Click on author name to view affiliation information
Wildflowers attract wildlife, increase pollinator habitat, and enhance the aesthetic value of the landscape. Wildflower establishment is increasingly part of an effort to reduce maintained turfgrass on golf courses, lawns, and other maintained environments. Weed competition decreases wildflower establishment and results in poor long-term stands. Research was conducted in a controlled environment to investigate the tolerance of wildflower species to common postemergence herbicides. Wildflower species included California poppy (Eschscholzia californica Cham.), common sunflower (Helianthus annuus L.), cornflower (Centaurea cyanus L.), garden coreopsis (Coreopsis lanceolata L.), partridge pea [Chamaecrista fasciculata (Michx.) Greene], plains coreopsis (Coreopsis tinctoria Nutt.), purple coneflower [Echinacea purpurea (L.) Moench], rosering gaillardia (Gaillardia pulchella Foug.), and violet prairie clover (Dalea purpurea Vent.). Herbicides evaluated were fluazifop at 0.28 kg⋅ha–1 a.i., mesotrione at 0.14 kg⋅ha–1 a.i., clopyralid at 0.29 kg⋅ha–1 a.i., bentazon at 0.56 kg⋅ha–1 a.i., halosulfuron at 0.053 kg⋅ha–1 a.i., and imazaquin at 0.42 kg⋅ha–1 a.i. An untreated check was included for comparison. Excessive damage (≥ 53% phytotoxicity) was observed on all wildflower species in response to clopyralid, except for California poppy. Fluazifop and bentazon were relatively safe (≤ 19% phytotoxicity, regardless of herbicide) on all wildflower species; however, bentazon resulted in ≥ 40% aboveground biomass reduction in several species. Common sunflower and garden coreopsis were susceptible to halosulfuron (37% and 73% phytotoxicity, respectively) and imazaquin (37% and 87% phytotoxicity, respectively), but on all other wildflower species, phytotoxicity was ≤ 18%. Although both halosulfuron and imazaquin only resulted in ≤ 18% phytotoxicity to purple coneflower, a 43% to 44% aboveground biomass reduction was recorded. Mesotrione was only safe on California poppy and cornflower (≤ 11% phytotoxicity and ≤ 24% aboveground biomass reduction). Results suggest high tolerance variability across herbicides and species considered, but may prompt new investigation of safety and utility within field and production scenarios.
Worldwide declines in the abundance and diversity of insect pollinators are well documented (Biesmeijer et al. 2006; Kleijn et al. 2015; Potts et al. 2010; Steffan-Dewenter et al. 2005; Thomann et al. 2013). Although focus is often placed on bees, populations of other insect pollinators such as butterflies, moths, and beetles have also decreased (Cameron et al. 2011; Fox 2013; Goulson et al. 2015; Haaland et al. 2011; Wallisdevries et al. 2012). Reductions in insect pollinators impact global biodiversity, and affect negatively the productivity of agronomic crops (Gallai et al. 2009; Garibaldi et al. 2009; Kevan and Phillips 2001; Thomann et al. 2013) and the reproduction capability of a substantial number of native plants (Ashman et al. 2004; Kearns et al. 1998; Thomann et al. 2013). Habitat loss is a key contributor to insect pollinator declines (Winfree et al. 2009).
Although numerous initiatives have been implemented in recent years to curb the continued decline of pollinator insects, the introduction of wildflower habitats in and around maintained landscapes has led to substantial increases in pollinator density, diversity, and pollination services (Blaauw and Isaacs 2014a; Braman et al. 2002; Frank and Shrewsbury 2004; Haaland et al. 2011; Korpela et al. 2013; Williams et al. 2015). However, the presence of wildflowers aids much more than just pollinators. Conservation of diverse landscapes attracts natural predators that contribute to biological-based pest management and landscape resilience (Blaauw and Isaacs 2014b; Braman et al. 2002; Frank and Shrewsbury 2004; Gontijo et al. 2013; Tooker and Hanks 2000) or become a food resource for birds and other wildlife (Vickery et al. 2002, 2009; Wratten et al. 2012). Floral plantings may also increase the aesthetic value of the landscape by adding color and texture, or create a transition between manicured landscapes and unmanaged ecosystems (Ahern et al. 1992; Benvenuti 2014; Bretzel et al. 2016; Scott 2008).
