The threat of glyphosate-resistant (GR) weeds, especially Palmer amaranth (Amaranthus palmeri S. Wats.), has motivated growers to consider the use of herbicides with alternative modes of action in cotton (Gossypium hirsutum L.) and soybean [Glycine max (L.) Merr.]. Glufosinate is a non-selective, broad-spectrum, postemergence herbicide that inhibits glutamine synthetase (Senseman 2007). Thus, it can be used to effectively control GR-Palmer amaranth in glufosinate-resistant crops if applied to plants less than 8 cm in height (Norsworthy et al. 2008; Wilson et al. 2007).

Glufosinate-resistant crops (Liberty-Link®) were developed by insertion of the bar gene isolated from the soil bacterium Streptomyces hygroscopicus (Duke 2005). The bar gene expresses the phosphinothricin acetyltransferase (pat) enzyme that acetylates L-phosphinothricin, and confers tolerance to glufosinate (Herouet et al. 2005; Lydon and Duke 1999). Liberty-Link® cotton cultivars were commercialized in 2004 (Duke 2005). Additionally, WideStrike® cotton cultivars, which also contain the pat gene, are commercially available but offer lower levels of resistance to glufosinate (Culpepper et al. 2009, Steckel et al. 2012). In soybean, regulatory approval for glufosinate-resistance in soybean occurred in 1996 (Duke 2005). However, wide-spread commercialization of LibertyLink® soybean cultivars did not occur until 2009. Glufosinate-resistant field corn (Zea mays L.) hybrids were commercialized in 1997 (Duke 2005).

Since peanut is grown in close proximity to field corn, cotton, and soybean in the US peanut belt, drift or sprayer contamination problems are likely to occur. Previous research has evaluated the unintentional effects of glyphosate and dicamba on peanut yield. When glyphosate was applied to peanut plants at 28 days after planting (DAP), rates of 280 g/ha or higher caused significant yield reductions (Lassiter et al. 2007). When applied between 75 to 105 DAP, glyphosate at 240, 320, and 470 g/ha reduced peanut yield by 12%, 24%, and 36%, respectively (Grey and Prostko 2010). Estimated peanut yield losses for dicamba applied at rates between 40 to 560 g/ha at 30, 60, or 90 DAP ranged between 2 and 100% (Prostko et al. 2011).

Limited studies have addressed the effects of glufosinate on peanut. In Texas, glufosinate applied at 470 to 580 g/ha provided 100% control of simulated volunteer runner peanut cultivars (Grichar and Dotray 2007). Early tests conducted in North Carolina indicated that peanut yield was reduced 14 to 74% when glufosinate was applied at rates ranging between 135 to 538 g/ha (Jordan et al. 2011). Subsequent North Carolina tests reported peanut yield was reduced 33 to 75% by 302 g/ha of glufosinate in 4 out of 4 site-years and 25% by 123 g/ha in 1 of 4 site-years (Johnson et al. 2012). However, both of these tests were conducted using only one application timing, approximately 25 to 28 DAP. Therefore, the objective of our research was to evaluate peanut yield response to glufosinate at 0, 41, 82, 164, 328, 656 g ai/ha applied 30, 60, or 90 DAP.

Small-plot field trials were conducted at 13 locations across the US peanut belt during 2010 to 2012. A complete description of these locations is presented in Table 1. Production and pest management practices were followed according to local Cooperative Extension recommendations.

At all locations, herbicide treatments were arranged in a split-plot design. Time of application was the main plot with glufosinate rate as the sub-plot. Application timings were 30, 60, and 90 DAP and glufosinate rates were 0, 41, 82, 164, 328, 656 g ai/ha. A logarithmic scale with a common multiplier (e.g. 2) is recommended to evaluate herbicide dose response relationships (Seefeldt et al. 1995). The typical use rate of glufosinate in cotton and and soybean is 656 g ai/ha (Anonymous 2013). Generally, peanut plants were in the R1, R4-R5, and R6 stages of growth at 30, 60, and 90 DAP timings, respectively (Boote 1982). All treatments were replicated four to six times. Glufosinate was applied using a CO₂-pressurized sprayer calibrated to deliver 94 to 140 L/ha. All plot areas were maintained weed-free throughout the season using a combination of herbicides (clethodim, diclosulam, flumioxazin, imazapic, pendimethalin, and 2,4-DB) and hand-weeding. Peanut yield data were obtained by mechanical harvesting at maturity.

Data were analyzed using a mixed model ANOVA. Rate and timing of glufosinate were fixed effects, while site-years, replications, and interactions with these factors were considered random effects. Peanut yield data were regressed on rate of glufosinate and fit to a log-logistic curve (Seefeldt et al. 1995). The mathematical expression relating peanut yield to glufosinate rate was:

where c  =  the mean yield response at very high glufosinate doses, d  =  mean yield response of the nontreated control, Y₅₀  =  the glufosinate rate causing a 50% reduction in yield, and b  =  slope of the curve around Y₅₀ . Differences in parameter estimates from the regression models were evaluated using t-tests with an alpha of 0.05 and t_critical  =  1.96 (Glantz and Slinker 2001). Similar to Askew and Wilcut (2001) and Jasieniuk et al. (1999), the nonlinear coefficient of determination (R² _nonlinear ) was calculated as:

There were significant interactions between time of application and glufosinate rate. Therefore, yield data were regressed on glufosinate rate by time of application.

These results provide growers across the U.S. Peanut Belt with an estimate of potential yield losses caused by a range of rates of glufosinate applied at three different growth stages. Consequently, decisions can be made to determine if a peanut crop that is unintentionally treated with glufosinate should be managed for production or terminated.