-ANOVA Tukey HSD tests, P < 0.05).
ment plots was 50–100% greater, and floret production 100% greater, than that of the total biomass of the comparable 12 focal individuals in the Festuca control plots where conspecific competition was more intense (Figures 1 and 2). Festuca plots were not
165
Figure 4. The relationship between root diameter and flower number for individual Centaurea maculosa plants in the ten plots with Centaurea, Festuca, and Agapeta. The upper regression line represents individuals with signs of root damage and the lower regression line represents individuals with no sign of root damage. The large symbols represent the means and 95% confidence limits for both root diameter (X axis) and flower number (Y axis). For individuals with no root damage, r 2 = 0.17, P = 0.006; for individuals with root damage, r 2 = 0.11, P = 0.003.
Figure 5. Biomass of the 12 focal Centaurea maculosa individuals as a function of the percent of the total number of Centaurea (24) plants damaged by Agapeta zoegana. P < 0.05.
Discussion more productive when Agapeta were applied to Centaurea neighbors, and may have actually been hurt by the response of Centaurea to mild root herbivory. The total number of Festuca florets on the 12 target plants decreased significantly in plots where Agapeta was applied to Centaurea (post-ANOVA Tukey HSD, P = 0.038, Figure 2). A similar trend, although not significant, was also seen for Festuca biomass (Figure 1). The negative response of Festuca to herbivory on Centaurea may have been due to a compensatory response by Centaurea. Root diameters of individual Centaurea plants with Agapeta damage were larger than those of plants without root damage, and more importantly, flower production per unit of root diameter on damaged individuals was almost twice that of undamaged individuals (Figure 4). For the ten plots with Centaurea, Festuca and Agapeta, Centaurea biomass was positively correlated with the number of Centaurea plants that had been damaged by Agapeta (Figure 5). Neither Festuca biomass nor floret number was significantly correlated with the number of Centaurea plants with Agapeta damage (r biomass = −0.11, P = 0.32; r floret = −0.22, P = 0.14, data not shown).
The fungal pathogen Sclerotinia eliminated Centaurea maculosa from experimental plots and substantially increased the biomass and reproduction of Festuca idahoensis. In contrast to the effect of Sclerotinia, root herbivory by Agapeta during the relatively short time period of our experiment did not have strong negative direct effects on Centaurea, and did not reduce the aboveground biomass of C. maculosa. In fact, for a similar root diameter, individuals with Agapeta damage had greater flowerhead production, suggesting reproductive over-compensation (Paige and Whitham 1987) and not simply Agapeta choosing larger plants. Furthermore, plots with higher numbers of Centaurea plants with Agapeta damage also had higher total aboveground biomass. These increases in Centaurea flower production and plot biomass with Agapeta damage suggest either a compensatory response or a shift in allocation with herbivoryinduced stress. However, because of the experimental design we cannot eliminate the possibility that the difference was due to the indirect effects of Festuca. The root diameters of damaged plants were also significantly larger than those of undamaged plants; however, this may have been due to the preference of Agapeta for larger host plants (see Story et al. (2000)).
