In Brazil, grain sorghum has expanded to some areas where farmers have the ability to grow two crops on the same field per year. Sorghum is usually grown after soybean crop as a successional crop. In season 2013/2014, the area sown with sorghum exceeded 800,000 hectares, especially in the “Cerrado” area which accounted for more than 482,000 ha (Conab, 2015). Sorghum crop has a very well adaption in this region because its temperature and rainfall requirements.

Among the factors that interfere in sorghum yield is the interference imposed by the presence of weeds in the crop. In studies described in the literature, yield losses due to this interference may reach 85% for grain sorghum and 81% for forage sorghum (Andres et al., 2009; Rodrigues et al., 2010).

Despite the importance that sorghum has taken in recent years, and the significant yield loss when this crop remains in coexistence with weeds, there are few herbicides registered for use in the crop in pre-emergence and/or post-emergence. Currently, only atrazine, simazine and 2,4-D are registered for use in Brazil (MAPA, 2015), which restricts the farmers to control weeds in sorghum crop.

The sorghum crop is very sensitive to herbicides and therefore herbicide residual activity studies use sorghum like a bioindicator plant for testing the behavior of herbicides in the soil (Guerra et al., 2014). On the other hand, some studies in the literature have reported herbicides with a potential for use in sorghum in post-emergence, such as tembotrione (Dan et al., 2010), mesotrione (Abit et al., 2011), bentazon (Stahlman and Wicks, 2000), fluroxypyr (Horky and Martin, 2005), pendimethalin, trifluralin (Grichar et al., 2005) and metsulfuron (Brown et al., 2004; Hennigh et al., 2010).

For pre-emergence applications, the lack of herbicide options with a potential for use is even greater. Geier et al. (2009) assessing the application effects of acetochlor and s-metolachlor, alone or in mixture with atrazine, found that these herbicides can be safely applied in sorghum only when the seeds are treated with the fluxofenim safener. In the United States, sorghum seeds are usually protected with safener to allow acetamide herbicide application.

Thus, it is important to conduct research to find new herbicide solutions that can be selective and effectively used in this crop. Another factor to consider is the degree of tolerance of each sorghum hybrid to herbicides and also the dose of the herbicides, which should be selective to the sorghum but sufficient to control weeds. Abit et al. (2011) have assessed the response of 85 sorghum hybrids to the mesotrione application in 0, 52, 105, 210 and 315 g a.i. ha^-1 when plants had three to four leaves. They found a differential response of sorghum hybrids to mesotrione.

Within this context, the aim of this study was to assess the potential of using of different herbicide treatments applied in pre-emergence or in post-emergence of the grain sorghum crop based on selectivity and weed control.

The experiments were performed in an agricultural area located in the Brazilian municipality of Mogi Mirim, SP, at 22°26’44’’S and 47°04’13’’O.

Two experiments were conducted in two different fields, one for each modality of application: Experiment 1: pre-emergence; Experiment 2: post-emergence. The soil experimental field had: pH (CaCl₂) of 5.2; 1.6 cmol_c of H⁺ + Al⁺³ dm^-3 of soil; 4.3 cmol_c dm^-3 of Ca⁺²; 1.3 cmol_c dm^-3 of Mg⁺²; 129 mg dm^-3 of K⁺; 46.7 mg dm^-3 of P; 1.2 dag kg^-1 of organic matter; 42% of sand; 5% of silt; and 53% of clay (clay texture). Before installation of the experiments, the emerged weeds present in the experimental area were controlled by an application of paraquat (N,N′-dimethyl-4,4′-bipyridinium dichloride) (400 g a.i. ha^-1).

Sowing was carried out distributing fifteen seeds per linear meter via grain drill, sown to a depth of 1 to 2 cm. The fertilizer used in the planting furrow was 200 kg ha^-1 of the commercial formula 5-20-20 (N‑P‑K). The experimental area was equipped with sprinkler type irrigation. Whenever necessary, an irrigation of approximately 10 mm was applied.

