Management Implications
Tradescantia fluminensis (small-leaf spiderwort) is a troublesome perennial ground cover and has become an increasingly damaging invasive plant in tropical and subtropical regions such as Florida. While previous research has identified several effective herbicide options, many practitioners report mixed results that have been attributed due to potential differences in optimal application timing. An additional component of this study was to reapply herbicide treatments when mean coverage ratings exceeded 25% following the initial application. Data indicated that triclopyr, regardless of formulation or application timing, was the most effective treatment and required no reapplication over the 12-mo study. The 2,4-D + triclopyr and glyphosate treatments were also highly effective, but glyphosate resulted in the need for a reapplication at 9 mo following spring applications. The 2,4-D and aminopyralid treatments were generally the least effective options and resulted in the need for reapplication. Across all six herbicide treatments, minimal differences were seen due to application timing, indicating that practitioners should expect similar results following either spring or fall applications if efficacious herbicides are chosen and applied at recommended rates.
Introduction
Small-leaf spiderwort (Tradescantia fluminensis Vell.) is a sprawling perennial herb of the spiderwort family (Commelinaceae) native to tropical and subtropical regions of Brazil and Argentina (Maule et al. Reference Maule, Andrews, Morton, Jones and Daly1995). It has naturalized in New Zealand (Standish Reference Standish2001), eastern Australia (Burns Reference Burns2004), Portugal (Aguiar et al. Reference Aguiar, Ferreira and Moreira2001), the Galapagos Islands (Guézou et al. Reference Guézou, Trueman, Buddenhagen, Chamorro, Guerrero, Pozo and Atkinson2010), Italy, Japan, the southeastern United States (FISC 2019), Puerto Rico, and Hawaii (Seitz and Clark Reference Seitz and Clark2016). It is currently listed as a Category I exotic plant pest in Florida (FISC 2019). Additionally, it is recognized as an environmental weed (i.e., ecologically damaging invasive exotic plant) in New Zealand due to its ability to spread rapidly to form dense mats of roots and stems that can smother native ground cover species and inhibit the establishment of seedlings of tree species (Esler Reference Esler1988; Kelly and Skipworth Reference Kelly and Skipworth1984).
In its native range, T. fluminensis typically inhabits tropical rainforests and other warm humid areas. In regions where it has been introduced, it demonstrates a broad adaptability to various climates, ranging from cool temperate to tropical environments (Dugdale et al. Reference Dugdale, McLaren and Conran2015). In Florida, it is mostly found in northern and central portions of the peninsula (Seitz and Clark Reference Seitz and Clark2016), but specimens have been vouchered in more than 30 counties across the state (EDDMapS 2024). Previous literature is limited on growth in response to climatic variables or induction of dormancy, but T. fluminensis has been shown to display a very seasonal growth pattern, with growth being positively correlated with temperature (Maule Reference Maule1991). In Florida, T. fluminensis growth is highest during the summer months during periods of both high temperatures and frequent rainfall. Growth typically slows during the winter months, with death and/or biomass reduction being reported at temperatures of −4.2 C (Gorchov Reference Gorchov2019).
Tradescantia fluminensis relies almost entirely on vegetative propagation for reproduction, demonstrating the ability to regrow from each broken stem fragment containing at least one node (Hurrell et al. Reference Hurrell, Belton, Lusk and Lamoureaux2012; Kelly and Skipworth Reference Kelly and Skipworth1984). The stems are easily dispersed in water currents, dumped garden waste, vehicles, and mowing equipment. The leaves of T. fluminensis are broadly ovate to oblong lanceolate, waxy, bright green in color, and arranged alternately on thin leafy shoots (Maule et al. Reference Maule, Andrews, Morton, Jones and Daly1995). The growth of the plant occurs at the stem apex, whereas the stem typically decays at the posterior end of the horizontal stem (Kelly and Skipworth Reference Kelly and Skipworth1984). A distinctive feature that sets T. fluminensis apart from other members of the Commelinaceae family is its unique flowers, making it relatively easy to differentiate. The flowers of T. fluminensis are characterized by their white color and centrifugal development, typically comprising five or six stamens, with each containing bright yellow anthers. Comparatively, other Tradescantia species typically have purple flowers, and while many Commelina species are similar to T. fluminensis in terms of vegetative characteristics, they will possess three stamens and have unequal-sized flower petals (Marble and Brown Reference Marble and Brown2021; Tucker Reference Tucker1989). Flowering is often sparse, but T. fluminensis plants may bloom during the spring and fall seasons in northern Florida (Seitz and Clark Reference Seitz and Clark2016).