166 A number of other studies support our finding that C. maculosa has a remarkable capacity to tolerate herbivory and defoliation, and under some conditions, may overcompensate (i.e. grow larger or reproduce more after herbivory [Paige and Whitham (1987) and Belsky et al. (1993), Trumble et al. (1993)]). Müller (1989) found that C. maculosa plants of German and Canadian origin increased fine root growth when infected by Agapeta and did not decrease in fecundity. Steinger and Müller-Schärer (1992) found that the biomass of Centaurea maculosa seedlings grown in pots were not affected by Agapeta feeding, and attributed the lack of effect to compensatory root growth. In other experiments, however, the root feeding weevil, Cyphocleonus achates reduced whole-plant biomass. In field experiments in Switzerland, MüllerSchärer (1991) found that low levels of Agapeta herbivory increased survival, shoot number, and fecundity of Centaurea maculosa, but the effects of herbivory were highly complex and were negative under other conditions. Callaway et al. (1999) reported (in addition to other analyses of some of the same field experiments reported here) that Centaurea plants experiencing leaf herbivory from Trichoplusia ni (cabbage looper) were stronger competitors against Festuca idahoensis. Kennett et al. (1992) found that defoliation of potted C. maculosa (up to 4 times in ⬇ 6 months and up to intensities of 75% of the leaves) had no effect on the final biomass of the defoliated plants. However, leaf defoliation decreased C. maculosa carbohydrate concentrations and pools substantially (Lacey et al. 1994). In our experimental conditions the effects of Agapeta on Centaurea were weak. However, repeated infections over a longer period of time, higher infection levels on individual plants, or more stressful abiotic conditions might result in completely different outcomes. Stronger doses of Agapeta, either through the eventual buildup of populations or local adaptation, would be expected to have stronger effects. However, the proportion of C. maculosa plants infested with Agapeta (30%) was within the range of proportions found in invasive populations in North America and natural populations in Europe. The infestation level in our experiment was also comparable to that at an Agapeta release site after 6–7 years (31.8%; Story et al. (2000)), and higher than that reported in some natural populations of C. maculosa in eastern Europe (15–36%; Müller et al. (1988)). Virtually all ecological interactions are conditional, or context-specific, and conditional effects of biological controls should
be expected. However, we should not exclude the possibility that the effects of Agapeta may be even weaker in some conditions. Our contrasting effects of Sclerotinia and Agapeta are confounded by different application intensities. We were able to apply substantial doses of Sclerotinia to Centaurea, but as noted only 30% of the Centaurea plants in Agapeta treatments were infected. Therefore a strict quantitative comparison is not justified, but 100% mortality for Sclerotinia compared to 0% mortality versus no significant effect of Agapeta on Centaurea growth does not require a strict quantitative comparison. Although little is known about the natural distribution of Sclerotinia in the field, our applications were undoubtedly far higher than Centaurea would experience in nature. Correspondingly, the efficacy of Sclerotinia as a biological control agent has been highly limited by the spread of the fungus and difficulties inherent to delivering a lethal dose. Story et al. (2000) compared C. maculosa biomass and reproduction at one field site where Agapeta had been released to one site where they had not been released. They found that Centaurea plants at release sites were smaller and had fewer flowers, but there were no data collected prior to Agapeta release. Furthermore, Agapeta were not excluded from the control site and by the final sampling date the percentages of Centaurea plants infested by Agapeta did not differ between the sites. Interestingly, individual Centaurea plants that were infested with Agapeta had 40% more flowerheads and 112% more aboveground biomass, a pattern that was attributed to the preference of Agapeta for larger plants. Our experiments (Figure 4), however, indicate that even when the size (root diameter) of Centaurea is accounted for, damage by Agapeta increases flower production substantially. Such compensatory growth may explain the slightly higher competitive effects of C. maculosa on F. idahoensis in the Agapeta treatment (also see Callaway et al. (1999)). The slightly higher competitive effects of C. maculosa on F. idahoensis in the Agapeta treatment were significant, but small in magnitude. Therefore, as a biological process the higher competitive effects after herbivory appear to be of little consequence relative to the competitive strength of C. maculosa in general. However, understanding the mechanisms behind such a counterintuitive response may provide new insight into complex interactions among plants and herbivores. Slightly higher competitive effects may have
