Timely and effective herbicide spray programs are essential for adequate weed control in crop production (Wilcut and Askew, 1999). However, the rapid growth of many problematic broadleaf weeds influences the effectiveness of control (Horak and Loughin, 2000). A major concern in the southeastern United States is crop interference from Palmer amaranth (Amaranthus palmeri S. Wats) and common ragweed (Ambrosia artemisiifolia L.). Weeds compete with crops for light and other resources and can complicate mechanical harvest (Horak et al., 1994; Morgan et al., 2001; Young et al., 1982). Weeds must be controlled throughout the growing season to avoid yield losses in peanut (Arachis hypogaea L.) (Jordan, 2018; Wilcut et al., 1995). As the duration of interference from weeds increases, yield of peanut and other crops often decreases (Zimdahl, 2004). Minimizing contributions of weed seed to the soil seedbank is important in managing weeds in succeeding crops and addressing evolved resistance to herbicides. One of the main concerns in peanut production is interference of common ragweed and Palmer amaranth with peanut and how in-season control practices affect weed density in the following rotational crop.

Palmer amaranth, which originated in southern California and Mexico, is difficult to control in peanut and crops rotated with peanut in the United States (Leon et al., 2019). Morphological features and capacity to produce flowers at almost any growth stage makes Palmer amaranth successful in exploiting diverse environments (Sauer, 1957). In California, Keeley et al. (1987) reported that Palmer amaranth can emerge in early March when soils have reached a temperature of 18 C and continue to infest fields into October. In Georgia, Webster and Grey (2015) found that glyphosate-resistant Palmer amaranth planted in early May produced 446,000 seeds per plant, while Palmer amaranth planted six weeks later produced 50 percent fewer seed. Palmer amaranth planted 9 and 12 weeks later had 89 and 99% fewer seeds, respectively (Webster and Grey, 2015). Horak and Loughin (2000) found that under natural growing conditions and in the absence of other weed or crop species, Palmer amaranth grew 0.18 to 0.21 cm/growing degree day (GDD). Because of its rapid growth rate, the application window for optimum control of Palmer amaranth may be relatively short (Horak and Loughin, 2000). The application interval for effective control of Palmer amaranth could be as soon as one week after weed emergence based on maximum height recommended for herbicides used on peanut (Jordan, 2018).

Morgan et al. (2001) reported that when 1.1 Palmer amaranth plants/m² were present, cotton (Gossypium hirsutum L.) yield was reduced by 54%. Yield loss of field corn (Zea mays L.) ranged from 11 to 91% at 0.66 to 10 plants/m² when Palmer amaranth emerged with the corn (Massinga et al., 2001). At a density of one plant/m of row, yield loss in peanut was predicted to be 28% (Burke et al., 2007). Yield of soybean [Glycine max (L.) Merr.] was reduced 79% when 10 plants/m² were present (Bensch et al., 2003). Palmer amaranth can also reduce harvesting efficiency, which is problematic for agronomic crops that are mechanically harvested (Young et al., 1982). Smith et al. (2000) reported that when 0.33 Palmer amaranth plants/m² were present in cotton, harvest time increased by 3.5 times compared to weed-free conditions. The increased harvest time was a result of the operator having to stop the machine to dislodge stalks from moving parts (Smith et al., 2000). When Palmer amaranth density exceeds 0.66 plants/m², damage to harvest equipment can occur (Morgan et al., 2001).

Common ragweed is a monoecious annual plant native to North America and is common to the temperate regions of the United States (Oosting, 1942). Common ragweed will generally dominate in the second year if soil disturbance is halted, shading out some of the species that dominated in the first year (Oosting, 1942), which relates to the increase of common ragweed problems in no-till or reduced tillage systems in soybean and corn (Jordan et al., 2014). Common ragweed can emerge in late April as soil warms and will continue to emerge until temperatures become unfavorable sending the seeds back into dormancy (Jordan et al., 2014).

Coble et al. (1981) reported that 0.11 common ragweed/m² reduced soybean yield by 32.5 kg/ha on average. When grown in 30% shade, common ragweed has been noted to grow just as well as sweet corn (Zea mays L. var. rugosa) and dry bean (Phaseolus vulgaris Herrm) but much poorer than these crops in 73% shade (Dickerson, 1968). Predicted yield loss in peanut is 40% when 1 common ragweed is present/m² for the entire season (Clewis et al., 2001). Everman et al. (2008) reported the critical period of weed control in peanut for mixed broadleaf populations that included common ragweed of 2.6 to 8 WAP.

