Home | Clemson University, South Carolina

5 downloads 321809 Views 753KB Size Report
Tyler Geer,a Clemson wildlife and fisheries biology graduate student, leads a ... Recent Clemson graduate uncovers a link between environmental toxicants and ...
Ecology, 85(2), 2004, pp. 471–483 q 2004 by the Ecological Society of America

NATURAL-ENEMY RELEASE FACILITATES HABITAT EXPANSION OF THE INVASIVE TROPICAL SHRUB CLIDEMIA HIRTA SAARA J. DEWALT,1 JULIE S. DENSLOW,2

AND

KALAN ICKES

Department of Biological Sciences, Louisiana State University, Baton Rouge, Louisiana 70803 USA

Abstract. Nonnative, invasive plant species often increase in growth, abundance, or habitat distribution in their introduced ranges. The enemy-release hypothesis, proposed to account for these changes, posits that herbivores and pathogens (natural enemies) limit growth or survival of plants in native areas, that natural enemies have less impact in the introduced than in the native range, and that the release from natural-enemy regulation in areas of introduction accounts in part for observed changes in plant abundance. We tested experimentally the enemy-release hypothesis with the invasive neotropical shrub Clidemia hirta (L.) D. Don (Melastomataceae). Clidemia hirta does not occur in forest in its native range but is a vigorous invader of tropical forest in its introduced range. Therefore, we tested the specific prediction that release from natural enemies has contributed to its expanded habitat distribution. We planted C. hirta into understory and open habitats where it is native (Costa Rica) and where it has been introduced (Hawaii) and applied pesticides to examine the effects of fungal pathogen and insect herbivore exclusion. In understory sites in Costa Rica, C. hirta survival increased by 12% if sprayed with insecticide, 19% with fungicide, and 41% with both insecticide and fungicide compared to control plants sprayed only with water. Exclusion of natural enemies had no effect on survival in open sites in Costa Rica or in either habitat in Hawaii. Fungicide application promoted relative growth rates of plants that survived to the end of the experiment in both habitats of Costa Rica but not in Hawaii, suggesting that fungal pathogens only limit growth of C. hirta where it is native. Galls, stem borers, weevils, and leaf rollers were prevalent in Costa Rica but absent in Hawaii. In addition, the standing percentage of leaf area missing on plants in the control (water only) treatment was five times greater on plants in Costa Rica than in Hawaii and did not differ between habitats. The results from this study suggest that significant effects of herbivores and fungal pathogens may be limited to particular habitats. For Clidemia hirta, its absence from forest understory in its native range likely results in part from the strong pressures of natural enemies. Its invasion into Hawaiian forests is apparently aided by a release from these herbivores and pathogens. Key words: abundance; biological control; Clidemia hirta; Costa Rica; enemy-release hypothesis; fungal pathogens; habitat distribution; Hawaii; herbivory; invasive species; neotropical shrub.

INTRODUCTION Insect herbivores and fungal pathogens can limit plant abundance through depression of plant growth, survival, or reproduction. Numerous studies of temperate and tropical plants have found that herbivores affect individual plants by decreasing lifetime fitness (Louda 1982a, b, Doak 1992, Louda and Potvin 1995, Louda and Rodman 1996, Root 1996) and growth (Marquis 1984, 1992, Aide and Zimmerman 1990, Schierenbeck et al. 1994), whereas fungal pathogens can cause substantial leaf damage (Coley and Barone 1996) and are implicated causal mechanisms of numerous tree seedling fatalities (Augspurger 1984, Augspurger and Manuscript received 18 November 2002; revised 12 May 2003; accepted 3 June 2003. Corresponding Editor: E. Menges. 1 Present address: Department of Ecology and Evolutionary Biology, Rice University, MS 170, Houston, Texas 77005 USA. E-mail: [email protected] 2 Present address: USDA Forest Service, Institute of Pacific Islands Forestry, 23 East Kawili Street, Hilo, Hawaii 96720 USA.

Kelly 1984, Stanosz 1994). Whether or not natural enemies affect plant demography likely depends on the type and extent of damage, life history traits of the plant species, and availability of resources (Whitham et al. 1991). The demographic consequences of insect herbivory and fungal pathogen attack may be conditional on the habitat in which the plant is growing (Whitham et al. 1991). Plants growing in adjacent habitats often experience different levels of herbivore damage (Huffaker and Kennett 1959, Harper 1969, Louda et al. 1987, Louda and Rodman 1996). Depending on the plant species, natural enemies may be more abundant or damaging in high light (e.g., Lincoln and Mooney 1984, Harrison 1987, Louda and Rodman 1996) or low light habitats (Maiorana 1981, MacGarvin et al. 1986, Denslow et al. 1990, Folgarait et al. 1995). Herbivores and pathogens may decrease individual growth rates and cause substantial mortality particularly in low light conditions where low carbon fixation rates can be further reduced by loss of leaf area and photosynthate.

471

472

SAARA J. DEWALT ET AL.

Chronic herbivore or pathogen attack could effectively exclude plants from such habitats. If natural enemies exhibit a strong selective pressure on plant abundance and habitat distribution, then species may increase dramatically in abundance and expand their habitat distribution in the absence of natural enemies. One such scenario occurs when species are introduced accidentally or intentionally to areas outside their native range. In this case, a nonnative plant’s natural enemies may be absent or scarce if generalist herbivores are rare in the area of introduction (as hypothesized for islands; Carlquist 1974), if specialist enemies are not introduced, and if host switching by native specialists does not occur (Keane and Crawley 2002). Generalist herbivores and pathogens present in the introduced range may prefer native to nonnative species and thus have limited impact on the nonnatives (Keane and Crawley 2002). In fact, several inventories of herbivores or pathogens have found fewer species and individuals attacking plants where they are introduced compared to where they are native (e.g., Bossard and Rejma´nek 1994, Szentesi 1999, Fenner and Lee 2001, Wolfe 2002, Mitchell and Power 2003), but few studies have examined the consequences of reduced enemy load in both areas. The release of plant populations from control by natural enemies in areas of introduction has been hypothesized to contribute to observed increases in plant abundance (Elton 1958, Crawley 1987, Fowler et al. 1996) and invasions of habitats in which species do not occur in their native range. This is called the enemy-release hypothesis (Keane and Crawley 2002). Classical biological control is predicated on the underlying assumptions that herbivores and pathogens limit population growth in native areas and that introduction of these enemies will therefore limit population growth in introduced areas. However, these assumptions have not been directly tested (Callaway et al. 1999). We used the tropical woody shrub Clidemia hirta (L.) D. Don (Melastomataceae) as a model organism to examine the role of natural enemies in areas of its native and introduced ranges. Clidemia hirta is native to lowlands of Central and South America and Caribbean Islands where it colonizes naturally and anthropogenically disturbed, relatively open areas such as pastures, riversides, roadsides, and tree plantations. However, in its native range, C. hirta apparently does not occur in old-growth forests (Cook 1929, Wester and Wood 1977). Clidemia hirta is naturalized now throughout the tropics including several islands in the Pacific and Indian Oceans, Peninsular Malaysia, the Indian subcontinent, and eastern Africa (Tanzania) (Wester and Wood 1977, Rejma´nek 1996, Strahm 1999). The introduction of C. hirta around the world likely was accidental (Simmonds 1933, Wester 1992). In its introduced range, C. hirta is abundant and invades open areas, as it does in its native range, as well as

