Using biological control agents to restore native habitats degraded by exotic plants should decrease the abundance of the invaders but should also result in a return toward preinvasion levels of native diversity. However, there are few long-term studies documenting changes in native biological diversity with the decline of an invader. We introduced a biocontrol flea beetle into three Montana grassland sites dominated by leafy spurge and monitored changes in leafy spurge abundance and frequency of associated vascular plants in 48 permanent microplots, in a 530 or 1,960 m2 macroplot immediately before and 14 yr after the release. Density and mass of leafy spurge declined 60 and 69%, respectively, over the 14 yr of the study across the three sites. Total species richness increased by 1.2 species/microplot (21%) between 1994 and 2008 across all three sites, but the increase differed among sites. Mean richness of exotic species was virtually unchanged across the three sites over the course of the study. Graminoid species richness was virtually unchanged across the three sites over the course of the study; most of the increase in diversity was due to the increase in forb richness at all three sites. Release of the biocontrol insects and a subsequent large reduction of leafy spurge were associated with an increase in native diversity after 14 yr, although causality cannot be confidently inferred from these associations because there were no controls. The increase in native diversity was small relative to the decline in leafy spurge abundance, suggesting that significant increases of native alpha diversity in semiarid grasslands may require many decades. Our results also suggest that the response to a decline of an invading species may depend on site quality and history.

Broom snakeweed (snakeweed) is an aggressive native range-weed found throughout arid and semiarid areas of the western United States, that increases following disturbances such as overgrazing, drought, or wildfire. Ecologically based strategies that include controlling snakeweed and reestablishing desirable herbaceous species are needed to restore productivity and diversity to invaded areas. The objective of this study was to compare the ability of selected introduced and native grass species and prostrate kochia (kochia) to prevent reinvasion of snakeweed, downy brome, and annual forbs following control. This field study was replicated at two sites (Howell and Nephi, Utah) within the sagebrush-steppe biome. Snakeweed and downy brome were controlled by picloram (0.25 kg/ha [0.22 lb/ac]) and glyphosate (1.5 kg/ha [1.3 lb/ac]). The seeding treatments were comprised of three introduced grasses and a mix of these species, three native grasses and a mix of these species, and kochia. The treatments were seeded into 3 by 15-m (10 by 50 ft) plots in October 2003. Frequency and biomass of seeded species, snakeweed, downy brome, other grasses, and annual forbs were measured in 2004, 2005, and 2008. Seeded species were evaluated for success of establishment and persistence, and their ability to restrict reinvasion of snakeweed, downy brome, and annual forbs. Crested wheatgrass and big squirreltail had the best initial establishment at both locations (> 75%). In contrast, kochia and Russian wildrye did not establish well at either site, and western wheatgrass did not establish well at Nephi. Snakeweed reestablishment was restricted in all treatments except the kochia treatments, in which kochia did not establish well. Frequency of downy brome increased at both sites, and annual forb frequency increased at Nephi to near 100%, but the better established grasses suppressed biomass production of these weedy species. Crested wheatgrass established best, had the greatest production, and provided greatest suppression of downy brome and annual weeds.

Protected area managers often face uncertainty when managing invasive plants at the landscape scale. Crested wheatgrass, a popular forage crop in the Great Plains since the 1930s, is an aggressive invader of native grassland and a problem for land managers in protected areas where seeded roadsides and abandoned fields encroach into the native mixed-grass prairie. Given limited resources, land managers need to determine the best strategy for reducing the cover of crested wheatgrass. However, there is a high degree of uncertainty associated with the dynamics of crested wheatgrass spread and control. To compare alternative management strategies for crested wheatgrass in the face of uncertainty, we conducted a decision analysis based on information from Grasslands National Park. Our analysis involves the use of a spatially explicit model that incorporates alternative management strategies and hypotheses about crested wheatgrass spread and control dynamics. Using a decision tree and assigning probabilities to our alternative hypotheses, we calculated the expected outcome of each management alternative and ranked these alternatives. Because the probabilities assigned to alternative hypotheses are also uncertain, we conducted a sensitivity analysis of the full probability space. Our results show that under current funding levels it is always best to prioritize the early detection and control of new infestations. Monitoring the effectiveness of control is paramount to long-term success, emphasising the need for adaptive approaches to invasive plant management. This type of decision analysis approach could be applied to other invasive plants where there is a need to find management strategies that are robust to uncertainty in the current understanding of how these plants are best managed.