The impact of T. fluminensis has been extensively studied in New Zealand, where it has been found to significantly affect the soil and native vegetation of forests in the region (Maule et al. Reference Maule, Andrews, Morton, Jones and Daly1995; McAlpine et al. Reference McAlpine, Lamoureaux and Westbrooke2015; Toft et al. Reference Toft, Harris and Williams2001). In New Zealand, it is listed as one of the threatening environmental weeds in several areas of the North Island (Hurrell et al. Reference Hurrell, James, Lamoureaux, Lusk and Trolove2009). Tradescantia fluminensis is shade tolerant and can grow in a wide range of light levels ranging from 1% to 100% (Maule et al. Reference Maule, Andrews, Morton, Jones and Daly1995). The species tends to invade forest remnants, particularly when a tree falls, resulting in increased light intensity on the forest floor (Kelly and Skipworth Reference Kelly and Skipworth1984). Tradescantia fluminensis typically does not invade undisturbed forests; instead, its presence is indicative of the degradation of forest remnants and potentially anthropogenic disturbances. It responds to an increase in resource availability, such as light or nutrients in these degraded areas (Standish Reference Standish2001). The species exhibits rapid growth, driven primarily by two key resources: light and nitrogen. For instance, a tree that falls can cause disturbance, leading to an increase in light and soil nitrogen favoring the invasion of T. fluminensis (Maule et al. Reference Maule, Andrews, Morton, Jones and Daly1995). Once introduced, it can rapidly spread to form a dense layer of interwoven stems that can grow up to 60 cm in length. According to a 2-yr study in New Zealand, T. fluminensis exhibited rates of stem growth ranging from 0.2 to 0.3 cm d−1 during the summer months, whereas in winter, the growth rates ranged from 0.04 to 0.06 cm d−1 (Maule et al. Reference Maule, Andrews, Morton, Jones and Daly1995). The dense growth habit of T. fluminensis as a ground cover can suppress the growth of other plants (Standish Reference Standish2002) and interfere with forest succession (Harden et al. Reference Harden, Marilyn and Barry2004). A study conducted in New Zealand reported that the survival of native kohekohe [Dysoxylum spectabile (G. Frost.) Hook. f.] was reduced with increasing T. fluminensis biomass (Standish Reference Standish2001). In a comparative study of three weed species sharing similar life forms [T. fluminensis, Asparagus fern (Asparagus scandens Thunb.), and speckled spur flower (Plectranthus ciliatus E. Mey.)]), McAlpine et al. (Reference McAlpine, Lamoureaux and Westbrooke2015) found that T. fluminensis had the most detrimental effect on native vegetation, leading to reductions in both native abundance and richness.
Tradescantia fluminensis has also shown to have an impact on invertebrate communities (Standish Reference Standish2004). Species diversity of Malaise-trapped beetles (Coleoptera) and fungus gnats (Diptera) was significantly reduced due to a decrease in floral diversity, owing to T. fluminensis invasion (Toft et al. Reference Toft, Harris and Williams2001). Furthermore, the litter produced by T. fluminensis decomposes more readily than the leaf litter of mixed-species forests. This accelerated decomposition rate leads to increased moisture and nutrient levels beneath the leaf litter compared with non-invaded forests (Standish Reference Standish2001, Reference Standish2004). This represents one of the ways T. fluminensis modifies ecosystem properties.