Ecology, Vol. 85, No. 2

gaps and understory of old-growth forest, where it does not occur in its native range (Smith 1992, Rejma´ nek 1996, Strahm 1999). Thus, C. hirta generally is weedy in open, disturbed habitats in both its native and introduced ranges, but appears to become more abundant in such habitats and appears more shade tolerant in its introduced range (Wester and Wood 1977). In this study, we conducted a natural-enemy exclusion experiment in replicated field sites in part of C. hirta’s native range (Costa Rica) and introduced range (Hawaii) to test whether there were differences in damage due to aboveground insect herbivores and fungal pathogens between the two areas and whether these differences affect growth and survival. Based on predictions of the enemy-release hypothesis, we expected to find high levels of herbivory on plants exposed to insect herbivores and fungal pathogens in the native range, but little to no damage in the introduced range. Given the lack of C. hirta in forest understory in Costa Rica but high abundance in both high- and low-light environments in Hawaii, we predicted enemy exclusion to have significant positive effects on plant growth and survival in the native but not introduced range, and to have greater positive effects in understory than open habitats in the native range. METHODS

Study species Clidemia hirta is a densely branching woody shrub that grows to a height of 2–3 m and occurs in mesic to wet areas from sea level to about 1500 m in both native and introduced ranges (Wester and Wood 1977). Clidemia hirta was first reported within the Hawaiian archipelago on Oahu in 1941 (Anonymous 1954) and spread in the early 1970s and 1980s to the islands of Kauai, Maui, Molokai, Lanai, and Hawaii (Smith 1992). It has been declared a noxious weed by the Hawaii Department of Agriculture, and since 1953 seven biological control agents have been introduced in an attempt to limit its spread (Nakahara et al. 1992; P. Conant, personal communication). At the time of our study, only Liothrips urichi Karny (Phlaeothripidae, a thrips which attacks terminal leaves and internodes), Lius poseidon Napp (Buprestidae, a leaf-mining beetle), and Colletotrichum gloeosporioides f.s. clidemiae (Melanconiaceae, a fungal pathogen of leaves) were established on the island of Hawaii, where this experiment was conducted. Liothrips urichi was first introduced in Hawaii in the 1950s and has reduced C. hirta populations in open habitats, but apparently has not affected populations in forest understory because the thrips lay eggs preferentially in the open (Reimer and Beardsley 1989). Lius poseidon, although established, is not widespread on the island of Hawaii, whereas Colletotrichum gloeosporioides f.s. clidemiae purportedly is established on all five Islands (P. Conant, personal communication). Several nonnative generalist in-

ENEMY RELEASE OF INVASIVE SHRUB

February 2004

FIG. 1.

473

Location of the three field sites in (A) Costa Rica and (B) Hawaii.

sects introduced accidentally or to control other plants also feed on C. hirta in Hawaii. There are no reports of vertebrate herbivores of C. hirta in either its native or introduced range.

Study sites At each site in the native and introduced areas, we planted C. hirta in paired open and forest understory habitats. Most habitat pairs were within 100 m of each other, and all were within 400 m of one another. Experimental sites in Costa Rica were located in the northeastern Caribbean lowlands at the La Selva Biological Station, El Bejuco Biological Station and adjoining pasture, and the Escuela de Agricultura de la Regio´ n Tropical Hu´meda (EARTH; Fig. 1A). Clidemia hirta occurred naturally at all three sites: wild populations were noted adjacent to experimental plots at El Bejuco and La Selva, and within 1 km of the EARTH site. Light levels in lowland forest understory in Hawaii likely are higher, on average, than in Costa Rica. Thus, understory sites in Costa Rica were placed in understory habitats with slightly greater light levels than found in old-growth forests. Understory sites at La Selva and EARTH were located in plantations of native and introduced trees. The plantations were not managed actively and understories of native plants were well developed at both sites. The understory site at El Bejuco was in secondary forest at least 20 years old. Open sites were abandoned pastures dominated by ferns at La Selva, and grass at El Bejuco and EARTH. Mean annual rainfall and temperature at La Selva are 4000 mm and 268C (Sanford et al. 1994). Soils are of volcanic origin and are ultisols or inceptisols. Experimental sites in Hawaii were located on the windward side of the island of Hawaii at the Waiakea Forest Reserve, University of Hawaii at Manoa (UHM) Agricultural Station, and Malama Ki Forest Reserve and Agricultural Experiment Station (Fig. 1B). Mean

annual rainfall and temperature in Hilo, which is close to the Agricultural Station and Waiakea sites, are 3300 mm and 238C. Soils at all Hawaiian sites are inceptisols. All understory sites were in lowland wet forest with native Metrosideros polymorpha Gaud. in the canopy. Clidemia hirta occurred at all three sites but was most common at Waiakea. Open sites at Waiakea and UHM were located in a power line right-of-way dominated by the pantropical fern Dicranopteris linearis (Burm.f.) Underw. and nonnative trees, shrubs, and grasses. The Malama Ki open site had been periodically mown and was dominated by nonnative grasses.

Experimental plants In Costa Rica, cuttings were taken from plants of four C. hirta populations in tree plantations, roadsides, and pastures. Collected material was trimmed to three nodes and a single small leaf or half of a large leaf and dipped in Daconil fungicide (active ingredient: chlorothalonil; ISK Biotech, Florida, USA). The first node was dipped in rooting hormone (Dip’N Grow 1.5 SL, Astoria Pacific, Clackamas, Oregon, USA) and the cutting placed in sand on a mist bench under neutraldensity shade cloth and clear plastic (15% incident radiation). After two weeks, rooted cuttings were transplanted to individual bags filled with alluvial soil and grown for two weeks under 15% incident radiation. Plants were 3.5–35.5 cm in total stem length when they were planted bare root at the three sites in May 1999. In Hawaii, small seedlings were collected from three populations along roadsides and in tree plantations on the eastern side of the island of Hawaii. We used seedlings rather than cuttings because cuttings proved difficult to root in Hawaii. Plants were kept in a solution of Vita Start (Lilly Miller, Clackamas, Oregon, USA) for two days to reduce transplant shock. Seedlings were planted in individual bags filled with Promix BX, which is a mixture of sphagnum moss, Vermiculite, and Per-

SAARA J. DEWALT ET AL.

474

lite (Premier Brands, Stamford, Connecticut, USA). The plants were kept on a mist bench within a greenhouse for two weeks. Seedlings were 6.5–33.0 cm in total stem length when outplanted in August 1999 with Promix included. By the end of the experiment, roots had grown beyond the potting mixture.

Exclusion of natural enemies We randomly assigned 624 plants each in Costa Rica and Hawaii to sites, habitat within each site (understory or open), and treatments within habitat (control, fungicide, insecticide, or dual application). Twenty-six seedlings or cuttings of each of the four treatments were planted within each habitat (104 plants per habitat per site). There were no initial differences in total stem length among treatments, habitat, or sites within each area (data not shown). Plants were positioned 1.5–2 m apart. In the open, aboveground vegetation was cleared prior to planting and periodically mowed or clipped around experimental plants throughout the experiment to create pasture-like conditions, to control for the effect of aboveground competition, and to ensure that plants received similar light levels. Understory vegetation was less dense and was not manipulated except that vines, fallen leaves, and branches were removed if they were touching experimental plants. Pesticides were used to exclude fungal pathogens and insect herbivores. Specific formulations of the pesticides differed in Costa Rica and Hawaii, but the same concentration of active ingredient was used in each area. Fungal pathogens were excluded by spraying plants with a 1.25% solution of the systemic fungicide benomyl [Methyl 1-(butylcarbamoyl)-2-benzimidazolecarbamate] (in Costa Rica, Benomil 50 WP, Helm AG, Hamburg, Germany; in Hawaii, Benlate 50 WP, DuPont, Wilmington, Delaware, USA). Our insecticide treatment contained a mixture of the systemic chloronicotinyl insecticide imidacloprid (1-[(6-Chloro-3-pyridinyl) methyl]-N-nitro-2-imidazolidinimine) to control sucking insects such as thrips and the synthetic, contact pyrethroid insecticide cyfluthrin (cyano(4-fluoro-3-phenoxyphenyl) methyl 3-(2,2-dichloroethenyl)-2,2-dimethylcyclopropanecarboxylate) to control chewing insects such as lepidopteran larvae. The percentage of the active ingredient after dilution was 17.4% for the imidacloprid and 0.00375% for the cyfluthrin. In Costa Rica these two chemicals were formulated as Confidor 70 WG and Baythroid 2.5 EC (Bayer AG, Leverkusen, Germany), respectively. In Hawaii they were Provado 1.6 F and Tempo 20 WP (Bayer, Kansas City, Missouri, USA). We attempted to exclude both insect herbivores and fungal pathogens (hereafter called ‘‘dual application’’) by applying a mixture of the fungicide and insecticides. Control plants were sprayed with water. Individual plants were sprayed with one of the four treatments using a hand sprayer approximately every two weeks. Stems and leaves were sprayed to the drip point so the total volume of solution applied was pro-