Broom snakeweed (snakeweed) is a native range shrub found throughout semiarid rangelands of the western United States, which increases and dominates plant communities after disturbances such as overgrazing, drought, or wildfire. The objective of this study was to compare the ability of selected grass species and prostrate kochia to restrict establishment and growth of snakeweed seedlings in potted-plant and replicated field studies within two sagebrush ecological sites. In the potted-plant studies, single snakeweed seedlings were grown with seedlings (seedling neighbor study) and established plants (established neighbor study) of three cool-season grasses (crested, pubescent, and bluebunch wheatgrass), prostrate kochia, and snakeweed at increasing densities (1, 3, 5 plants/pot). Interference from crested wheatgrass in the seedling neighbor study, and both crested and bluebunch wheatgrass in the established neighbor study, induced the greatest mortality of snakeweed seedlings, and snakeweed growth was suppressed more by interspecific (grass) than intraspecific (snakeweed) neighbors in both potted-plant studies. Snakeweed establishment was also evaluated at two field sites: Howell and Nephi, UT. Snakeweed and downy brome were controlled by picloram (0.25 kg ae/ha) and glyphosate (1.5 kg ae/ha), then three native and three introduced grasses were drill-seeded, and prostrate kochia was dribble-seeded in replicated plots (3 m by 15 m) at both sites in October 2003. Snakeweed seedlings were transplanted into seeded plots and a bare ground control plot in autumn 2004. Snakeweed mortality was greatest (73%) in crested wheatgrass plots at Howell, but there were few differences among species treatments at Nephi. Of the snakeweed seedlings that survived, there was relatively little growth in any of the seeded plots compared to those in the bare ground control plots. These results indicate that seeded cool-season grasses interfered with and reduced establishment of snakeweed seedlings.

Manipulating plant litter to direct successional trajectories is rarely considered as a management strategy. Our objective was to determine the influence of litter from an intact native plant community on a community dominated by an invasive species within the same habitat type as well as the influence of litter from a community dominated by an invasive species on an intact native plant community. We hypothesized that litter amount, type (source), and fragment size would influence various functional groups within a native plant community differently than within a weed-dominated plant community. We used reciprocal plant litter exchanges between native and invasive plant–dominated grasslands to gain an initial understanding of litter's influence on the density and biomass of native grasses, native forbs, common St. Johnswort, and downy brome. Common St. Johnswort was not influenced by any treatment. Native grass density increased with application of low (454 g/m2) amounts of litter where the grasses were subordinate to common St. Johnswort, and adding native plant litter to the weedy site nearly doubled native grass biomass. Low amounts of finely fragmented litter and high amounts of coarse litter induced native forbs to produce about twice the biomass as found in the non–litter-amended controls. Our study suggests that plant litter may be a component of vegetation that can be managed to shift the plant community toward those plants that are desired.

Management methods for invasive species vary in their restoration success in the presence or absence of herbivores. We investigated the performance of understory plants after management of the invasive shrub Amur honeysuckle using two herbicide-based methods (cut/paint and basal application) in fenced and unfenced plots. The cut/paint method resulted in the removal of above-ground stems, while the basal application method resulted in the dead stems remaining in place. Light level in the cut/paint treatment was higher than in the basal application treatment, which was higher than in the control (no management) treatment. Across fencing treatments, fruit production, height, and subsequent recruitment of transplanted jewelweed were greater in the cut/paint treatment. Across management treatments, jewelweed plants were taller in the fenced treatment. Native species richness was generally higher in the cut/paint and basal application treatments than in the control treatment. There were more jewelweed recruits, more jewelweed fruits, and greater native species richness in the cut/paint treatment than in the basal application treatment in fenced plots, but these measures were similar in both management treatments in unfenced plots. Thus standing dead stems of Amur honeysuckle offered protection from damage in the presence of herbivores, offsetting the advantage of the cut/paint method seen in the fenced plots. There was a trend for more leaves of transplanted wild ginger in the basal application treatment. There were more invasive garlic mustard and more Amur honeysuckle seedlings in the cut/paint treatment than in the control treatment. Our results illustrate the complexities involved in selecting appropriate restoration management techniques given herbivore pressure, differential species response, and presence of multiple invasive species. In our study, we demonstrated that the basal application honeysuckle management method (and therefore perhaps similar methods that leave the dead stems standing) led to increased understory plant restoration success when compared to the cut/paint method.

