CHEMICAL WEED MANAGEMENT IN GRAIN SORGHUM AND SELECTIVITY OF ATRAZINE + S-METOLACHLOR TO DIFFERENT HYBRIDS

Grain sorghum (Sorghum bicolor) is one cereal crop that faces huge problems with weed interference mostly because the lack of selective herbicides. This study aimed to assess the efficacy and safety of herbicide alternatives for weed control in grain sorghum as well as the selectivity of atrazine + s-metolachlor to different hybrids. Three field trials were designed as a randomized complete block with four replications. All experiments were conducted in Jardinópolis-SP and Mogi Mirim-SP during the 2015/16 growing season. Two trials included acetochlor, flumioxazin, fluroxypyr, mesotrione and s-metolachlor, applied in pre or post-emergence, in association or not with atrazine. A third trial was carried out with rates of the premix containing atrazine + s-metolachlor applied to the following hybrids: 1G100, 1G220, 1G230, 1G244, 1G282, 50A10, 50A40, 50A50 and 50A70. The pre-emergence herbicides that exhibited satisfactory efficacy of weed control and selectivity to sorghum crop were flumioxazin, atrazine + mesotrione and atrazine + s-metolachlor. For post-emergence, atrazine, atrazine + acetochlor, atrazine + s-metolachlor and atrazine + fluroxypyr were the best treatments for both efficacy and selectivity. The application of atrazine + s-metolachlor at the evaluated rates was considered selective to the nine hybrids assessed.

Grain sorghum (Sorghum bicolor) is one of the five most important cereal crops in Brazil and worldwide especially because its use as an energy source in human and animal nutrition (Mutisyat et al., 2009).World production of this cereal has been around 43 million tons, most in the United States, while Brazil contributes with about 4.5% of production (USDA, 2018).In Brazil, grain sorghum is usually grown with low levels of investment, resulting in low average grain yields (2.8 tons ha -1 ) (Acompanhamento da Safra Brasileira [de] Grãos, 2018).Yet, in the center-west region, this crop has gained special attention as a second crop, being considered an important alternative in crop rotation with soybean (Fonseca et al., 2008).This is due to the grain sorghum advantages as a wide sowing window and greater resistance to water stress, when compared to corn, which is commonly used as second crop (Borghi et al., 2016).
Among the summer cereals, grain sorghum is one that faces the greatest problems with weeds due to the negative interference of these species and lack of herbicides to control them.Estimates indicate that weeds can reduce up to 97% of the grain yield potential when there is no control (Rodrigues et al., 2010;Tamado et al., 2002).However, even using herbicides the outcome is not always as expected, particularly for grass species (Archangelo et al., 2002;Dan et al., 2010).In fact, only atrazine, simazine and 2,4-D are currently registered herbicides for grain sorghum cropping in Brazil, but the main target of these herbicides are broadleaved species (Brasil, 2003).Although atrazine applications may help in the control of some weeds, its efficacy is variable due to several factors, especially the weed species and its growth stage (Marchesan et al., 2013).
In recent years, several studies have investigated other herbicide alternatives for use in grain sorghum crops in order to provide new viable weed chemical management strategies.A wellstudied and technically viable alternatives are the chloroacetamides, associated or not with atrazine, for post-emergence control (Archangelo et al., 2002;Takano et al., 2016).In a different way, other herbicides such as metsulfuron (Brown et al., 2004), fluroxypyr (Takano et al., 2016), flumioxazin (Galon et al., 2016), mesotrione (Abit et al., 2009) and tembotrione (Dan et al., 2010) have shown potential for use in grain sorghum crops, but they still lack more details regarding crop selectivity and technical positioning.
So, additional information on herbicide rate, plant stage and interactions with climate and soil is required for a safe use of these herbicides in grain sorghum.
Other key aspect when one considers other chemical alternatives for the control of weeds in grain sorghum is the selectivity stability to their distinct hybrids.In corn, tolerance of commercial hybrids to sulfonylurea herbicides varies considerably, being high for some hybrids and low for others (Cavalieri et al., 2008).In addition, tolerance of maize hybrids may vary according to the environment, crop management system, herbicide rate and the plant stage during the herbicide application (Cavalieri et al., 2012;Guerra et al., 2010).In a study comparing five grain sorghum hybrids, the hybrid 50A50 was the least sensitive and recovered more rapidly on the evaluation of 18 herbicide treatments (Takano et al., 2016).Despite clearly showing different tolerances for different hybrid, this study was preliminary and did not investigate other variables relating to crop development and performance.
The objectives of the present study were to assess the efficacy and safety of herbicide alternatives The experimental sites were sowed from March 15 to 27, 2016, using a seeder-for weed control in grain sorghum, applied in pre-or post-emergence of the crop, as well as regarding the selectivity of atrazine + s-metolachlor to different hybrids of this cereal.The experimental sites were sowed from March 15 to 27, 2016, using a seeder-fertilizer machine, simulating a conventional second-crop planting in Brazil.The hybrid 50A40 of grain sorghum was sowed with 18 seeds m -1 in rows, spaced 0.5 m, aiming to reach a population of 320,000 plants ha -1 .The agricultural practices used were those recommended for the region, with investment for high grain yield (Borghi et al., 2016).The experimental sites were spray-irrigated to a water level of 10 mm when occurred drought for a period longer than seven days.The experimental units consisted of field plots measuring 3 x 4 m (12 m 2 ) of useful land, containing six crop lines.
The herbicides were applied using a CO 2 backpack sprayer equipped with 3-m long boom and six nozzles (AIXR 110.015) and calibrated for 100-L ha and atrazine + s-metolachlor (2,000 + 576 and 770).In this experiment, it was also added an untreated checkplot without herbicides application.The herbicides were applied as described in the previous experiment.