Ecology, Vol. 85, No. 2

portional to the size of the plant. Pesticide application was initiated two weeks after planting in Costa Rica and one month after planting in Hawaii. In December 1999 in Costa Rica, heavy rains occurred daily, and pesticides were not applied during that month. Care was taken to avoid spraying the ground because the fungicide has been shown to inhibit functioning of arbuscular mycorrhiza when used as a soil drench (Pedersen and Sylvia 1997). Translocation of benomyl from leaves to the active site of mycorrhizal infection in the roots has been found to be minimal (Larsen et al. 1996, Pedersen and Sylvia 1997). Some pesticides are phytotoxic under certain conditions, while others may stimulate growth through the addition of nutrients such as nitrogen (e.g., Brown et al. 1987, Paul et al. 1989, Root 1996). We assume the pesticides did not directly affect C. hirta growth in Costa Rica or Hawaii. This assumption was supported for the effect of Benlate fungicide, whose active ingredient, benomyl, contains four molecules of nitrogen, in a test using Hawaiian seedlings in a greenhouse study. We found that relative growth rates (total stem length) of C. hirta seedlings grown for eight months did not differ when Benlate or water only was sprayed every two weeks (F1,17 5 0.75, P 5 0.40). The cyfluthrin and imidacloprid insecticides, with one and five nitrogen molecules respectively, also likely had little effect on plant growth or survival as the amount of nitrogen applied was small and there were no differences in growth rates between plants sprayed with these insecticides and control plants (see under Results: Effects of natural-enemy exclusion on survival and growth). In addition, previous studies have shown that Costa Rican seedlings of related taxa are not limited by nitrogen availability (Vitousek and Denslow 1986, Denslow et al. 1998).

Measurements Total stem length was measured on all plants at the time of initial planting. Mortality during the first two months was attributed to transplant shock, and those plants were excluded from all analyses. Costa Rican plants were measured and harvested after 450 days. Hawaiian plants in the open sites were harvested ;300 days after planting because some individuals had become reproductive and removal of reproductive plants was a condition of our research permits. (Most plants in open sites in Costa Rica also were flowering by 300 days.) We continued the experiment with understory plants in Hawaii for 410 days because these plants did not become reproductive. Relative growth rates and survival patterns in understory plants in Hawaii were similar at 300 and 410 days (data not shown); therefore harvest data only are presented. At harvest, total stem length was measured on a randomly selected subset of plants in the open (N 5 5–8 plants per treatment per site) and all plants in the understory. Relative growth rate (RGR) was calculated

February 2004

ENEMY RELEASE OF INVASIVE SHRUB

as [ln(L1) 2 ln(L0)]/t, in which L0 and L1 were the total cumulative stem lengths in centimeters at time of planting and harvest, respectively, and t is the number of days between measurements (Evans 1972). RGR reflects the net effect of growth, loss to enemies, and damage from falling debris, on total stem length. Herbivory was assessed as proportion of leaf area missing measured at harvest. Following Coley and Barone (1996), we use the term ‘‘herbivory’’ to include damage due to both insects and pathogens. We did not follow individual leaves and therefore cannot estimate whole leaf loss. Estimates of leaf damage are therefore conservative (Lowman 1984, Filip et al. 1995). At harvest, leaf area and leaf area missing were determined using a LI-COR 3100 leaf area meter (LI-COR, Lincoln, Nebraska, USA). Proportion of leaf area missing was calculated from the difference between actual leaf area and total leaf area when holes were covered with opaque tape divided by total leaf area. Necrotic tissue was also treated as leaf area missing and was removed prior to measuring the actual leaf area. For open habitats, total leaf area and leaf area missing were measured on 20 randomly chosen leaves for 5–8 plants per treatment. For understory habitats, leaf area was determined for all plants alive at harvest at El Bejuco and La Selva and 12 plants per treatment at EARTH and each of the three Hawaiian sites. We also examined plants for insects such as thrips, stem borers, galls, and leaf rollers that caused damage not quantified by leaf area missing. The objective of the experiment was to examine relative effects of reduction in pest load in Costa Rica and Hawaii, not to account for differences in growth rates between the two areas. Several factors, such as the source of plants (cutting or seedling), soil fertility, climate, and forest structure, differed between the two areas and likely affected growth and survival. We therefore limit our analyses for these two variables to differences among treatments within areas. Nevertheless, we think the understory and open habitats were relatively similar between the sites in Costa Rica and Hawaii for two reasons. First, we evaluated understory light environments by measuring photon flux density (PFD) above .30 randomly chosen plants at each site with quantum sensors (LI-190SA; LI-COR, Lincoln, Nebraska, USA) mounted on a LAI-2000 canopy analyzer operated in the two-instrument mode. Abovecanopy PFD was estimated from measurements taken in a nearby clearing. Mean percentage of diffuse transmittance (PFD understory/PFD clearing 3 100) did not differ significantly between Hawaiian (2.6%) and Costa Rican sites (2.7%; F1,4 5 0.02, P 5 0.89). (Percentage transmittance was natural log-transformed for analysis to normalize residuals.) Thus, we expect that the difference between understory and open habitats did not differ significantly between the two areas. Second, we found indirect evidence that habitats were similar in that cutting and seedling responses to habitat differ-

475

ences were comparable. Relative growth rates of plants that survived to the final harvest were similar for both open and understory habitats between Costa Rica and Hawaii (understory: F1, 244 5 0.02, P 5 0.89; open: F1, 171 5 0.53, P 5 0.47). This is also evidence that the choice of cuttings or seedling likely did not affect the outcome of the experiment.

Data analysis All statistical analyses were conducted in SAS version 8 (SAS Institute 2000). The design within each area (Costa Rica and Hawaii) was a split plot with a completely randomized design at the within-plot (habitat) level. We examined the effects of habitat and natural enemies on percentage survival, relative growth rate (RGR), leaf area missing, number of leaves rolled, and presence of galls. ANOVA tests for percentage survival, RGR, percentage leaf area missing, and number of leaves rolled were performed using PROC MIXED with site and site 3 habitat treated as random effects and insecticide and fungicide treatments as fixed effects. Pesticide treatments were analyzed as a full factorial testing the two main effects of application of fungicide and insecticide and their interaction. Percentage survival per treatment per habitat per site was calculated and analyzed as a continuous variable with N 5 3 sites per habitat within each area. Percentage leaf area missing was square-root transformed to normalize residuals, but survival and RGR did not need to be transformed. Proportion of plants with galls per pesticide treatment was analyzed with PROC GENMOD using a binomial distribution and logit link function; variation due to site differences could not be examined in that analysis. As stated previously (Methods: Measurements), analyses of growth and survival were conducted separately within each area because of the methodological and abiotic differences between the areas. Analyzing growth and survival separately by area has the advantage of focusing on relative performance of plants among pesticide treatments and controlling for differences in overall performance of plants that could have arisen because of source of plants or other factors. The percentage leaf area missing on control plants additionally was compared between Costa Rica and Hawaii. We expected this measure to be affected less by source of plants and abiotic variables than survival and growth. RESULTS

Effects of natural-enemy exclusion on survival and growth In Costa Rica, part of the native range, survival of Clidemia hirta differed markedly between open and understory habitats and among treatments. Percentage survival to harvest of plants regardless of treatment was lower in the understory (44.7%) than in the open

SAARA J. DEWALT ET AL.

