Myrtle spurge (Euphorbia myrsinites) is an invasive plant in North America that was introduced from Eurasia as a drought-tolerant ornamental. It has escaped cultivation and is widespread in western North America. Infestations can be reduced by either mechanical or chemical means, but the two approaches have not been compared directly. One study recommended mechanical treatment in the spring and chemical treatment in the fall, but neither recommendation was supported by data. To test the effectiveness of mechanical versus chemical treatment in the fall, we located six sites in the foothills of the Wasatch Mountains in Utah that were uniformly infested with myrtle spurge. At each of the sites we randomly assigned a 2.25 m2 quadrat to one of three treatments (i.e., mechanical, chemical, or control). The mechanical treatment entailed hand pulling and the chemical treatment involved applying the herbicide Roundup to each myrtle spurge. The controls were unmanipulated. We photographed each quadrat pretreatment and one year post-treatment. Both mechanical and chemical approaches significantly reduced the percent cover and stalk count of myrtle spurge compared to the controls. Additionally, the effects of the two treatments did not differ from each other. These results indicate that land managers can control myrtle spurge in the fall with either treatment and do not have to wait until the spring to employ mechanical methods. Mechanical removal of myrtle spurge is better suited to events that involve members of the general public such as “Purge Your Spurge” activities because they reduce the public's exposure to herbicide.
INTRODUCTION
Myrtle spurge (Euphorbia myrsinites L.) is an herbaceous, short-lived perennial that has become invasive in western North America. It was introduced from Eurasia as a drought-tolerant ornamental plant for rock gardens, but it has escaped cultivation and is widespread as an invasive plant. Myrtle spurge has a high potential to continue to spread throughout western North America (Alexander 2013) for several reasons: (1) it has hybrid vigor, having been sourced from a wide area in Eurasia (Alexander 2013), (2) it produces seeds that are ejected up to 4.5 m from seed capsules, (3) it can reproduce asexually from root fragments, and (4) it tolerates both dry and nutrient-poor soil conditions (Lowry et al. 2017).
Myrtle spurge is particularly problematic in the foothills of the Wasatch Mountains in Utah, in which the northern mountain-brush community (Van Buren et al. 2011) is characterized by native shrubs including Gambel oak (Quercus gambelii Nutt.) and netleaf hackberry (Celtis reticulata Torr.). The study area (elevation = 1573 m) is rocky and dry, receiving an annual average of 43.7 cm of precipitation and having an annual average temperature of 12.6 °C (1991–2020) (NOAA-NWS 2020).
In terms of invasive-species management in the area, myrtle spurge was listed as a Salt Lake County (Utah) Noxious Weed in 2007 (Salt Lake County Weed Control Program 2007). That same year, the Bonneville Cooperative Weed Management Area started a program to offer native plants in exchange for bags of myrtle spurge that Salt Lake County residents removed from their yards in a “Purge Your Spurge” event that brought in more than 3000 kg of myrtle spurge in one year (Bureau of Land Management et al. 2007). Purge Your Spurge events continue in Utah and public awareness of the problems associated with myrtle spurge have increased over time and expanded beyond one county. Currently, myrtle spurge is prohibited in the state of Utah and has been included on the state's noxious weed list (Utah Department of Agriculture and Food 2019), in part because it outcompetes native plants such as sego lily (Calochortus nuttalli Torr. & A. Gray), Utah's state flower. In adjacent Colorado, myrtle spurge must be eradicated by state law (Colorado Department of Agriculture 2005).
Multiple studies have indicated that myrtle spurge infestations can be controlled by mechanical and chemical means (Salt Lake County Weed Control Program 2007; Rondeau et al. 2010; Lowry et al. 2017), but the two treatments have not been compared directly. Additionally, the Salt Lake County Weed Control Program (2007) recommended controlling infestations through multiple years of digging up at least 10 cm of root, especially in the spring when the soil is moist and the plants have not yet produced seeds. Chemically, they recommend applying herbicides containing either 2, 4-D or glyphosate (one of the active ingredients in Roundup), in the late fall. However, neither of these recommendations is backed up with data from the primary scientific literature.