Black greasewood (greasewood) and gray rabbitbrush are important shrub components of several plant communities throughout western North America. Land managers may view these species differently depending on their goals. Greasewood and gray rabbitbrush shrub communities may be invaded by several invasive plants including Russian knapweed, perennial pepperweed, hoary cress, halogeton, and several annual mustards. Metsulfuron and chlorsulfuron are commonly used for controlling these and other invasive plants, but little is known regarding their impacts on greasewood and gray rabbitbrush. Our objective was to quantify the impact of these herbicides on greasewood and gray rabbitbrush communities from both an efficacy and nontarget impact perspective. Field studies were established in the spring of 2004 and repeated twice in 2005 near Laramie, WY, in a pasture with mixed stands of greasewood and gray rabbitbrush. Treatments included metsulfuron applied at 21, 42, 63, 84, 126, and 168 g ai/ha (0.3, 0.6, 0.9, 1.2, 1.8 and 2.4 oz ai/A), chlorsulfuron applied at 52, 105, and 157 g ai/ha (0.75, 1.5 and 2.25 oz ai/A), and an untreated control. All treatments contained methylated seed oil at 2% v/v. Treatments were applied in mid-June to 3.3 by 9-m (10 by 30 ft) plots with a handheld broadcast sprayer delivering 187 L/ha (20 gal/ac) in a randomized complete block, with three blocks per study. Plots were sampled 12 and 24 mo after treatment (MAT), utilizing visual control estimates and point frame sampling for vegetative cover of greasewood, gray rabbitbrush, perennial grasses, and bare ground. Metsulfuron at 42 g/ha and chlorsulfuron at 105 g/ha provided > 75% visual control of greasewood 24 MAT. For gray rabbitbrush, metsulfuron at 63 g/ha provided approximately 60% control 24 MAT, while chlorsulfuron provided negligible control at any rate. These results suggest differential impacts of these herbicides on greasewood and gray rabbitbrush, and provide land managers with a decision tool for noxious and invasive weed control when managing for or against greasewood and gray rabbitbrush.

Two field experiments were conducted at the Agronomy Research Station near Stillwater, OK in 2004 and 2005 to identify and evaluate herbicides applied preplant incorporated, preemergence, early postemergence, or late postemergence for the control of seedling sericea lespedeza. Trifluralin, applied preplant incorporated, controlled seedlings 77 (15 wk after emergence [WAE]) and 63% (16 WAE) in 2004 and 2005, respectively. Flumioxazin, imazapic, fluometuron, diuron, sulfentrazone, atrazine, metribuzin, and metolachlor applied preemergence all provided greater than 86% seedling control at 15 and 16 WAE in both years. Diclosulam, applied preemergent, controlled seedlings 47% at 15 WAE in 2004 and 91% control at 16 WAE in 2005. In 2004, triclopyr, metsulfuron-methyl, glyphosate, picloram, dicamba, and 2,4-D amine plus picloram (tank-mix) applied early postemergence controlled 90 to 100% of the sericea lespedeza seedlings at 15 WAE. However, in 2005, only triclopyr, metsulfuron-methyl, glyphosate, and 2,4-D amine plus picloram (tank-mix) showed greater than 80% control at 16 WAE. Triclopyr, applied late postemergence, controlled seedlings 100% at 15 and 16 WAE, in both years. In 2004, dicamba plus 2,4-D (premix) and glyphosate were the only other herbicides which provided greater than 75% control at 15 WAE. These data suggest that there were preemergence-applied herbicides that were effective for the control of seedling sericea lespedeza. These data also suggest that triclopyr was the most effective postemergence-applied herbicide for the control of seedling sericea lespedeza. The data also showed that the overall level of control of seedling sericea lespedeza decreased as the season progressed and the plant matured.