Tradescantia fluminensis clearly exerts negative impacts on native species and communities, altering ecosystems. However, there is a notable absence of effective management practices for controlling its spread, especially when compared with other invasive plants with comparable impacts. The most practical method for controlling large infestations of T. fluminensis has been through herbicide application (Standish Reference Standish2002). Triclopyr has been found to be most effective at various application rates, resulting in 98% to 100% control (Hurrell et al. Reference Hurrell, James, Lusk and Trolove2008; McCluggage Reference McCluggage1998; Standish Reference Standish2002; Yu et al. Reference Yu, Marble and Minogue2022) and has also demonstrated efficacy on other Commelinaceae species such as Asiatic dayflower (Commelina communis L.) (Aulakh Reference Aulakh2023). Similarly, Marble and Chandler (Reference Marble and Chandler2021) reported 89% and higher control rates with three different formulations of triclopyr (ester, amine, and choline) in greenhouse and field evaluations. Several studies found that glyphosate was much less effective compared with other herbicides (Hurrell et al. Reference Hurrell, James, Lusk and Trolove2008; Marble and Chandler Reference Marble and Chandler2021) for control of T. fluminensis and other Commelinaceae plants such as doveweed [Murdannia nudiflora (L.) Brenan] (Atkinson et al. Reference Atkinson, McCarty, Powell, McElroy, Yelverton and Estes2017). However, McCluggage (Reference McCluggage1998) found that 3% glyphosate (v/v) achieved nearly 100% control, albeit requiring at least three treatments. Other herbicide options that have shown mixed or poor results include metsulfuron-methyl (McCluggage Reference McCluggage1998), fluroxypyr (Dugdale et al. Reference Dugdale, McLaren and Conran2015; Marble and Chandler Reference Marble and Chandler2021), 2,4-D, aminopyralid, and glufosinate, among others (Marble and Chandler Reference Marble and Chandler2021).
While a small number of effective herbicide options have been established, optimal application timing is unknown and could be contributing to differences in efficacy observed by researchers, particularly with herbicides such as glyphosate (Hurrell et al. Reference Hurrell, James, Lusk and Trolove2008; McCluggage Reference McCluggage1998). Application timing with regard to time of year and/or season has been shown to have a significant impact on herbicide efficacy (Gous Reference Gous1997). In Monterey pine (Pinus radiata D. Don) plantations, glyphosate provided significantly higher control of a mixed species of herbaceous plants when applied in summer compared with spring, fall, or winter applications (Gous Reference Gous1997). Other research has also shown that herbicide efficacy can increase as temperature increases (Adkins et al. Reference Adkins, Tanpipat, Swarbrick and Boersma1998), but this is highly herbicide and species specific (Mudge and Sartain Reference Mudge and Sartain2018; Reddy Reference Reddy2000; Zhou et al. Reference Zhou, Tao, Messersmith and Nalewaja2007). For example, in studies examining the translocation of dicamba under different temperature regimes, translocation decreased at temperatures of 32.5 C compared with temperatures of 17.5 or 25 C (Ou et al. Reference Ou, Stahlman and Jugulam2016). Examining seasonal effects on herbicide efficacy for control of Tradescantia has not been previously conducted, as early experiments only contained treatments applied at one application timing such as winter (Marble and Chandler Reference Marble and Chandler2021) or spring (McCluggage Reference McCluggage1998). Hurrell et al. (Reference Hurrell, James, Lusk and Trolove2008) conducted a series of experiments on T. fluminensis control with herbicides in spring and summer seasons in New Zealand and generally saw similar results, with triclopyr providing the highest level of control across all experiments. However, in testing conducted by Hurrell et al. (Reference Hurrell, James, Lusk and Trolove2008), plants were tested at different growth stages and under different conditions across the four experiments (field, glasshouse, in trays, etc.), and thus seasonal differences are difficult to decipher. The importance of understanding temporal dynamics in invasive plant management has been established; thus, further information on how herbicide application timing influences T. fluminensis control is needed to develop optimal management strategies. Therefore, the objective of this study was to evaluate the efficacy of selected postemergence herbicide applications applied in either the spring or fall seasons in central Florida for control of T. fluminensis and to understand differences in these timings relating to effective control and rate of regrowth.
Materials and Methods
Field experiments were initiated in the spring of 2022 at two central Florida locations in Gainesville, FL, including Payne’s Prairie State Preserve (29.6102°N, 82.2992°W) and a local city park, 29th Road Nature Park (29.6811°N, 82.3417°W). Both sites were heavily shaded by mature tree canopy and contained dense, uniform, and fully covered areas of T. fluminensis in the forest understory (Figure 1). At each site, fifty-eight 5-m2 treatment plots were delineated by marking four corners of the plot with wire flags or wood stakes using a square 5-m2 PVC frame. Distance between individual treatment plots was variable due to native vegetation, trees, fallen tree limbs, and other natural barriers, but plots were separated by at least 3 m on each side to avoid inadvertent drift or foot traffic within the plots. After plots were established, each was randomly assigned one of seven herbicide treatments (Table 1) and a treatment timing of either spring or fall. Herbicides were then applied to the spring treatment plots using a CO2 backpack sprayer calibrated to deliver 233 L ha−1 at 241 kPa using a TeeJet® 8006 flat-fan nozzle (TeeJet® Technologies, Glendale Heights, IL). A single nozzle was used to make three uniform passes over each plot. Spring-applied treatments were made on March 14 (18.9 C, winds 8 km h−1, relative humidity 59%, cloudy skies) and fall-applied treatments were made on October 18 (21.7 C, winds 12 km h−1, relative humidity 43%, sunny skies) following the same procedures.