476

Ecology, Vol. 85, No. 2

TABLE 1. ANOVA tests of effects of habitat (open vs. understory), fungicide application, and insecticide application on survivorship, relative stem growth rate (RGR), and percentage leaf area missing on C. hirta in Costa Rica and Hawaii. Survivorship Source

df

F

Costa Rica Habitat (H) Fungicide (F) Insecticide (I) H3F H3I F3I H3F3I

1, 1, 1, 1, 1, 1, 1,

2 12 12 12 12 12 12

19.41 16.74 7.94 12.13 6.34 0.00 2.06

Hawaii Habitat (H) Fungicide (F) Insecticide (I) H3F H3I F3I H3F3I

1, 1, 1, 1, 1, 1, 1,

2 12 12 12 12 12 12

3.71 3.48 0.14 0.17 0.14 0.14 1.29

RGR

P

df

Leaf area missing

F

P

df

F

P

0.048* 0.002** 0.016* 0.005** 0.027* 0.947 0.176

1, 1, 1, 1, 1, 1, 1,

2 175 175 175 175 175 175

108.6 8.16 1.09 0.34 0.37 0.01 1.94

0.009** 0.005** 0.297 0.321 0.543 0.907 0.165

1, 1, 1, 1, 1, 1, 1,

2 176 176 176 176 176 176

1.16 3.43 1.97 0.00 1.94 0.55 0.38

0.395 0.066 0.162 0.967 0.165 0.461 0.540

0.194 0.087 0.711 0.689 0.711 0.711 0.277

1, 1, 1, 1, 1, 1, 1,

2 228 228 228 228 228 228

31.21 0.06 2.51 1.33 1.30 0.35 0.03

0.031* 0.804 0.115 0.249 0.255 0.554 0.855

1, 1, 1, 1, 1, 1, 1,

2 228 228 228 228 228 228

0.49 0.87 5.75 0.17 1.37 2.83 0.17

0.558 0.353 0.017* 0.679 0.243 0.094 0.679

* P , 0.05; ** P , 0.01.

in Costa Rica (91.3%; Table 1). Application of insecticide and fungicide significantly affected survivorship of C. hirta depending on the habitat (significant habitat 3 insecticide and habitat 3 fungicide interactions; Table 1). Survival was significantly higher when either insecticide or fungicide was sprayed on plants in the understory but not in the open habitats (Fig. 2A, B). Neither the second-order interaction of insecticide 3 fungicide nor the third-order interaction of habitat 3 insecticide 3 fungicide was significant for survivorship, suggesting that the effect of one type of pesticide did not depend on use of the other type of pesticide (Table 1). The effects of each pesticide treatment, therefore, were additive in the understory in Costa Rica. In relation to control plants in the understory, C. hirta survival increased by 12% if sprayed with insecticide, 19% with fungicide, and 41% with dual application.

Thus, both insect herbivores and fungal pathogens restricted C. hirta survival in one area of its native range, but only in understory habitats. Fungicide application had an overall positive effect on relative growth rate (RGR) on plants that survived to harvest regardless of habitat, but insecticide application had no effect on RGR in either habitat (Fig. 3A, C; Table 1). In Hawaii, part of the introduced range, C. hirta survival in both open and understory sites was high (Fig. 2B). We found no difference in percentage survival between open (98.7%) and understory sites (99.2%; Table 1). Pesticide application had no effect on survival or RGR of plants in either habitat in Hawaii (Figs. 2B and 3B, D; Table 1).

Natural-enemy damage We found that natural-enemy damage to leaves differed between Costa Rica and Hawaii (see Plate 1). At

FIG. 2. Survival of Clidemia hirta plants to harvest in understory and open habitats in (A) Costa Rica and (B) Hawaii in four natural-enemy exclusion treatments. Survivorship was higher in Costa Rican understory sites when plants were sprayed either with fungicide (F1,6 5 17.1, P 5 0.01) or insecticide (F1,6 5 8.7, P 5 0.03). No effect of either type of pesticide was found in the open in Costa Rica or in either habitat of Hawaii. Least-squares means 1 1 SE are shown.

February 2004

ENEMY RELEASE OF INVASIVE SHRUB

477

FIG. 3. Relative total stem growth rates of Clidemia hirta in understory and open habitats in Costa Rica and Hawaii in four natural-enemy exclusion treatments. Fungicide application significantly increased growth of plants in both habitats in Costa Rica. Relative growth rate (RGR) was calculated as [ln(L1) 2 ln(L0)]/t, where L0 and L1 are the total cumulative stem lengths in centimeters at time of planting and harvest. For presentation, RGR values have been multiplied by 10 3. Leastsquares means 1 1 SE are shown.

harvest, the mean percentage of leaf area missing on control plants that lived to harvest was roughly five times greater on Costa Rican plants (4.4%) than Hawaiian plants (0.9%; area effect: F1,4 5 9.18, P 5 0.03). Levels of leaf damage did not differ between understory- and open-grown plants in either area (habitat effect: F1,4 5 0.16, P 5 0.71). Herbivore damage in Costa Rica was attributed to gall-makers, stem borers, leaf-rolling moth larvae, and leaf-sucking curculionid weevils. Other leaf-chewing damage was evident, but the insects responsible could not be determined. The galls were thought to be formed by cecidomyiid fly larvae (P. Hanson, personal communication). At least one species of leafroller was found at all sites and was identified as a microlepidoptera (Compsolechia sp., Gelechiidae; D. Wagner and R. Hodges, personal communication). The type of herbivory differed dramatically but not consistently between paired open and understory habitats. For example, most plants in the EARTH understory had galls, whereas only 2% had galls in the open. In contrast, at El Bejuco none had galls in the understory, but most plants had galls in the open. Despite frequent application of two types of insecticides, neither leaf chewers nor other herbivores were completely excluded in Costa Rica. By harvest, con-

siderable leaf area was missing even on plants sprayed with insecticide (Fig. 4A). Mean percentage of leaf area missing did not differ significantly among plants in the insecticide vs. non-insecticide treatments in Costa Rica in the open or understory (Fig. 4A; Table 1). Leaf rollers also appeared undeterred by insecticide application. Among plants in open habitats with rolled leaves, there was a significantly higher proportion of leaves rolled at harvest among plants sprayed with insecticide than among those not sprayed with insecticide (F1,89 5 9.26, P , 0.01). Few leaves were rolled in the understory. We were unable to estimate the damage caused by the galls and stem borers, but we observed that both promoted branch death. Probability of gall presence was significantly lower for open-grown plants that received insecticide (x21 5 4.6, P 5 0.03) but not for plants in the understory (x21 5 1.2, P 5 0.28). What little herbivory was found on plants in Hawaii could be attributed to introduced insects. In open habitats, Liothrips urichi, an introduced biological control agent, was found on 10 out of 312 plants and caused some damage. A nonnative generalist moth, the Mexican leafroller Amorbia emigratella Busck, caused substantial damage, but only on a few plants at one site. We found no insects on plants in Hawaiian understory sites. However, in Hawaii, insecticide application sig-

SAARA J. DEWALT ET AL.

