Description of the study area

The present study was developed in 2017 under rainfed conditions in the community of Ojo de Agua, in the municipality of Zinacantepec, State of Mexico, located at 99° 46’ 13’’ north latitude and 19° 12’ 58’’ west longitude, to 3 037 masl. In this region the temperate sub-humid climate predominates, with an average temperature of 28 °C in summer and 5 °C in winter. The Andosol soil is a migajon type, with clay-sandy to silty texture, dark gray-brown in color and with a depth of 4 to 9 m, very fertile (^{IGECEM, 2014}).

Genetic material

The potato cultivars Fianna, Ágata, Rosita and Lucero, provided by producers from the Toluca-Zinacantepec region (Table 1) were used.

Design and size of the experimental plot

The 16 treatments that were formed by combining the levels of four cultivars and four types of fertilization (Table 2) were evaluated in the field in a randomized complete block design in a 4x4 factorial arrangement, with three replications. The experimental plot had three rows of 4 m long and 0.9 m wide, a tuber was deposited every 20 cm. The useful experimental unit was the central row.

Physical-chemical analysis of the soil where the experiment was made

The soils that predominate in the community of Ojo de Agua, in the municipality of Zinacantepec, State of Mexico have the characteristics shown in Table 3.

Composition of chicken manure

His physical and chemical analysis was carried out in a laboratory of the Faculty of Agricultural Sciences (Table 4).

Development of the field experiment

A fallow and a harrow pass were made on June 21, 2017. The 90 cm wide furrows were formed with a yoke of horses. On June 1st the tubers were immersed for five minutes in water+Celeste (fludioxanil, control of Rizoctonia solani)+Bactrol (streptomycin, control of Erwinia carotovora) and Tecto 60 (Thiabendazole, control of Fusarium spp.). The manual seeding was carried out on July 1st and consisted of depositing the tubers at a depth of 15 cm and at a distance between them of 20 cm, then covered using the yoke. Fertilization was carried out in two stages: when planted, 2, 3 and 4 t ha^-1 of chicken manure and 154 N-356 P-60 K were applied, in the latter they were used for 600 kg ha^-1 of diammonium phosphate (18 N-46 P- 00 K), 100 kg of urea, 400 kg of simple calcium phosphate superphosphate and 100 kg of potassium chloride.

On September 2, 200 kg ha^-1 of urea, 200 kg of potassium chloride and 400 kg of simple calcium phosphate were applied per ha. Post-emergence weed control was done on July 20 with Titus CE (Rimsulfurom), applying in a band of 75 to 100 g ha^-1. The two hives were carried out on August 12 and September 2, to control weeds and prevent pests such as blind hen (Phyllophaga sp.). To prevent damage by nematodes (Globodera rostochiensis and Meloidogyne sp.) CYREN 5% G (Chlorpyrifos ethyl) was applied, mixed with fertilizers.

To combat whitefly (Bemisia tobacco and Trialeurodes vaporariorum), Paratrioza (Bactericera cockerelli) and cutworms (Agrotis sp., Trichoplusia ni, Spodeptera sp.), alternate sprays of insecticides were carried out. To prevent and control diseases such as late blight (Phytophthora infestans), early blight (Alternaria solani) and black crust (Rizoctonia solani), 12 applications were made with different systemic and contact fungicides. The foliage was cut 100 days after sowing and the harvest was made 20 days after the foliage was removed.

In 10 plants of the useful plot plant height was recorded (AP, in cm, from the base of the main stem to the apex), stems per plant, tubers per plant (NT), tuber weight per plant (PT, in g), foliage weight in fresh (PFF, in g), harvest index (IC, quotient between agronomic and biological performance), tuber length and diameter (LT and DT, in cm), and yield per hectare (Rha, in t ha^-1).

Statistical analysis

An analysis of variance was performed and when the F values were significant, a comparison of treatment means was made with the Tukey test (p= 0.01). Both techniques were described by ^{Martínez (1988)}. The analysis of principal components, also called biplot treatment x variable, was described by ^{Sánchez (1995)}; ^{Pérez et al. (2010)}.

Variance analysis

In the present work it was observed that the experimental area was homogeneous (Table 5), because the effects between repetitions were only significant in fresh foliage weight (PFF) and harvest index (IC). The effects caused by chicken manure (factor A) were only detected in tuber weights per plant (PT), PFF and yield per hectare (Rha), which coincides with that published by ^{Seminario et al. (2017)}. Among cultivars (factor B) in plant height (AP), stems per plant (NTP), number of tubers (NT), PT, PFF, IC and tuber length (LT), there were significant effects (p= 0.01). The AxB interaction was significant in tuber weight (p= 0.01) and PFF (p= 0.05). These results are similar to those reported by ^{Pérez et al. (2009)}; ^{Pérez et al. (2010)}; ^{Seminario et al. (2017)}. The coefficients of variation changed from 6.7 (LT) to 31.48% (Rha).

Comparison of chicken stockings (factor A)

The effect of the chicken manure was not significant in plant height (AP), stems per plant (NTP), tubers per plant (NT), harvest index (IC), diameter (DT) and tuber length (LT) (Table 6).

For tuber weight per plant (734.1 g, PT), fresh weight of the foliage (400.71g, PFF) and yield per hectare (22.61 t ha^-1, Rha) there was better response when applying 4 t ha^-1 of chicken manure and this dose differed statistically from minor and chemical fertilization. The latter and the highest dose differed by 8.38 t ha^-1 (Table 6). ^{Carter et al. (2004)} noted that the increase in yield is due to a higher production of tubers, which is favored by a higher content of organic matter, phosphorus and cation exchange available in chicken manure.