A 2015 survey conducted by the Environmental Institute for Golf (funded by the United States Golf Association) concluded that 46% of all respondents reported an increase in the total acreage of natural areas on their golf courses during the past 10 years, whereas only 5% reported a decrease (Gelernter et al. 2017). These natural areas are typically comprised of heterogeneous grasslands that often contain combinations of native or commercially available grasses, forbs, and wildflowers (Gelernter et al. 2017). Efforts to convert managed turfgrass or disturbed habitats to natural areas on golf courses have been propelled by the numerous environmental benefits that can be conferred through their conservation. Managing areas as pollinator habitats may improve soil quality by limiting soil erosion and decreasing leaching of excess nitrogen (Nearing et al. 2005; Wratten et al. 2012), while protecting water quality by mitigating runoff of soils, fertilizers, and pesticides (Environmental Institute for Golf 2016; Wratten et al. 2012). Additional benefits may include reduced labor, fuel, and other inputs associated with intensively managed turfgrass (Ahern et al. 1992; Environmental Institute for Golf 2016; Nelson 1997).
Wildflowers are often established on previously disturbed land containing mixtures of native, nonnative, and often invasive vegetation. Site preparation typically consists of a postemergence application of a nonselective herbicide, followed weeks later by cultivation and debris removal (Frances et al. 2010). Wildflowers are often propagated by seed, so use of preemergence herbicides at seeding can reduce establishment and decrease stand density. The use of postemergence herbicides during establishment may limit weed interference; however, the overall tolerance of wildflowers to applied herbicides is largely unexplored.
Wildflower growth and establishment was most successful when competition for light, space, water, and nutrients from invasive grasses was reduced (Ahern et al. 1992). Minimal research has been conducted on the tolerance of wildflowers to postemergence herbicides, and studies have focused predominantly on monocot-specific chemistries such as the acetyl coenzyme A carboxylase inhibiting aryloxyphenoxypropionates and cyclohexanediones, which are uniquely well tolerated by most dicotyledonous species. Evaluation of wildflower tolerance to herbicides when grown in a greenhouse would eliminate negative effects of weed competition or environmental conditions. Therefore, the objective of our research was to evaluate the tolerance of several commonly planted wildflower species to both monocot- and dicot-specific herbicides in a controlled environment.
Research was conducted at the Athens Turfgrass Research and Education Center greenhouse complex (lat. 33.54°N, long. 83.22°W) in Athens, GA, during Summer and Fall 2017. California poppy (Eschscholzia californica Cham.), common sunflower (Helianthus annuus L.), cornflower (Centaurea cyanus L.), garden coreopsis (Coreopsis lanceolata L.), partridge pea [Chamaecrista fasciculata (Michx.) Greene], plains coreopsis (Coreopsis tinctoria Nutt.), purple coneflower [Echinacea purpurea (L.) Moench], rosering gaillardia (Gaillardia pulchella Foug.), and violet prairie clover (Dalea purpurea Vent.) (Applewood Seed Co., Golden, CO, USA) were seeded at 25 kg⋅ha–1 into 15.2-cm circular pots containing a perlite-free grower-grade Canadian sphagnum peatmoss potting media (Sun Gro Horticulture, Agawam, MA, USA) on 12 May 2017 and 14 Sep 2017. Wildflower species were selected based on their high frequency of incorporation into common pollinator seed mixtures. Only common sunflower was thinned to six plants per pot after germination. Fertilizer (7N–7P2O5–7K2O) (Grigg Seven Iron; Brandt Consolidated, Inc., Springfield, IL, USA) was applied at the time of seeding at a rate of 25 kg⋅ha–1 nitrogen. Greenhouse temperatures were maintained at 29/25 °C (day/night) with average midday (1200 and 1300 HR) solar radiation ranging from 636 to 754 µmol⋅m–2⋅s–1. Irrigation was supplied through an overhead irrigation system calibrated to deliver ∼3.8 cm of water per week. Wildflower species were allowed to grow for ∼5 weeks before herbicide application.
