Seed Sources and Research Site

Three Palmer amaranth accessions were included in the experiments: a putative herbicide-resistant accession (BRO) identified near Broadhead, WI (42.6183°N, 89.3762°W) in 2018 and two control accessions from Nebraska, a glyphosate-resistant accession (KEI2) and a glyphosate-susceptible accession (KEI3), both from Keith County, NE (for complete information regarding the control accessions, see Oliveira et al. 2020). Seeds from the BRO accession were collected from a field cultivated with soybean, whereas the KEI2 and KEI3 accession was from a field cultivated with soybean and corn, respectively (Oliveira et al. 2020); herbicide use records of all accessions were not available. After collection from the field, seeds were threshed, cleaned, and stored at 5 C until the onset of the experiments, which were conducted at the University of Wisconsin–Madison Walnut Street Greenhouses (43.076194°N, 89.423611°W), Madison, WI.

Palmer Amaranth Response to POST Herbicides

The experiment was organized in a randomized complete block design (RCBD) with eight replications per treatment and repeated over time (two experimental runs). Treatments were arranged as 3 × 8 × 2 factorial consisting of three accessions (BRO, KEI2, and KEI3), eight herbicides (Table 1), and two herbicide rates (1X and 3X the recommended label rates). A nontreated control (NTC) of each accession was included.

Palmer amaranth seeds were planted at 1.5-cm depth in potting mix (PRO-MIX^® HP MYCORRHIZAE^™, Premier Tech Horticulture, Rivière-du-Loup, QC, Canada) in 23-cm-diameter disposable aluminum pans. Seedlings at the 2-true-leaf stage were transplanted into potting mix as described, contained in 656-mL pots (D40H Deepot^™, Stuewe and Sons Inc., Tangent, OR, USA). The experimental unit was one seedling per pot. Postemergence herbicide treatments were applied when plants reached 5 to 10 cm in height (4- to 6-true-leaf stage) using a single-nozzle research track spray chamber (DeVries Manufacturing, Hollandale, MN, USA) equipped with a TP8002EVS nozzle (TeeJet^® Technologies, Wheaton, IL, USA). Owing to vapor drift concerns within an enclosed environment (greenhouse) with the presence of several sensitive broadleaf species, the dicamba and 2,4-D herbicide treatments were applied at the University of Wisconsin–Madison Arlington Agricultural Research Station (43.302631°N, 89.345367°W). Palmer amaranth plants were transported to this field location on the morning of the application and returned to the greenhouse at the end of the day to allow for better herbicide absorption while minimizing potential unintended vapor drift issues. A CO₂-pressurized backpack spray boom with four TTI110015 nozzles (TeeJet^® Technologies) was used for the application. A carrier volume of 140 L ha^–1 was used in all applications (spray chamber and backpack). Plants were maintained in the greenhouse at 20 to 35 C with a natural ventilation system. Natural lighting was supplemented with 400 W high-pressure sodium lightbulbs simulating a 16-h photoperiod. The soil was watered daily and fertigated weekly with 20-10-20 water-soluble fertilizer (Peters^® Professional, ICL Fertilizers, Dublin, OH, USA) delivering 500 ppm of both N and K and 250 ppm of P.

At 21 days after treatment (DAT), plant survival was assessed visually as dead (no green tissue; assessed value of 0) or alive (green tissue and evidence of regrowth; assessed value of 1; Figure 1). Accessions with ≥50% (± standard error) plant survival were classified as resistant to each herbicide × rate treatment (adapted from Schultz et al. 2015; Vennapusa et al. 2018). Aboveground biomass was harvested and force air-dried at 52 C to constant mass. The biomass data were converted into percent biomass reduction compared to the NTC using Equation 1 (adapted from Wortman 2014):

Table 1.

Postemergence herbicide treatments used to evaluate the response of three Palmer amaranth accessions.^a

Figure 1.

Plant survival rating used for herbicide resistance classification for Palmer amaranth response to POST herbicides.

where BEU represents the biomass of the experimental unit and represents the biomass mean of the NTC for the respective accession. Seed production of survivor plants was not determined.

Palmer Amaranth Response to PRE Herbicides

The experiment was organized in a RCBD with four replications per treatment and repeated over time (two experimental runs). Treatments were arranged as 3 × 5 × 3 factorial consisting of three accessions (BRO, KEI2, and KEI3), five herbicides (Table 2), and three herbicide rates (0.5X, 1X, and 3X the recommended label rate). A NTC of each accession was included.