167 been due to compensatory resource uptake, or an induction of greater production of defensive secondary metabolites that also functioned as allelopathic chemicals. Centaurea maculosa contains the chemical cnicin, a sesquiterpene lactone that has been implicated in both anti-herbivore and allelopathic interactions (Kelsey and Locken 1987; Landau et al. 1994), but to our knowledge there is no evidence that cnicin defenses are inducible. Others have found dual anti-herbivore/allelopathic roles in inducible plant metabolites, which increase under stress (Lovett and Holt (1995) and Tang et al. (1995), Siemens et al., in review). Root observation chamber experiments suggest that C. maculosa may be allelopathic (Ridenour and Callaway 2001). In these experiments Festuca roots grew more slowly near Centaurea roots than when near conspecific roots, and activated carbon, which absorbs organic compounds, reduced the negative effects of Centaurea roots. This indicates that Centaurea may be allelopathically interfering with its neighbors. Other studies indicate that C. maculosa and C. diffusa outcompete North American natives for resources (Callaway and Aschehoug 2000), but in our experiments the interfering effect of C. maculosa on F. idahoensis was not clearly manifest through depletion of soil resources. A third hypothesis for Centaurea’s compensatory response to herbivory and slightly higher competitive effects on Festuca involves mycorrhizae. Centaurea maculosa and other Centaurea species may benefit from a form of mycorrhizae-mediated parasitism through common mycorrhizal networks (Grime et al. 1987; Marler et al. 1999; Carey and Callaway (1999, 1999)) or a shift in the relative abundance of mutualistic and pathogenic fungi in the presence of Festuca so that Centaurea is favored (Callaway et al. (in press)). Carey and Callaway (1999) found that the stable carbon isotope concentration of C. maculosa shoot tissue was significantly more similar to that of F. idahoensis in the presence of mycorrhizae than without mycorrhizae, indicating that carbon was transferred from the Festuca to the Centaurea. In contrast, experiments conducted by Zabinski et al. (unpublished) suggest that AM-mediated neighbor effects are the result of mycorrhizal networks that increase species’ access to phosphorus. In general, the role of carbon transfer among plants via AM fungi remains controversial (Robinson and Fitter 1999). In contrast to the weak competitive effect of the North American native Festuca idahoensis, MüllerSchärer (1991) found that competition from the Eu-
ropean native, Festuca pratensis, was the single most important factor influencing the success of Centaurea maculosa in field experiments, and the most important factor influencing the effect of Agapeta herbivory on Centaurea. He also found that “in the absence of grass competition, Agapeta herbivory showed no significant impact on plant height, biomass, and fecundity”. The finding that Festuca pratensis, commonly associated with Centaurea in Europe, is a much stronger competitor with Centaurea than with F. idahoensis, may explain some of the contradictory results between our study and Müller-Schärer’s and suggests that escape from natural consumers (the theoretical basis for biocontrols) is not the only mechanism driving the success of this weed in North America. Further support for the role of plant communities themselves in determining the success of invaders was published by Callaway and Aschehoug (2000) who found that Eurasian grass species had much stronger competitive effects on Centaurea diffusa, a close relative of C. maculosa, than closely related North American grass species. These differences appeared to be due to different responses to allelopathic root exudates from C. diffusa. Their results indicated that the success of some invasive exotics might not be due only to escaping consumers, but also to novel competitive mechanisms not previously experienced by indigenous species. In a review of studies of the effects of natural enemies and competitors, Sheppard (1996) reported that the dominant factor in 10 of 12 studies in natural grasslands was competition. However, unlike we observed for C. maculosa, most effects of competitors and natural enemies were “multiplicative”; each factor had an impact but without an interaction. Classical biological control can be defined as the introduction of enemies of invasive, exotic plants in order to control their spread (Harris 1991). Many biological control efforts involve the release of one biological control agent (usually an insect herbivore, Wurtz (1995) and Julien (1992)), and some of these have been successful (Huffaker et al. 1961; Cullen et al. 1973; Kok and Surles 1975). Our study clearly demonstrates the efficacy of a single fungal pathogen. However, species abundances are often regulated by factors other than consumers. Even though our experiments found exceptionally strong negative direct and positive indirect effects of Sclerotinia on a native grass, Agapeta had weak to positive (compensatory response) direct effects and negative indirect effects. These results are not desired of a biocontrol and em-
168 phasize the importance of achieving strong negative effects of biocontrols on target species in order to achieve strong indirect positive effects of biocontrols on native plants. In the case of Centaurea maculosa, other researchers have shown that large reductions in the weed’s abundance are required to shift the balance of competition in favor of native grasses (Sheley and Jacobs 1997), which corroborates our finding of weak biocontrol effects. Considered together, the compensatory responses of C. maculosa to biocontrol herbivory and defoliation shown in this study and in a number of others, and the ability of C. maculosa to outcompete native grasses even after large reductions in its abundance, suggest that criteria applied prior to introducing biological control species for this weed should include convincing experimental evidence that the biocontrols will have strong direct effects on their hosts.
