A total of twenty-two field experiments (six maize and five soybean using low glyphosate doses to assess weed control and six maize and five soybean using high glyphosate doses to assess tolerance) were conducted over a three-year period (2010 to 2012) at Ridgetown (42.4406˚N, 81.8842˚W) and Exeter (43.3500˚N, 81.4833˚W), Ontario, Canada. All experiments were designed as a randomized complete block, replicated four times. Treatments in the weed control and tolerance experiments conducted in both maize and soybean included glyphosate formulated as a potassium salt in WeatherMAX and glyphosate formulated as an isopropylamine salt in Clearout and WiseUp. In the weed control experiments, treatments included glyphosate applied at 225, 450, and 900 g·ae·ha⁻¹ in addition to weedy and weed-free controls. In the tolerance experiments conducted in maize, treatments included glyphosate applied at 1800 and 3600 g·ae·ha⁻¹ in addition to an untreated, weed-free control whereas in soybean, the glyphosate doses used were 2700 and 5400 g·ae·ha⁻¹. In maize, 1800 g·ae·ha⁻¹ repre- sented the maximum single application glyphosate dose [3] whereas 2700 g·ae·ha⁻¹ used in soybean, while slightly more than the maximum allowed dose in a single application [10] , represented the cumulative glyphosate dose (900 + 1800 g·ae·ha⁻¹) that could be received in sequential applications [3] .

In both the weed control and tolerance experiments, each treatment plot was 2 m wide by 8 m (Ridgetown) to 10 m (Exeter) long. Glyphosate-resistant maize hybrids were seeded 4 to 5 cm deep at a rate of 80,000 seeds·ha⁻¹ in rows spaced 0.75 m apart. Glyphosate-resistant soybean cultivars were seeded 3 to 4 cm deep at a rate of 370,000 to 480,000 seeds·ha⁻¹ in rows spaced 0.75 m apart. In the weed control experiments, glyphosate treatments were applied to maize at the 7- to 9-leaf stage (approximately June 13) and to soybean at the 3^rd trifoliate leaf stage (approximately July 2). In the tolerance experiments, glyphosate treatments were applied to maize to at the 9- to 11-leaf stage (approximately June 22) and to soybean at the flower initiation stage (approximately July 11); both application timings were generally later than recommended [3] . All glyphosate treatments were applied using a CO₂ pressurized backpack sprayer calibrated to deliver 200 L·ha⁻¹ of water at 210 kPa through Ultra-Low Drift 120-02 nozzles (Hypro, New Brighton, MN). Plots within the tolerance experiments were maintained weed-free for the entire growing season using preemergence herbicides and hand weeding as needed.

In the maize and soybean weed control experiments, visible crop injury and weed control were estimated visually on a scale of 0% (no injury/control) to 100% (complete plant death) relative to untreated control plants. Maize and soybean injury was rated 1, 2, and 4 weeks after glyphosate application (WAT) whereas control of velvetleaf, pigweed species, common ragweed, common lambsquarters, and green foxtail was rated 4 and 8 WAT. In the maize and soybean tolerance experiments, visible crop injury was estimated visually on a scale of 0% (no injury) to 100% (complete plant death) relative to untreated control plants 1, 2, and 4 WAT in maize and 1, 2, 3, 4, and 8 WAT in soybean. At physiological maturity of maize, cob length was assessed by measuring ten fully exposed, randomly selected ears from end to end within each plot. The same cobs were rated for cob deformity on a scale from 1 (completely deformed) to 10 (no deformity) similar to methods used previously [1] . Both crops were harvested at maturity using a small plot combine and crop moisture and weight were recorded; final yields were adjusted to 15.5% and 13% moisture content for maize and soybean, respectively.

Maize and soybean injury and weed control data were analyzed separately by crop within the weed control and tolerance experiments using PROC MIXED (SAS Ver. 9.2, SAS Institute Inc., Cary, NC). In the individual analysis of the experiments, variances were divided into fixed (glyphosate treatment) and random effects [block; environment (i.e., year or location-year combinations); block nested within environment; and the herbicide treatment × environment interaction]. Significance of the fixed effect was tested using F-test and random effects were tested using a Z-test of the variance estimate. The herbicide treatment × environment interactions in maize and soybean experiments were not significant and the data for each set of experiments were pooled across environments within each crop for the weed control and tolerance experiments. PROC UNIVARIATE in SAS was used to test data for normality and homogeneity of variance. For all crop injury and weed control ratings, the untreated control treatment (assigned a value of zero) was excluded from the analyses. However, all values were compared independently to zero to evaluate treatment differences with the untreated control. To satisfy the assumptions of the variance analyses, data were transformed with the appropriate transformation as needed to improve normality. Means were separated using Fisher’s Protected LSD at P < 0.05 and any data compared on the transformed scale were converted back to the original scale for presentation of results.

