Management of Palmer Amaranth (Amaranthus palmeri ) in Glufosinate-Resistant Soybean (Glycine max) with Sequential Applications of Herbicides

Palmer amaranth (Amaranthus palmeri S. Wats.) is one of the most difficult weeds to control in soybean (Glycine max (L.) Merr.) in North Carolina. Research was conducted during 2010 and 2011 to determine if Palmer amaranth control and soybean yield were affected by soybean plant population and combinations of preemergence (PRE) herbicides followed by a single application of glufosinate postemergence (POST) versus multiple applications of glufosinate POST. Palmer amaranth was controlled more, and soybean yield was greater when soybean was established at 483,000 plants ha−1 in 3 of 4 experiments compared with soybean at 178,000 plants ha−1 irrespective of herbicide treatments. In separate experiments, application of PRE herbicides followed by POST application of glufosinate or multiple POST applications of glufosinate provided variable Palmer amaranth control, although combinations of PRE and POST herbicides controlled Palmer amaranth the most and provided the greatest soybean yield. In 1 of 3 experiments, sequential applications of glufosinate were more effective than a single application. Yield was higher in 2 of 3 experiments when glufosinate was applied irrespective of timing of application when compared with the nontreated control. In the experiment where glufosinate was applied at various POST timings, multiple applications of the herbicide provided the best control and the greatest yield compared with single applications.


Introduction
Resistance of Palmer amaranth to acetolactate synthase (ALS-) inhibiting herbicides and glyphosate is well documented in North Carolina [1][2][3]. Controlling herbicideresistant Palmer amaranth can be challenging because alternatives to glyphosate and ALS-inhibiting herbicides require timely application [4,5]. Row pattern and plant population can influence weed management in soybean [6,7]. Herbicides that applied PRE or POST other than glyphosate are important in managing glyphosate-resistant Palmer amaranth in soybean [8][9][10]. Under weed-free conditions, soybean at populations below those established in many fields in North Carolina yield as well as soybean at higher populations considered "standard" (exceeding 300,000 plants ha −1 ) (E. J. Dunphy, personal communication). Establishing soybean at lower populations can reduce production costs if yields are maintained. Results comparing the economic value of high soybean populations versus increasing the number of POST herbicides are mixed [11][12][13][14][15][16]. Additional research is needed to determine if Palmer amaranth control is more effective when soybean is established at higher populations in glufosinate-resistant soybean.
Glufosinate-resistant soybean allows producers to apply glufosinate to control problematic weeds like Palmer amaranth [17]. Sequential programs, either as PRE followed by POST applications or multiple POST applications of herbicides, can be effective in controlling weeds, especially glyphosate-resistant Palmer amaranth [18]. Timely applications of glufosinate in glufosinate-resistant soybean can be effective in controlling glyphosate-resistant and glyphosatesusceptible Palmer amaranth [17]. Reed et al. [19] reported that multiple applications of glufosinate in a total POST herbicide program controlled Palmer amaranth effectively. However, most recommendations by Cooperative Extension Service representatives include sequential applications of PRE herbicides followed by timely POST herbicides, especially in fields where glyphosate-resistant Palmer amaranth is suspected.
Given that Palmer amaranth has become more prevalent in North Carolina, especially herbicide-resistant biotypes, research is needed to better define the role of seeding rate on weed management programs for soybean. Determining if sequential applications of PRE and POST herbicides are as effective as multiple POST applications of glufosinate alone will be important, especially in terms of response to soybean population. Research is also needed to better define the time between multiple applications of glufosinate. Therefore, research was conducted in North Carolina to compare Palmer amaranth control and soybean yield with combinations of PRE and POST herbicides applied to glufosinate-resistant soybean at two soybean populations and to determine the most effective timing of multiple POST applications of glufosinate.

Interactions of Plant Population and Herbicide Program.
