Introduction
Main objectives
Long-term goals
Study system
Scientific and educational significance
Integrating active learning

Introduction
A fundamental assumption within invasion biology is that introduced populations of invasive species reach higher densities than native populations, and that individuals in the native range are are larger and more fecund than individuals from the native range. This assumption underlies most hypotheses to explain invasion and also many management approaches (e.g. biological control). However, little research has been conducted to examine this claim, even for most of the world’s worst invaders. Testing the validity of this assumption in the field is vital to developing sound, long-term approaches to managing invaded ecosystems. For example, if a species tends to be dominant in both the native and introduced range, then the inevitable costs and risks associated with management (via chemical approaches, biological control, or other means) may simply be too great when compared to the relatively small benefits likely to be accrued. Alternatively, if performance of a species is dramatically enhanced in the introduced range, then evaluating the mechanisms that underlie the increase should provide insight into which management approaches might be useful. Additionally, the performance of introduced and native populations of a given species is likely to vary substantially across both the introduced and native ranges (Colautti et al. 2009), and understanding that variation can also lend insight into the invasion process and potential management approaches.

One of the most problematic exotic weeds in the U.S. is the European herb garlic mustard (Alliaria petiolata). It dominates deciduous forests where it outcompetes native understory species and limits tree establishment (McCarthy 1997, Stinson et al. 2006, 2007). Despite the quantification of impacts and abundance at a few locations, relatively few data exist on the abundance and performance of this species across its native and introduced ranges, limiting our ability to make informed management decisions.

We have initiated a research project to formally evaluate whether garlic mustard performance (i.e. size and fecundity of individuals, area of coverage and density of plants) differs between the native and introduced range (www.GarlicMustard.org), and also how herbivory and abiotic environmental factors affect performance measures within each range. The key element of our project is a simple, standardized sampling protocol, which we have started to spread among scientists, teachers and citizen-scientists. The simplicity of the protocol allows involvement of a broad range of people, including students and others with little scientific training or specialized equipment. The standardization of methods creates comparability and scientific value of the collected data.

Main objectives
1. Test the fundamental assumption that invaders are more robust in the introduced range, with data that university students, citizen-scientists, and traditional scientists collect on garlic mustard size, density, and reproduction in the native and introduced range. The samples will be stratified by disturbance regime (roadside, forest edge, understory), allowing us to evaluate the influence of disturbance on invasion.

2. Improve the effectiveness of garlic mustard management strategies. Land-managers and volunteers who assist in garlic mustard control efforts are able to participate in the survey. This data will allow for multi-year monitoring of the effectiveness of different control efforts. These 'test sites' can be compared 'control sites' that are measured but not managed; together these sites will allow us to evaluate different control options (e.g. herbicide, manual removal, biocontrol) and test whether the effectiveness depends on environmental factors such as season length, altitude, moisture, etc.

3. Curriculum development, educational outreach and citizen-science. We are developing a flexible undergraduate laboratory protocol to be integrated into field courses. This will enable students to contribute to 'real' scientific research, which can be used as the starting point to teach important concepts in plant ecology, evolution and resource management. We are also planning to involve regional high school teachers and students through workshops.  Other citizens will be involved through a citizen-science program using the www.GarlicMustard.org site.

Long-term goals:
We see the work described above as Stage One of a larger on-going project. After both we and the participants (other scientists, land managers, citizen scientists, teachers and students) gain experience with coordinating this cross-continental sampling protocol, Stage Two would have participants implement scientific-based management of garlic mustard, based on results from Objective 2, again, focusing on garlic mustard.  In Stage Three, we would expand both the sampling and management efforts to include other widespread invaders.

Study system:
Garlic mustard (Alliaria petiolata) is a biennial herb in the mustard family (Brassicaceae). The species is native to the Eurasian temperate zone where it grows in nutrient-rich, semi-shade habitats such as forest edges or moist woodlands. Plants typically germinate in early spring, form a rosette in the first year, overwinter as a rosette, develop flowering stems the following spring, produce seeds in June/July and die. The species' common name comes from the characteristic garlic-like odor of its crushed leaves.