478

Ecology, Vol. 85, No. 2

FIG. 4. Leaf area missing on Clidemia hirta in understory and open habitats in (A) Costa Rica and (B) Hawaii in four natural-enemy exclusion treatments. Note the different scales for Costa Rican and Hawaiian plants. Hawaiian plants sprayed with insecticide had significantly less leaf area missing than plants not sprayed with insecticide. Back-transformed leastsquares means 6 1 SE are shown.

nificantly decreased the amount of leaf area missing in both habitats (Fig. 4B; Table 1). Fungicide had no effect on proportion of leaf area missing in either habitat. DISCUSSION

Evidence for the enemy-release hypothesis is habitat specific We found that that percentage of leaf area missing was greater in native (Costa Rican) than introduced (Hawaiian) areas of C. hirta’s range and that natural enemies may limit the survival, growth, and habitat distribution of C. hirta in its native range. Regardless of habitat, natural enemies that attacked C. hirta caused greater leaf area loss in Costa Rica than Hawaii, with missing leaf area five times greater on control plants in Costa Rica than in Hawaii. Galls and stem borers caused further, and potentially more detrimental, damage on plants than leaf-chewing insects in the native Costa Rican sites. In contrast, plants in Hawaii sustained very little damage (see Plate 1). The fitness effects of natural enemies in the native range were apparent as well because application of insecticide or fungicide significantly increased survivorship in the understory and fungicide application increased relative growth rates in both habitats in Costa Rica. In contrast, exclusion had no effect on survivorship or growth in Hawaii, where regardless of treatment, almost 100% of plants survived. These results are consistent with the enemy-release hypothesis but point to a greater role of natural enemies in restricting the abundance of C. hirta in forest understory than high light environments. Our results suggest that enemy release may contribute to the success of C. hirta in Hawaii, but does not preclude the effects of other factors on its spread. Suitable climate and adequate resource availability are certainly prerequisites for invasion. Higher resource availability (Denslow 2003), less dispersal or recruitment

limitation (Tilman 1997, Hubbell 2001), or the appearance of more vigorous genotypes (Blossey and No¨tzold 1995, Blossey and Kamil 1996) in introduced than native areas may also account for differences in abundance and habitat distribution. Indeed, the percentage survival of plants in the dual application treatment in Costa Rican understory sites was not equivalent to the percentage survival of plants in any pesticide treatment in Hawaiian understory sites. We attribute some of this discrepancy to the incomplete efficacy of our pesticide treatments in Costa Rica. Galls purportedly made by cecidomyiid fly larvae had substantial effects on plant vigor but seemed undeterred by the insecticides in the understory habitats of Costa Rica. In addition, standing leaf area missing did not differ among plants alive at harvest in the insecticide vs. noninsecticide treatments in Costa Rica. This may be an artifact of the low survivorship of plants not sprayed with insecticide, but also shows that standing leaf area missing was not a good proxy for the amount of herbivore damage on plants. Undoubtedly, it underrepresented the effects of sucking and boring herbivores on C. hirta. We assume that the percentage survival of plants in the dual application treatment would have been closer to levels in the understory of Hawaii or open of Costa Rica if complete natural-enemy exclusion had been achieved. Lack of aboveground natural enemies also does not seem to account for the greater proliferation of C. hirta in high light environments of Hawaii than Costa Rica. In fact, had we conducted this study only in open areas, we would have found no support for the hypothesis that natural enemies limit survival of this species in its native range. Mechanisms for the increased abundance of C. hirta in open areas in its introduced range compared to its native range should be examined further. In addition, effects of belowground herbivores and

February 2004

ENEMY RELEASE OF INVASIVE SHRUB

479

PLATE. 1. Representative Costa Rican (left) and Hawaiian (right) Clidemia hirta plants roughly 13 months after planting into understory habitats. Note the higher levels of insect damage (galls and stem borers) on the Costa Rican plant. Photo credit: Saara J. DeWalt.

pathogens should be examined because they may also have negative effects on fitness (Blossey 1993, Maron 1998, Packer and Clay 2000, 2003). Several other studies have found evidence that natural enemies are either more abundant or have greater effects on demographic parameters of plants in their native than in their invasive ranges. Mitchell and Power (2003) examined 473 plant species introduced to the United States from Europe and found 84% fewer fungal pathogen and 24% fewer viral pathogen species on the plants in the United States (Mitchell and Power 2003). In addition, Fenner and Lee (2001) examined flowerheads of 13 species of Asteraceae in Britain where they are native and in New Zealand where they have become naturalized and found that seed-eating insect larvae infected six times more flowerheads in Britain than New Zealand. Two studies of Cytisus scoparius (L.) Link (Fabaceae), one from the native range (Waloff and Richards 1977) and one from the introduced range conducted almost 20 years later (Bossard and Rejma´ nek 1994), found results similar to ours. Adults sprayed regularly with an insecticide in Europe, where C. scoparius is native, had higher survivorship and growth than unsprayed plants, suggesting that chronic insect herbivory may contribute to smaller population densities where the species is native. In California, where this species was introduced, Bossard and Rejma´ nek (1994) also sprayed plants with insecticide but detected no effect on growth. The number of phytophagous species found on plants in California was lower than in sites within the native range (Bossard and Rejma´ nek 1994). Interestingly, another study of this species in its native range detected no impact of invertebrates or pathogens on seedling survival, growth, or minimum age of reproduction (Paynter et al. 1998). Natural enemies therefore may affect only some life history stages, may vary in impact across the native and introduced ranges, or may have episodic rather than chronic impacts (e.g., Carson and Root 2000).

A complimentary study to ours would be to exclude herbivores and pathogens from both native Hawaiian vegetation and C. hirta to determine whether lower natural-enemy loads are responsible for the competitive advantage of the invasive over native species (Keane and Crawley 2002). The prediction based on the enemy-release hypothesis is that C. hirta survival and growth would be unaffected by natural-enemy exclusion (as found in our study), while the competitive ability of native species would be increased. To our knowledge, the only study that has conducted such an experiment found evidence contrary to this prediction. Instead, suppression of insect herbivores increased survivorship and growth of seedlings of the exotic invasive tree Sapium sebiferum L. in Texas but did not affect seedlings of two native trees (Siemann and Rogers 2003b). The success of S. sebiferum in forest understory and coastal prairies in Texas was attributed to genetically determined increases in growth and reproduction at the expense of defense (Siemann and Rogers 2001, 2003a). In contrast, we found no such evidence of post-invasion evolution of increased competitive ability in a common garden study with Costa Rican and Hawaiian genotypes of C. hirta (DeWalt 2003). For C. hirta, the proliferation in forest understory, at least, seems to be a plastic response to reduced pest load and not a genetic response. In our study system, the absence of native Melastomataceae makes it unlikely that specialist natural enemies native to Hawaii could limit C. hirta growth, survival, or habitat distribution (i.e., there are few biotic barriers to naturalization; Mack 1996). Clidemia hirta is a relatively recent introduction to the Hawaiian archipelago (70 years) and is even more recent on the island of Hawaii (30 years) where our study was conducted (Smith 1992). Thus, any potential host switching by specialist herbivores may not yet have occurred. Fenner and Lee (2001) hypothesized that the low incidence of colonization of nonnative Asteraceae by in-

480

SAARA J. DEWALT ET AL.