Hybridization between species has the potential to change invasion dynamics. Field observations suggest that spotted knapweed and diffuse knapweed, two ecologically and economically destructive invasive plants, hybridize in their introduced range. As a first step towards understanding whether hybridization has affected the dynamics of the invasion of these species, we conducted field surveys in the introduced (North American) and native (European) ranges to discern patterns of hybridization and measured fitness-related traits among field hybrids and parental species. In North America we detected plants with hybrid morphology in 97% of the diffuse knapweed sites (n = 40); such hybrid plants were taller and more often exhibited polycarpy than plants with typical diffuse knapweed morphology. Hybrids were not detected in North American spotted knapweed sites (n = 22). In most regions surveyed in Europe, diffuse knapweed and spotted knapweed were isolated from each other and existed as distinct, nonhybridizing species. However, in Ukraine, the two species frequently coexisted within a site, resulting in hybrid swarms. On average, the plants from the North American diffuse knapweed sites (including plants with both diffuse and hybrid morphology), were larger than the apparently pure diffuse knapweed in the native range. The cross-continental patterns of hybridization likely are explained by differences in cytology. It recently has been confirmed that the spotted knapweed in North America is tetraploid whereas the diffuse knapweed is diploid. Genetic incompatibilities associated with these two cytotypes likely prevent ongoing hybridization. We hypothesize that hybrid individuals were introduced to North America along with diffuse knapweed. Because plants with hybrid morphology are found in nearly all North American diffuse knapweed sites, the introduction of hybrids likely occurred early in the invasion of diffuse knapweed. Thus, although the presence of hybrids might facilitate the ongoing invasion of diffuse knapweed into North America, elevated concern regarding their presence might not be warranted. Because such individuals are not likely to represent a new hybridization event, currently effective management strategies used in diffuse knapweed sites should not need alteration.

When exposed to native or introduced herbivores and pathogens, invasive plants may become weaker competitors with more benign impacts on individual plants and plant communities. In a greenhouse pot study, we tested whether the presence of powdery mildew disease caused by Erysiphe cruciferarum could alter the competitive impact of garlic mustard on Impatiens pallida, a North American native understory plant. Target I. pallida plants were grown alone or with one, two, or three garlic mustard neighbors. Half of the pots exposed to garlic mustard were inoculated with conidia of E. cruciferarum. Competition with garlic mustard moderately affected aboveground growth of I. pallida, particularly at high garlic mustard density, but it strongly reduced seed output across all densities. In contrast, inoculation of garlic mustard plants with E. cruciferarum completely abolished their competitive impact on seed output of I. pallida across all densities, independent of effects on aboveground growth of target plants. This effect was likely due to alteration in the ability of garlic mustard to compete for belowground resources. Even without killing garlic mustard, these results indicate that the presence of powdery mildew disease in the field will likely dampen the competitive impact of garlic mustard on individual plants and plant communities. Escape from such attackers has likely contributed to the invasiveness and impacts of garlic mustard in North America.

The magnitude of the invasive plant species problem necessitates prioritization of species for control, regulatory, and public-education programs. Many such priority lists exist but few have been developed according to specified procedures and criteria. We reviewed approaches to assessing the status of nonnative plant species currently occurring in natural areas (status assessments). We identify four generalized types of status assessments, which reflect a gradation from those that simply adopt existing lists from elsewhere (type 1), to those with relatively easy and rapid development and implementation (type 2), to those that are more time-consuming and costly but may be more robust in the face of challenges (types 3 and 4). These latter assessments explicitly have greater transparency, objectivity, and consistency than the other types. We use a matrix of assessment characteristics to distinguish the types of 17 example status assessments. We also review the factors related to assessment intent, scope, structure, content, and implementation that must be considered during the development of new status assessments so that the resulting tool and its products are appropriate for the user's purposes. These analyses should facilitate evaluation of different assessment methods and provide a basis for development of improved assessments. Identification of the relatively low percentage of nonnative plant species that are inflicting ecological and economic harm using well-understood and accepted assessment methods should facilitate a more comprehensive, collective approach to implementation of effective management efforts.

Basal bark treatments involve the application of concentrated herbicide solution on each individual stem of targeted plants. When applied to stands of invasive plants with high stem densities, basal bark treatments may result in the use of large quantities of herbicide in a given area. The effect of basal bark treatments using a solution of 25% triclopyr herbicide and 75% methylated seed oil was tested on research plots located in six different groves of invasive fig, a densely stemmed, problematic invader of riparian forests in California. The experimental treatments resulted in application rates that were equivalent to 28 to 44 kg ae/ha, greatly in excess of the labeled maximum use rate of 9 kg ae/ha (8 lb/ac). At 175 d after application, soils near the fig trunks contained high levels of triclopyr residues (up to 6.6 ppmw), suggesting that the chemical made its way into soils during this period and was not completely degraded. Although the mortality of native plants transplanted into treated fig groves was low (up to 16%), it was significantly greater than the mortality experienced by native plants transplanted into untreated control sites (0%). Although effective in controlling invasive fig trees (> 99% mortality), the high herbicide application rates from basal bark treatment preclude the use of this treatment in large fig groves. These treatments may be appropriate, however, when fig groves are small or isolated enough to prevent overapplication on a per-area basis. In addition, neither limited basal bark applications of triclopyr (< 40% of stems treated) nor foliar spray treatments of 2% glyphosate were effective control measures. Further investigation is needed on ways to control large invasive fig groves.