Acknowledgements We thank Jim Story for providing Agapeta and David Sands for providing Sclerotinia. Many thanks are due to Steve Baker and Jim Plummer for their valuable assistance with soil nutrient analysis and contribution of laboratory space; Jennifer Costich, Michael Wojdylak, Paul Ridenour, Todd Wojtowicz, Erik Aschehoug and Raven Stevens for their assistance during the course of the research; and Dr Colin Henderson for helpful discussions, especially regarding statistical analysis. This study was funded by The University of Montana Grant Program, a grant to Ragan Callaway and Cathy Zabinski from the National Science Foundation, DEB-9726829, and a grant to Ragan Callaway from the Andrew W. Mellon Foundation.
References Adler F.R. and Morris W.F. 1994. A general test for interaction modification. Ecology 75: 1552–1559. Belsky A.J., Carlson W.P., Jensen C.L. and Fox G.A. 1993. Overcompensation by plants - herbivore optimization or red herring? Evolutionary Ecology 7: 109–121. Berlow E.L. 1999. Strong effects of weak interactions in ecological communities. Nature 398: 330–334. Binkley D. and Vitousek P. 1991. Soil nutrient availability. In: Pearcy R.W., Ehleringer J.R., Mooney H.A. and Rundel P.W. (eds), Plant Physiological Ecology. Chapman and Hall, London, pp. 75–96.
Callaway R.M. and Aschehoug E.T. 2000. Invasive plants versus their new and old neighbors: a mechanism for exotic invasion. Science 290: 521–523. Callaway R.M., DeLuca T. and Belliveau W.M. 1999. Herbivores used for biological control may increase the competitive ability of the noxious weed Centaurea maculosa. Ecology 80: 1196– 1201. Callaway R.M., Newingham B., Zabinski C.A. and Mahall B.E. Compensatory growth and competitive ability of and invasive weed are enhanced by soil fungi and native neighbors. Ecology Letters (in press). Carey E.V. and Callaway R.M. 1999. Stable isotopes provide evidence for carbon transfer from the invasive weed Centaurea maculosa to a native bunchgrass. Annual Meeting of the Ecological Society of America., Spokane, Washington, USA. Charudattan R. 1986. Integrated control of water hyacinth (Eichhornia crassipes) with a pathogen, insects, and herbicides. Weed Science 34: 26–30. Chicoine T.K., Fay P.K. and Nielson G.A. 1985. Predicting weed migration from soil and climate maps. Weed Science 34: 57– 61. Clay K. 1990. The impact of parasitic and mutualistic fungi on competitive interactions among plants. In: Grace J.B. and Tilman D. (eds), Perspectives on Plant Competition. Academic Press Inc, San Diego, USA, pp. 391–412. Clay K., Marks S. and Cheplick G.P. 1993. Effects of insect herbivory and fungal endophyte infection on competitive interactions among grasses. Ecology 74: 1767–1777. Crawley M.J. 1992. Natural Enemies: the Population Biology of Predators, Parasites, and Diseases. Blackwell Science, Oxford, UK. Cullen J.M., Kable P.F. and Catt M. 1973. Epidemic spread of a rust imported for biological control. Nature 244: 462–464. Goeden R.D. and Andrés L.A. 1999. Biological control of weeds in terrestrial and aquatic environments. In: Bellows T.S. and Fisher T.W. (eds), Handbook of Biological Control. Academic Press, New York, pp. 871–890. Griffith D. and Lucey J.R. 1991. Economic evaluation of spotted knapweed (Centaurea maculosa) control using picloran. Journal of Range Management 44: 43–47. Grime J.P., Mackey J.M.L., Hillier S.H. and Read D.J. 1987. Floristic diversity in a model system using experimental microcosms. Nature 328: 420–422. Harris P. 1984. Current approaches to the biological control of weeds. In: Kelleher J.S. and Hulme M.A. (eds), Biological Control Programmes Against Insects and Weeds in Canada 1969–1980. Commonwealth Agricultural Bureau, Farnham Royal, England. Harris P. 1991. Invitation paper (C.P. Alexander Fund): classical biocontrol of weeds: its definition, selection of effective agents, and administrative, political problems. Canadian Entomologist 123: 827–849. Huffaker C.B., Ricker D. and Kennett C. 1961. Biological control of puncture vine with imported weevils. California Agriculture 15: 11–12. Jacobs J.S., Sheley R.L. and Maxwell B.D. 1996. Effect of Sclerotina sclerotiorum on the interference between bluebunch wheatgrass (Agropyron spicatum) and spotted knapweed (Centaurea maculosa). Weed Technology 10: 13–21.