Treatments were arranged in a 9 × 7 factorial (nine wildflower species × seven herbicide treatments) within a randomized complete block design with five replications repeated twice in time (two experimental runs). Experimental blocks were arranged along a gradient created by the greenhouse cooling pads and associated fans. Herbicides were applied once per experimental run using a carbon dioxide pressurized spray chamber (Generation III Spray Booth; DeVries Manufacturing, Inc., Hollandale, MN, USA) equipped with an XR8002VS nozzle tips (Teejet, flat-fan extended-range spray tips; Spraying Systems Company, Glendale Heights, IL, USA) calibrated to deliver 375 L⋅ha–1 at 221 kPa. Treatments were applied on 16 Jun 2017 (run 1) and 19 Oct 2017 (run 2) and consisted of fluazifop (Fusilade II; Syngenta Crop Protection, LLC, Greensboro, NC, USA) at 0.28 kg⋅ha–1 a.i., mesotrione (Tenacity, Syngenta Crop Protection, LLC) at 0.14 kg⋅ha–1 a.i., clopyralid (Lontrel; Corteva Agriscience, Wilmington, DE, USA) at 0.29 kg⋅ha–1 a.i., bentazon (Basagran; BASF Corp., Research Triangle Park, NC, USA) at 0.56 kg⋅ha–1 a.i., halosulfuron (Sedgehammer; Gowan Company, Yuma, AZ, USA) at 0.053 kg⋅ha–1 a.i., and imazaquin (Scepter T&O 70 WDG; AMVAC Chemical Corp., Newport Beach, CA, USA) at 0.42 kg⋅ha–1 a.i. Herbicide rates were chosen based on label restrictions and weed control spectrum. No surfactants were added to any treatment to minimize potential phytotoxicity. A untreated check was included for comparison. Wildflower species were at the following heights at the time of herbicide application: California poppy, 15.2 to 16.5 cm; common sunflower, 25.4 to 35.6 cm; cornflower, 17.8 to 20.3 cm; garden coreopsis, 15.2 to 19.1 cm; partridge pea, 14.0 to 15.2 cm; plains coreopsis, 30.5 to 35.6 cm; purple coneflower, 11.4 to 14.0 cm; rosering gaillardia, 14.0 to 15.2 cm; and violet prairie clover, 6.4 to 7.6 cm.
Visual ratings of percent wildflower phytotoxicity were recorded 3 weeks after treatment (WAT) on a scale of 0% (no phytotoxicity) to 100% (complete necrosis) (Frans et al. 1986). Plant pots were harvested on 7 Jul 2017 (run 1) and 9 Nov 2017 (run 2) 3 WAT by cutting all aboveground foliage at the soil surface with scissors. Plant tissue was dried at 50 °C for 7 d and weighed to obtain biomass (measured in grams). Percent aboveground biomass reduction was determined for each treatment by comparing treated pots with the untreated check within each trial replication using Eq. [1]: [1] where BR is the aboveground biomass reduction, C is the aboveground biomass of the untreated check pots 3 weeks after trial initiation, and T is the aboveground biomass of treated pots 3 WAT. Percentage of aboveground biomass reduction used a scale of 0% to 100%, where 0% was no aboveground biomass reduction and 100% was complete aboveground biomass reduction.
Analysis of variance was performed using the general linear models procedure (PROC GLM) in SAS (version 9.2; SAS Institute, Cary, NC, USA) using the appropriate expected mean square values described by McIntosh (1983). Means were separated using Fisher’s protected least significant difference test at α = 0.05. All data were arcsine square root–transformed before analysis to stabilize variance (Bowley 2008). Interpretations were not different from nontransformed data; therefore, nontransformed means are presented for clarity.
