Results and discussion

SPAD values, leaf area index and plant height

The greenness index (SPAD) of the genotypes was relatively constant for each of the clones throughout the crop cycle and in the three planting seasons, in general, the Citlali variety had statistically higher SPAD values (p≤ 0.05) a those registered for clone 99-39, which is due to the phenotypic characteristics of both cultivars, Citlali has more intense green color in its foliage (46.25 ^a, 46.96 ^a and 46.96 ^a) than the hue of clone 99-39 (44.68 ^b, 44.85 ^b and 44.41 ^b), in the combined analysis.

In the Table 1 shows the values of the green index for the interaction cycle x substrate x genotype; while the combined analysis indicated higher SPAD values in the perlite substrate than, in aggregates, at 90 days after sowing, on average (Table 2), this result was probably due to the fact that the nutrient supply was constant in perlite, while in aggregates it was provided every two days, with values of 46.79 ^a for perlite and 43.48 ^b for aggregates.

The average foliar area index of the two genotypes was not statistically different, in the combined analysis throughout the crop cycle; however, when the substrates evaluated were compared, the value of IAF in perlite was higher (p≤ 0.05) in perlite (5.96 ^a) than in aggregates (4.67 ^b), at 90 dds (Table 1).

In the Table 3 shows the response of genotypes in the three growing seasons, it is evident that spring-summer, when the average temperature was higher (22 °C) the plants of both genetic materials showed greater leaf area, especially when they were planted in perlite.

Plant height had a differential response for genotypes; clone 99-39 had an average height of 31 ^a, 55 ^a and 72 ^a cm, against 28 ^b, 50 ^b and 64 ^b cm, at 30, 60 and 90 days after sowing (Table 4); that length, in centimeters, was statistically greater (p≤ 0.05) when the culture was done in perlite (80 ^a) than when it was carried out in aggregates (56 ^b), according to the results of the combined analysis shown in Table 1. Both Foliar area index, such as plant height were statistically higher in perlite, compared to the values recorded in the substrate mix, probably due to the frequency of supply of the nutrient solution.

Tuber production per plant

In general, there was greater potato production when it was grown in horticultural grade perlite for both cultivars (Table 2). The number of tubers smaller than 15 mm in diameter did not show statistical differences between production cycles, but there was a genotype effect, since clone 99-39 produced, on average, 3.8 ^a mini-tubers and the Citlali variety 2 ^b. In the cycle x substrate interaction, more small tubers were produced in perlite than in aggregates, considering the three cycles.

The fact that in perlite the plant produced more tubers with a diameter of less than 15 mm, was probably due to the availability of nutrients continuously, which was quantified with the number of total tubers per plant. The foregoing coincides with that indicated by ^{Minhas et al. (2004)}, who report higher tuber production with greater availability of assimilates.

Genotypes grown in horticultural grade perlite, in the hydroponic system, produced an average of 25.08 tubers equal to or greater than 15 mm in diameter per plant, 83% higher than that recorded in the aggregate substrate (13.67 tubers), which coincides with as reported by ^{Adams (2004)}, who reports that yields are lower in organic substrates than in inorganic substrates in hydroponics, in addition to considering the continuous supply of the nutrient solution in the pelite substrate and the supply of nutrients every two days in the substrate aggregates. Clone 99-39 had a productivity per plant of 24.83 minitubers larger than 15 mm in diameter, statistically superior yield (p≤ 0.05) to that obtained by the Citlali variety with 13.92 tubers per plant; all this, on average of the three cycles and two substrates.

On the other hand, there were more mini-tubers of size equal to or greater than 15 mm in diameter in the autumn-winter cycle 35.06 ^a, compared to that obtained in spring-summer (12 ^b) and summer-autumn (11.06 ^b), on average of the two cultivars in the two substrates. The potato is known to have a higher yield when the difference between the maximum and minimum temperatures is high and the night temperature is less than 18 °C, which occurred in the autumn-winter cycle, with an average of 15 °C, a maximum from 35 °C and minimum of -2 °C as well as the night temperature is cool in November and December; which coincides with that indicated by ^{Beukema and Van der Zaag (1990)}.

When comparing the tubers per plant produced by the genotypes included in this study, it was found that clone 99-39 produced statistically (p≤ 0.05) more tubers per plant than the Citlali variety, with 24.83 and 13.92, respectively. Likewise, the response of the genotypes in the three production cycles, it was found that clone 99-39 generated, on average per plant, 34 minitubers equal to or greater than 15 mm in diameter, with perlite; however, the productivity of that clone was significantly reduced (p≤ 0.05) when grown in aggregates, where it only produced 15.41 tubers per plant of the indicated dimensions (Figure 1), the Citlali variety produced 15.9 tubers of equal size per plant or greater than 15 mm in diameter, when cultured in perlite and 11.9 in aggregates; these figures, for the second clone were statistically different (p≤ 0.05).

Figure 1 Average production of tubers equal to or greater than 15 mm in diameter in three growing cycles and two substrates for the Citlali variety and the mexican clone 99-39. 

It is important to indicate that tubers with a diameter equal to or greater than 15 mm are used for planting in the field, where between 60 and 80 thousand units are required to plant one hectare (^{Flores-López et al., 2014}).

The cycle x system interaction was the one with the greatest impact, since the average response in the production of mini-tubers with a diameter equal to or greater than 15 mm was remarkable; thus, in the autumn and perlite cycle, genotypes produced 42.25 tubers per plant, a figure significantly higher than the rest of the combinations. The substrates added in the spring and summer cycles showed the lowest yields, with 6.75 and 6.38 tubers per plant, respectively; that is, 6.5 times less.