169 Julien M.H. 1992. Biological control of weeds: a world catalogue of agents. Commonwealth Agricultural Bureau International, Wallingford, Oxon, UK. Kareiva P. 1994. Higher order interactions as a foil to reductionist ecology. Ecology 75: 1527–1528. Kelsey R.G. and Locken R.J. 1987. Phytotoxic properties of cnicin, a sesquiterpene lactone from Centaurea maculosa (spotted knapweed). Journal Chemical Ecology 13: 19–33. Kennett G.A., Lacey J.R., Butt C.A., Olson-Rutz K.M. and Haferkamp M.R. 1992. Effects of defoliation, shading and competition on spotted knapweed and bluebunch wheatgrass. Journal of Range Management 45: 363–369. Kok L.T. and Surles W.W. 1975. Successful biological control of musk thistle by an introduced weevil. Environmental Entomology 4: 1025–1027. Lacey J.R., Olson-Rutz K.M., Haferkamp M.R. and Kennet G.A. 1994. Effects of defoliation and competition on total nonstructural carbohydrates of spotted knapweed. Journal of Range Management 47: 481–484. Landau I., Müller-Schärer H. and Ward P.I. 1994. Influence of cnicin, a sesquiterpene lactone of Centaurea maculosa (Asteraceae) on specialist and generalist insect herbivores. Journal Chemical Ecology 20: 929–942. Lesica P. and Shelley J.S. 1996. Competitive effects of Centaurea maculosa on the population dynamics of Arabis fecundis. Bulletin Torrey Botanical Club 123: 111–121. Lovett J.V. and Holt A.C.H. 1995. Allelopathy and self-defense in barley. In: Derjit K.M.M., Dakshini and Einhellig F.A. (eds), Allelopathy. American Chemical Society, Washington, DC, USA, pp. 170–183. Louda S.M., Keeler K.H. and Holt R.D. 1990. Herbivore influences on plant performance and competitive interactions. In: Grace J.B. and Tilman D. (eds), Perspectives on Plant Competition. Academic Press Inc, San Diego, pp. 413–444. Marler M.J., Zabinski C.A., and Callaway R.M. 1999. Mycorrhizae indirectly enhance competitive effects of an invasive forb on a native bunchgrass. Ecology 80: 1180–1186. Miller T.E. 1994. Direct and indirect species interactions in an early old-field plant community. The American Naturalist 143: 1007– 1025. Muir A.D. and Majak W. 1983. Allelopathic potential of diffuse knapweed (Centaurea diffusa) extracts. Canadian Journal of Plant Science 63: 989–996. Müller H. 1989. Growth pattern of diploid and tetraploid spotted knapweed, Centaurea maculosa Lam. (Compositae) and effects of the root mining moth Agapeta zoegana (L.) (Lepidoptera:Cochylidae). Weed Research 29: 103–111. Müller H., Schroeder D. and Gassmann A. 1988. Agapeta zoegana (L.) (Lepidoptera:Cochylidae), a suitable prospect for biological control of spotted and diffuse knapweed, Centaurea maculosa Monnet de la Marck and Centaurea diffusa Monnet de la Marck (Compositae) in North America. Canadian Entomologist 120: 109–124. Müller-Schärer H. 1991. The impact of root herbivory as a function of plant density and competition: survival, growth and fecundity of Centaurea maculosa in field plots. Journal Applied Ecology 28: 759–776. Müller-Schärer H. and Schroeder D. 1993. The biological control of Centaurea spp. in North America: do insects solve the problem? Pesticide Science 37: 343–353.
