Bread wheat (Triticum aestivum) is the most grown cereal in the world (^{FAO, 2018}). In Mexico, in 2018, bread wheat was the third most important cereal produced in terms of area sown (541,789 ha) (^{SIAP, 2019}). Demand for wheat by 2050 is predicted to increase by 70% due to population growth and changes in people’s diet. However, in recent years, wheat production has been affected by biotic factors (pests and diseases) and mostly by abiotic factors (soil and weather) (^{Beddow et al., 2015}). Among the biotic factors, diseases mainly caused by fungi, the most important are those caused by Puccinia, which is considered the most devastating genus causing the greatest crop losses in the world (^{Singh et al., 2016}). Currently, phytopathogenic bacteria are responsible for considerable wheat yield losses (^{Adhikari et al., 2011}; ^{Kandel et al., 2012}). Nowadays, bacterial stripe caused by Xanthomonas translucens is the most common wheat disease (^{Adhikari et al., 2011}). On the other hand, Xanthomonas campestris pv. translucens (Syn. Xanthomonas translucens pv. undulosa) and Pseudomonas syringae pv. syringae have re-emerged as a potential threat to bread wheat production in the United States and other countries in the world (^{Kandel et al., 2012}). There has been speculation that the re-emergence of diseases has to do with the low level of wheat varieties resistance, changes in the environment, especially temperature increases during the crop cycle, cultural practices and, probably, the evolution of pathogenic populations, which has turned them more virulent (^{Kandel et al., 2012}; ^{Ávila-Quezada et al., 2018}). Diverse studies mention that this could be a problem caused by seed contamination, a factor that restricts germplasm exchange (^{Adhikari et al., 2011}; ^{Mezzalama, 2013}). According to ^{Gitaitis and Walcott (2007}) and ^{Navarrete-Maya et al. (2014}), seed movement is the source of pathogen’s inoculum not only in new areas but also among countries. The same authors mention that exclusion and evasion measures such as quarantine and seed production in semiarid regions are effective up to a certain level, since these practices have not totally controlled bacterial diseases transmitted by seed. Over the years, the main measures for controlling diseases caused by phytobacteria have been non-host crop rotations, use of fluorescent antagonistic bacteria of Pseudomonas genus, seed production in disease-free regions and genetic improvement to obtain resistant varieties (^{Sapkota et al., 2020}; ^{Valencia-Botín et al., 2012}). The elimination of the pathogen from the seed could help to control bacterial diseases, but the methods developed so far are not effective. In recent years, the use of seed chemicals treatments has quickly accelerated and evolved due to the integration of diverse factors that have contributed to the development of strategies for controlling the disease (^{Munkvold, 2009}). For this reason, the objective of this study was to try different chemical treatments for controlling bacteria in wheat seed that was produced under rainfed conditions and to determine the effect of these products on germination.

The experiment was established in May 2018 under controlled conditions in the National Laboratory for Rusts and Other Cereal Diseases (LANAREC), at the Experiment Station Valle de México of the National Institute of Forestry, Agricultural and Livestock Research (INIFAP). The experiment was conducted using a completely randomized design under a factorial arrangement with three replications (3x16x4x3): varieties, chemical treatments, locations and replications. Seed lots of three bread wheat varieties recommended for rainfed conditions were evaluated: Gálvez M87, Tlaxcala F2000 and Nana F2007. These varieties are representative of different commercial release cycles. The seed lots (100 g) of the three seed genotypes came from four locations: Terrenate and Nanacamilpa, Tlaxcala; Chapingo, State of Mexico; and Cuyoaco, Puebla. Previous evaluations showed that these genotypes were saprophytic and phytopathogenic bacteria carriers, mainly of Pseudomonas genus (data not published). The 16 treatments used (including the check) were hydrogen peroxide (3% oxygenated water) for 1, 2 and 3 min; 2% and 5% sodium hypochlorite (Cloralex) for 1, 2 and 3 min, Methoxychlor+Captan (Metacaptan) to 80, 90 and 100 g/100 kg of seed, and copper hydroxide (Cupravit Hidro) to 200, 300 and 400 g/100 kg of seed; and seed with no chemical treatment. Hydrogen peroxide was applied to seed by immersion. In the case of Metacaptan and Cupravit Hidro, the products were first mixed with water and then applied to the seed. Once the chemical treatment was applied, the seed of each of the 576 treatments, which resulted from the combination of all the factors involved in the experiment design, were sown in plates containing nutritive agar. The experiment unit consisted of 10 seeds taken at random per plate (^{Warham et al., 1997}). They were then incubated at 25 to 28 °C for six days. The evaluated variables were infected seed percentage (PSI; seeds with bacterial exudate) and germination percentage (PG), which were obtained by visually counting each variable in relation to the total number of seeds in each plate. Two evaluations were carried out. The first was made three days after sowing, and the second, six days after sowing. The obtained data were subjected to statistical analysis using SAS 9.1.3 (SAS Institute®, United States) and the means of the variables were compared by applying Tukey’s test (p≤0.05).