The experiment was conducted in North Carolina in two separate fields during each year (2010 and 2011) at the Upper Coastal Plain Research Station located near Rocky Mount with natural and relatively high populations of Palmer amaranth. Populations of weed species other than Palmer amaranth were low and inconsistent across fields compared with density and distribution of Palmer amaranth. One field consisted of a Lynchburg fine sandy loam soil (fine-loamy, siliceous, semiactive, thermic, and Aeric Paleaquults) while the second field was a Nahunta loam soil (fine-silty, siliceous, thermic, and Aeric Paleaquults). Corn (Zea mays L.) and cotton (Gossypium hirsutum L.) preceded soybean on these respective soil series. Soybean cultivar LL 595 N (Southern States Cooperative, Inc., Farmville, NC 27828, USA) was planted between May 24 and June 7 after disking and field cultivation in rows spaced 20 cm apart in plots 12 rows wide by 9 m.
Treatments consisted of two levels of soybean population (178,000 plants ha −1 versus 483,000 plants ha −1 ) and four levels of herbicide program. Herbicide treatments included glufosinate POST applied 2 weeks after planting (WAP), glufosinate applied sequentially 2 and 4 WAP, S-metolachlor plus fomesafen applied PRE, and S-metolachlor plus fomesafen applied PRE followed by glufosinate POST 4 WAP. A nontreated control also was included. S-metolachlor plus fomesafen (Prefix herbicide, Syngenta Crop Protection, Greensboro, NC 27709, USA) was applied at 1200 + 270 g ae ha −1 while glufosinate-ammonium (Ignite 280 herbicide, Bayer CropScience, Research Triangle Park, NC 27709) was applied at 670 g ae ha −1 . Herbicides were applied using a CO 2 -pressurized backpack sprayer calibrated to deliver 145 L ha −1 using 8002 regular flat-fan nozzles (Teejet Corporation, Wheaton, IL 60187, USA) at 275 kPa.
Palmer amaranth was 2 to 6 cm in height when glufosinate was applied POST. Visible estimates of percent Palmer amaranth control were recorded 6 WAP using a scale of 0 to 100, where 0 is no control and 100 is complete control. Soybean was machine harvested, and final yield was adjusted to 15% moisture.
The experimental design was a randomized complete block with treatments replicated four times. Data for Palmer amaranth control and soybean yield were subjected to ANOVA using the Proc GLM procedure in SAS (SAS Institute, GLM Procedure, Cary, NC 27513, USA) for a four (site/year combinations) by a four (herbicide program) factorial arrangement of treatments. The nontreated control was not included in this analysis because soybean was not harvestable when herbicide was not included. Means of significant main effects and interactions were separated using Fisher's Protected LSD test at P ≤ 0.05.

Efficacy of Sequential Herbicide Programs in a Single Plant
Population. Two experiments were conducted with either sequential applications of PRE followed by POST herbicides or multiple applications of glufosinate only. The experiment with sequential applications of PRE and POST herbicides was conducted during 2010 and 2011 at the Upper Coastal Plain Research Station near Rocky Mount in the fields discussed previously (Nahunta soil series during 2010 and Lynchburg soil series during 2011). The experiment with multiple applications of glufosinate was conducted during 2010 on a Norfolk loam sand soil (fine-loamy, kaolinitic, thermic, and Typic Kandiuldults) and in two separate fields during 2011 on both the Nahunta and Lynchburg soil series described previously. The glufosinate-resistant soybean cultivar LL 595N (Southern States Cooperative, Inc., Farmville, NC 27828, USA) was planted after disking and field cultivation in rows spaced 20 cm apart with a final plant population of 480,000 plants ha −1 . Plot size was the same as described previously.