The experimental design was a randomized block design with four replications, in split plots. For experiment 1, the different sorghum hybrids (50A10, 50A40, 50A50, 1G100, 1G233 and SS318) constituted the subplots, and the plots constituted by pre-emergence herbicides treatments (g a.i. or a.e. ha^-1 only for acetochlor): atrazine (1000), atrazine (2000), mesotrione (100), tembotrione (75), nicosulfuron (50), chlorimuron (20), s-metolachlor (800), acetochlor (2300), atrazine + trifluralin (1000+1000), atrazine + mesotrione (1000+100), atrazine + tembotrione (1000+75), atrazine + s-metolachlor (1000+800), atrazine + nicosulfuron (1000+50), atrazine + trifluralin (2000+1000), atrazine + mesotrione (2000+100), atrazine + tembotrione (2000+75), atrazine + s-metolachlor (2000+800), atrazine + nicosulfuron (2000+50) and untreated.

For experiment 2, the different sorghum hybrids (50A10, 50A40, 50A50, 1G100, 1G233 and SS318) constituted the subplots, and the plots constituted by post-emergence herbicides treatments (g a.i. or a.e. ha^-1 only for fluroxypir-meptyl): atrazine (1000), atrazine (2000), mesotrione (50), mesotrione (100), tembotrione (37.5), tembotrione (75), nicosulfuron (50), fluroxypir-meptyl (100), bentazon (720), metsulfuron (2), mesotrione + atrazine (50+1000), mesotrione + fluroxypyr-meptyl (50+100), mesotrione + nicosulfuron (50+50), tembotrione + atrazine (37.5+1000), tembotrione + fluroxypyr-meptyl (37.5+100), tembotrione + nicosulfuron (37.5+50), atrazine + nicosulfuron (1000+50), cloransulam (33.6) and untreated.

For both experiments, the experimental units comprised two rows of each hybrid (subplot), totaling twelve sowing rows (plot) spaced 0.45 m, 4 m long, with a total area of 21.60 m² per plot. Each plot corresponded to twelve rows except 0.5 meters at the ends of the sowing rows.

The applications were done with a CO₂ backpack sprayer at a constant pressure (45 psi), fitted with six AIXR 110.015 type spray nozzles, spaced 0.5 m, providing an application volume equivalent to 100 L ha^-1 of spray solution.

For experiment 1, application was done one day after sowing, in pre-emergence of the crop and of the weeds. The application conditions were: moist soil; average temperature of 26°C; average relative humidity of air of 78%; average wind speed of 0.5 km h^-1 and clear sky with few clouds.

For experiment 2, application occurred sixteen days after sowing, in post-emergence of the crop and of the weeds. At the time of applying, the crop had 3-4 fully expanded leaves, while weeds Panicum maximum and Bidens pilosa were at the 2 to 4 leaf stage. These two weed species were seeded on the field because the low infestation of weeds. The application conditions were: moist soil; average temperature of 25.8°C; average relative humidity of air of 71.0%; average wind speed of 0.8 km h^-1 and clear sky with few clouds.

Phytotoxicity assessments for each hybrid were conducted in both experiments, in which 0% means no plant injury, and 100% means all plants death. For experiment 1, the pre-emergence control of Eleusine indica, Brachiaria plantaginea, Euphorbia heterophylla and Ipomoea grandifolia was assessed at 7, 14 and 21 days after emergence of the sorghum (DAE). A stand assessment at 3 DAE has also taken place, counting the number of emerged sorghum plants in 2 linear meters of each hybrid within each plot. For purposes of analysis, the average value per meter sampled was considered. As for experiment 2, the control in post-emergence of weeds present at the time of application at 7, 14 and 21 days after application (DAA) was assessed.

All data were submitted to analysis of variance and, when detecting a significant effect among the tested factors or the levels of each factor, the means comparison Tukey test at 5% significance was applied.

Experiment 1: Pre-emergence

According to the variance analysis, the herbicide versus hybrid interaction was not considered statistically significant for any of the response variables. Thus, only the effects of herbicides were compared within each hybrid.

The application of the different herbicide treatments did not affect the emergence of the sorghum plants at 3 DAE (Table 1). What was observed was a delayed emergence of some hybrids when subjected to the application of some of the herbicide treatments, such as, for example, hybrids 50A50 and SS318, subjected to application of acetochlor and nicosulfuron, respectively.

 

 

At 7 DAE, the phytotoxicity in the treatments containing acetochlor, nicosulfuron, chlorimuron and s-metolachlor was extremely severe, especially when these same herbicides were applied in association with atrazine (Table 2). In a second level of phytotoxicity it is possible to include trifluralin + atrazine, mesotrione, tembotrione and their combinations with atrazine. Applying atrazine alone caused a little injury in the sorghum plants. Hybrids 1G233 and SS318A were considered sensitive to the application of high doses of atrazine + s-metolachlor.