Experimental units were approximately 130 seeds (measured by volume) planted 1.5 cm deep in 360-mL pots (8.9 cm Kord Traditional Square Pot, HC Companies, Twinsburg, OH, USA) filled with nonsterilized field soil (silt loam; 7.0 pH; 2.8% organic matter; 21% sand, 57% silt, 22% clay by weight). The soil was watered immediately after planting and before herbicide application to facilitate seed germination and herbicide activation. Preemergence herbicide treatments were applied using the spray chamber and carrier volume described earlier, equipped with a AI9502EVS nozzle (TeeJet^® Technologies). Plants were maintained in a greenhouse under the same conditions described previously.

At 25 DAT, emerged plants per experimental unit were counted. The count data were converted into percent plant density reduction compared with the NTC using Equation 2 (adapted from Wortman 2014):

where PCEU represents the plant counts of the experimental unit and represents the plant counts mean of the NTC for the respective accession.

Herbicide × rate treatments that provided <90% (± standard error) plant density reduction were classified as ineffective for each accession (adapted from Vennapusa et al. 2018).

Table 2.

Preemergence herbicide treatments used to evaluate the response of three Palmer amaranth accessions.^a

Statistical Analyses

A generalized linear mixed model with Gaussian distribution was fitted to the biomass reduction data (POST) and plant density reduction data (PRE) using the glmmTMB package version 1.0.2.1 (Brooks et al. 2017). Analysis of variance (ANOVA) type II Wald chi-square was performed followed by Tukey's honestly significant difference (α = 0.05) pairwise comparisons using the emmeans package version 1.5.4 (Lenth 2020). Both response variables were logit-transformed to improve normality assumptions (Barnes et al. 2020; Davies et al. 2019, 2020; Striegel et al. 2020; Warton and Hui 2011). Back-transformed means are presented. Accession, herbicide, and rate were considered as fixed effects, whereas experimental run was considered as a random effect. Statistical analyses were performed using R version 4.0.3 (R Core Team 2020).

Palmer Amaranth Response to POST Herbicides

Plant survival of each accession was ≥63% after exposure to imazethapyr POST 3X rate (Figure 2). Survival of BRO and KEI2 was 44% (±13) and 50% (±13), respectively, after exposure to atrazine POST 3X rate. Survival of BRO was 69% (±12) after exposure to glyphosate POST 1X rate, whereas survival of KEI2 was 44% (±13) after exposure to glyphosate POST 3X rate. After exposure to 2,4-D POST 1X rate, KEI2 and KEI3 survival was 38% (±13) and 50% (±13), respectively. Survival of all accessions was ≤31% after exposure to 2,4-D POST 3X rate or dicamba, glufosinate, lactofen, and mesotrione POST at either rate evaluated in this study. No plants of any accession survived exposure to glufosinate at either rate. Our glyphosate results for KEI2 and KEI3 accessions corroborate the findings of Oliveira et al. (2020), who reported these accessions as glyphosate-resistant and glyphosate-susceptible, respectively.

The ANOVA exhibited a significant three-way interaction among accession, herbicide, and rate for biomass reduction (P value ≤ 0.0001). For imazethapyr 1X rate, the biomass reduction did not differ between KEI3 (67%) and KEI2 (50%) nor between KEI2 and BRO (30%; Figure 3). For imazethapyr 3X rate, the biomass reduction did not differ between KEI2 (88%) and KEI3 (81%), which was greater than for BRO (43%). For glyphosate 1X rate, the biomass reduction was greater for KEI3 (98%) than for BRO (65%), which was greater than it was for KEI2 (33%). The biomass reduction did not differ among accessions for glyphosate 3X rate (each ≥ 96%). For atrazine 1X rate, biomass reduction was greater for KEI3 (97%) than for BRO (89%) and KEI2 (80%), which did not differ. For atrazine 3X rate, biomass reduction did not differ between KEI3 (97%) and BRO (95%) but was greater than it was for KEI2 (88%). The biomass reduction did not differ among accessions for 2,4-D, dicamba, glufosinate, lactofen, and mesotrione at either rate (≥91%).