Experimental run-by-treatment interactions were not detected (F = 0.49, P = 0.87) for wildflower species phytotoxicity; therefore, data were pooled across experimental runs. A significant wildflower species (F = 43.5, P ≤ 0.0001) and herbicide main effect (F = 1.17, P = 0.02) was observed for percent phytotoxicity 3 WAT (Table 1). Experimental run-by-treatment interactions were not detected (F = 1.06, P = 0.63) for wildflower species aboveground biomass; therefore, data were pooled across experimental runs. A significant wildflower species (F = 2.23, P = 0.001) and herbicide main effect (F = 27.8, P ≤ 0.0001) was observed for aboveground biomass 3 WAT (Table 2). Experimental run-by-treatment interactions were not detected (F = 0.86, P = 0.74) for wildflower species aboveground biomass reduction; therefore, data were pooled across experimental runs. A significant wildflower species (F = 15.21, P ≤ 0.0001) and herbicide main effect (F = 108.68, P ≤ 0.0001) was observed for aboveground biomass reduction 3 WAT (Table 3).
Fluazifop applications resulted in ≤ 3% phytotoxicity 3 WAT, regardless of wildflower species (Table 1). This coincided with similar aboveground biomass as the untreated check (except for garden coreopsis and common sunflower) (Table 2) and only resulted in ≤ 8% aboveground biomass reduction compared with the untreated check for all wildflower species except for California poppy (Table 3). This was expected because fluazifop is typically used for the control of annual and perennial grass weeds in dicot cropping systems. Similarly, Olszyk et al. (2013) observed little to no response of several forb, species including western buttercup (Ranunculus occidentalis Nutt.), sicklekeel lupine (Lupinus albicaulis Douglas), farewell to spring [Clarkia amoena (Lehm.) A. Nelson & J.F. Macbr.], common madia (Madia elegans D. Don ex Lindl.), heal-all (Prunella vulgaris L.), and slender cinquefoil (Potentilla gracilis Douglas ex Hook.) to fluazifop at 0.28 kg⋅ha–1 a.i. Blake et al. (2012) also reported minimal to no phytotoxicity as well as no reduction in plant biomass of nine wildflower species when treated with a similar rate of fluazifop postemergence.
California poppy and cornflower were the only wildflower species that exhibited tolerance to mesotrione 3 WAT (5% and 11% phytotoxicity, respectively) (Table 1). However, mesotrione applications still resulted in 24% to 27% aboveground biomass reductions within these same species, respectively (Table 3). Contrarily, York and Seagroves (2008) reported 62% injury for cornflower and 53% injury for California poppy 3 WAT in response to mesotrione at 0.11 kg⋅ha–1 a.i. Although mesotrione in their study was applied at a lower rate (0.11 kg⋅ha–1 a.i. vs. 0.14 kg⋅ha–1 a.i. in our study), the addition of a crop oil concentrate may have caused more injury to plants examined in their research. All other wildflower species expressed 20% to 45% phytotoxicity and reductions in aboveground biomass of 39% to 58% (Tables 1–3). York and Seagroves (2008) noted even greater injury (59%–71%) for plains coreopsis, Gaillardia sp., and garden coreopsis in response to mesotrione 3 to 5 WAT.
Clopyralid only caused 3% phytotoxicity 3 WAT when applied to California poppy, but resulted in 53% to 94% phytotoxicity (almost complete death) for all other wildflower species (Table 1). Aboveground biomass was only reduced 9% in California poppy, but ≥ 41% in the other wildflower species (Table 3). York and Seagroves (2008) observed slightly higher phytotoxicity (19%) of California poppy 4 WAT when subjected to a lower rate of clopyralid (0.11 kg⋅ha–1 a.i. vs. 0.29 kg⋅ha–1 a.i. in our study); however, a nonionic surfactant was added to the mixture. They also observed unacceptable phytotoxicity (56%–75%) in a Gaillardia sp., garden coreopsis, cornflower, and plains coreopsis. Partridge pea was the most sensitive wildflower species to clopyralid (94% phytotoxicity and 72% aboveground biomass reduction) in our research (Table 3). Similarly, Meyer and Bovey (1991) observed reductions in partridge pea cover of 83% to 100% in multiple trial locations in response to clopyralid at 0.28 kg⋅ha–1 a.i.