The 2- to 3-week period after flowering is the critical time for weed management in most of the weed species to prevent weeds from producing viable seeds (Hill et al., 2016). Most weed management studies only evaluate a single growing season failing to quantify the impact these weed management strategies may have on contributions to the soil seedbank (Inman et al., 2016). Information relative the impact of weed management in peanut on succeeding crops is limited. This study was designed to evaluate the effects that timing of weed management have on weed control, peanut yield, estimated economic returns, and weed populations and cotton lint yield in the following growing season.

The experiments were conducted in North Carolina near Lewiston-Woodville (36.135 N, -77.177 W) and Rocky Mount (35.897 N, -77.675 W) during 2016 and 2017 growing seasons. Adjacent areas of the same fields were used for different years at both locations. Soils included a Norfolk sandy loam (fine-loamy, siliceous, thermic Typic Paleudult) at Lewiston-Woodville and an Aycock very fine sandy loam (fine-silty, siliceous, subactive, thermic Typic Paleudult) at Rocky Mount. Soil pH ranged from 5.7 to 6.1 at the two locations with soil organic matter content of 1.8% and 2.4% at these respective locations. The Lewiston-Woodville and Rocky Mount sites were naturally infested with common ragweed (129 plants/m² in 2016 and 61 plants/m² in 2017) and Palmer amaranth (54 plants/m² in 2016 and 65 plants/m² in 2017), respectively. Peanut cultivar 'Bailey' (Isleib et al., 2011) was planted into conventionally-tilled, raised seedbeds at a seeding rate designed to provide a final in-row population of 15 plants/m. Plot size was 4 rows (spaced 91 cm apart) by 9 m at both locations. Other than herbicide treatments for the experiment, peanut was managed according to North Carolina Cooperative Extension Service recommendations (Jordan et al., 2018).

Treatments were arranged in a randomized complete block design and replicated 4 times at Lewiston-Woodville and 3 times at Rocky Mount. Herbicide treatments consisted of a single application of paraquat plus bentazon at 2 weeks after planting (WAP); a single application of imazapic plus lactofen at 6 WAP; applications of paraquat plus bentazon at 2 WAP followed by bentazon plus lactofen at 4 WAP; bentazon plus lactofen at 4 WAP followed by imazapic plus lactofen at 6 WAP; and paraquat plus bentazon at 2 WAP followed by bentazon plus lactofen at 4 WAP followed by imazapic plus lactofen at 6 WAP. A NTC was included. Clethodim was applied over the entire test area 6 WAP to control annual grasses. At 2 and 4 WAP, peanut was in the vegetative growth stage while at 6 WAP peanut was in the flower (R1) growth stage (Balota, 2019; Boote, 1982). Herbicide rates and manufacturer details are listed in Table 1. Nonionic surfactant (Induce^®, Helena Chemical Co., Collierville, TN) at 0.125% (v/v) was applied with paraquat plus bentazon. Crop oil concentrate (Agri-Dex^®, Helena Chemical Co., Collierville, TN) at 1.0% (v/v) was applied with bentazon plus lactofen, imazapic plus lactofen, and clethodim. Herbicides were applied using a CO₂-pressurized backpack sprayer equipped with flat-fan nozzles (AIXR 11002 TeeJet^® Air Induction XR flat-spray nozzles, TeeJet Technologies, Wheaton, IL) calibrated to deliver 140 L/ha at 152 kPa.

Prior to the first herbicide application, initial weed densities were recorded from 15 randomly determined sections across the entire test area using a 0.3 by 0.3 m square. After herbicides were applied, weed control was visually estimated 7, 10, and 20 WAP using a 0 to 100% scale where 0 was equal to no control and 100 to complete control. Foliar chlorosis, necrosis, and density were considered when recording the visual estimates. Peanut pods were dug and vines inverted at optimum maturity based on pod mesocarp color (Williams and Drexler, 1981). Within two weeks prior to digging and vine inversion, weeds were mowed above the canopy to improve efficiency of digging and vine inversion.

