Results and discussion

Rainfall and soil moisture balance of the experimental sites

During the autumn-winter cycle of 2013 in the sites located in Ocozocoautla and Medellin, rainfall accumulated until the physiological maturity of the culture of 97.7 and 52.1 mm, respectively (Figure 1), with low humidity availability during most of the year. the reproductive phase, while in the Tlapacoyan, frequent rains were recorded, with an accumulated rainfall during the crop cycle of 391.6 mm, so that the bean plants developed without stress due to humidity.

Figure 1 Rainfall occurred and available moisture in the soil profile in Ocozocoautla, Medellin and Tlapacoyan, during the autumn-winter cycle of 2013. 

The soils in which the trial was established showed a field capacity (CC) of 23.6, 20.3 and 13% at the sites of Medellin, Tlapacoyan and Ocozocoautla, respectively, with a permanent wilting point (PMP) of 12.5, 10.8 and 6%, in the same places, so the maximum usable humidity was 11.1, 9.5 and 7% (^{Hillel, 1980}).

In Medellin, a gradual decrease in the available soil moisture was observed as the crop advanced to physiological maturity, which was associated with the low amount of rains that occurred during the vegetative phase. In Ocozocoautla, there was high variation in the usable humidity, especially before the flowering of the genotypes; after that period until the stage of physiological maturity, water deficit was observed, which was more severe than in Medellin, mainly due to a lower capacity of soil moisture retention (7% vs 9.5%), associated with its more sandy texture (52.4% vs 33.2% of sand) and lower content of organic matter (1.1% vs 2.5%). In turn, in Tlapacoyan, although the usable humidity was variable during the cultivation cycle, it remained above the critical limit (45%), so that the bean plants did not suffer from water stress throughout their development.

The usable humidity from sowing to flowering of most of the genotypes was 73.2, 67 and 61.9%, at the sites of Medellin, Tlapacoyan and Ocozocoautla, respectively. After that period, towards the final stages of the crop, the usable humidity in the respective sites was 34.3, 65.9 and 15.5%.

In the spring-summer cycle of 2014, no rainfall was recorded during the entire crop cycle at the Medellín and Cintalapa sites, so that the beans were developed with the moisture provided by the irrigation that was applied in pre-sowing and during the vegetative phase, after this phase, both sites showed a gradual decrease in the usable humidity (Figure 2), with occurrence of drought from the beginning of the pod filling stage (between 50 and 52 d after from sowing) to physiological maturity, which occurred on average at 70.7 and 68.1 d, respectively.

Figure 2 Rainfall occurred and usable soil moisture in Cintalapa, Medellín and Tlapacoyan, during the spring-summer 2014 cycle. 

In turn, in Tlapacoyan rainfall was recorded that accumulated 187.2 mm during the cycle, with adequate distribution and more during the reproductive phase, so, in this case, the crop was developed under better conditions of usable humidity than in the other two test sites (mainly after flowering), which was sufficient for the bean plants to be without moisture stress until the end of the pod filling stage (69 d). Subsequently, the usable humidity decreased until exceeding the critical limit of 45% (^{Allen et al., 2006}) towards the stage of physiological maturity. It is known that the occurrence of terminal drought at the end of pod filling has no effect on grain yield, because bean plants begin their maturity (^{Mouhouche et al., 1998}; ^{Nielsen and Nelson, 1998}).

In the sites of Medellín, Tlapacoyan and Cintalapa the soils presented a CC of 23.6, 20.2 and 13.4% and a PMP of 12.5, 9.9 and 7.6%, respectively, which allowed a usable humidity of 11.1, 10.3 and 5.8%. The above, indicates that Cintalapa soil has less capacity to store moisture, than that of Medellin and Tlapacoyan, mainly due to its lower content of organic matter (1.36% vs 2.44 and 2.24%, respectively) and its more sandy texture (73.2% vs 66.2 and 33.2%). In the Tlapacoyan, Medellin and Cintalapa sites, the available humidity from sowing to flowering was 92.2, 90.5 and 82.4%, respectively; while, towards the terminal stage of the crop, the values in the respective sites were 66.6, 30 and 32.5%.

The usable humidity of the soil observed in the two production cycles indicate that in Medellin and in Ocozocoautla, in AW of 2013 and in Medellin and Cintalapa in SS of 2014, there were terminal drought conditions, since the usable humidity of the soil was lower to 45% during the reproductive phase of the crop (^{Allen et al., 2006}), while, in Tlapacoyan, this moisture condition only prevailed from the end of the pod filling stage in spring-summer of 2014 (Figure 2 ), and that this did not affect the grain yield, it can be asserted that in this evaluation site, in both cycles, moisture conditions prevailed that did not cause water stress to the bean genotypes.

Physiological maturity and grain yield

Significant differences (p≤ 0.01) were detected between environments and genotypes, under conditions of terminal drought, without drought and in the combined analysis of the six test environments. In Tlapacoyan in the autumn-winter cycle of 2013, physiological maturity occurred on average 81 days after sowing (dds), significantly later than in the other evaluation sites (Table 2), which was due mainly to the availability of higher humidity during the crop cycle (Figure 1), which favored a longer duration of the reproductive period. In contrast, in the four environments in which there were no rains during the reproductive phase of the crop, the maturity occurred on average before 71 dds (Table 2), due to the lack of moisture after the flowering stage of the crop, which provoked an advance in physiological maturity, which according to ^{Rosales-Serna et al. (2001)}, the acceleration of maturity occurs when the drought is prolonged during the reproductive phase and there are no favorable moisture conditions for recovery.