The analysis of variance of the evaluated variables showed highly significant statistical differences for seed origin (locations), varieties and chemical treatments at the corresponding doses and/or time exposure. For the variety*product interaction, only the PSI variable was statistically significant, and though no significant statistical differences were found in PG, it showed negative tendencies by increasing the exposure time.

The comparison of the average values of the infected seed percentage per location (Table 1) showed that the seed of the different varieties harvested in Terrenate, Tlaxcala, had a higher level of bacteria incidence with 62% of infected seed. Seed collected in Nanacamilpa, Tlaxcala, showed a lower level of bacteria incidence (27% of infected seeds). These results may be attributed to the environmental conditions prevailing in each location, and that, according to ^{Villaseñor and Espitia (2000}), the first location is considered as an intermediate production environment where rainfall ranges from 400 to 600 mm during the crop growth stage and the main problems are poorly distributed rainfall, high temperatures and foliar diseases. On the other hand, Nanacamilpa is considered a location with favorable production conditions with amounts of rainfall greater than 600 mm, usually well distributed, and mild temperature.

Regarding PG, the greatest problems were found in Terrenate, which only reached 55% germination, a fact that could indicate that there is a relation between seed infection and seed germination. The best average value of PG (80.6%) was observed in Chapingo, followed by Nanacamilpa (73.0%) (Table 1), whose response is probably influenced by the prevailing conditions in this region, which are ideal for other diseases but limited for bacteria.

The varieties mean indicates that there is a higher level of variation and that, for this reason, their performance varies under the environmental conditions in the evaluated sites. The differences identified in PSI were not statistically significant between varieties Tlaxcala F2000 and Nana F2007, but they were in Gálvez M87 (Table 1). This result could indicate that of the three evaluated varieties, Gálvez M87 probably has genes that confer resistance to bacteria, and/or because it is an early cycle variety (45 days to heading), this trait helps it escape the disease, as mentioned by ^{Hortelano-Santa Rosa et al. (2016)}. On the other hand, when the PG results were compared, each variety also showed different performance. Tlaxcala F2000 had the highest PG, whose value was statistically higher than the value of Nana F2007 and Gálvez M87, as shown in Table 1.

The results of means comparison of the different products and exposure time (chemical treatments) in Table 2, show statistically contrasting differences among the values observed in each treatment applied to the evaluated varieties. The most effective treatment to reduce PSI was 5% sodium hypochlorite at 2 and 3 min, whose values were statistically like the values obtained when using 2% sodium hypochlorite. Although the values do not show statistical differences, there is a numerical superiority in the case of 5% sodium hypochlorite at different exposure times. For the seed treated with Metacaptan, the values were statistically similar to those obtained using 2% sodium hypochlorite, except at a 90 g dose, where the mean value was statistically similar that obtained using 5% sodium hypochlorite (Table 2). The treatments with copper hydroxide were more effective at low doses (200 g). However, our results are not satisfactory due to the high percentage of infected seeds (55% of infected seed). In the case of seed treated with hydrogen peroxide, the infection values were very high and the three-time intervals for the treatment were statistically similar. The check treatment had the highest percentage of infected seeds (93.9%), and as expected, the control of bacterial infection was statistically null. However, the PG was more than 10% higher compared to the PG of the treatment with copper hydroxide (300 g/100 kg of seed) (Table 2).