The reduced performance of imazethapyr and glyphosate POST on the three Palmer amaranth accessions evaluated in our study is consistent with previous findings (Chahal et al. 2017; Drewitz et al. 2016; Kumar et al. 2020; Norsworthy et al. 2008; Oliveira et al. 2020; Schwartz-Lazaro et al. 2017). The adoption of genetically modified herbicide-resistant crops substantially reduced herbicide SOA diversity in cotton and soybean cropping systems in past decades (Kniss 2018), and the overreliance on a single herbicide, such as glyphosate, contributed to rapid resistance evolution (Culpepper et al. 2006; Legleiter and Bradley 2008; Norsworthy et al. 2012; VanGessel 2001). Recently, field escapes and greenhouse screenings have identified Palmer amaranth accessions resistant to dicamba and glufosinate in TN and AR, respectively (Barber et al. 2021; Steckel 2020), threatening the sustainability of recently introduced herbicide-tolerant soybean traits in the market. Additionally, several Palmer amaranth accessions have been confirmed resistant to multiple SOAs (Kohrt et al. 2017; Kumar et al. 2018; Schwartz-Lazaro et al. 2017), with one known accession confirmed resistant to five SOAs: ALS, PSII, AM, EPSPS, and HPPD (Kumar et al. 2019).

Palmer Amaranth Response to PRE Herbicides

Plant density reduction of KEI2 was 77% (±13) after exposure to atrazine PRE 1X rate, whereas density reduction of BRO was 56% (±13) after exposure to atrazine PRE 3X rate (Figure 4). After exposure to mesotrione PRE 0.5X rate, BRO and KEI plant density reduction was 83% (±8) and 83% (±12), respectively. Plant density reduction of all accessions was ≥94% after exposure to mesotrione PRE 1X and 3X rates or metribuzin, S-metolachlor, and sulfentrazone PRE at either rate evaluated in this study.

The three-way interaction among accession, herbicide, and rate was not significant for plant density reduction (P value = 0.75). The ANOVA exhibited a significant two-way interaction between accession and herbicide for plant density reduction (P value < 0.0001). For atrazine, plant density reduction was greater for KEI3 (95%) than it was for KEI 2 (83%), which were greater than it was for BRO (34%; Figure 5). Plant density reduction did not differ among accessions for mesotrione, metribuzin, S-metolachlor, or sulfentrazone (≥95%). Comparing atrazine and metribuzin, both PSII inhibitors but from different chemical families (triazine and triazinone, respectively), we observed different responses when applied PRE. Similarly, Vennapusa et al. (2018) reported higher efficacy of metribuzin than atrazine for control of waterhemp [Amaranthus tuberculatus (Moq.) J. D. Sauer] accessions from Nebraska, both applied PRE. In contrast, Schwartz-Lazaro et al. (2017) reported higher mortality of Palmer amaranth accessions from Arizona with atrazine compared to metribuzin, both applied PRE. Additionally, Fuerst et al. (1986) observed cross-resistance between atrazine and metribuzin applied PRE to smooth pigweed (Amaranthus hybridus L.).

Figure 2.

Palmer amaranth plant survival (± standard error) of accessions from Wisconsin (BRO) and Nebraska (KEI2 and KEI 3) in response to POST herbicides. Accessions with survival ≥50% (represented by the red line) were classified as ineffectively controlled by each herbicide × rate treatment.

Figure 3.

Palmer amaranth biomass reduction of accessions from Wisconsin (BRO) and Nebraska (KEI2 and KEI 3) represented by the three-way interaction among accession, POST herbicide, and rate. The blue boxes represent the 95% confidence intervals. Treatments with the same letters did not differ according to Tukey's honestly significant difference, α = 0.05.

Figure 4.

Palmer amaranth plant density reduction (± standard error) of accessions from Wisconsin (BRO) and Nebraska (KEI2 and KEI 3) in response to PRE herbicides. Treatments with plant density reduction <90% (represented by the red line) were classified as ineffective.

The ANOVA also exhibited a significant two-way interaction between herbicide and rate for plant density reduction (P value = 0.0001). At the 0.5X rate, plant density reduction for sulfentrazone did not differ compared to S-metolachlor and metribuzin (each ≥95%) and was greater than for mesotrione (92%) and atrazine (52%; Figure 6). At the 1X and 3X rates, plant density reductions for sulfentrazone, S-metolachlor, metribuzin, and mesotrione (each ≥97%) were greater than it was for atrazine (≤90%). The use of reduced PRE herbicide rates as an attempt to reduce costs, herbicide carryover, and/or environmental impacts may increase the selection pressure and lead to rapid herbicide-resistance evolution (Belz 2020; Manalil et al. 2011; Maxwell and Mortimer 1994; Norsworthy 2012; Tehranchian et al. 2017; Vieira et al. 2020). Our results suggest that herbicides applied PRE at the 0.5X label rate may provide reduced Palmer amaranth control, particularly for atrazine and mesotrione. Consequently, the reliance on herbicides applied POST may increase, and in the end, the short-term economic benefits associated with using reduced herbicide rates are quickly outweighed by the future costs related to herbicide-resistance evolution and spread (Gressel 1997).