The Verdín variety, together with the SEN-56 line, were the earliest genotypes to reach maturity, both under drought conditions and without humidity stress, although under this latter condition, the genotypes showed a lower precocity, due to a higher better distribution of rainfall during the crop cycle. On average of the six test environments, these two genotypes reached maturity before 69 dds, a significantly shorter time than that of seven other lines and the commercial varieties Negro Tacaná and Negro Jamapa (Table 2). The precocity of Verdín and SEN-56, has been documented in evaluation studies conducted in the state of Veracruz (^{Tosquy et al., 2014}), which is important to avoid terminal drought, since it allows bean plants complete its cycle and escape the drought before critical problems arise due to lack of humidity (^{Acosta-Díaz et al., 2004}; ²⁰¹¹; ^{Acosta Gallegos and Kelly, 2012}). In turn, the genotypes with the greatest number of days to reach maturity under moisture-limiting conditions generally show lower productivity (^{Zilio et al., 2013}).

The grain yield showed a significant difference (p≤ 0.01) between environments and genotypes, in the two humidity conditions and in the joint analysis of the six evaluation environments. The highest averages were obtained in Tlapacoyan, in the autumn-winter cycle of 2013 and spring-summer of 2014, which were higher than those obtained in the four environments where terminal drought occurred (Table 3). This was due to the fact that in Tlapacoyan, in both production cycles, there was an adequate distribution of rainfall, mainly from the stages of flowering to pod filling, during which time the crop received 92 and 91.7 mm of rainfall in autumn-winter of 2013 and spring-summer of 2014, respectively, sufficient quantities for the filling of pods (^{López et al., 2008}), as well as there was more usable soil moisture for the bean plants (Figures 1 and 2).

The lowest average yield was obtained in Ocozocoautla Autumn-Winter cycle of 2013 (Table 3), due to the severe terminal drought and the soil has low moisture retention capacity, due to its high percentage of sand and low content of organic matter. These soil characteristics imply low usable humidity for the crop, and in this case the occurrence of drought during the reproductive stage of the crop (Figure 1), affected grain yield (^{Acosta-Díaz et al., 2009}). In Medellin, in both production cycles, there was less rainfall than in Ocozocoautla in Autumn-Winter of 2013; however, the yield was higher, mainly due to the fact that the soil where the trial was established is a river vega, with a high content of organic matter and a frank texture, which gives it a greater moisture retention capacity than that of Ocozocoautla.

The variety Verdín, together with the lines NGO-17-99 and SCN-2, obtained an average performance significantly outstanding in conditions of terminal drought, while, without the presence of drought, Verdín was the most productive. This variety obtained the highest average yield of the six test environments, which was statistically similar to that of NGO-17-99 and SCN-2 and superior to those of the rest of the genotypes (Table 3). According to the parameters of stability, all the genotypes were stable in their performance; that is, they presented a low interaction with the environments in which they developed, since their regression coefficients and regression deviations were: bi= 1 and S²di= 0 (^{Eberhart y Russell, 1966}). The stability characteristic of the Negro Tacaná and Negro Jamapa commercial varieties had already been documented previously (López et al., 2001; 2007b).

These results indicate that Verdín, NGO-17-99 and SCN-2 adapt to the environmental conditions of the humid tropics of Veracruz and Chiapas and have greater yield potential than the commercial varieties of current use in both entities. However, the NGO-17-99 line has a significantly longer maturity cycle than that of the Verdín variety, while the grain of the SCN-2 line is larger (>25 g per 100 seeds) than the demanded by producers and consumers in the states of Veracruz and Chiapas (^{López et al., 2012}). Therefore, the new Verdín variety is considered superior.

The Verdín variety, due to its characteristics of precocity and tolerance to conditions of humidity stress (^{Tosquy et al., 2014}; ^{López et al., 2015}; ^{Tosquy et al., 2016a}), will allow the producer to reduce risks of yield loss for periods of terminal drought, as well as harvesting up to 10 days earlier than with other varieties and obtaining greater grain yield, with and without the presence of terminal drought stress, than with Negro Jamapa and Negro Tacaná varieties and other materials commonly used in Veracruz and Chiapas (^{Tosquy et al., 2016b}). In addition, the grain of the variety Verdín, gathers the characteristics of bean type, black, opaque and of small size (<25 g per 100 seeds), which consumers demand in the southeast of Mexico and its technological quality of grain (data not shown), is similar to that of the Negro Jamapa variety (^{Tosquy et al., 2016b}).

Conclusions

The differences in moisture availability during the reproductive stage; through, the test sites affected the productive response and the duration of the crop cycle of the evaluated genotypes, to higher stress due to drought, lower yield and shorter cycle.

The variety Verdín, together with the lines NGO-17-99 and SCN-2, obtained an average yield significantly superior to the rest of the genotypes in the test locations with conditions of terminal drought, while, in the localities without presence of drought, Verdín was the most productive.

The Verdín variety showed a shorter cultivation cycle than the rest of the genotypes under evaluation, a phenological characteristic that partially allowed it to escape the terminal drought.