In the PG variable, the performance was very similar to the control in PSI. The highest germination percentage was obtained with 5% sodium hypochlorite at 1 min, with an average value of 82%. This value is statistically similar to that obtained with sodium hypochlorite at 2 min, which, when compared to the PSI values, showed the best bacterial control. Although the variations in the average response to the PG treatments are more contrasting between and among the groups of products, it can be generally observed that treatments with hydrogen peroxide and copper hydroxide have the lowest germination percentages and are directly related to the control of the incidence; they are also the treatments with the lowest control in this study.

In the case of the control, the mean germination percentage is statistically similar to that of all the chemical treatments, when all the treatments were conjunctly analyzed; showed that the time interval or the product dose has a significant influence in the response of seed germination. The germination value of the check is not affected as it happens in the treatments to which some chemical product is applied though there is no control of bacteria incidence in the check. These results agree with the results reported by ^{Lozano-Ramírez et al. (2006}), who indicated that when a mixture of four active ingredients (carboxine+captan+thiram+tebuconazole) was used to control Fusarium graminearum, the germination percentage was affected and reduced to 85.5% compared to that of the control (97%).

Currently, with the re-emergence of phytopathological problems caused by bacteria, as pointed out by ^{Adhikari et al. (2011}) and ^{Kandel et al. (2012}); considering that one of the main means of bacterial diseases dispersal is seed, germplasm exchange and seed movement from one site to another, as indicated by ^{Navarrete-Maya et al. (2014}), ^{Gitaitis and Walcott (2007}); and, because of the short availability of genetically resistant wheat germplasm (^{Sapkota et al., 2020}), chemical treatments can be one of the most viable measures for seed treatment, as indicated by ^{Munkvold (2009}), which entails the decrease of the level of incidence and dispersal of bacterial and phytopathogenic fungi in wheat crops, since these factors limit the yield potential of the genotype.

Although the best bacterial control was achieved with 5% sodium hypochlorite (Cloralex commercial product), when the seed is exposed for long periods, the percentage germination is affected. ^{Maeso and Walasek (2012}) reported an effective control (99%) using sodium hypochlorite 1% of active chlorine in a treatment applied to tomato seed to protect it against bacterial canker caused by Clavibacter michiganensis subsp. michiganensis, without affecting germination. Currently, the main objective of seed treatments is to control fungi. Yet, it is also important to consider bacteria, and, for this reason, a treatment with Metacaptan (Metoxicloro+Captan) at a higher dose (90 g/100 kg/seed) than the recommended (80 g/100 kg/ of seed) is a viable option. Regarding copper hydroxide, ^{Forster and Schaad (1988}) reported an effective control in Xanthomonas campestris pv. translucens (Syn. Xanthomonas translucens pv. undulosa)) in seed treatments at a dose of 260 mL/100 kg of seed. However, in this study, the obtained results were not satisfactory because other products were more effective (5% sodium hypochlorite). It is important to consider the combination of chemotherapy and thermotherapy studies in further research since these could provide a more effective control.

Based on the results of this study, the best product (effective and low-cost) for controlling bacteria was 5% sodium hypochlorite at 2 and 3 min, followed by 2% sodium hypochlorite at 2 min, and Metacaptan at 90 g. However, when using 5% sodium hypochlorite at 3 min, germination was reduced by 17%. The least effective product was hydrogen peroxide at 2 min, but none of the products used was 100% effective for controlling bacteria. Currently, it is important to highlight that the genetic improvement programs should take into consideration the diseases caused by bacteria given that the crop yield and economic losses are significant and there is no genetic resistance to address this problem.