Yield of fresh tubers and related variables

Highly significant statistical differences (p> 0.01) were found for cultivars in the six variables studied (Table 5); that is, there are real differences between the averages of these variables. These differences are probably explained by the genetic variability of the cultivars under study.

Statistical significance (p> 0.01) was also found for the environments (repetitions) in four yield-related variables (NTT, NTC, AP and NT), which suggests that the environments where the trials were carried out (different years and months) influenced the evaluated components, in a differentiated way. The coefficients of variation varied between 13.5% and 27.6%, these values are considered acceptable for the field conditions where the trials were carried out (^{Vásquez, 2014}). This influence was to be expected because sowings were carried out in different seasons (rain and dry) and where temperatures, rainfall and hours of sunlight varied significantly (Table 4).

In previous tests with cultivars of the same group, statistical differences were found for most of the variables evaluated here (^{Pérez et al., 2008}; ^{Rozo and Ñustez, 2011}; Rojas and Seminario, 2014; ^{Seminario et al., 2017}; ^{Seminario et al., 2018}). These similarities are important because they allow identifying the quantitative variables that differentiate the cultivars and also to identify the basic components of the yield such as the number of tubers and the average weight of tubers per plant, which combined with the density of planting allow estimating the yield per hectare (^{Hay and Walker, 1989}).

According to Duncan’s test (α= 0.05) (Table 6), it is observed that the cultivars ‘Chachapoyana’, ‘Montañera’, ‘Blanca’ and ‘Shoga’ are statistically similar and stand out for the total weight of tubers (PTT) and weight of commercial tubers (PTC) with respect to the remaining eleven cultivars.

The cultivar ‘Chachapoyana’ ranks first in terms of the components total number of tubers (NTT), number of commercial tubers (NTC) and number of stems (NT). The cultivars ‘Huagalina’, ‘Amarilla’ and ‘Yuquilla’ do not show significant statistical differences for the variables NTT, NTC and NT. With regard to plant height (AP), the Cultivars Amarilla and Mulla do not differ statistically, but exceed the remaining thirteen cultivars.

The superiority shown by certain cultivars of the study group, with respect to the main components of the yield such as the number and total weight of tubers and, number and weight of commercial tubers constitutes an advantage that must be used to carry out specific yield tests, in order to identify the most productive cultivars in the region.

In previous studies with cultivars of the same group and from the Cajamarca region, other cultivars have stood out in the basic components of yield (NTT, NTC, PTT and PTC). Thus, they are mentioned as the best: ‘Roja-2’, ‘Piña amarilla’ and ‘Montañera-3’ (Rojas and Seminario, 2014). ‘Roja-2’, ‘Amarilla mahuay’, ‘Piña amarilla’, ‘Amarilla’, ‘Montañera-3’ and ‘Porpora’ (^{Seminario et al., 2017}). ‘Blanca amarilla’, ‘Limeña huachuma’, ‘Llanqueja’, ‘Amarilla redonda’ and ‘Piña amarilla’ (^{Seminario et al., 2018}). These elite cultivars are properly morphologically characterized (^{Seminario et al., 2019}) and must be analyzed through specific studies.

Dry matter yield (g plant^-1) of cultivars

High statistical significance was found for the source of cultivar variation, meaning that there are real differences between the dry matter averages of the 15 cultivars. Similarly, there was statistical significance at 5% probability for blocks (Table 5), indicating that there was an environmental effect on the variability of dry matter content in g plant^-1. That is, the dry matter content is a genetically controlled character and is affected by the conditions of the environment and by the genotype x environment interaction (^{Vásquez, 1988}; ^{Hay and Walker, 1989}). The coefficient of variation was 21.9%, an acceptable value for the field conditions where the experiment was conducted.

The Tukey multiple-range test (α= 0.05) (Table 6) indicated that there are no statistical differences between the cultivars ‘Montañera’, ‘Chachapoyana-1’, ‘Shoga’ and ‘Blanca’, whose averages vary between 74.28 g and 91.88 g, the cultivar ‘Blanca’ with the highest yield of dry matter (91.888 g plant^-1) stands out. A second group of cultivars was found, which consists of ‘Piña amarilla’, ‘Poropora’, ‘Amarilla mahuay’, ‘Mulla’ and ‘Limeña’, with averages varying between 66.613 g plant^-1 and 58. 638 g plant^-1 of dry matter, but no statistical difference between them.

The variation in dry matter yield among the 15 cultivars (37.58 g plant^-1 to 91.88 g plant^-1) is explained because these are different genotypes and also by the environmental effect. It is known that the dry matter content is a polygenic character, highly heritable and of low interaction with the environment (^{Vásquez, 1988}; ^{Martínez and Ligarreto, 2005}; Bautista et al., 2012). Several studies have shown that in these cultivars, it is expected that the highest proportion of dry matter is allocated to the tubers, which is demonstrated with the harvest index (IC). In general, IC values are higher than 40% (Rojas and Seminario, 2014; ^{Saldaña et al., 2015}; ^{Seminario et al., 2019}).