In our study, we directly compared the effectiveness of mechanical versus chemical control of myrtle spurge in the fall. Our two research questions were (1) In the fall, which is a more effective treatment method against myrtle spurge, mechanical or chemical? and (2) Although mechanical treatments are recommended in the spring, can they also be useful when applied in the fall?
METHODS
To test the effectiveness of mechanical versus chemical means to eradicate this invasive species, we located six sites in the foothills above Provo, Utah, USA, that were uniformly infested with myrtle spurge in terms of both percent cover (P = 0.602) and stalk count (P = 0.803). At each of the six sites we randomly assigned a 2.25 m2 quadrat to one of three treatments (control, mechanical, and chemical). There were 18 quadrats in total. The control quadrats were unmanipulated. The mechanical treatment entailed the hand pulling of each myrtle spurge plant (including at least 10 cm of root) with the assistance of shovels. The chemical treatment involved applying Roundup Weed & Grass Killer (Monsanto Lawn & Garden Products, Marysville, Ohio, USA) to each myrtle spurge plant following the manufacturer's directions for use. The active ingredients in this commercially available formulation of Roundup were glyphosate (2.0%) and pelargonic acid (2.0%).
We photographed each quadrat pre-treatment on 16 August 2017, applied the treatments on 1 September 2017, and rephotographed the quadrats one year post-treatment on 21 August 2018. The percent cover for each quadrat was calculated using Adobe Photoshop CS4 (Adobe, San Jose, California, USA). The images were first uploaded to Photoshop and the pixel count for each quadrat was obtained by opening the histogram window and selecting the plot area with the Polygonal Lasso Tool. To determine the percent cover, the pixel count for the myrtle spurge within each quadrat was selected. This was done by using the select color range function in Photoshop. The selection for the color range was complete once all stalks from the original images were accounted for in the selection. The selected pixels were then copied and pasted onto a black background. At this point, all selected pixels that were identifiable as non–myrtle spurge were removed from the selection. The pixel count of the selection was recorded, and the percent cover was calculated by dividing the selection pixel count by the quadrat pixel count. Myrtle spurge stalk counts were also determined from these photographs.
Digital photography has been shown to be an effective method for estimating plant cover that is objective, accurate, and repeatable in a variety of habitats (Luscier et al. 2006; Chen et al. 2010; Salas-Aguilar et al. 2017). Imagery analysis has also been used to count both individual plants (Rominger et al. 2021) and plant parts (Crimmins and Crimmins 2008; Stevens et al. 2016). Using digital photography was especially appropriate for our study because we were working with a plant that contains caustic sap (Lowry et al. 2017) and in quadrats that we treated with herbicide.
For statistical analyses, we used ANOVA (SAS Institute 2021) for overall comparisons between the three treatments (i.e., control, mechanical, and chemical). We then followed up with ttests (SAS Institute 2021) for comparisons between pairs of treatments (i.e., control versus mechanical, control versus chemical, and mechanical versus chemical). To meet the assumptions of ANOVA, we logit transformed the percent cover data (Warton and Hui 2011) and log transformed the stalk count data.
Figure 1.
Bars show the mean (standard error) for percent cover by myrtle spurge from 2.25 m2 quadrats (n = 6) in the Wasatch Mountain foothills, Utah, USA, one year after one of three treatments was applied to areas uniformly infested with myrtle spurge. Bars with different letters are significantly different. See text for treatment details.
RESULTS
Both mechanical and chemical treatments significantly reduced the percent cover (Figure 1) of myrtle spurge in our quadrats compared to the controls (P = 0.001) and did not differ from each other (P = 0.708). The mechanical treatment reduced percent cover by 84% and the chemical treatment reduced percent cover by 88%, compared to the controls.