In the experiment with sequential application of PRE and POST herbicides, treatments consisted of seven levels of PRE herbicides, including no PRE, flumioxazin (Valor SX, Valent USA Corporation, Walnut Creek, CA 94596, USA) alone at 70 g ai ha −1 or with chlorimuron (Valor XLT herbicide, E. I. DuPont de Nemours and Co., Wilmington, DE 19898, USA) at 21 g ai ha −1 , the sodium salt of fomesafen (Reflex herbicide, Syngenta Crop Protection, Greensboro, NC 27709, USA) at 280 g ae ha −1 , S-metolachlor (Dual II Magnum, Syngenta Crop Protection, Greensboro, NC 27709, USA) at 1500 g ai ha −1 , S-metolachlor plus fomesafen (Prefix herbicide, Syngenta Crop Protection, Greensboro, NC 27709, USA) at 1200 + 270 g ha −1 , and sulfentrazone plus cloransulam-methyl (Authority First herbicide, FMC Corporation, Philadelphia, PA 19103, USA) at 280 + 36 g ai ha −1 , respectively. Each of these PRE herbicides was followed by no POST herbicide or a single application of glufosinate ammonium (Ignite 280 herbicide, Bayer CropScience, Research Triangle Park, NC 27709, USA) at 670 g ha −1 2 WAP or sequential applications of glufosinate at 400 g ha −1 at 2 and 4 WAP. In the experiment with only POST applications of glufosinate, treatments consisted of glufosinate applied 2 WAP at 400 and 670 g ha −1 (referred to as POST 1) followed by no additional applications of glufosinate or followed by a second application of glufosinate at 400 g ha −1 at 3 WAP (referred to as POST 2) or 4 WAP (referred to as POST 3). Treatments also included three sequential applications of glufosinate at 400 g ha −1 .
Palmer amaranth size ranged from 5 to 10 cm in height when glufosinate was applied in the experiment with PRE and POST herbicides. In the experiment with multiple applications of glufosinate only, Palmer amaranth height was 5 to 10 cm at POST 1, 5 to 20 cm at POST 2, and 5 to 40 cm at POST 3. The range in size of Palmer amaranth reflected differences in emergence of weeds caused by PRE herbicides prior to the POST application in experiment.
In both experiments, glufosinate was applied using the procedures described previously.
Percent Palmer amaranth control was estimated visually and was recorded 10 WAP in both experiments as described previously. Soybean yield was determined in each plot and adjusted to a final moisture of 15%.
The experimental design was a randomized complete block with four replications. Data for Palmer amaranth control and soybean yield were subjected to ANOVA using the Proc GLM procedure in SAS (SAS Institute, Cary, NC 27513, USA) considering the factorial arrangement of treatments. Means of significant main effects and interactions were separated using Fisher's Protected LSD test at P ≤ 0.05.

Interactions of Plant Population and Herbicide Program.
Interactions of experiment (site/year combination) by soybean population and experiment by herbicide program were significant for both Palmer amaranth control (P = 0.0525 and ≤ 0.0001, resp.) and soybean yield (P ≤ 0.0001 and 0.0344, resp.). The interaction of experiment by soybean population by herbicide program was not significant for Palmer amaranth control (P = 0.5630) and soybean yield (P = 0.8912). Additionally, the interaction of soybean population by herbicide program was not significant for these parameters (P = 0.4373 and 0.2719, resp.).
Palmer amaranth was controlled more effectively when soybean was established at the higher seeding rate by 11 to 13 percentage points at 3 of 4 locations (Table 1). At a fourth location control was numerically higher by 6 percentage points at the higher plant population. Soybean yield was higher in 3 of 4 experiments when established at the higher population (Table 1). However, there was some discrepancy in Palmer amaranth control and soybean yield among experiments when comparing soybean population. The relatively low soybean yields on the Nahunta soil series during 2010 due to dry weather most likely minimized the possibility of separating yields based on Palmer amaranth control. The higher soybean yield during 2010 on the Lynchburg soil series or during 2011 on the Nahunta soil series most likely reflected differences in Palmer amaranth control. On the Lynchburg soil series during 2011 there was a slight but nonstatistical difference in Palmer amaranth control but a clear increase in yield at the higher plant population. Other factors could have contributed to higher yields at a higher soybean population that are not related to weed control.
Considerable variation in Palmer amaranth control was noted among herbicide treatments across experiments ( Table 2). Control by multiple applications of glufosinate exceeded that of a single application of glufosinate in 3 of 4 experiments. Sequential applications of glufosinate only and S-metolachlor plus fomesafen PRE controlled Palmer amaranth similarly in 3 of 4 experiments. Applying glufosinate following S-metolachlor plus fomesafen was more effective than S-metolachlor plus fomesafen alone, sequential applications of glufosinate, and a single application of glufosinate in 1, 3, and 4 experiments, respectively.