Figure 1. Examples of Tradescantia fluminensis density at the time of treatment at 29th Road Nature Park (right) and Payne’s Prairie State Preserve (left) in Gainesville, FL, in 2022.
Table 1. Herbicides evaluated for Tradescantia fluminensis control in spring- and fall-timed experiments in Florida.
a A non-ionic surfactant (Capsil®, Aquatrols, Paulsboro, NJ) was added to each treatment at 0.25% v/v based on manufacturer recommendations, with the exception of the aminopyralid treatment, which contained no surfactant.
b Annual maximum shows the annual maximum rate allowed by each product label for the use described in this study.
Following treatment, plots were assessed each month by recording visual coverage ratings of T. fluminensis based on a 0% to 100% scale, where 0% = 0% of the treated plot contained T. fluminensis (or 100% control if treated) and 100% = 100% of the plot area contained T. fluminensis. At 3, 6, and 9 mo after the initial herbicide treatment (MAT), treatment averages were calculated at each location, and herbicides were reapplied following the procedures described earlier to plots in which mean coverage ratings exceeded 25% and had also exceeded 25% at the preceding rating period (at 2, 5, or 8 MAT). Using two consecutive rating periods to determine retreatment was done to account for natural minor fluctuations in T. fluminensis coverage due to insect herbivory, wildlife disturbances, and other related incidences that were noted in some plots. The use of a ≥25% threshold was based on conversations with land managers and biologists regarding when reapplications would ideally be made in lands they managed. The overall objective of this treatment schedule was to allow determination of regrowth in each treatment and efficacy of each herbicide based on the length of control from initial application and to estimate the number of treatments that may be required within one calendar year depending upon the herbicide option chosen. Ratings were collected monthly for 12 mo following the spring application, but only 11 mo after the fall application due to scheduled land maintenance that required the termination of the experiment.
The experiment was a completely randomized design with four replications for each herbicide treatment and treatment timing at each location. Due to the use of two locations and the use of retreatment thresholds, two levels of analysis were performed on the data. First, an analysis was performed by location to obtain means and SEs to describe the development of vegetation posttreatment for each application timing and active ingredient. Second, a combined analysis was performed with location incorporated as a random effect with herbicide treatment and application timing (spring or fall) as fixed effects. All statistical analysis was performed using the PROC GLIMMIX (Littell et al. Reference Littell, Milliken, Stroup, Wolfinger and Schabenberger2006) package of SAS® (v. 9.4, SAS Institute, Cary, NC). Percent coverage ratings were arcsine square-root transformed to achieve normality but were back-transformed for presentation. Individual treatment means were compared using the Tukey adjustment for multiplicity. The individual location and the combined location analyses were performed separately by month after treatment (MAT) due to significant interactions of MAT with all fixed effects (P < 0.0001) in a preliminary longitudinal analysis.
Results and Discussion
When examining T. fluminensis regrowth following treatment at each of the two locations, 2,4-D + triclopyr, glyphosate, and both triclopyr acid and triclopyr amine performed similarly, providing a high level of control at both sites, and no definitive application timing differences were observed for the length of control observed between spring and fall applications (Figure 2). Of these herbicides, only glyphosate required a reapplication, which occurred at 9 MAT at the Payne’s Prairie location following an initial spring application. While only one reapplication was needed, it should be noted that glyphosate-treated plots had increasingly high levels of coverage at 10 and 11 MAT (32% and 45% coverage, respectively), which decreased precipitously at 12 MAT despite no herbicide being applied. While the reason for this is unclear, it was likely due to a combination of cooler and drier weather and insect herbivory as well as plant disturbances due to wildlife activity in these plots. Fall applications of 2,4-D + triclopyr, glyphosate, or either triclopyr formulation resulted in coverage ratings of less than 25% at 3, 6, and 9 MAT, and thus no reapplications were made.