Müller H. and Steinger 1990. Separate and joint effects of root herbivores, plant competition, and nitrogen shortage on resource allocation and components of reproduction in Centaurea maculosa (Compositae). In: Szentesi A. (ed.), Proceedings of the 7 th International Symposium on Insect-Plant relationships. Hungarian academy of Sciences., Budapest, pp. 215– 224. Paige K.N. and Whitham T.G. 1987. Overcompensation in response to mammalian herbivory: the advantage of being eaten. American Naturalist 129: 407–416. Rice P.M., Bedunah D.J. and Carlson C.E. 1992. Plant community diversity after herbicide control of spotted knapweed. US Forest Service Res. Paper. USDA. Rice P.M. 1994. Invaders Database. University of Montana, Missoula, Montana, USA. Ridenour W.M. and Callaway R.M. 2001. The relative importance of allelopathy in interference: the effects of an invasive weed on a native bunchgrass. Oecologia 126: 444–450. Robinson D. and Fitter A. 1999. The magnitude and control of carbon transfer between plants linked by a common mycorrhizal network. Journal Experimental Botany 50: 9–13. Sheley R.L. and Jacobs J.S. 1997. “Acceptable” levels of spotted knapweed (Centaurea maculosa control. Weed Technology 11: 363–368. Sheley R.L., Jacobs J.S. and Carpinelli M.F. 1998. Distribution, biology, and management of diffuse knapweed (Centaurea diffusa) and spotted knapweed (Centaurea maculosa). Weed Technology 12: 353–362. Sheppard A.W. 1996. The interaction between natural enemies and interspecific plant competition in the control of invasive pasture weeds. In: Moran V.C. and Hoffman J.H. (eds), Proceedings of the IX Inernational Symposium on Biological Control of Weeds., Stellenbosch, South Africa, pp. 19–26. Siemens D., Garner S., Mitchell-Olds T. and Callaway R.M. 2002. The cost of defense in the context of competition. Brassica rapa may grow and defend. Ecology 83: 505–517. Steinger T. and Müller-Schärer H. 1992. Physiological and growth responses of Centaurea maculosa (Asteraceae) to root herbivory under varying levels of interspecific competition and soil nitrogen availability. Oecologia 91: 141–149. Story J.M., Good W.R., White L.J. and Smith L. 2000. Effects of the interaction of the biocontrol agent Agapeta zoegana (L.) (Lepidoptera:Cochylidae) and grass competition on spotted knapweed. Biological Control 17: 182–190. Tang C., Cai W., Kohl K. and Nishimoto R.K. 1995. Plant stress and allelopathy. In: Derjit K.M.M., Dakshini and Einhellig F.A. (eds), Allelopathy. American Chemical Society, Washington, DC, USA, pp. 142–157. Trumble J.T., Kolodny-Hirsch D.M. and Ting I.P. 1993. Plant compensation for herbivory. Annual Review Entomology 38: 93– 119. Tyser R.W. and Key C.H. 1989. Spotted knapweed in natural area fescue grasslands: an ecological assessment. Northwest Science 62: 151–160. Van der Putten W.H., Vet L.E.M., Harvey J.A. and Wäckers F.L. 2001. Linking above- and belowground multitrophic interactions of plants, herbivores, pathogens, and their antagonists. Trends in Ecology and Evolution 16: 547–554.
170 Weiner J., Martinez S., Müller-Schärer H., Stoll P. and Schmid B. 1997. How important are environmental maternal effects in plants? A study with Centaurea maculosa. Journal of Ecology 85: 133–142. Wootton J.T. 1994. Putting the pieces together: testing the independence of interactions among organisms. Ecology 75: 1544–1551.
Wurtz T.L. 1995. Domestic geese: biological weed control in an agricultural setting. Ecological Applications 5: 570–578. Zabinski C.A., Quinn L. and Callaway R.M. Phosphorus uptake, not carbon transfer, explains arbuscular mycorrhizal enhancement of Centaurea maculosa in the presence of native grassland species. Functional Ecology (unpublished).