Garlic mustard is thought to have been introduced to North America as a food and medicinal plant, in the early 19th century, most likely many times (Durka et al. 2005). After a lag time of 100 years, the species started to spread rapidly and is now present in at least 37 US states and five Canadian provinces. It has been declared a prohibited or noxious weed in eight states.
Garlic mustard is particularly problematic because it invades and dominates the understory of native North American deciduous forests where it outcompetes understory species and reduces native plant diversity (McCarthy 1997, Carlson & Gorchov 2004, Stinson et al. 2007). It produces abundant secondary compounds including including glucosinolates and cyanide (Haribal & Renwick 1998, Haribal et al. 2001, Cipollini 2002, Cipollini & Gruner 2007). These compounds are toxic to native invertebrate and vertebrate herbivores alike (Haribal & Renwick 1998, Cipollini & Grunner 2007), and may also have allelopathic effects (Prati & Bossdorf). Garlic mustard has been shown to disrupt plant-pollinator associations (Porter 1994; Huang et al. 1995) as well as mutualistic belowground interactions between native plants and mycorrhiza fungi (Stinson et al. 2006, Wolfe et al. 2008), and it strongly reduces understory diversity and tree establishment (Stinson et al. 2007). Through accelerating the rate of litter decomposition, garlic mustard also increases the availability of soil nitrogen and phosphorous (Rodgers et al. 2008).  This N mobilization puts native plants, many of which perform well at low nitrogen levels, at a competitive disadvantage relative to nitrophilic invaders like garlic mustard (Rodgers et al. 2008), and may also increase nitrogen leaching into surface waters.

Because of these detrimental effects on native ecosystems, garlic mustard is not only frequently managed with hand removal or herbicides, but it is also the target of a biological control program (Blossey et al. 2001, Gerber et al. 2007a,b, 2008, 2009). Petitions for permission to release the first agents have been submitted to APHIS PPQs Technical Advisory Group (known as TAG). Garlic mustard does appear to have escaped from most of its natural enemies (Hinz & Gerber 2000, 2001, Lewis et al. 2006), but whether that escape has released it from top-down population regulation is unknown. As with other biological control programs, the program against garlic mustard is based on the idea that because of a release from natural enemies, garlic mustard plants grow more vigorously in North America than they do in Europe, and that introduction of a biological control agent will impose top-down regulation of population sizes, suppressing the North American populations to levels comparable to those found in Europe. But do plants really grow better in North America? There is actually some evidence that individual plants are smaller in North America than the native range (Lewis et al. 2006). However, Lewis et al. (2006) were able to include just two regions from the native and introduced ranges, preventing the discernment of broader geographical patterns and generalizations about this invasion.

Scientific and educational significance:
Do plants in introduced populations perform differently (i.e., more aggressively) than plants in native populations?

Many of the mechanisms proposed to explain the success of invasive species (disturbance, propagule pressure, hybridization, facilitation by other invaders, and enemy release) assume that introduced populations generally have higher abundances with more vigorous plants than native populations. Few studies have directly tested this assumption, and the few data available do not show a clear pattern (Bossdorf et al. 2005). Furthermore, recent work suggests that most genetic variation in growth and reproduction may occur among populations within geographic ranges, rather than between native and introduced ranges, perhaps as a result of local adaptation to abiotic environmental conditions (Colautti et al. 2009). As a result, differences observed between native and introduced ranges can vary dramatically depending on the number and geographic location of populations sampled. The best solution to this problem is extensive sampling of native and introduced populations, but that is difficult for most individual scientists or even small collaborative groups to achieve.

Current management strategies for invasive plants also generally assume that the biotic and abiotic factors that govern the distribution and abundance of individuals (which we will here call simply a species’ ecology) fundamentally differ between the native and introduced ranges. For example, classical biological control of weeds is likely to be most effective if invaders benefit from a loss of specialist herbivores (‘enemy release’) during the invasion process. Release from herbivory is hypothesized to result in uncontrolled population growth that is reversed through the introduction of biological control agents (which are themselves released from their own natural enemies; Müller-Schärer & Schaffner 2008). However, the degree to which the enemy release hypothesis explains the success of invasive species remains controversial (Colautti et al. 2004). If enemy release is not occurring, or more fundamentally, if similar biotic and abiotic processes regulate populations in the native and introduced ranges (e.g. their ecologies do not differ substantively), then the risk of releasing exotic biological control agent may not be warranted and other approaches to management will need to be sought.