sects in New Zealand could be attributed either to the relatively recent introduction of the plants (100–200 years) or the lack of closely related native flora. These two reasons may also contribute to the lack of colonization of C. hirta by native insects in Hawaii. In contrast, on Silhouette, the primary island of the Seychelles where there are native Melastomataceae, an endemic cricket, Pelerinus rostratus, was found feeding on Clidemia hirta within 10 years of its introduction (C. Awmack, unpublished data). It would be valuable to assess the impact of herbivores on C. hirta where it is invasive in the presence of native Melastomataceae (e.g., Malaysia, India, or Indonesia) to determine if the role of enemy release in invasion success differs across its introduced range. We expect herbivore and pathogen impacts to be higher where there are closely related species. In addition, areas with high plant species diversity may have a more diverse herbivore assemblage, increasing the probability of host switching (Crawley 1983, Prieur-Richard et al. 2002). The lack of an effect of either the fungicide or insecticide on growth or survival of plants in Hawaii suggests that introduced biological control agents are not having a substantial impact on C. hirta in Hawaii. The introduced Liothrips urichi was found on only 3% of plants in open sites. In addition, if the fungal pathogen Colletotrichum gleosporiodes f.s. clidemiae were affecting plants, we likely would have found an effect of fungicidal treatment on Clidemia hirta growth or survival. Benomyl, the active ingredient in the fungicide treatment, has been shown to be effective against other varieties of Colletotrichum gloeosporioides on other species (Childers 1992, Elmer et al. 2001). We found no effect in Hawaii of the fungicide treatment. Interestingly, results from our study suggest that introduction of pests from our study areas in Costa Rica would not reduce survival of C. hirta in high light environments in Hawaii because we found no evidence that natural enemies affected survival in open sites. However, insect herbivores and pathogens may be more effective in the introduced range, even in open sites, if the herbivores themselves are released from their own natural enemies, such as parasites and parasitoids. To reduce the impacts of C. hirta in forested environments in its introduced range, biological control agents active in shaded humid conditions should be sought. However, surveys of natural enemies of C. hirta where it is common in its native range, i.e., open and disturbed habitats such as roadsides and pastures, may not yield biocontrol agents effective across the habitat range where it is invasive. The best biocontrol agents might be those collected from habitats where the plant is rare or fails to establish. Where exclusion by natural enemies has been most complete, the effective agents may be detected only by use of outplants and experimental pest exclusions. We suspect that control of C. hirta in both open and understory habitats may require intro-

Ecology, Vol. 85, No. 2

duction of a suite of herbivores or pathogens that attack plants in a variety of habitats.

Natural enemies as determinants of habitat distribution There are at least two mechanisms by which herbivory or fungal damage may restrict the habitat distribution of C. hirta where it is native. First, spatial variation in extent of damage could drive habitat distribution. Habitat and host plant selection by some herbivores is affected by variation in temperature, irradiation, and humidity among open and shady light environments (Sipura and Tahvanainen 2000). Preference for particular light environments has been found for insect herbivores of plants in their native range (Collinge and Louda 1988, Louda and Rodman 1996, Sipura and Tahvanainen 2000) as well as for particular biocontrol agents released against species in introduced areas (Huffaker and Kennett 1959, Harper 1969, Reimer and Beardsley 1989). Restriction of Hypericum perforatum L. to shaded areas in California where it is an exotic pest was attributed to significant consumption of leaves by a biocontrol agent in open but not shaded areas (Huffaker and Kennett 1959). Although individual herbivore species may have narrow habitat preferences, the net impact of all natural enemies affects habitat distribution of the species. In this study, we found no evidence for differences in habitat preferences among the suite of natural enemies of C. hirta in Costa Rica. We found similar one-time estimates of leaf area missing between C. hirta in open and understory habitats, and the occurrence of galls and weevils did not differ systematically between habitats. These results differ from those of Denslow et al. (1990) who found that Miconia (Melastomataceae) and Piper (Piperaceae) cuttings transplanted into understory had a higher percentage of leaf area missing than those planted into gaps in Costa Rica. In our study, the suite of herbivores (such as galls and weevils) varied, even among relatively nearby habitats. We found no evidence that spatial variation in natural-enemy habitat preferences drives the observed light-related habitat distribution of C. hirta. Second, the consequences of herbivory and pathogen attack may differ in shaded and open habitats. In the understory, carbon assimilation rates may be too low to replace tissue or photosynthate lost to herbivores and pathogens. Natural enemies may not reduce growth or survival appreciably in open areas if plants growing in high light can compensate for high rates of leaf area loss (Whitham et al. 1991). In a greenhouse study, Anten and Ackerly (2001) found that photosynthetic rates of an understory palm grown at high light levels increased sufficiently to compensate for the loss of leaf area when defoliated. In contrast, individuals grown in low light were unable to compensate for the loss in photosynthetic area (Anten and Ackerly 2001). This seems to be the case with C. hirta as well.

ENEMY RELEASE OF INVASIVE SHRUB

February 2004

Clidemia hirta is not the only case of habitat expansion into shaded areas in the introduced range from open areas in the native range. This phenomenon has been observed in at least seven other tropical woody species (DeWalt 2003). Whether reduction in pest loads in the introduced range also explains their seemingly greater shade tolerance has not been examined. Little information about pest loads is available for most invasive species in their native range. There is increasing evidence, however, that rankings in shade tolerance among native temperate trees of North America vary with herbivore pressure (Z. T. Long and W. P. Carson, unpublished manuscript) and pathogen attack (Vaartaja 1962, Packer and Clay 2003). Conclusions Physiologically, C. hirta appears relatively shade tolerant (able to persist in low light conditions); it occurs in understory in Hawaii and elsewhere in its introduced range and had high survival and positive relative growth rates under low light levels in a greenhouse study in Hawaii (Baruch et al. 2000). In the presence of natural enemies, however, it is effectively shade intolerant. Genetic differences in shade tolerance between native and introduced genotypes do not seem to explain the changed habitat distribution of this invasive shrub between Costa Rica and Hawaii (DeWalt 2003). Thus, this study provides some of the first experimental evidence that natural-enemy regulation can be an important factor in plant distributions at both geographic and local scales. It further highlights the risks of predicting habitat distributions of introduced exotic species based on information from their native ranges. At least some tropical invasive plants may be more tolerant of shaded conditions than their habitat distribution in their native range would indicate. ACKNOWLEDGMENTS We extend special thanks to R. Nagata in Hawaii and P. Cascante Arce and D. Zamora in Costa Rica for maintaining the experiments during our absence. We appreciate the field assistance of G. DeWalt, R. Cabin, S. Cordell, D. Goo, and A. Urakami. Logistical support in Costa Rica was provided by R. Matlock and the Organization for Tropical Studies, J. Alvarado and T. Ray at El Bejuco Biological Station, Juan Enriquez, and C. Sandı´ and R. de la Cruz at EARTH. In Hawaii, we thank the USDA Forest Service, Hawaii County Administrator, Hawaii Division of Forestry and Wildlife, Hawaii Department of Agriculture, the University of Hawaii at Manoa, and the Hawaiian Electric Company for their help finding field sites and obtaining permits. Advice on pesticides was provided by C. Hollier, F. Elango, W. Carson, R. Matlock, and S. Johnson. Information on the distribution of Clidemia hirta and biological control agents in Hawaii was provided by P. Conant. Discussions with D. O’Dowd and J. Ewel at the inception of this project helped refine the ideas. Comments by J. Ewel, L. Poorter, S. Schnitzer, and two anonymous reviewers improved the manuscript. Research was funded by grants and awards to S. J. DeWalt by the USDA Forest Service Pacific Southwest Research Station, Organization for Tropical Studies, Sigma Xi Grants-in-Aid-of-Research, Bernard Lowy Fund at Louisiana State University, Association for Women