For maize and soybean, no crop injury symptoms were noted at any of the observation timings, regardless of glyphosate formulation or dose (data not shown), consistent with previous research [4] [6] [14] .

In general, weed control in maize increased with glyphosate dose at 4 and 8 WAT (Table 1, Table 2); however, the level of control varied by formulation and weed species. For example, velvetleaf control 4 WAT with 900 g·ae·ha⁻¹ applied at the 7- to 9-leaf stage was similar to the weed-free control for all three of the formulations tested (Table 1). Control of pigweed species, common ragweed, common lambsquarters, and green foxtail 4 WAT using 900 g·ae·ha⁻¹ ranged from 89% to 99%, regardless of glyphosate formulation (Table 1); however, in most instances, this level of control was not equivalent to the weed-free control. The exceptions to this were control of pigweed species using Clearout or WiseUp and control of green foxtail using WeatherMAX or WiseUp (Table 1). In the literature, glyphosate has been highly efficacious (e.g., greater than 90% control) on these weed species found in maize, but across years and/or locations this level of control still may not be equivalent to a weed-free control [7] [15] [16] . At 4 WAT, WeatherMAX, Clearout, and WiseUp provided equivalent control of all weed species with two exceptions: WiseUp provided better control of common ragweed than Clearout at 450 g·ae·ha⁻¹ and WiseUp provided better control of common lambsquarters than Clearout at 225 g·ae·ha⁻¹. Similar to the velvetleaf control at 4 WAT, control 8 WAT using 900 g·ae·ha⁻¹ was equivalent to the weed-free control for all formulations tested (Table 2). For the remaining weed species, especially common lambsquarters and green foxtail, the level of control 8 WAT using 900 g·ae·ha⁻¹ ranged from 82% to 89% (Table 2), an overall decrease in control compared to observations 4 WAT (Table 1). There were no differences in control between using 450 or 900 g·ae·ha⁻¹ for common lambsquarters and green foxtail for any of glyphosate formulations tested; furthermore, control of green foxtail was similar across all WiseUp doses (Table 2), indicative of the potential for a late-season weed escape [2] and for yield losses [3] . At 8 WAT, WeatherMAX, Clearout, and WiseUp provided equivalent control of all weed species with two exceptions: WiseUp provided better control of both velvetleaf and common ragweed than Clearout at 450 g·ae·ha⁻¹. At harvest, maize yields were statistically equivalent with WeatherMAX, Clearout, and WiseUp. Using 900 g·ae·ha⁻¹ of glyphosate applied as WeatherMAX or Clearout, maize yield was equivalent to the weed-free control; but with WiseUp, maize yield was 8.8% less than the weed-free control (Table 2).

^aAbbreviations: ABUTH, velvetleaf; AMASS, pigweed species; AMBEL, common ragweed; CHEAL, common lambsquarters; SETVI, green foxtail; WAT, weeks after glyphosate treatment. ^bMeans followed by the same letter within a column are not significantly different according to Fisher’s Protected LSD (P < 0.05). ^cIncluded N-Tank at 0.5% v/v.

^aAbbreviations: ABUTH, velvetleaf; AMASS, pigweed species; AMBEL, common ragweed; CHEAL, common lambsquarters; SETVI, green foxtail; WAT, weeks after glyphosate treatment. ^bMeans followed by the same letter within a column are not significantly different according to Fisher’s Protected LSD (P < 0.05). ^cIncluded N-Tank at 0.5% v/v.

For soybean, weed control increased with glyphosate dose and excellent control was observed 4 and 8 WAT when 900 g·ae·ha⁻¹ was applied at the 3^rd trifoliate leaf stage for nearly every weed species tested, regardless of glyphosate formulation (Table 3, Table 4). For example, control of velvetleaf, pigweed species, and green foxtail 4 WAT using 900 g·ae·ha⁻¹ was similar to the weed-free control for all glyphosate formulations tested whereas control common lambsquarters ranged from 93% to 97% (Table 3). At 4 WAT, WeatherMAX, Clearout, and WiseUp provided equivalent control of all weed species with three exceptions: WiseUp provided better control of pigweed species than WeatherMAX at 225 g·ae·ha⁻¹, WiseUp provided better control of velvetleaf than Clearout at 450 g·ae·ha⁻¹, and both WeatherMAX and WiseUp provided better control of common lambs- quarters than Clearout at 450 g·ae·ha⁻¹. By 8 WAT, control of velvetleaf, pigweed species, common lambsquarters, and green foxtail using 900 g·ae·ha⁻¹ of glyphosate as WeatherMAX were all similar to the weed-free control (Table 4). For the Clearout and WiseUp formulations, control of velvetleaf, pigweed species, and green foxtail 8 WAT with 900 g·ae·ha⁻¹ was also similar to the weed-free control; however, control of common lambsquarters, while rated at over 95%, differed from the weed-free control (Table 4). Control of common lambsquarters can be difficult as this species tends to respond better to glyphosate doses that are greater than 900 g·ae·ha⁻¹ [17] . At 8 WAT, WeatherMAX, Clearout, and WiseUp provided equivalent control of all weed species with four exceptions at the 225 g·ae·ha⁻¹ dose: WiseUp provided better control of pigweed species than WeatherMAX, WiseUp provided better control of velvetleaf than Clearout or WeatherMAX, and WiseUp provided better control of common lambsquarters than Clearout. At harvest, the excellent season-long control of the weeds present in this study, regardless of the glyphosate dose or formulation used, resulted in crop moisture levels and yields that were generally similar to the weed-free control (Table 4), consistent with the literature [5] [18] -[21] .