Thus, it is necessary to know which cultivars have the highest production of dry matter in tubers, which have an advantage for fresh consumption and processing. The higher dry matter content implies floury consistency, appreciated by the consumer, lower absorption of oil in frying and higher yield in starch, mash, powder, flakes and potato chips (^{Andrade, 1997}; ^{Ligarreto and Suarez, 2003}).

Yield of fresh tubers (t ha^-1) and dry matter of tubers (t ha^-1)

Fresh tuber yields per plant (g plant^-1) indicate 5 to 11 t ha^-1 of fresh tubers (Figure 1). These yields can be considered low to moderate if compared with the maximum experimental yields in the region (27 t ha^-1 to 28 t ha^-1). Nevertheless, they are close to the averages obtained in other studies with materials of the same group (Rojas and Seminario, 2014; ^{Seminario et al., 2016}; ^{Seminario et al., 2017}). It is important to note the high variability shown by tuber yield:

Figure 1 Yield of fresh tubers (t ha^-1) and dry matter (t ha^-1) of 15 potato cultivars of the Phureja group from Cajamarca, in eight environments. 

Thus, for example, in harvest II (dry season, 2011), the cultivar ‘Chachapoyana’ and ‘Montañera’ yielded 14.7 t ha^-1 and 14.2 t ha^-1 and the lowest yield corresponded to ‘Limeña huachuma’ with 5.3 t ha^-1. On the other hand, in harvest VI (2014), the lowest yield (3 t ha^-1) was recorded in the cultivar ‘Porpora’ and the highest yield (15 t ha^-1) corresponded to the cultivar ‘Chachapoyana-1’.

Both harvests were produced in the rainy season, which shows that the yield is a multifactorial character, in which the factors of the environment have a differentiated impact on the genotypes. It can be inferred that the components number and total weight of tubers per plant and especially the number and weight of commercial tubers per plant played a decisive role in the definition of the total yield. Overall, for all cultivars, only 33% of tubers were commercial but this proportion ranged from 12.8% to 66.9% and their average weight was 37.5 g, but ranged from 21.4 g to 75.3 g. Likewise, it can be inferred that the differences between the values of the meteorological variables of the rainy and dry seasons played a compensatory role between them. For example, the lower average and minimum temperatures of the dry season were offset, in a way, by the higher hours of sunshine of this season and the lower hours of sunshine in the rainy season were offset by the higher temperatures. It is evident that, in field crops, it is almost impossible to find optimal conditions (environmental and management) for plants, as recommended by ^{Navas and Díaz (2012)}.

The low yields are also explained by the plants only receiving a low dose of humus (5 t ha^-1) and a deep fertilization, also in low doses (15-24-14 of N-P-K). On the other hand, the backgrounds of experimental studies indicates that the Phureja potato has moderate to low yields, even when high levels of chemical fertilizer or organic fertilizer are applied (^{Pérez et al., 2008}; Muñoz and Lucero, 2008; ^{Gómez-García, 2017}). However, these yields are offset by their precocity and high price in the market.

The dry matter yields in the tubers (g plant^-1) recorded in the 15 cultivars are equivalent to yields of 1 to 2.6 t ha^-1 (Figure 1). These values are an expression of the genotype, due to the different percentage of dry matter in the tubers in each cultivar, which was determined during the previous characterization (^{Seminario et al., 2019}) and except in one case (cultivar ‘Shoga’), the best cultivars in fresh weight were not the best in dry matter (Figure 1). This information is important because it allows comparing and selecting cultivars for their efficiency in the use of external inputs and light energy and their transformation into organic matter, as an expression of net photosynthesis.

Correlations and regressions between yield-related variables

A positive, highly significant correlation was found between PTC and PTT (r= 0.9387) The regression equation was: Y= 55.33 + 1.2x. The correlations between NTT, NTC, NT and AP, with PTT were low and not significant. The high positive correlation between PTC and PTT and the lack of correlation between NT and PTT is explained by the fact that crop produced abundant small tubers, smaller than 2 cm in diameter or smaller than 5 g (non-commercial) and commercial tubers, according to the classification of ^{Pérez et al. (2008)} (commercial ≥ 2 cm in diameter or greater than 5 g), were few (30% ± 10% of the total) and were the ones that most influenced the total weight of tubers per plant (Table 5). Likewise, the lack of correlation between NTT, NT and AP with PTT indicates that their relationships are not allometric or stable (^{Gardner et al., 1985}) and on the contrary, these vary according to environmental and management conditions. Rojas and Seminario (2014) analyzing 10 cultivars from the same group also found a positive correlation (r= 0.074) between PTC and PTT.