Selectivity of atrazine + S-metolachlor to sorghum hybrids (third trial)
The experimental design and number of replicates were identical to the previous experiments; however, in this case, the experiment was conducted only in Jardinópolis-SP.The experiments were arranged in factorial design with subdivided plots (A = 3 x B = 9), with Factor A being allocated to the plots, and Factor B to the sub-plots.Factor A consisted of three rates of the pre-mixture atrazine + s-metolachlor applied in the crop post-emergence (3 to 4 leaves) (g a.i.ha -1 ): 0, 1,480 + 1,160 and 2,960 + 2,320 was the only weed evaluated.

Evaluations and statistical analysis
The efficacy of weed control was assessed 21 and 42 days after application (daa) and crop injury 14, 21 and 42 daa, using a visual scale ranging from 0 to 100%.In addition, the number of days between the plants emergence and 50% flowering (F 50 ), the crop population and height during full blossom were also evaluated.Finally, crop grain yield was assessed by harvesting the four central lines of the plot, and normalizing the values found to the standard grain moisture of 13%.The data were subjected to analysis of variance by the F-test for significant effects (p<0.05) and Tukey's pairwise comparison test for mean separation of the tested treatments.When the effect of sites and its interactions was not significant, the locations were considered as a random variable because there was homogeneity of error variances (Zimmermann, 2004).

Results and Discussion
A joint analysis of the data of Jardinópolis-SP and Mogi Mirim-SP was made for the majority of the variables, except for efficacy of control, where significant differences were found at the sites (p<0.05).So, the treatments values for this variable are shown in separate by location, while for the other variables are considered the means derived from the means of both places.