Soybean yield during 2010 on the Nahunta soil series was low and did not differ among herbicide programs, most likely because yield was relatively low compared to the Lynchburg soil series during 2010 and both soil series during 2011 (Table 2). On the Lynchburg series during 2010, soybean yield was higher following sequential applications of glufosinate compared with a single application; yield with S-metolachlor plus fomesafen alone or followed by glufosinate was intermediate. On the Nahunta soil series during 2011, soybean yield with the more intensive program of S-metolachlor plus fomesafen followed by glufosinate yield exceeded yield following the other herbicide programs. On the Lynchburg soil series, yield with this program also was higher than a single application of glufosinate.
Although some variation in Palmer amaranth control and soybean yield was noted when comparing herbicide programs, lack of a soybean population by herbicide program interaction for these parameters suggests that higher soybean populations will increase Palmer amaranth control and soybean yield irrespective of herbicide program.

Efficacy of Sequential Herbicide Programs in a Single
Plant Population. The interaction of year by PRE by POST herbicide was significant for Palmer amaranth control (P ≤ 0.0001) and for soybean yield (P ≤ 0.0001). Therefore, data were analyzed by year. The interaction of PRE by POST was not significant for Palmer amaranth control (P = 0.1466) but was significant for soybean yield (P = 0.0268) during 2010. However, the interaction of PRE by POST herbicides was significant for both Palmer amaranth control (P = 0.0011) and soybean yield (P = 0.0049).
Palmer amaranth control by flumioxazin alone or with chlorimuron was 93 to 98% while fomesafen alone or with S-metolachlor provided 77 to 80% control (Table 3). Sulfentrazone plus cloransulam methyl controlled Palmer amaranth 66%, while S-metolachlor provided only 56% control. Whitaker et al. [20] found similar results for Palmer amaranth control with flumioxazin and fomesafen PRE.
Palmer amaranth control was similar when glufosinate was applied either once or sequentially ( Table 3). Lack of a difference in single and sequential applications most likely  reflects limited weed emergence after the first POST application and rapid closure of the soybean canopy.
All PRE herbicides increased soybean yield during 2010 compared with the no-PRE herbicide control ( Table 3). Yield of soybean receiving S-metolachlor plus fomesafen exceeded that of soybean treated with flumioxazin alone, sulfentrazone plus cloransulam methyl, or flumioxazin plus chlorimuron. Surprisingly, yield following a single application of glufosinate was lower than soybean yield without a POST application of glufosinate or soybean receiving multiple applications of glufosinate.
Several differences in Palmer amaranth control among PRE herbicides were observed during 2011, especially in absence of glufosinate (Table 4). Flumioxazin, flumioxazin plus chlorimuron, S-metolachlor plus fomesafen, and fomesafen alone provided the best control. Sulfentrazone plus cloransulam-methyl and S-metolachlor alone provided intermediate control between the most effective PRE herbicides and the nontreated control. When glufosinate was applied only once, no difference in control among PRE herbicides was observed.
There were no significant differences in soybean yield among PRE herbicide treatments when glufosinate was applied, regardless of the number of glufosinate applications (Table 5). In the absence of glufosinate, soybean yield following fomesafen, S-metolachlor, and sulfentrazone plus cloransulam methyl was intermediate between the nontreated control and flumioxazin alone or with chlorimuron or S-metolachlor plus fomesafen.
In the second experiment with total POST applications of glufosinate applied sequentially 2, 3, and 4 WAP, the interaction of glufosinate rate with sequence of glufosinate application was not significant for Palmer amaranth control (P = 0.1064) or soybean yield (P = 0.7318). Although glufosinate rate as a main effect or the interaction with year or sequence of glufosinate application was not significant (P ≤ 0.05), the interaction of experiment by sequence of glufosinate application was significant for Palmer amaranth control (P = 0.0077) and soybean yield (P = 0.0100). In 2010, multiple applications of glufosinate controlled Palmer amaranth 97 to 99% than a single application 82% (Table 6). In systems with two applications, timing of the second application had no effect on Palmer amaranth control. Also, Palmer amaranth control was similar with two and three applications. Soybean yield was not affected by the number of glufosinate applications. However, glufosinate POST increased yield over nontreated soybean. During 2011, on the Nahunta soil series, Palmer amaranth control was similar with all glufosinate applications and was at least 97%. In the second field on a Lynchburg soil series, control was greater where glufosinate was applied sequentially compared with a single application.