Figure 2. Means and SEs for T. fluminensis coverage (%) in experimental plots following a spring (March) or fall (October) application timing at two locations in Gainesville, FL, including 29th Road Nature Park and Payne’s Prairie State Preserve. An “×” indicates when reapplications were made of specific herbicides at each location, which occurred at approximately 3, 6, and 9 mo after treatment (MAT) when mean coverage ratings for an individual herbicide treatment exceeded 25%.
Results with 2,4-D and, to a lesser extent, aminopyralid were more variable (Figure 2). For 2,4-D, reapplications were required at 3 MAT at both sites following spring applications, as the 25% target threshold level of coverage was not reached following the initial application. Ratings at 4 MAT were still above 25% at both sites, but notable injury (i.e., control) was observed, and by 5 MAT, coverage ratings fell below 10% at both locations and remained below the 25% threshold for the remainder of the experiment. Fall 2,4-D treatments required a reapplication at 6 MAT at both sites. At Payne’s Prairie, an additional application was required at 9 MAT to decrease coverage below the target threshold. It is unclear why poor control resulted from the 6 MAT application at Payne’s Prairie and not at the Nature Park, but this in in line with previous research in which inconsistent results have been reported (Marble and Chandler Reference Marble and Chandler2021). For aminopyralid, reapplications were needed following the spring timing at the Nature Park. While this application was originally scheduled to be made at 3 MAT, it was delayed due to weather conditions, and only the 2,4-D applications could be made. Following the fall aminopyralid application, no reapplications were required at the Nature Park, while one reapplication was needed at Payne’s Prairie at 9 MAT.
Across both locations, differences were detected between herbicide treatments on all 12 evaluation dates, while timing (spring vs. fall) affected T. fluminensis coverage only for the first three MAT evaluations (Table 2). The lowest T. fluminensis coverage (i.e., best control) was generally observed in plots treated with triclopyr (acid or amine), but 2,4-D + triclopyr and glyphosate resulted in similar coverage ratings on most evaluation dates. In reference to timing, fall-treated plots had lower levels of T. fluminensis coverage for the first 3 MAT compared with spring-treated plots. This was not surprising, as the spring applications occurred in March, with 3 MAT occurring in June, compared with fall-applied treatments, in which 3 MAT was in December. Temperature and rainfall were both lower (Figure 3) and daylength was shorter following the fall application, which would coincide with slower T. fluminensis regrowth potential (Maule Reference Maule1991).
Table 2. ANOVA summary and main effects of herbicide timing (spring vs. fall), herbicide treatment, and interactions of those effects on mean percent coverage of Tradescantia fluminensis by months after treatment (MAT) following herbicide application at two locations in Gainesville, FL.
a MAT, months after treatment. Spring treatments were applied March 14, 2022, and fall treatments on October 18, 2022. Asterisks indicate MAT in which there were significant herbicide × timing interactions.
b Main effects included herbicide and timing of herbicide applications including a spring (March) or fall (October) application in central Florida.
c P-values derived from ANOVA using PROC GLIMMIX in SAS® v. 9.4. Effects were considered significant at P ≤ 0.05 and are bold for presentation.
d Means within main effects or interactions (herbicide × timing) followed by the same letter are not significantly different (Tukey’s adjustment for multiplicity; P = 0.05).
e Herbicide control was assessed by taking visual coverage ratings of T. fluminensis on a 0% to 100% scale where 0 = no coverage and 100 = complete coverage.
Figure 3. Average monthly temperature (C) and rainfall (cm) in the Gainesville, FL, area over the study period. Weather data were not collected at each individual experimental site but were obtained from a Gainesville area weather station to provide general weather patterns during the experiment (Florida Automated Weather Network, http://fawn.ifas.ufl.edu).
There was significant timing by herbicide interactions, which occurred at 1, 5, 6, and 7 MAT. Interaction effects at 1 MAT showed that while five of the herbicides evaluated performed equally well regardless of timing, 2,4-D applied alone provided significantly better control when applied in the fall, resulting in 19% coverage versus 75% coverage for spring application, which was not different from the control. However, at 5, 6, and 7 MAT, lower coverage ratings were observed for 2,4-D plots treated in the spring, which was due to a reapplication that occurred at 3 MAT at both locations (Figure 2).