Efficacy and selectivity of pre-emergence herbicides (first trial)
In general, the efficacy of weed control in the treatments had mean values of at least 90%, especially when atrazine was associated with other residual herbicides (Table 1).In fact, these associations provided a high control of all species examined, except Portulaca oleraceae, which indicated 83-86% of control with atrazine and mesotrione.However, when applied alone the effect of the herbicides was not always satisfactory, such as of atrazine applied in both sites, and s-metolachlor in the experimental site in Mogi Mirim-SP.Control was higher than 90% only with flumioxazin applied alone, for most of the weeds in both assessments.
Regarding crop injury, a great variability of response was found in the treatments tested, atrazine being the only herbicide that caused less than 5% of injury in the three assessments (Table 2).Mesotrione, flumioxazin and their mixtures with atrazine caused maximum injury levels of 20%, which declined over time and remained below 10% at 42 daa.In the case of s-metolachlor, an effect of the dose-response was observed for the magnitude of this variable since the grain sorghum injury increased as the rate of the   herbicide increased.Less than 20% of injury was observed in the plots treated with 576 g ha -1 , while in the plots with 770 and 1,150 g ha -1 injury levels were higher than 50%.The population of grain sorghum was reduced only in the plots treated with s-metolachlor at 770 and 1,150 g ha -1 , which were statistically different from the untreated check-plot (Table 2).The crop height was only lower with application of s-metolachlor at 1,150 g ha -1 , associated or not with atrazine, when compared to the control.The number of days between the plants emergence and 50% of flowered sorghum plants (F 50 ) was smaller for the treatment with atrazine only and for the untreated check-plot, when compared to the s-metolachlor (1,150 g ha -1 ) alone or mixed with atrazine.The crop grain yield was only reduced with application of s-metolachlor (1,150 g ha -1 ), associated or not with atrazine, while the yields in the other treatments were not reduced.

Efficacy and selectivity of post-emergence herbicides (second trial)
In most cases, the efficacy of control values was higher than 95% in the plots treated with atrazine and its associations with other herbicides, in both assessments (Table 3).However, when applied alone (not associated with other herbicides), the effect was unsatisfactory (<90%), although no significant statistical difference was always found.
In fact, acetochlor, s-metolachlor and fluroxypyr did not achieve a satisfactory efficacy for Amaranthus spp., P. oleracea and R. raphanistrum, regardless of herbicide rate and experimental site.For mesotrione, the effect was satisfactory for Amaranthus spp., in both sites, and for R. raphanistrum in Jardinópolis-SP, but not for P. oleracea in Jardinópolis-SP.
Concerning crop injury, most of the treatments indicated high selectivity to the grain sorghum hybrid 50A40, with mean values lower than 5%, in the three assessment dates (Table 4).Acetochlor associated with atrazine with the highest rate tested caused injuries ranging from 6 to 10%, but did not differ statistically from the other treatments with the herbicides tested.Mesotrione and its associations with atrazine caused the greatest injury levels to the crop, either at 50 g ha -1 or 100 g ha -1 , in the three dates of visual assessment of the symptoms.While with the exclusive application of mesotrione the injury levels ranged from 12 to 27%, the magnitude of the effect of its association with atrazine was greater, with values ranging from 40 to 55%.
The crop population and height did not vary as a result of the treatments studied (data not presented).
The F 50 value was higher for the association of the atrazine + mesotrione (2,000 + 100 g ha -1 ) herbicides than for the untreated check-plot, as described in Table 4. Crop grain yield was lower only for the treatment with atrazine + mesotrione (2,000 + 100 g ha -1 ), when compared with the untreated check-plot.Therefore, in the plots treated with atrazine + mesotrione (2000 + 50 g ha -1 ), as well as with the other herbicides tested, there was not a decrease in yields that might be caused by the treatments.