^{Lima et al. (2000)} when evaluating organic fertilizers and chemical fertilizers reported that for each ton of chicken manure applied there was an increase in tuber yield of 1.468 t over 40.752 t. ^{Zamora and Rodríguez (2008)} estimated 32 t ha^-1 when fertilizing with organic fertilizers and ^{Rodríguez and Ortuño (2007)} recorded from 11.53 to 15.93 t ha^-1 with application of arbuscular mycorrhizae in interaction with organic fertilizers. ^{Corzo and Moreno (2003)} mentioned that the main element responsible for the mobilization of the starch from the leaves towards the tuber is potassium, in such a way that a good availability of this nutrient is decisive for obtaining a high yield and better quality.

Comparison of means between cultivars (Factor B)

In Fianna, the highest plant height (41.89 cm) was recorded. These results are contrasting with those of ^{Zamora and Rodríguez (2008)} and ^{Buchanan (1993)} (Table 7). Rosita, Creole variety and Ágata, improved variety, produced two stems per plant and differed statistically from Fianna and Lucero. These results are similar to those of ^{Devaux et al. (1997)}. Number of stems is closely related to tuber size and population density (^{Van der Zaag et al., 1990}; ^{Masarirambi et al., 2012}). Rosita produced 17.15 tubers per plant. This characteristic is genetically controlled and depends on the number of stems per plant (^{Estrada, 2000}), in this study this variety obtained more stems per plant, but shorter tuber length. The results are contrasting with those of ^{Luna et al. (2015)}, who reported 5.91 for Rosita and 9.35 for Libertad. ^{Pérez et al. (2009)} reported quantified from 3.75 to 5.23 in 17 potato genotypes (Table 7).

The tuber weight per plant (PT) in Ágata and Rosita was 655.04 and 651.49 g, respectively, and both differed statistically from Lucero and Fianna. This behavior could be explained by the genetic differences between cultivars, especially those related to their biological cycle; Rosita is late and the other three are precocious. ^{Pérez et al. (2010)} reported values of 627.38 to 204.13 g and ^{Seminario et al. (2017)} from 68.3 to 987.3 g (Table 7). Rosita had the highest fresh weight of foliage (548.74 g) and statistically differed from the rest. These results contrast with those of ^{Pérez et al. (2009)}, who registered from 111 to 280.2 g.

Fianna, Lucero and Ágata had harvest rates (66, 70 and 71%) statistically higher than Rosita (50%). These results are similar to those obtained by ^{Rojas and Seminario (2014}, between 53 and 77%). Currently, plant breeders have given greater importance to the increase of the harvest index since a greater partition of biomass to the tubers will contribute to a higher yield per hectare. ^{Pérez et al. (2009)} reported values of 38 to 72.22%.

The previous results show that potatoes are very efficient in the transformation of inputs, water and light to green or dry matter and specifically, to the production of tubers per plant and per hectare. In yield per hectare the four cultivars showed a similar behavior. These yields were obtained under temporary conditions but were lower than the national average (24.40 t ha^-1) and the state average (29.71 t ha^-1). In the High Sierras and Valleys of Mexico, the yield potential is lower in relation to those obtained in the North of the country and in the Bajio region (Table 7).

The four cultivars were statistically equal in tuber diameter, results similar to those reported by ^{Luna Ágata, (2015)}. Fianna (8.83 cm) had a longer tuber length, but it was similar to that of Ágata (8.50 cm) and both differed statistically from Rosita (6.87 cm) and Lucero (7.1 cm). This characteristic is very important for the fresh market and for the agroindustry; in the first, dimensions between 4.5 and 5 cm are required and for fried foods they must be greater than 6.5 cm. These values agree with those obtained in the present work and Ágata and Lucero are recommended for industrial use, Fianna is dual purpose and Rosita is preferred for fresh consumption. With the application of the scale proposed by ^{Macías et al. (2006)}, the tubers obtained in the present study were classified in first category.

Principal component analysis

Major components 1 (43.21%) and 2 (25.98) explained 69.19% of the original total variability (Figure 1). These percentages are desirable to reliably interpret the approximate correlations observed in the biplot, as suggested by ^{Sánchez (1995)}; ^{Pérez et al. (2014)}, among others, the closest approximation between cultivars and variables indicates a close relationship between them.

Figure 1 Biplot between treatments (cultivar + fertilizer in number) and agronomic variables (in letter) based on components 1 and 2. 

Tuber yield correlated positively and significantly with most of the variables. CP1 was determined mainly by tuber weight (PT), stems per plant (NTP), fresh foliage weight (PFF), tuber yield (Rha) and tubers per plant (NTP). In other studies, similar results have been observed (^{Rousselle et al., 1999}; ^{Estrada, 2000}; ^{Caceres, 1971}). In Quadrant 1 it was observed that treatment 15 (cv. Rosita in 4 t ha^-1 of chicken manure) showed a positive association with yield and with its main performance components (^{Romero et al., 2000}). This fact partially corroborated the results that were previously shown when their arithmetic averages were compared with the Tukey test. The chicken manure is one of the organic fertilizers that help to improve the biological properties of the soil (^{Estrada et al., 2004}; ^{Orozco and Thienhaus, 1997}), but also contributes to the increase in the acidity of the soils, particularly those of volcanic origin, as those that predominate in the study area. The adaptation of the potatoes to the High Valleys of Central Mexico has made this species an important alternative for producers.