Our goal is to thoroughly test this assumption, which lays the foundation for the field of invasion biology, for one of the most problematic plant invaders in the U.S., garlic mustard, by gathering extensive data on its abundance and distribution in the native and introduced range, and incorporating measurements of each population’s size, density, age structure, fecundity, and herbivory with scientifically rigorous standards. Previous studies have randomly sampled a relatively small number of native and introduced populations but ignored spatial structure due to abiotic environmental variables such as season length, temperature variability, day length, canopy cover and disturbance. Our analysis will incorporate climate data, latitude, altitude, disturbance (e.g. roadside vs. forest interior) and in situ measurements of canopy cover. Only after the influence of these factors is taken into account can we adequately address our objective.

How intensive does management need to be to effectivly reduce population sizes and restore ecosystem function?

Our broad sampling program sets the stage for evaluating consequences of management. One of the main ways to manage garlic mustard is simply by pulling the weed. This is commonly done both by land managers and by groups of citizens during organized community “weed pulls”. However, invasive weeds such as garlic mustard can reach densities where the strongest limiting factor is intraspecific competition. When weed density is reduced, intraspecific competition is also reduced and individual plants can produce more seed. This has been called the ‘hydra effect’ after the mythical beast that produces two heads for every one cut off (Abrams & Matsuda 2005; Abrams 2009). There is strong theoretical and empirical evidence for hydra effects, but its occurrence has historically been treated as an interesting phenomenon in basic ecology, with little attention paid to the profound implications for management (but see Pardini et al. 2008). Thus, the common assumption that removing individuals will decrease population size cannot be assumed to be true, and there is good evidence that it is not true for garlic mustard (Pardini et al 2008, 2009). Thus, the critical question is, what removal level is needed to actually reduce population size? For the garlic mustard system, research can be guided by a demographic model (Pardini et al. 2009), which makes specific predictions regarding percent control. Pardini et al. (2009) predict that 75% removal will not be effective, and may even increase population size in the following year, and that 95% removal will be effective in reducing population size. This has not been evaluated empirically.

Clearly, reducing population sizes of invaders is a critical goal of invasive species management.  However, we must also look beyond simply suppressing invaders and actively manage for desirable aspects of the community or ecosystem affected by the invader (Ditomaso et al. 1999; Ditomaso 2000). One of the most critical negative impacts garlic mustard has on ecosystem services is its reduction of tree recruitment, as trees provide food (e.g. maple syrup), fiber, fuel, and building materials. Garlic mustard is associated with decreased tree recruitment, and has been shown to inhibit micorrhizal associations of trees (Stinson et al. 2006, 2007); thus tree seedlings may need to be planted in to restore the ecosystem services provided by trees.

Active Learning:
Integrating science into active learning of students and citizens.

Active modes of learning have been shown repeatedly to be more effective than passive modes (National Research Council, 2000). For example, students will learn more about mark-recapture techniques for monitoring population sizes by marking and ‘recapturing’ candy bars in a classroom, than by listening to a lecture on the topic. However, students tend to shut down when they view the activities that faculty try to get them involved in as ‘busy work’, i.e., routine work that is assigned to keep students busy but appears to be of little value. Laboratory and field courses focus on active modes of learning. However, the work done in those courses is very often seen as busy work by students. With this project, we have the opportunity to connect students to a real, on-going scientific project. The students will gain many skills, and learn about the process of science. Our protocol focuses first on the basic science, and the hypotheses we are testing. Knowing that their data are contributing to answering an important question that is being addressed across the globe will be very inspiring for students. The work requires cooperation, learning about invasive species, appropriate sampling in biology, population-level processes, comparative biology, and how aspects of the environment influence species.

Citizens also continue to learn, and through this project they can focus their energy on contributing to solving a visible problem, that of biological invasions in general, and garlic mustard in particular.