481

in Science, and Board of Regents of Louisiana State University. LITERATURE CITED Aide, T. M., and J. K. Zimmerman. 1990. Patterns of insect herbivory, growth, and survivorship in juveniles of a neotropical liana. Ecology 71:1412–1421. Anonymous. 1954. Notes and exhibitions. Proceedings of the Hawaiian Entomological Society 15:263–265. Anten, N. P. R., and D. D. Ackerly. 2001. Canopy-level photosynthetic compensation after defoliation in a tropical understorey palm. Functional Ecology 15:252–262. Augspurger, C. K. 1984. Seedling survival of tropical tree species: interactions of dispersal distance, light-gaps, and pathogens. Ecology 65:1705–1712. Augspurger, C. K., and C. K. Kelly. 1984. Pathogen mortality of tropical tree seedlings: experimental studies of effects of dispersal distance, seedling density, and light conditions. Oecologia 61:211–217. Baruch, Z., R. R. Pattison, and G. Goldstein. 2000. Responses to light and water availability of four invasive Melastomataceae in the Hawaiian Islands. International Journal of Plant Sciences 161:107–118. Blossey, B. 1993. Herbivory below ground and biological weed control: life history of a root boring weevil on purple loosestrife. Oecologia 94:380–387. Blossey, B., and J. Kamil. 1996. What determines the increased competitive ability of invasive non-indigenous plants? Pages 3–9 in V. C. Moran and J. H. Hoffmann, editors. IX International Symposium on Biological Control. University of Cape Town, Stellenbosch, South Africa. Blossey, B., and R. No¨tzold. 1995. Evolution of increased competitive ability in invasive nonindigenous plants: a hypothesis. Journal of Ecology 83:887–889. Bossard, C. C., and M. Rejma´nek. 1994. Herbivory, growth, seed production, and resprouting of an exotic invasive shrub. Biological Conservation 67:193–200. Brown, V. K., M. Leijn, and C. S. A. Stinson. 1987. The experimental manipulation of insect herbivore load by the use of an insecticide (malathion): the effect of application on plant growth. Oecologia 72:377–381. Callaway, R. M., T. H. DeLuca, and W. M. Belliveau. 1999. Biological-control herbivores may increase competitive ability of the noxious weed Centaurea maculosa. Ecology 80:1196–1201. Carlquist, S. 1974. Island biology. Columbia University Press, New York, New York, USA. Carson, W. P., and R. B. Root. 2000. Herbivory and plant species coexistence: community regulation by an outbreaking phytophagous insect. Ecological Monographs 70:73– 99. Childers, C. C. 1992. Suppression of Frankliniella bispinosa (Thysanoptera, Thripidae) and the fungal pathogen Colletotrichum gloeosporioides, with pesticides during the bloom cycle and improved fruit-set on Naval orange in Florida. Journal of Economic Entomology 85:1330–1339. Coley, P. D., and J. A. Barone. 1996. Herbivory and plant defenses in tropical forests. Annual Review of Ecology and Systematics 27:305–335. Collinge, S. K., and S. M. Louda. 1988. Herbivory by leaf miners in response to experimental shading of a native crucifer. Oecologia 75:559–566. Cook, B. A. 1929. Some notes on the plant associates and habitat of Clidemia hirta (L.) D. Don in Trinidad. Fiji Agricultural Journal 2:92–93. Crawley, M. J. 1983. Herbivory: the dynamics of animal– plant interactions. University of California Press, Berkeley, California, USA. Crawley, M. J. 1987. What makes a community invasible? Pages 429–453 in A. J. Gray, M. J. Crawley, and P. J.

482

SAARA J. DEWALT ET AL.

Edwards, editors. Colonization, succession, and stability. Blackwell, Oxford, UK. Denslow, J. S. 2003. Weeds in paradise: thoughts on the invasibility of tropical islands. Annals of the Missouri Botanical Garden 90:119–127. Denslow, J. S., A. M. Ellison, and R. E. Sanford. 1998. Treefall gap size effects on above- and below-ground processes in a tropical wet forest. Journal of Ecology 86:597–609. Denslow, J. S., J. C. Schultz, P. M. Vitousek, and B. R. Strain. 1990. Growth responses of tropical shrubs to treefall gap environments. Ecology 71:165–179. DeWalt, S. J. 2003. The invasive tropical shrub Clidemia hirta (Melastomataceae) in its native and introduced ranges: tests of hypotheses of invasion. Dissertation. Louisiana State University, Baton Rouge, Louisiana, USA. Doak, D. F. 1992. Lifetime impacts of herbivory for a perennial plant. Ecology 73:2086–2099. Elmer, W. H., H. A. Yang, and M. W. Sweetingham. 2001. Characterization of Colletotrichum gloeosporioides isolates for ornamental lupines in Connecticut. Plant Disease 85: 216–219. Elton, C. S. 1958. The ecology of invasions of animals and plants. Methuen, London, UK. Evans, G. C. 1972. The quantitative analysis of plant growth. Blackwell, Oxford, UK. Fenner, M., and W. G. Lee. 2001. Lack of pre-dispersal seed predators in introduced Asteraceae in New Zealand. New Zealand Journal of Ecology 25:95–99. Filip, V., R. Dirzo, J. M. Maass, and J. Sarukha´n. 1995. Within- and among-year variation in the levels of herbivory on the foliage of trees from a Mexican tropical deciduous forest. Biotropica 27:78–86. Folgarait, P. J., R. J. Marquis, P. Inguarsson, H. E. Braker, and M. Arguedas. 1995. Patterns of attack by insect herbivores and fungus on saplings in a tropical tree plantation. Environmental Entomology 24:1487–1494. Fowler, S. V., H. M. Harman, J. Memmott, Q. Paynter, R. Shaw, A. W. Sheppard, and P. Syrett. 1996. Comparing the population dynamics of broom, Cytisus scoparius, as a native plant in the United Kingdom and France and as an invasive alien weed in Australia and New Zealand. Pages 19–26 in V. C. Moran and J. H. Hoffmann, editors. IX International Symposium on Biological Control. University of Cape Town, Stellenbosch, South Africa. Harper, J. L. 1969. The role of predation in vegetational diversity. Brookhaven Symposia in Biology 22:48–62. Harrison, S. 1987. Treefall gaps versus forest understory as environments for a defoliating moth on a tropical forest shrub. Oecologia 72:65–68. Hubbell, S. P. 2001. The unified neutral theory of biodiversity and biogeography. Princeton University Press, Princeton, New Jersey, USA. Huffaker, C. B., and C. E. Kennett. 1959. A ten-year study of vegetational changes associated with biological control of Klamath weed. Journal of Range Management 12:69– 82. Keane, R. M., and M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis. Trends in Ecology and Evolution 17:164–170. Larsen, J., I. Thingstrup, I. Jakobsen, and S. Rosendahl. 1996. Benomyl inhibits phosphorus transport but not fungal alkaline phosphatase activity in a Glomus–cucumber symbiosis. New Phytologist 132:127–133. Lincoln, D. E., and H. A. Mooney. 1984. Herbivory on Diplacus aurantiacus shrubs in sun and shade. Oecologia 64: 173–176. Louda, S. M. 1982a. Distribution ecology: variation in plant recruitment over a gradient in relation to insect seed predation. Ecological Monographs 52:25–41.