Figure 5.

Palmer amaranth plant density reduction of accessions from Wisconsin (BRO) and Nebraska (KEI2 and KEI 3) represented by the two-way interaction between accession and PRE herbicide. The blue boxes represent the 95% confidence intervals. Treatments with the same letters did not differ according to Tukey's honestly significant difference, α = 0.05.

Figure 6.

Palmer amaranth plant density reduction represented by the two-way interaction between PRE herbicide and rate. The blue boxes represent the 95% confidence intervals. Treatments with the same letters did not differ according to Tukey's honestly significant difference, α = 0.05.

The Concerns of Palmer Amaranth Introduction in Wisconsin

The indication that this recently introduced Palmer amaranth accession (BRO) in Wisconsin is likely to carry multiple herbicide-resistance traits is cause for great concern. The most notable source of new Palmer amaranth infestations in Iowa, Ohio, Illinois, and Minnesota was credited to the use of Palmer amaranth–contaminated seeds for CRP (Hartzler and Anderson 2016; Yu et al. 2021). The 2021 State-Noxious-Weed Seed Requirements Recognized in the Administration of the Federal Seed Act (USDA 2021) designates Palmer amaranth as a prohibited noxious weed seed in Wisconsin, prohibiting the sale of agricultural seeds contaminated with Palmer amaranth seed. Similarly, Iowa and Minnesota designate Palmer amaranth as a noxious weed, whereas Illinois and Michigan do not. Minnesota went beyond and now requires a genetic test of any Amaranthus contaminant to determine if Palmer amaranth is present in agricultural seeds (USDA 2021; Yu et al. 2021).

Animal feed contaminated with Palmer amaranth seeds and subsequent manure applications have been reported as a possible cause of Palmer amaranth spread. In 2018, the Minnesota Department of Agriculture identified animal feed and manure as pathways for the introduction of Palmer amaranth in the state, after contaminated sunflower feed was used for cattle (Yu et al. 2021). Whole cottonseed is another example of a low-cost by-product with good nutritional value commonly used in dairy diets (Warner et al. 2020). If not properly monitored, it may become a pathway for Palmer amaranth introduction in new areas, particularly because the Cotton Belt is one of the areas in the United States most harshly affected by Palmer amaranth (Norsworthy et al. 2014; Ward et al. 2013; Webster and Nichols 2012). Kellog et al. (2001) reported that from 133 dairy farms surveyed across the United States, 71% used whole cottonseed as a feed source, with the greatest use in the western United States. The 2017 to 2018 Wisconsin Statutes and Annotations, in chapter 94.72, “Commercial Feed” (Wisconsin Statutes 2020), do not list Palmer amaranth as a noxious weed seed in commercial feed, which is cause for concern. More research is needed to evaluate the impact of animal feed sources on dispersal of noxious weed seeds in Wisconsin, the second-largest dairy state in the United States, with a production of 13.88 million tons of milk in 2018 and a herd size of 1.28 million cows distributed among approximately 9,037 farms (USDA 2020).

In conclusion, our results suggest that each accession is resistant (≥50% survival) to imazethapyr POST, that BRO and KEI2 accessions are resistant to atrazine and glyphosate POST, and that KEI2 and KEI3 are resistant to 2,4-D POST. In contrast, each accession was susceptible (<50% survival) to dicamba, glufosinate, lactofen, and mesotrione POST. The recently introduced BRO accession exhibited multiple resistance to imazethapyr, atrazine, and glyphosate POST. In addition, atrazine PRE was ineffective (<90% plant density reduction) for BRO control. Metribuzin, sulfentrazone, S-metolachlor, and mesotrione PRE effectively controlled (≥90% plant density reduction) each accession at 1X and 3X rates. Atrazine and mesotrione PRE at 0.5X rate provided reduced Palmer amaranth control and may impose selection pressure on POST herbicides. Community efforts, training, economic incentives, policies, and proactive scouting to prevent new Palmer amaranth infestations, which, according to our findings, are likely to carry herbicide resistance, and the use of effective PRE and POST herbicides as part of an IWM are vital for Palmer amaranth management in Wisconsin.