Evaluation dates in which no timing by herbicide interaction occurred included 2 through 4 MAT and 8 through 11 MAT (Table 2). At early evaluation dates, triclopyr-containing treatments (acid, amine, and in combination with 2,4-D) and glyphosate generally provided the highest level of control, followed by 2,4-D and aminopyralid, which were less efficacious. Fewer treatment differences were observed among herbicides at later evaluation dates, but it should be noted that by 9 MAT, 2,4-D had been reapplied at both locations following the initial spring application and twice following the fall application at Payne’s Prairie (Figure 2). Aminopyralid had also been reapplied following the spring application at the Nature Park and fall application at Payne’s Prairie, while glyphosate was reapplied once, at Payne’s Prairie following the initial spring application. None of the triclopyr-containing treatments required reapplication, as the initial application resulted in plots with <25% T. fluminensis coverage throughout the experiment, regardless of initial timing.
No previous literature on the influence of temperature or season with regard to herbicide efficacy for controlling T. fluminensis could be found in our review. However, this topic has been studied extensively with other invasive plants and weed species. Glyphosate and 2,4-D have both been shown to be more effective when controlling ragweed species (Ambrosia artemisiifolia L. and Ambrosia trifida L.) when applied at times of high temperature (29 C) compared with the same application at lower temperature (20 C) due to increased absorption and translocation (Gaine et al. Reference Gaine, Jugulam and Jhala2017). In this study, the fall application (22 C) was made under a higher temperature compared with the spring application (19 C), which might explain the higher efficacy observed following the fall application of 2,4-D alone, but the difference in temperature was minimal compared with more climate-focused studies. However, for some weed species, increased efficacy may be short-lived, and few differences in efficacy between high versus low temperatures have been reported when rated several weeks after application (Derr and Serensits Reference Derr and Serensits2016). Consistently high levels of control were achieved with triclopyr-containing herbicides following either application timing, similar to previous reports on triclopyr efficacy for control of swinecress [Coronopus didymus (L.) Sm.], in which control was high and consistent following February and April applications in Griffin, GA (Reed and McCullough Reference Reed and McCullough2012). Glyphosate tended to provide a high level of control and similar results regardless of initial application timing. While a reapplication was needed at one site following the spring application, it was not until 9 MAT. The growth cycle of T. fluminensis, in which growth tends to increase with increasing temperature and rainfall, likely contributed to regrowth more significantly than herbicide efficacy.
Overall, results from this experiment indicate that triclopyr was the most effective herbicide evaluated for control of T. fluminensis, and neither treatment timing nor formulation had any impact on efficacy. On certain evaluation dates, triclopyr applied alone as an acid or amine formulation provided higher control than the 2,4-D + triclopyr combination, but in the combination, the rate of triclopyr was about 16% of that when triclopyr was applied solely (Table 1). Aminopyralid and 2,4-D were the least effective options and required retreatment at both sites. It should also be noted that the initial application rate of aminopyralid was the maximum allowable dose; thus, in management scenarios, a repeated application could not be made. Similarly, at Payne’s Prairie following the fall application, the three applications of 2,4-D would have exceeded the label maximum annual dose had it been broadcast over the entire area. Both herbicides provided some control and significant reduction in coverage relative to the non-treated plots; our results suggest they would not be suitable as a management option. In contrast to our findings, 2,4-D has been shown to provide a high level of control of other Commelinaceae species such as Commelina benghalensis L. (benghal dayflower), but these studies focused on smaller plants and were short-term evaluations (Filus et al. Reference Filus, Barroso, Albrecht, Silva, Albrecht and Concatto2024). Based on our results, use of triclopyr would be the recommended management option for T. fluminensis, and application timing could be based more upon manager preference than season. Glyphosate would also likely be a useful herbicide and would likely perform similarly to triclopyr in the early months after application, but reapplications may be needed sooner to prevent regrowth. Further work is necessary to determine the optimum retreatment thresholds that could potentially result in no to minimal regrowth and optimum native plant restoration. As T. fluminensis is a ground cover species that may infest large and difficult to treat and access terrain, research should also address the feasibility and success of single selective herbicide applications followed by monitoring. None of the herbicides evaluated resulted in complete eradication; thus future research should focus on monitoring and retreatment frequency, as well as methods of restoring native plant communities to prevent regrowth of T. fluminensis.
Acknowledgments
The authors wish to acknowledge Keith Morin and Brian Law (Payne’s Prairie Preserve State Park) and Nicole Barbieri (Habitat Naturalist, City of Gainesville) for their assistance and allowing research to be conducted on lands they manage.
Funding statement
This research was funded by the Florida Fish and Wildlife Conservation Commission Invasive Plant Management Research grant program.
Competing interests
The authors declare no conflicts of interest.
Open access