 

 

The phytotoxicity observed at 21 DAE, in most treatments, was less than that observed in the first assessment, which indicates a recovery of these plants to the symptoms of injuries (Table 3). On the other hand, in the treatments containing nicosulfuron alone or mixed with atrazine there were very high levels of crop injury (>90%).

 

 

Some treatments such as atrazine, mesotrione, tembotrione, atrazine + trifluralin and atrazine + mesotrione had low levels of phytotoxicity at 21 DAE, which can be a potential indication for use in this crop. Importantly, no treatment showed a high degree of selectivity, other than the application of atrazine alone.

In general, hybrids did not show marked differences in crop response to the application of these herbicides. However, it is noteworthy that hybrids 1G233 and SS318 showed greater sensitivity to S-metolachlor application applied both alone and in combination with atrazine. Abit et al. (2011) have assessed the response of 85 sorghum hybrids to the application of different herbicides and also found differential responses across sorghum hybrids, which corroborates the results obtained in this experiment.

At 14 DAE, the treatments with atrazine effectively controlled the assessed weeds, especially when this herbicide was applied at higher doses (Table 4). It is noteworthy that the control of grasses was better when atrazine was applied in combination with other herbicides, especially for B. plantaginea. Acetochlor and S-metolachlor gave a satisfactory control of grass only, whereas chlorimuron and nicosulfuron controlled (>80%), especially for broadleaves. Mesotrione and tembotrione alone did not provide acceptable levels of control.

 

 

At 21 DAE, it was observed that the lowest dose of atrazine maintained control of the assessed weeds, except for B. plantaginea. However, control of grasses was better when this herbicide was applied at higher doses (2000 g a.i. ha^-1) or in combination with mesotrione and tembotrione (Table 4). Nicosulfuron provided excellent levels of control of E. heterophylla and I. grandifolia, while chlorimuron also provided a satisfactory control of E. heterophylla only. The other herbicides applied alone were not effective in controlling these weeds in pre-emergence.

 

Experiment 2: Post-emergence

The hybrid factor of the hybrid versus herbicides interaction showed no significant interaction in the assessments performed. This indicates that there was no differential response of the assessed hybrids to the application of the herbicides during post-emergence. On the other hand, the hybrid factor showed significance, which means different levels of selectivity of these herbicides for the sorghum hybrids used.

At 7 DAA, treatments atrazine (1000 and 2000), bentazon and fluroxypyr showed very low levels of phytotoxicity (<5%) (Table 5). In a second level of selectivity are mesotrione (50), mesotrione (100), mesotrione + atrazine and mesotrione + fluroxypyr. These treatments caused mild symptoms of chlorosis in plants, which accounted for these percentages of phytotoxicity. The other treatments gave high percentages of phytotoxicity, especially with nicosulfuron (>60%). Treatments containing tembotrione alone or in combination showed severe chlorosis.

 

 

Chlorosis observed in treatments containing herbicides from the group of carotenoid synthesis inhibitors has been minimized in the course of time, being observed only in older leaves (Table 6). It should be remembered that this chlorosis was more severe for tembotrione than for mesotrione. The symptoms of chlorosis in these treatments are due to the oxidative degradation of the chlorophyll and of the photosynthetic membranes, since carotenoid synthesis that protect them does not occur (Grossmann and Ehrhardt, 2007; Pataky et al., 2008).

 

 

For maize, what is observed is the opposite, because when assessing mesotrione and tembotrione herbicides in the selectivity for maize crop, Bollman et al. (2008) have found that mesotrione was what caused most phytotoxicity compared to tembotrione.Another point noted in this assessment is that the hybrids have shown high sensitivity to treatments containing ALS-inhibiting herbicides such as nicosulfuron, metsulfuron and cloransulam, which prevents their use in crops. At the last assessment of phytotoxicity (21 DAA), it was observed that most treatments recovered from the injuries seen inother assessments, indicating a potential use of these treatments in crops postemergence (Table 6). Herbicides that showed potential for use in crops were: atrazine, mesotrione, bentazon, fluroxypyr, mesotrione + atrazine and mesotrione + fluroxypyr. Bentazon and fluroxypyr are shown as viable options for controlling broadleaves, which decreases the dependence of atrazine and 2,4-D. Moreover, under the conditions of this study, fluroxypyr can be applied in more advanced stages of crops (V3-V4), which does not occur in the case of 2,4-D, which has the limitation of being applied before the crop reaches the V2 stage. In the other treatments, despite this reduction in symptoms of phytotoxicity, phytotoxicity levels can be an indication that these herbicides cause reductions in crop yield.