During the year following peanut, cotton cultivar DP 1522 (Monsanto Co., St. Louis, MO) was planted into the previous year's plots on 4 May and 11 May during 2017 at Lewiston-Woodville and Rocky Mount, respectively. In 2018, the cotton cultivar DP 1646 (Monsanto Co., St. Louis, MO) was planted 10 May and 15 May at Lewiston-Woodville and Rocky Mount, respectively. Cotton was planted in conventionally-tilled, raised seedbeds. Within each plot, density of common ragweed at Lewiston-Woodville and density of Palmer amaranth at Rocky Mount were determined 3, 7, and 20 WAP. Density at 3 and 7 WAP was measured using a 0.3 by 0.3 m square in three randomly determined sections of each plot and then the average of each plot was converted to a m² basis. Due to reduced weed densities at 20 WAP, populations were determined by counting common ragweed or Palmer amaranth present in the center two rows of the entire plot area and then converted to a m² unit. After initial weed densities were recorded 3 and 7 WAP in cotton, glyphosate was applied at Lewiston-Woodville. Glyphosate plus dicamba was applied 3 WAP followed by glyphosate only 7 WAP at Rocky Mount. These herbicides were applied over the entire test area.

Economic returns were estimated using a modified North Carolina State University enterprise budget for conventional Virginia market type peanut to accommodate a 400-ha operation in North Carolina (Washburn, 2019). Costs were calculated for the different treatments with respect to varying costs of herbicides and adjuvants quoted from major suppliers in the region. Costs associated with hauling, drying and cleaning peanut, state check-off fee, and national assessment costs based on yield varied by treatment. A base cost of $1502/ha for peanut production before these respective inputs was used. A ten-year average price of $0.57/kg received for Virginia market type peanuts in the United States was used for estimated economic return analysis (USDA-NASS, 2018).

Data for common ragweed and Palmer amaranth control, peanut yield, estimated economic return, broadleaf densities in cotton the following year, and cotton lint yield were subjected to analysis of variance using PROC GLIMMIX and PROC CORR procedures of SAS (version 9.4; SAS Institute INC., Cary, NC). Data for each location with the same broadleaf weeds were also subjected to ANOVA for a 2 (year) by 6 (timing of herbicide application) factorial arrangement of treatments. Timing of herbicide application treatments and locations were fixed factors whereas replications were treated as random. Means for significant main effects and interactions were separated using Fisher's Protected LSD at 𝜶 = 0.05.

The interaction of year by timing of herbicide application was not significant for common ragweed control 7, 10, and 20 WAP; pod yield; and estimated economic return at Lewiston-Woodville or for Palmer amaranth control 10 and 20 WAP; pod yield; and estimated economic return at Rocky Mount (Tables 2 and 3). The interaction of year by timing of herbicide application was significant for Palmer amaranth control 7 WAP. Therefore, data for common ragweed control 7, 10, and 20 WAP, pod yield, and estimated economic return at Lewiston-Woodville (Table 4) and Palmer amaranth control 10 and 20 WAP, pod yield, and estimated economic return at Rocky Mount (Table 5) are presented pooled over years. Palmer amaranth control 7 WAP was analyzed by year at Rocky Mount.

Common ragweed control at Lewiston-Woodville and peanut pod yield as well as common ragweed control and estimated economic returns were positively correlated (p ≤ 0.0001; R = 0.60 to 0.68) regardless of when the visual estimates of weed control were recorded (Table 6). As was noted for the experiments at Lewiston-Woodville for common ragweed, peanut yield and estimated economic returns were positively correlated with Palmer amaranth control 7, 10, and 20 WAP at Rocky Mount (p ≤ 0.0001, R = 0.60 to 0.68) (Table 7).

The presented research reinforces the importance of timely and effective weed management in peanut. Effective control of common ragweed and Palmer amaranth can be achieved with proper timing of herbicide applications. Usually, the more intensive herbicide programs generated the greatest weed control, yield, and estimated economic returns. In the case of common ragweed, an intensive weed management approach in peanut the previous season showed a reduction in weed emergence the following year in cotton. Due to the abundance of Palmer amaranth seed in the seedbank, multiple years of implementation are likely needed to observe the full benefits of an intensive weed management program.