At all observation timings, visible injury levels of maize treated with 1800 g·ae·ha⁻¹ of glyphosate applied at the 9- to 11-leaf stage were similar to the untreated control, regardless of formulation (Table 5). For the WeatherMAX and WiseUp glyphosate formulations however, visible injury tended to increase with dose, but decrease over time. Yet, 2.1% and 2.8% injury was still observed 4 WAT for maize treated with 3600 g·ae·ha⁻¹ of glyphosate

^aAbbreviations: ABUTH, velvetleaf; AMASS, pigweed species; CHEAL, common lambsquarters; SETVI, green foxtail; WAT, weeks after glyphosate treatment. ^bMeans followed by the same letter within a column are not significantly different according to Fisher’s Protected LSD (P < 0.05). ^cIncluded N-Tank at 0.5% v/v.

^aAbbreviations: ABUTH, velvetleaf; AMASS, pigweed species; CHEAL, common lambsquarters; SETVI, green foxtail; WAT, weeks after glyphosate treatment. ^bMeans followed by the same letter within a column are not significantly different according to Fisher’s Protected LSD (P < 0.05). ^cIncluded N-Tank at 0.5% v/v.

^aAbbreviation: WAT, weeks after glyphosate treatment. ^bMeans followed by the same letter within a column are not significantly different according to Fisher’s Protected LSD (P < 0.05). ^cRidgetown 2011 (4 WAT) and Exeter 2012 (1, 2, and 4 WAT) injury data were not included in the analysis as no injury was detected. ^dN-Tank at 0.5% v/v. ^eN-Tank at 1.0% v/v.

as WeatherMAX and WiseUp, respectively (Table 5). Maize injury levels and symptoms were similar with previous research [1] [22] , such as chlorosis, degradation of the leaves present in the whorl at the time of application, wrapped leaves, and overall reduced growth compared to the untreated control. Similar to visible injury, cob deformity tended to increase with glyphosate dose, especially for maize treated with WeatherMAX and WiseUp (Table 5). However, maize demonstrated excellent tolerance as cob length and final yields were unaffected by glyphosate dose or formulation (Table 5), consistent with prior studies [1] [6] [14] [22] .

For soybean, visible injury 1 WAT with 2700 g·ae·ha⁻¹ of glyphosate applied at the flower initiation stage was 3.2%, 1.4%, and 1.0% for WeatherMAX, WiseUp, and Clearout formulations, respectively (Table 6). Visible soybean injury increased more than two-fold as 8.5%, 4.5%, and 3.7% injury was observed 1 WAT with 5400 g·ae·ha⁻¹ of glyphosate formulated as WeatherMAX, WiseUp, and Clearout, respectively (Table 6). The WeatherMAX label warns of short term yellowing in spray overlap areas if soybean is treated with 2522 g·ae·ha⁻¹, but according to the label this injury should not influence yield [10] . In the literature, significant injury to soybean has been reported when glyphosate was applied at doses greater than 2000 g·ae·ha⁻¹ [23] -[26] . Using similarly high doses to those in the current study, other research demonstrated that an isopropylamine salt formulation was more injurious than a potassium salt formulation [27] and was equally injurious as a trimethylsulfonium salt formulation [28] . Conversely, related research in Ontario reported no injury in soybean treated with 3600 g·ae·ha⁻¹ of glyphosate formulated as WeatherMAX [11] , contrary to the results of the current study (Table 6). Yet, visible soybean injury was transient. For example, by 4 WAT, visible injury of soybean treated with 2700 g·ae·ha⁻¹ of glyphosate as WeatherMAX and WiseUp was similar to the untreated control while visible injury of soybean treated with glyphosate as Clearout was similar to the untreated regardless of dose; by 8 WAT, visible injury was similar to the untreated control, regardless of formulation or dose (Table 6). At harvest, yield of soybean treated with 2700 g·ae·ha⁻¹ of glyphosate, regardless of formulation, was similar to the untreated control (Table 6), consistent with other research which used doses ranging from 2240 to 4480 g·ae·ha⁻¹ [28] . Unfortunately, the persistent observations of injury from 5400 g·ae·ha⁻¹ of glyphosate resulted in 8.5%, 4.6%, and 5.5% yield loss for WeatherMAX, WiseUp, and Clearout formulations, respectively (Table 6).