Japanese clematis, a herbaceous to woody vine native to Asia, has been widely used for landscaping in the southeastern United States and is now a naturalized and invasive in Florida. Herbicides that contain the active ingredients fluroxypyr, glyphosate, imazapic, metsulfuron, or triclopyr, applied alone or in combination on a spray-to-wet basis, were evaluated for their effectiveness to control the plant in a natural area of Gainesville, FL. All herbicide treatments provided some level of control 30, 60, and 90 d after application. The most effective herbicide treatments, imazapic 0.6 g ae/L, glyphosate as low as 3.6 g ae/L, triclopyr as low as 3.6 g ae/L, metsulfuron 0.07 g ai/L, and fluroxypyr as low as 1.8 g ae/L, resulted in 72% to 99% control 90 d after application, but were not significantly different from each other. Natural area managers throughout the range of Japanese clematis should be vigilant for this species and remove it whenever encountered. Our results indicate that it can be controlled by foliar application of several herbicides commonly used for management of pest plants in natural areas, but repeat applications will be necessary.

Perennial pepperweed is an invasive plant species that occurs throughout the western United States. This study develops a predictive model for perennial pepperweed distribution for the San Francisco Bay Area, based on spatial variables. Distribution data were developed by mapping perennial pepperweed along the shoreline of the South San Francisco Bay, using geographic positioning system units. Spatial relationships between its distribution and spatial variables were tested using binomial logistic regression. Predictive models were mapped using geographic information systems (GIS), and high risk areas within the San Francisco Bay Area were identified. Perennial pepperweed was found to occur within marsh habitats with full tidal action and near open water. This study demonstrates that habitat variables from widely available GIS layers can be used to predict distribution patterns for perennial pepperweed. The model results were compared to land ownership within the study area to demonstrate a management application of the model.

Cabomba is a submersed aquatic plant native to the southeastern United States that is commonly sold worldwide through the aquarium trade. While infrequently managed in its native range, cabomba has recently been reported as invasive and tolerant to management efforts in the northern areas of the United States and in other countries. Invasive populations of cabomba are characterized by a phenotype that is bright green. In contrast, cabomba native to the southeastern United States is characterized by a red phenotype, while plants sold through the aquarium trade have intermediate characteristics of both the green and red phenotypes. The response of the three cabomba phenotypes to selected herbicides was evaluated by measuring photosynthetic response over the course of a static 144-hr exposure. Plants were exposed to the maximum recommended use-rates of 2,4-D, carfentrazone, copper, diquat, endothall (amine and dipotasium salt formulation), flumioxazin, quinclorac, triclopyr, and a combination of diquat and copper. A submersed plant species known to be sensitive to each of these herbicides was also included to compare photosynthetic response of the cabomba to a susceptible plant. The photosynthetic response of the red and green phenotypes differed following exposure to carfentrazone, diquat, 2,4-D, triclopyr, and flumioxazin. Diquat, diquat plus copper, endothall (amine salt), and flumioxazin were the only products that resulted in a greater than 50% reduction of photosynthesis in all three phenotypes of cabomba. A second experiment was conducted where all three phenotypes of cabomba were exposed to these four herbicides for 24 hr, and photosynthesis was evaluated. Following the 24-hr exposure, results further documented distinct response differences between the green and red phenotypes, with the green phenotype demonstrating a reduced sensitivity to the herbicides evaluated. Results demonstrate clear phenotypic differences in response to herbicide treatments and lack of susceptibility of cabomba to most herbicides.

