In the Andean region of Peru, potato (Solanum tuberosum L.) is the main source of food and economic income, and its cultivation is conducted mainly by family farmers located between 3 500 and 4 200 masl, who still preserve ancestral technology, based on reduced soil mechanization and restriction on the use of agrochemicals (^{Caycho et al., 2009}). Under these management conditions, they obtain yields of 5 t ha^-1, a value lower than the national average of 16.1 t ha^-1 (^{FAO, 2020}), so it is urgent to develop technologies according to their reality, which promote competitiveness and sustainability in the use of resources.

In this sense, among the different existing alternatives, the incorporation of different organic sources is an interesting option to favor the increase of yield and preserve their ancestral technology, in addition to improving the physical, chemical and biological properties of the soil (^{Alaluna and Villagarcía, 2000}; ^{Bolo et al., 2020}). Thus, various studies show positive results in potato cultivation, such as those evidenced by ^{Ríos et al. (2015}) with the incorporation of island guano (1 and 2 t ha^-1), by obtaining yields (46.7 and 55.9 t ha^-1, respectively) higher than the control (24.9 t ha^-1), ^{Peñaloza et al. (2019}) with chicken manure (2, 3 and 4 t ha^-1), by observing increasing yields and finding no significant differences for the number of tubers and stems per plant, ^{Villagaray et al (2021}) with sheep manure (0; 0.25; 0.5; 1 and 1.50 kg plant^-1), by reaching the highest yield (20.31 t ha^-1) with the highest dose (1.5 kg plant^-1), significantly surpassing the control (11.89 t ha^-1). In this context, the present study aimed to evaluate the effect of three organic sources with three levels of application on the production of potato in the Andean region of Peru.

The research was carried out in a family property, located at 3 425 masl, in the province of Huari, Ancash-Peru region, located at the coordinates 9° 22’ 10” SL and 77° 01’ 04” WL, from October 2019 to March 2020. The climate of the region is classified as Dwb according to the Köppen classification, characterized by cold or temperate and dry winters with cool and rainy summers. The soil with loam texture has a pH of 6.64, organic matter of 1.248%, nitrogen of 0.062%, phosphorus of 13 mg kg^-1 and potassium of 78 mg kg^-1. Prior to the installation of the experiment, the analyses of the organic sources were carried out and the results are shown in Table 1.

The statistical design of randomized complete blocks was implemented with a 3 x 3 factorial arrangement plus one additional treatment (control) and in four blocks. Two factors were studied: organic sources (sheep manure, chicken manure and island guano) and three levels of application (5, 10 and 15 t ha^-1). The additional treatment (control) did not receive any type of fertilization. The area of the experimental unit was 10.94 m². The preparation of the land was carried out with a team, the weeding was manual, and the presence of pests and diseases was not noticeable. The experiment was conducted under rainfed conditions.

In five plants of the central furrow per each experimental unit, the number of stems and tubers per plant, total yield and the category extra (tubers with weights between 321 and 710 g) were evaluated. The data obtained for the variables evaluated were subjected to the analysis of variance by the F test (p< 0.05) and the means were compared with the Tukey test (p< 0.05). The data were analyzed with the statistical software R.

According to the analysis of variance, Table 2, there is interaction between organic sources and levels of application for number of stems per plant, yield and category extra. For the number of tubers per plant, significant differences have only been shown between organic sources, but not between levels of application. Likewise, for the set of variables evaluated, there have been significant differences between the control treatment and the mean of the factorial.

Comparing the control treatment and the factorial treatment, Table 3, the application of organic sources promoted a greater number of stems, tubers per plant and yield.

With regard to the interaction between organic sources and levels of application for the number of stems per plant, Table 4 shows that, at the three levels of application, island guano and chicken manure were significantly superior to sheep manure; while within each source, the application of 10 and 15 t ha^-1 of chicken manure and sheep manure promoted greater stem formation, a response that was not observed in island guano. The increase in the number of stems per plant can be explained by the increase in the availability of nutrients in the soil (^{Ghemam et al., 2016}), thus favoring the growth of new dormant shoots in the seed tuber.

For the number of tubers per plant, Table 5, the highest values were obtained with the application of island guano, which were significantly higher than the other organic sources.

Evaluating the interaction between organic sources and levels of application for yield and category extra, Tables 6 and 7 show that, at the levels of 5 and 10 t ha^-1, island guano surpassed the other organic sources; while within organic sources, in island guano, the highest yield was obtained with the application of 5 t ha^-1, while in chicken manure and sheep manure, yields increased as the amount applied increased.

In general, it can be seen that the application of 5 t ha^-1 of island guano has promoted better results for the variables under study. This result is explainable because this source has a high concentration of nutrients compared to the other sources; however, it shows as a disadvantage the high content of salts (Table 1) which raise the electrical conductivity of the soil and end up negatively affecting the yields, as observed in the application levels of 10 and 15 t ha^-1. The fact that island guano has stood out above the other organic sources is due to the immediate availability of nutrients for the crop (Minagri, 2018).

It is possible that the high content of nutrients, mainly nitrogen, has favored to keep the leaf area functional for a longer time, directly causing a greater photosynthetic activity and consequently, a greater transfer of photosynthates to the storage organs, a situation that has not been observed with the application of sheep manure, which due to its low coefficient of mineralization (^{Gaham et al., 2016}) obtained the lowest yield. Chicken manure stood out in second place, obtaining better yield characteristics with the application of 15 t ha^-1, this result being explainable by the lower nutrient content compared to island guano (Table 1).

In the case of sheep manure, the lower values obtained for the set of characteristics evaluated can be explained by the low nutrient content in its composition (Table 2) and the resistance it presents to decomposition due to the presence of a membrane that covers it (^{Célio et al., 2012}).