Similar to our results for percent cover, both mechanical and chemical treatments reduced the stalk count (Figure 2) of myrtle spurge in our quadrats compared to the controls (P = 0.001) and did not differ from each other (P = 0.623). The mechanical treatment caused a 78% reduction in the number of myrtle spurge stalks and the chemical treatment resulted in an 85% reduction in the number of stalks, compared to the controls.
DISCUSSION
We have shown that myrtle spurge infestations can be reduced in terms of percent cover and stalk count by either mechanical or chemical means in the fall. Due to its responsiveness to mechanical and chemical treatments in a study done at the U.S. Air Force Academy in Colorado, Rondeau et al. (2010) recommended that myrtle spurge be a high priority for management. Additionally, its eradication is required by Colorado state law (Colorado Department of Agriculture 2005). In Utah, Lowry et al. (2017) also recommended either mechanical or chemical treatments.
Figure 2.
Bars show the mean (standard error) for stalk count of myrtle spurge plants from 2.25 m2 quadrats (n = 6) in the Wasatch Mountain foothills, Utah, USA, one year after one of three treatments was applied to areas uniformly infested with myrtle spurge. Bars with different letters are significantly different. See text for treatment details.
While both treatments can yield effective results, they both come with their own pros and cons. The pros for mechanical treatments include their low cost (if volunteer labor is available) and not needing specialized equipment or permitting. Mechanical treatments can be focused on the plant of interest with very few adverse effects to nontarget plants or the environment. The downsides include the time and effort needed to implement the treatments. Additionally, because myrtle spurge produces a caustic milky sap, Lowry et al. (2017) recommended wearing gloves and eye protection when engaged in mechanical treatments. Chemical treatments are easy to apply, but there are several cons associated with them including cost, negative public perceptions of herbicides, the prohibition of herbicides in some areas, and difficulties associated with accessing natural areas with spray equipment. Additionally, chemical treatments can damage nontarget plants and contaminate groundwater, if not used according to label (Bourchier et al. 2006). In some cases, plants may be able to block the translocation of herbicides, as is the case for leafy spurge (E. esula L.), a congener of myrtle spurge (Bourchier et al. 2006). Although Bourchier et al. (2006) recommended using multiple methods to control invasive species, referred to as Integrated Weed Management, we recommend evaluating the longer list of pros associated with mechanical treatments and the longer list of cons associated with chemical treatments, in light of the lack of differences we observed in the effectiveness of the two treatments for myrtle spurge. Although mechanical treatments may be effective for small patches of myrtle spurge, such treatments may be impractical for large-scale infestations.
Although both treatment types can be effective initially, total eradication is difficult to achieve. For example, Rondeau et al. (2010) were able to reduce myrtle spurge density with both mechanical and chemical treatments, but after their treatments ended, they found many small myrtle spurge plants that were either coming from seeds or unexcavated roots. If eradication is the management goal, then ongoing management and consistency is required.
With respect to consistency, a useful takeaway from this study is that land managers can effectively use mechanical treatments in either the spring or the fall. The Salt Lake County Weed Control Program (2007) recommended mechanical treatments in the spring and chemical treatments in the fall. However, based on the recommendations from the Salt Lake County Weed Control Program (2007) and evidence from our study, mechanical treatments can be used effectively in either the spring or the fall, and perhaps even twice a year. Furthermore, mechanical treatments can be implemented immediately in the fall without having to wait until the spring. These findings make Purge Your Spurge events easier to schedule. Additionally, the application of mechanical treatments is more appropriate for members of the general public because the use of herbicides is not required. A further benefit of fall mechanical treatments is that the caustic sap that myrtle spurge produces is less profuse in the fall versus the spring.
ACKNOWLEDGMENTS
We thank Gary Stone for allowing us to undertake this study on his private land. Funding for D.T.Y. was provided by the Undergraduate Research, Scholarly, and Creative Activities program through the Office of Engaged Learning at Utah Valley University. Comments from two anonymous reviewers and the Associate Editor helped us improve the manuscript.