Garlic mustard seeds are dormant at maturity, and 90 to 105 d of cold-moist stratification at 4 C have been used to induce germination. We studied methods for breaking dormancy and inducing germination without cold stratification, for use in laboratory and greenhouse experiments with garlic mustard. Seeds were collected from large infestations, stored at room temperature, and subjected to chemical and mechanical scarification treatments. For chemical scarification, seeds were immersed in 3% (v/v) H2O2 for 12, 24, or 48 h with constant stirring, or immersed in concentrated (95 to 97%) H2SO4 for 1 or 5 min with stirring. For mechanical scarification, seeds were placed in a sandpaper-lined tumbler for 1 or 3 s. Scarified seeds, along with non-scarified seeds, were placed in petri dishes on germination blotters saturated with gibberellic acid (GA3, 10−3 M) or deionized water, and incubated for 35 d at either 20/10 C or 15/6 C (12 hr/12 hr). None of the non-scarified seeds germinated, regardless of germination solution or temperature. Seeds germinated only following scarification, and only when imbibed in GA3 solutions. Seeds immersed in H2SO4 for 5 min or mechanically scarified for 3 s had the highest level of germination in GA3. Cumulative percent germination after 35 d was greater for seeds stored 30 mo (44 to 83%), than for seeds stored 6 (2 to 60%) or 18 mo (35 to 79%), regardless of scarification treatment. The germination results, along with scanning electron micrographs of seed coats, suggest that the intact garlic mustard seed coat is permeable to water but not GA3; therefore, both scarification and GA3 are needed to break dormancy and induce germination without cold stratification.

White sweetclover is invading the Alaska glacial river floodplains and roadsides adjacent to natural areas, and control methods are needed. Chlorsulfuron, 2,4-DB, clopyralid, triclopyr, and 2,4-D controlled white sweetclover seedlings below recommended rates in the greenhouse. Biomass of established plants in the field was reduced by chlorsulfuron at recommended (17.6 g ai/ha), 1/2, and 1/4 rates and was reduced by triclopyr and 2,4-D at recommended rates (1,260 and 1,600 g ai/ha). Herbicides were more effective at reducing white sweetclover viable seed production in 2007 than in 2006. Only chlorsulfuron at 17.6 g ai/ha (recommended rate) eliminated seed production in both years. Flaming killed first-year plants, but some second-year plants resprouted and produced viable seed. Cutting at the 2.5 or 10 cm height did not control first-year plants because of regrowth, and second-year plant density and seed production was reduced by cutting at 2.5 cm but not by cutting at 10 cm.

Geographic information systems (GIS) analysis considering size and spatial distribution of infestations of invasive species on a landbase can assist with developing appropriate control strategies for that species. We used GIS to evaluate strategies that placed highest priority on: smallest patches of cogongrass, largest patches of cogongrass, and cogongrass patches on land considered high-value under military training and conservation priorities on Camp Shelby Joint Forces Training Center (CSJFTC), MS. Strategies were first evaluated using predicted increase in area of new growth prevented, if prioritized patches were treated with herbicide. Travel and other time needed to implement strategies were then assessed as they affected total area treated. Assuming equivalent areas treated, predicted total reduction in area was consistently highest for the strategy prioritizing smallest patches and lowest for the strategy prioritizing largest patches. However, travel and other time was sufficiently reduced for the strategy prioritizing patches on high-value land compared to the other strategies to suggest that this strategy would result in the greatest reduction in area infested with cogongrass. Analyses of spatial distribution and size of infestations can assist land managers with selecting the appropriate strategy for controlling invasive species.

There has been recent interest in determining the upper limits to the feasibility of weed eradication. Although a number of disparate factors determine the success of an eradication program, ultimately eradication feasibility must be viewed in the context of the amount of investment that can be made. The latter should reflect the hazard posed by an invasion, with greater investment justified by greater threats. In simplest terms, the effort (and hence investment) to achieve weed eradication comprises the detection effort required to delimit an invasion plus the search and control effort required to prevent reproduction until extirpation occurs over the entire infested area. The difficulty of estimating the required investment at the commencement of a weed eradication program (as well as during periodic reviews) is a serious problem. Bioeconomics show promise in determining the optimal approach to managing weed invasions, notwithstanding ongoing difficulties in estimating the costs and benefits of eradication and alternative invasion management strategies. A flexible approach to the management of weed invasions is needed, allowing for the adoption of another strategy when it becomes clear that the probability of eradication is low, owing to resourcing or intractable technical issues. Whether the considerable progress that has been achieved towards eradication of the once massive witchweed invasion can be duplicated for other weeds of agricultural systems will depend to a large extent upon investment (> $250 million over 50 yr in this instance). Weeds of natural ecosystems seem destined to remain more difficult eradication targets for a variety of reasons, including higher impedance to eradication, more difficulty in valuing the benefits arising from eradication, and possibly less willingness to pay from society at large.