Bentazon was well tolerated by most wildflower species (partridge pea, purple coneflower, violet prairie clover, rosering gaillardia, and California poppy were similar to the untreated check; ≤ 5% phytotoxicity) whereas the greatest phytotoxicity was observed on common sunflower and garden coreopsis (14% and 19%, respectively) 3 WAT (Table 1). Moderate aboveground biomass reductions (40%–44%) were observed on plains coreopsis, purple coneflower, and common sunflower, whereas ≤ 26% reductions were observed in the remainder of wildflower species (Table 3). Although bentazon resulted in no phytotoxicity to violet prairie clover, a 26% reduction in aboveground biomass was observed. This was slightly higher than the stunting (≤ 14%) recorded on ‘Durana’ white clover (Trifolium repens L.) by Basinger and Hill (2021) 10 WAT with a higher rate of bentazon (0.84 kg⋅ha–1 a.i. vs. 0.56 kg⋅ha–1 a.i. in our study).
Halosulfuron was most injurious to garden coreopsis (73% phytotoxicity) and common sunflower (37% phytotoxicity) 3 WAT (Table 1). This resulted in a 45% and 37% reduction in aboveground biomass of these same species, respectively (Table 3). All other species resulted in ≤ 16% phytotoxicity in response to halosulfuron. This coincided with ≤ 13% aboveground biomass reduction for those same species, except purple coneflower, which resulted in a 43% reduction. Although applied at a higher rate (0.14 kg⋅ha–1 a.i. vs. 0.053 kg⋅ha–1 a.i. in our study), Wiese et al. (2011) also reported halosulfuron safety on white prairie clover (Dalea candida Michx. Ex Willd.) 20 d after treatment. However, a 90% chance of phytotoxicity in response to halosulfuron was predicted (according to the Monte Carlo simulation) for blanketflower (Gaillardia aristata Pursh), as well as a reduction in biomass compared with the untreated check (Wiese et al. 2011). This was much greater than the phytotoxicity observed on rosering gaillardia (19%) in our research, but may also be related to the higher halosulfuron use rate.
Imazaquin resulted in a similar response as halosulfuron, with the highest phytotoxicity (87% and 37%) observed in garden coreopsis and common sunflower, respectively; however, the greatest aboveground biomass reduction (57% and 44%) was recorded in garden coreopsis and purple coneflower, respectively, 3 WAT (Tables 1–3). All other wildflower species exhibited ≤ 18% phytotoxicity and ≤ 23% aboveground biomass reduction in response to imazaquin. McCurdy et al. (2013) reported a differential response among Trifolium species to imazaquin at 5.6 g⋅100 m–2 a.i. plus a nonionic surfactant. Control of white clover, crimson clover (Trifolium incarnatum L.), ball clover (Trifolium nigrescens Viv.), and small hop clover (Trifolium dubium Sibth) in the field was 50%, 62%, 80%, and 91%, respectively, 6 WAT with imazaquin. This coincided with a height reduction in small hop clover of 88%, whereas height was reduced ≤ 47% for all other Trifolium species Although applied preemergence, Beran et al. (1999) reported no significant reduction in percent emergence and flower density of purple coneflower and blanketflower compared with the untreated check in response to imazaquin at 70 g⋅ha–1 a.i.
Clopyralid phytotoxicity was ≥ 53% for all but one wildflower sp., California poppy, which was tolerant to all herbicides tested. Fluazifop and bentazon were relatively safe (≤ 3% and ≤ 19% phytotoxicity, respectively) on all wildflower species; however, bentazon resulted in ≥ 40% aboveground biomass reduction in several species. These two herbicides may provide effective grass and immature broadleaf weed control, respectively. Common sunflower and garden coreopsis were susceptible to halosulfuron (37%–73% phytotoxicity) and imazaquin (37%–87% phytotoxicity), but all other wildflower species resulted in ≤ 18% phytotoxicity to these same chemistries. Although both halosulfuron and imazaquin only resulted in ≤ 18% phytotoxicity to purple coneflower, a 43% to 44% aboveground biomass reduction was recorded in response to both herbicides. Annual and perennial sedge and broadleaf weed species are often controlled with these two herbicides. Mesotrione was only safe on California poppy and cornflower (≤ 11% phytotoxicity and ≤ 24% aboveground biomass reduction); however, this herbicide may provide preemergence and postemergence control of weeds.