Dan et al. (2010) report that herbicide tembotrione showed high levels of phytotoxicity to the sorghum crop. They also state that there was greater potential for phytotoxicity when this herbicide was applied in the earlier stages of sorghum cultivar AG-1040. The same authors also state that there are different levels of selectivity, which can vary depending on dose and time used for application. In the case of this experiment, the studied sorghum hybrids showed great tolerance to tembotrione, demonstrating potential for use in that crop. A plausible explanation for this discrepancy may be the differentiated response that hybrids have to the application of tembotrione. Nevertheless, studies to assess the crop yield in response to these variables are needed to infer conclusions about the selectivity of these herbicides. In the control of B. pilosa and P. maximum, it was observed that at 7 DAA, some treatments such as atrazine, metsulfuron, bentazon, mesotrione + atrazine, tembotrione + atrazine, atrazine + nicosulfuron and cloransulam already had a satisfactory control of B. pilosa (Table 7). On the other hand, only tembotrione + atrazine provided a satisfactory control of P. maximum at 7 DAA, indicating a high difficulty of control of grasses by these treatments.

 

 

The mesotrione + fluroxypyr treatment provided a percentage of control of B. pilosa above 80% in this assessment. It should be noticed that for the treatments containing mesotrione or tembotrione, when the control of P. maximum is unsatisfactory (<80%) there is a significant suppression of this grass imposed by these herbicides. This suppression is even higher when these treatments are in combination with atrazine and nicosulfuron.

At 21 DAA, some treatments provided control that was higher than that observed in the previous assessment (Table 7). For the control of B. pilosa, treatments such as atrazine, bentazon, metsulfuron, mesotrione + atrazine, mesotrione + fluroxypyr, tembotrione + atrazine, atrazine + nicosulfuron and cloransulam were considered satisfactory. As for P. maximum, the number of options is smaller and only tembotrione + atrazine and atrazine + nicosulfuron are the effective treatments for this species.

Importantly, the treatments with mesotrione in general were selective, except when this herbicide was applied in combination with nicosulfuron. In addition, the combination of this herbicide with a post-emergence for broadleaves is an excellent option for the tillage of monocotyledon and dicotyledon weeds. Although these treatments have not obtained a satisfactory control of P. maximum, the effectiveness of mesotrione in early post-emergence has been observed in several grasses such as Digitaria horizontalis and Cenchrurs echinatus (Dan et al., 2011). Furthermore, the synergism between the pre- emergence and post-emergence applications of atrazine plus HPPD inhibitors on others weed species have already been related for other authors (Armel et al., 2005; Williams et al., 2011). The results obtained in this study indicate a number of treatments with potential use in crops, both as regards to the selectivity, as to the control of weeds during post-emergence. Nevertheless, it is noteworthy the recommendation that, for these treatments to be safe, further studies are needed to quantify the yield of grains of the crops when subjected to these applications.

The control of grasses remains a problem, since atrazine alone is not sufficient to ensure that the crop is in the clean during its development. To this end, as discussed in this paper, there are viable alternatives to complement the control in post-emergence. Thus, studies that assess "managements systems" are essential for progress in the weed management in sorghum crop.

Based on the results obtained, the herbicides and their respectively doses that had potential for use in sorghum crop in pre-emergence were: atrazine (1000 and 2000), mesotrione (100), tembotrione (75), atrazine + mesotrione (1000+100 and 2000+100) and atrazine + trifluralin (1000+1000 and 2000+1000). Meanwhile in post-emergence the best options were: atrazine (1000 and 2000), mesotrione (50 and 100), bentazon (720), fluroxypyr (100), mesotrione + atrazine (50+1000) and mesotrione + fluroxypyr (50+100). All of those treatments provided a satisfactory control of weeds, and presented less than 25% of plant injury which means less potential to reduce the sorghum grain yield.

 

The authors have not declared any conflict of interests.