Ecology, Vol. 85, No. 2

Louda, S. M. 1982b. Limitation of the recruitment of the shrub Haplopappus squarrosus (Asteraceae) by flower- and seed-feeding insects. Journal of Ecology 70:43–53. Louda, S. M., P. M. Dixon, and N. J. Huntly. 1987. Herbivory in sun versus shade at a natural meadow–woodland ecotone in the Rocky Mountains. Vegetatio 72:141–149. Louda, S. M., and M. A. Potvin. 1995. Effect of inflorescence-feeding insects in the demography and lifetime fitness of a native plant. Ecology 76:229–245. Louda, S. M., and J. E. Rodman. 1996. Insect herbivory as a major factor in the shade distribution of a native crucifer (Cardamine cordifolia A. Gray, bittercress). Journal of Ecology 84:229–237. Lowman, M. D. 1984. An assessment of techniques for measuring herbivory: is rainforest defoliation more intense than we thought? Biotropica 16:264–268. MacGarvin, M., J. H. Lawton, and P. H. Heads. 1986. The herbivorous insect communities of open and woodland bracken: observations, experiments and habitat manipulations. Oikos 47:135–148. Mack, R. N. 1996. Biotic barriers to plant naturalization. Pages 39–46 in V. C. Moran and J. H. Hoffmann, editors. IX International Symposium on Biological Control. University of Cape Town, Stellenbosch, South Africa. Maiorana, V. C. 1981. Herbivory in sun and shade. Biological Journal of the Linnean Society 15:151–156. Maron, J. L. 1998. Insect herbivory above- and belowground: individual and joint effects on plant fitness. Ecology 79: 1281–1293. Marquis, R. J. 1984. Leaf herbivores decrease fitness of a tropical plant. Science 226:537–539. Marquis, R. J. 1992. A bite is a bite is a bite? Constraints on response to folivory in Piper arieianum (Piperaceae). Ecology 73:143–152. Mitchell, C. E., and A. G. Power. 2003. Release of invasive plants from fungal and viral pathogens. Nature 421:625– 627. Nakahara, L. M., R. M. Burkhart, and G. Y. Funasaki. 1992. Review and status of biological control of Clidemia in Hawaii. Pages 452–465 in C. P. Stone, J. T. Tunison, and C. W. Smith, editors. Alien plant invasions in native ecosystems of Hawaii: management and research. University of Hawaii Cooperative National Park Resources Studies Unit, Honolulu, Hawaii, USA. Packer, A., and K. Clay. 2000. Soil pathogens and spatial patterns of seedling mortality in a temperate tree. Nature 404:278–281. Packer, A., and K. Clay. 2003. Soil pathogens and Prunus serotina seedling and sapling growth near conspecific trees. Ecology 84:108–119. Paul, N. D., P. G. Ayres, and L. E. Wyness. 1989. On the use of fungicides for experimentation in natural vegetation. Functional Ecology 3:759–769. Paynter, Q., S. V. Fowler, J. Memmott, and A. W. Sheppard. 1998. Factors affecting the establishment of Cytisus scoparius in southern France: implications for managing both native and exotic populations. Journal of Applied Ecology 35:582–595. Pedersen, C. T., and D. M. Sylvia. 1997. Limitations to using benomyl in evaluating mycorrhizal functioning. Biology and Fertility of Soils 25:163–168. Prieur-Richard, A.-H., S. Lavorel, Y. B. Linhart, and A. Dos Santos. 2002. Plant diversity, herbivory and resistance of a plant community to invasion in Mediterranean annual communities. Oecologia 130:96–104. Reimer, N. J., and J. W. Beardsley, Jr. 1989. Effectiveness of Liothrips urichi (Thysanoptera: Phlaeothripidae) introduced for biological control of Clidemia hirta in Hawaii. Environmental Entomology 18:1141–1146.

February 2004

ENEMY RELEASE OF INVASIVE SHRUB

Rejma´nek, M. 1996. Species richness and resistance to invasions. Pages 153–172 in G. Orians, R. Dirzo, and J. H. Cushman, editors. Biodiversity and ecosystem processes in tropical forests. Springer, New York, New York, USA. Root, R. B. 1996. Herbivore pressure on goldenrods (Solidago altissima): its variation and cumulative effects. Ecology 77:1074–1087. Sanford, R. L., Jr., P. Paaby, J. C. Luvall, and E. Phillips. 1994. Climate, geomorphology, and aquatic systems. Pages 3–33 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: ecology and natural history of a neotropical rain forest. University of Chicago Press, Chicago, Illinois, USA. SAS Institute. 2000. SAS/STAT user’s guide. Version 8. Volumes 1, 2, and 3. Cary, North Carolina, USA. Schierenbeck, K. A., R. N. Mack, and R. R. Sharitz. 1994. Effects of herbivory on growth and biomass allocation in native and introduced species of Lonicera. Ecology 75: 1661–1672. Siemann, E., and W. E. Rogers. 2001. Genetic differences in growth of an invasive tree species. Ecology Letters 4:514– 518. Siemann, E., and W. E. Rogers. 2003a. Reduced resistance of invasive varieties of the alien tree Sapium sebiferum to a generalist herbivore. Oecologia 135:451–457. Siemann, E., and W. E. Rogers. 2003b. Herbivory, disease, recruitment limitation, and the success of alien and native tree species. Ecology 84:1489–1505. Simmonds, H. W. 1933. The biological control of the weed Clidemia hirta, D. Don., in Fiji. Bulletin of Entomological Research 24:345–348. Sipura, M., and J. Tahvanainen. 2000. Shading enhances the quality of willow leaves to leaf beetles—but does it matter? Oikos 91:550–558. Smith, C. W. 1992. Distribution, status, phenology, rate of spread, and management of Clidemia in Hawaii. Pages 241– 253 in C. P. Stone, J. T. Tunison, and C. W. Smith, editors. Alien plant invasions in native ecosystems of Hawaii: management and research. University of Hawaii Cooperative National Park Resources Studies Unit, Honolulu, Hawaii, USA.

483

Stanosz, G. R. 1994. Benomyl and acephate applications increase survival of sugar maple seedlings during their first growing season in northern Pennsylvania. Canadian Journal of Forest Research 24:1107–1111. Strahm, W. 1999. Invasive species in Mauritius: examining the past and charting the future. Pages 325–347 in O. T. Sandlund, P. J. Schei, and A. Viken, editors. Invasive species and biodiversity management. Kluwer, Dordrecht, The Netherlands. Szentesi, A. 1999. Predispersal seed predation of the introduced false indigo, Amorpha fruticosa L. in Hungary. Acta Zoologica Academiae Scientiarum Hungaricae 45:125– 141. Tilman, D. 1997. Community invasibility, recruitment limitation, and grassland biodiversity. Ecology 78:81–92. Vaartaja, O. 1962. The relationship of fungi to survival of shaded tree seedlings. Ecology 43:547–549. Vitousek, P. M., and J. S. Denslow. 1986. Nitrogen and phosphorus availability in treefall gaps of a lowland tropical rainforest. Journal of Ecology 74:1167–1178. Waloff, N., and O. W. Richards. 1977. The effect of insect fauna on growth, mortality and natality of broom, Sarothamnus scoparius. Journal of Applied Ecology 14:787– 798. Wester, L. 1992. Origin and distribution of adventive alien flowering plants in Hawaii. Pages 99–154 in C. P. Stone, J. T. Tunison, and C. W. Smith, editors. Alien plant invasions in native ecosystems of Hawaii: management and research. University of Hawaii Cooperative National Park Resources Studies Unit, Honolulu, Hawaii, USA. Wester, L. L., and H. B. Wood. 1977. Koster’s curse (Clidemia hirta), a weed pest in Hawaiian forests. Environmental Conservation 4:35–41. Whitham, T. G., J. Maschinski, K. C. Larson, and K. N. Paige. 1991. Plant responses to herbivory: the continuum from negative to positive and underlying physiological mechanisms. Pages 227–256 in P. W. Price, T. M. Lewinsohn, G. W. Fernandes, and W. W. Benson, editors. Plant–animal interactions: evolutionary ecology in tropical and temperate regions. Wiley, New York, New York, USA. Wolfe, L. M. 2002. Why alien invaders succeed: support for the escape-from-enemy hypothesis. American Naturalist 160:705–711.