Wildflower plants that exist in nature are often subjected to different environmental conditions (moisture, temperature, light intensity/quality, etc.) than those propagated in a controlled environment. Greenhouse conditions are maintained to promote optimal plant growth and health; therefore, wildflower species grown in controlled environments may exhibit morphological and physiological characteristics that are more conducive to herbicide absorption, translocation, and subsequent activity (phytotoxicity and growth reduction). For example, clopyralid applied postemergence to Canada thistle [Cirsium arvense (L.) Scop] at 65% relative humidity was absorbed and translocated more (80% and 55%, respectively) than when it was applied at 35% relative humidity (60% and 45%, respectively) (Kloppenburg and Hall 1990). Therefore, less phytotoxicity and plant biomass reduction may occur when wildflower species are subjected to herbicides in field plantings. Wirestem muhly [Muhlenbergia frondosa (Poir.) Fern.] control was 98% 4 WAT in response to glyphosate (0.42 and 0.84 kg⋅ha–1) applied in the greenhouse, but only 60% to 87% in the field (Lingenfelter and Curran 2007). Similarly, Cooper et al. (2016) reported 100% bermudagrass [Cynodon dactylon (L.) Pers.] control after sequential applications of metamifop (0.3–0.5 kg⋅ha–1 a.i.) applied in the greenhouse, whereas Doroh et al. (2011) only observed 36% bermudagrass control in the field in response to sequential applications of metamifop at 0.4 kg⋅ha–1 a.i. Future research should evaluate herbicide response in the field when wildflower species are grown as monocultures and mixed stands under a variety of environmental conditions and/or geographies.
Initial wildflower establishment from seed is often unsuccessful; therefore, herbicides are important for the reduction of weed competition and subsequent promotion of successional wildflower establishment through seed and vegetative propagules. Angelella et al. (2019) observed 9% wildflower establishment of a pollinator refuge habitat during the first year, but 49% wildflower cover by the end of year 2. In general, little is known regarding herbicide use within mixed-wildflower meadow establishment. Practitioners must consider whether existing stands of turfgrass species will be included intentionally as a component of the meadow, or whether they should be removed before or during succession toward a native or naturalized stand. Weed species communities within these locations may also shift over time, further complicating control and herbicide selection and safety. Grasses dominated the landscape during the first year of pollinator habitat establishment by Angelella et al. (2019) and averaged 89% cover. However, a shift in weed species dynamics during year 2 resulted in 55% broadleaf cover and only 15% grass cover.
Fluazifop, as with other acetyl coenzyme A carboxylase–inhibiting “grass herbicides” is a novel means of controlling or suppressing grasses within broadleaf crops and meadow mixtures. Love et al. (2016), and others, have proposed a “grass-first” strategy for the establishment of urban wildflower meadows, whereby grasses are established before the introduction of wildflower species This strategy allows for suppression of nuisance weeds with broadleaf herbicides before introduction of sensitive, intentionally included wildflowers. The mixture of grasses should be considered for tolerance to a chosen grass herbicide, but the utility of grass removal and thinning is an added value of grass herbicides to manipulate succession and stand establishment.
Other herbicides considered within this study offer a range of options for manipulation of sward complexity, but none are a panacea for general weed control within mixed meadows. A tailored approach for each scenario will be required and, undoubtedly, nonselective herbicides or hand-removal will be required for the control of some nuisance weeds.
Contributor Notes
We thank Caydee Savinelli and Syngenta Crop Protection, LLC, for donating the wildflower seed. Additional thanks to Ross Shilling for technical assistance with trial harvest.
G.H. is the corresponding author. E-mail: gmhenry@uga.edu.