This response was evidenced by analyzing the interaction cycle x substrate x genotype (Figure 2) where clone 99-39 planted in the autumn-winter cycle yielded 10 times more than the Citlali variety planted in aggregates, both in spring and summer; however, clone 99-39 planted in aggregates, in spring and summer did not show high potential for tuber production, so it is possible to point out that the effect of temperature is fundamental in the productivity of the materials and coincides with what is mentioned by ^{Beukema and Van der Zaag (1990)} and it is also evident that the response of the cultivars was diverse in this work carried out in the hydroponic system and is consistent with that indicated by ^{Muro et al. (1997)}; ^{Corrêa et al. (2008)}; ^{Tierno et al. (2013)}.

Figure 2 Effect of the interaction cycle x substrate x genotype on the production per plant of minitubers greater than or equal to 15 mm. 

The high production of clone 99-39 in the greenhouse in both cultivation conditions compared to the Citlali variety, is consistent with what happened in field conditions, where this clone has a yield potential greater than the Citlali variety, indicated by (^{Rubio et al., 2017}), who mention that clone 99-39 has an intermediate to late production cycle, under field conditions, with a potential yield of up to 80 t ha^-1, while the Citlali variety is an early genotype, with a potential yield in the field of 54 t ha^-1.

When the total production of tubers per plant (all sizes) was considered, clone 99-39 showed the best yield (p≤ 0.05) in the autumn-winter cycle, season in which the Citlali variety also obtained its best yield (34.5 tubers), although significantly less than that obtained by clone 99-39 (58 tubers); that is, 68% more tubers for 99-39 (Table 5).

Genotype 99-39 in autumn and planted in perlite produced 58 tubers per plant, the same clone planted in spring with perlite obtained 34.3 and in 26.3. That clone planted in aggregates yielded 33.5 tubers per plant in autumn, however, its response in aggregates was 33.5 tubers per plant in autumn, but only 10.8 in spring and 8.5 in summer. The Citlali variety showed a production of 34.5 tubers per plant when cultivated in perlite, in the autumn-winter cycle, 8 in spring and 6.3 in summer, while in aggregates it yielded 27.5 tubers per plant, when it was grown in aggregates in autumn and 6.3 in spring and summer.

In this study, the most productive crop cycle was that of autumn-winter 2015-2016, the conditions of cold temperatures at night and the difference between the maximum and minimum daily favored the mobilization of photoassimilates towards the tubers, in accordance with the noted by ^{Beukema and Van der Zaag (1990)} and the largest production of stolons.

Tuber weight and harvest index

The fresh weight of tubers per plant showed contrasting results with respect to the number of tubers per plant, the value of Pearson’s correlation coefficient between these variables was 0.08 and without statistical significance. In that sense, the comparison of means for crop cycles showed that in spring-summer the highest weight of tubers per plant was obtained, with 283 ^a gram, in summer-autumn the average was 185 ^b g and in autumn-spring 141 ^c grams per plant. Among substrates the response was similar to that observed in the number of tubers with 236 ^a gram in perlite and 170 ^b grams in aggregates.

When considering the interaction cycle x genotype, the response in fresh weight was significantly (p≤ 0.05) higher for spring and clone 99-39 (325 ^a g), which surpassed the Citlali variety planted in the autumn cycle that obtained 119^d grams. The interaction cycle x substrate x genotype (Figure 3) shows the differential behavior of the materials under study. With this, it is evident that the production of more tubers does not necessarily have a greater impact on them, a fact that contrasts with what was pointed out by ^{Beukema and Van der Zaag (1990)}, but which apparently is related to the differential response of the genotypes of potato and its interactions in hydroponic cultivation (^{Muro et al., 1997}; ^{Corrêa et al., 2008}; ^{Tierno et al., 2013}.

Figure 3 Effect of the interaction cycle x substrate x genotype on the fresh weight of tubers per plant of two potato genotypes. 

In addition, it should be considered that the tubers of the Citlali variety are larger, due to the precocity of that cultivar. The weight of the tubers of clone 99-39, in general, was lower, probably because they were only grown for 90 days, when their cycle is at least 110 days, therefore, their tubers, although in good numbers, could still gain more weight; especially in the autumn-winter cycle, when plants tend to have a later cycle and this fact does coincide with what was reported by ^{Beukema and Van der Zaag (1990)}.

Regarding the dry weight of tubers per plant, the combined analysis showed results similar to those of fresh tuber weight, therefore, Pearson's correlation coefficient between these variables was 0.94 (p< 0.0001). In relation to the harvest index, the greatest statistical differences were found when comparing the values obtained in the three production cycles (Figure 4). In the autumn-winter cycle, the highest harvest index (0.80) was obtained, followed by the summer-autumn cycle (0.76), and the lowest was the spring-summer cycle (0.73).

Figure 4 Average harvest index of two Mexican potato genotypes in three crop cycles and two substrates. 

These values were closely associated with the temperatures recorded in the different production cycles; at a lower temperature, higher harvest rate; this corresponds to what was observed by ^{Flores-Gutiérrez et al. (2018)} and mentioned by ^{Beukema and van der Zaag (1990)}. The harvest index values of the Citlali variety and clone 99-39 on two substrates and three crop cycles are presented in Figure 5. Although in the analysis of substrates and genotypes there was no statistical significance, if there was in the interaction indicated, so that the environment is probably the factor with the greatest effect on the accumulation of dry weight in the tubers.

Figure 5 Crop index in the genotype x cycle per substrate interaction.