Selectivity of atrazine + s-metolachor to sorghum hybrids (third trial)
The efficacy of control of Amaranthus spp. at 28 daa varied only as a function of the rate of the atrazine + s-metolachor, applied in a pre-formulated mixture, as shown in Figure 2. It can be seen that the efficacy of control provided by both dosages of herbicide was over 98%, and the control obtained        33.9 28.9 2.5 2.076.9 1/ daa: days after application; 2/ Mean values of both sites by the joint analysis.Means followed by the same letter do not differ from each other by the Tukey's test (p<0.05).
with the greater rate was significantly higher than that of the lower rate.
Regarding crop injury, there was a variation resulting from the interaction between the factors related to the rate of atrazine + s-metolachor and sorghum hybrids (    1 and 2) and post-emergence (Tables 3 and   4) crops, both with good results in weed control and crop selectivity.In addition, the study shows that the pre-mix atrazine + s-metolachlor has a high level of selectivity to the most hybrid varieties of sorghum (Table 5, Figure 2).
Flumioxazin and mesotrione, applied alone or in combination with atrazine, achieved satisfactory levels of efficacy and selectivity (Tables 1 and 2).
Added to these herbicides, s-metolachlor (576 g ha -

1
) was also another selective alternative to the crop, although its efficacy has not always been satisfactory, being required its association with atrazine (Tables 1 and 2).However, with rates of 770 g ha -1 and 1,150 g ha -1 , the use of this herbicide should not be recommended because of the high levels of injury caused to the crop, especially in sandy soils.In contrast, atrazine + s-metolachlor with rates of 1,120 + 1,400 g ha -1 (Geier et al., 2009) or 1,000 + 800 g ha -1 ( Takano et al., 2016) exhibited selectivity to grain in pre-emergence condition.
In the case of application of post-emergence herbicides, diverse combinations with atrazine and acetochlor, s-metolachlor and fluroxypyr indicated satisfactory efficacy of control and selectivity to the crop (Tables 3 and 4).These combinations also exhibited promising results in preliminary studies (Takano et al., 2016), and their use prevents the selection of triazine-resistant weed biotypes (Geier et al., 2009).In addition to these herbicides, mesotrione (50 g ha -1 ) was also an interesting option, even causing low-to-moderate injuries, which however did not affect the crop development and crop grain yield.These results corroborate those found by Hugie et al. (2008) and Abit et al. (2009), where mesotrione was a major option for the control of Amaranthus spp.Furthermore, this herbicide exhibited a high level of control of grass weeds such as Brachiaria spp., Digitaria spp., among others, and can be used with this purpose in grain sorghum crops.
The results obtained in the study with different grain sorghum hybrids showed that there is a special tolerance to atrazine + s-metolachlor, and that selectivity is associated with the herbicide rate (Table 5).These differences were also found in a preliminary study with grain sorghum (Takano et al., 2016) and have been widely known and studied for different maize hybrid (Cavalieri et al., 2008(Cavalieri et al., , 2012)).Therefore, the choice of the herbicide and rate to be applied, the hybrid variety to be planted and the crop management system should be taken into account in the strategy of weed control.For the particular case of the pre-formulated mixture of atrazine + s-metolachlor, with dose of 1,480 + 1,160 g a.i ha -1 , there is high selectivity stability among the commercial hybrids of this crop.This ensures the safe use of this new option for weed chemical control, and the effect of these herbicides will be a high level of control of a wide range of species (Table 4, Figure 2).
Although it is the most used herbicide for weed control in grain sorghum crops in Brazil, atrazine applied alone did not indicate good control of part of the weeds under evaluation in this study (Table 2).This shows the importance of its association with other residual herbicides in pre-and postemergence conditions, which benefits the effect obtained with the use of chemical control.The herbicides combinations that exhibited satisfactory weed control and selectivity to grain sorghum in the crop pre-emergence were: atrazine + flumioxazin (2,000 + 60 g ha -1 ), atrazine + mesotrione (2,000 + 100 g ha -1 ) and atrazine + S-metolachlor (2,000 + 576 g ha -1 ).In the crop post-emergence, the best herbicide combinations were: atrazine + acetochlor (2,000 + 1150 and 2300 g ha -1 ), atrazine + s-metolachlor (2,000 + 576 and 770 g ha -1 ) and atrazine + fluroxipyr (2,000 + 100 and 120 g ha - 1 ).These management strategies would provide benefits such as the rotation of herbicides action mechanisms, an increased range of weeds control, especially grasses, in the case of acetochlor, s-metolachlor and mesotrione.
The application of atrazine + s-metolachlor at the evaluated rates was selective to the nine hybrids assessed.

Figure 1 -
Figure 1 -Climate data collected during the conduction of the studies in Jardinópolis-SP and Mogi Mirim-SP, 2016.
Three trials were carried out during the 2015/2016 growth season in the municipalities of Mogi Mirim-SP (22°26'43"S; 47°04'02"W, 690 m altitude) and/or Jardinópolis-SP (20°54'39"S; 47°53'45"W, 560 m altitude), Brazil.The weather conditions during the period are illustrated in Figure 1.Both areas had a history of summer crops, especially soybean and corn, and have used no-till system for at least 10 years.Prior to the installation of the trials, soil samples were collected for physicochemical analysis and recommended fertilization.The soils of the experimental sites were made of 15 and 57% of clay, 73 and 22% of sand, 12 and 21% of silt, with 5.5 and 5.4 pH (CaCl 2 ) values, and 1.6 and 2.6% of organic matter in Mogi Mirim-SP and Jardinópolis-SP, respectively.The spontaneous vegetation grown in the sites was eliminated fi ve days before implementation of the crop with application of 400 g a.i.ha -1 of paraquat.

Figure 1 -
Figure 1 -Climate data collected during the conduction of the studies in Jardinópolis-SP and Mogi Mirim-SP, 2016.
. In this case, besides serving as a comparison pattern for the herbicides treatments, the untreated check-plot was also one of the levels of treatments tested for Factor A. Factor B consisted of nine sorghum hybrids with different levels of sensitivity to herbicides(Takano et al., 2016): 1G100, 1G220, 1G230, 1G244, 1G282, 50A10, 50A40, 50A50 and 50A70.Application of the treatments was carried out as described for the other experiments, and Amaranthus spp.(33 plants m -2 ) both sites by the joint analysis.Means followed by the same letter do not differ from each other by the Tukey's test (p<0.05).

Efficacy and selectivity of post-emergence herbicides (second trial)
-1of spray volume.The weeds assessed were Amaranthus spp., Bidens pilosa and Portulaca oleraceae in Jardinópolis-SP and Amaranthus spp.,

Table 2 .
Visual injury, crop height, crop stand, flowering time of 50% of the plants (F Mean values of both sites by the joint analysis.Means followed by the same letter do not differ from each other by the Tukey's test (p<0.05).

Table 4 .
Visual injury, flowering time of 50% of the plants (F 50 ) and crop grain yield after application of post-emergence herbicides on grain sorghum (50A40).Jardinópolis-SP and Mogi Mirim-SP.2016.Mean values of both sites by the joint analysis.Means followed by the same letter do not differ from each other by the Tukey's test (p<0.05).

Table 4 .
Visual injury, flowering time of 50% of the plants (F 50 ) and crop grain yield after application of post-emergence herbicides on grain sorghum (50A40).Jardinópolis-SP and Mogi Mirim-SP.2016.

Table 5 .
Visual injury 1/ and crop grain yield as a result of different hybrids and [atrazine + S-metolachlor] rates on grain sorghum crop.Jardinópolis-SP, 2016.Means followed by the same lowercase letter in the column and uppercase in the row not differ from each other by the Tukey's test (p<0.05). 2/ Evaluated at 28 days after application.

Table 5 .
Visual injury 1/ and crop grain yield as a result of different hybrids and [atrazine + S-metolachlor] rates on grain sorghum crop.Jardinópolis-SP, 2016.
1/ Means followed by the same lowercase letter in the column and uppercase in the row not differ from each ot.