INTRODUCTION

Lettuce (Lactuca sativa L.) is the most cultivated leafy vegetable in Brazil. The States of São Paulo and Minas Gerais are the major producers, where green-leaf let tuce enjoys consumer preference, reaching 70% of the market, followed by American lettuce (15%), flat let tuce (10%) and 5% for other commercial types (^{Sala and Costa, 2012}). In recent years, the consumption of processed lettuce has grown significantly, mainly serving fast-food restaurants, industries and hospi tals (^{Mota et al, 2012}).

As a short cycle plant, lettuce uses high quantities of mineral fertilizers to meet plant development de mands (^{Queiroz et al., 2017}), mainly for nitrogen (N). This nutrient is the most required macronutrient in horticultural crops; however, this nutrient presents great losses through leaching, volatilization, and im mobilization, requiring special care in its manage ment (^{Carvalho and Zabot, 2012}).

The frequent increase in the cost of mineral fertilizers and increasing environmental pollution as a result of inappropriate agricultural activities highlight the use of organic fertilizers as an attractive option. Organic fertilizers are attractive for all farm sizes because of nutrient cycling and organic matter addition, which may modify a soil's physical, chemical and biological attributes, thus improving soil fertility (^{Pereira et al., 2013}). These potential effects of organic fertilizers have generated new demands on researchers for as sessing the technical and economic feasibility of such fertilizers (^{Melo et al., 2008}).

Agricultural activities produce plant and animal (ma nure) waste in large quantities, which has been used in agriculture for thousands of years but requires more efficient management in the fertilization of agricultural crops, mainly in vegetable crops with a short cycle (^{Figueiredo et al., 2012}; ^{Vergel et al., 2016}). Knowledge on the dynamics of nutrient mineraliza tion must be fully understood to improve nutrient availability in the soil of short-cycle crops, avoiding immobilization or rapid mineralization of nutrients during periods of high or low demand (^{Peixoto Filho et al., 2013}; ^{Vaz et al., 2019}).

Bovine (cattle) manure (CaM) is widely used as an organic fertilizer for the production of vegetables. When this fertilizer is incorporated, it improves soil aeration and water absorption, as well as the chemi cal, physical and biological properties of the soil, generating a more balanced nutrient availability for plants (^{Cunha et al., 2012}). ^{Abreu et al. (2010)} em phasized that CaM has a water pH of 6.7, 26.9 g kg^-1 of organic matter, N, P, K, Ca, Mg and S at 11.4, 16.1, 8.3, 62, 13 and 6.7 g kg^-1; and B, Cu, Fe, Mn and Zn at 7.24, 67.8, 147, 146 and 119 mg kg^-1 (dry basis), respectively. However, this composition depends on animal feed because, when done exclusively in pas tures, these values are different than where there is supplementation with concentrates (^{Peixoto Filho et al., 2013}). In contrast, chicken manure (ChM) from intensive farming (poultry) is richer in nutrients, wa ter pH of 8.4, 31.1 g kg^-1 of organic matter (OM), N, P, K, Ca, Mg and S at 32.2; 22.6, 30, 171, 8.9 and 6.9 g kg^-1; and B, Cu, Fe, Mn and Zn at 36.5, 69.6, 272, 583 and 631 mg kg^-1 of in dry basis, respectively (^{Abreu et al., 2010}).

In addition, manure fertilizers contain high levels of cellulose, which make residue decomposition and nu trient release slowly, generating positive consequenc es in vegetable crops (^{Souza, 2007}). The contents of N, P and K in ChM are at higher concentrations than in other species of domestic animals since it contains 5 to 15% water, while other manures have 65 to 85% (^{Tedesco et al., 2008}). Using ChM as a form of fertil izer, ^{Abreu et al. (2010)} and ^{Peixoto Filho et al. (2013)} demonstrated that lettuce presented better root development and improved production of dry mass and yield when compared to other forms of organic fertilization.

The majority of studies assess the productivity of let tuce after the application of different sources of or ganic fertilization, but results vary according to the region, soil type and form of irrigation; however, only a few studies have assessed the residual effect of or ganic fertilization on the production of lettuce in the Cerrado biome. The objective of this study was to evaluate the influence of different sources and doses of organic and mineral fertilization on the production of green-leaf lettuce.

MATERIAL AND METHODS

The study was conducted in the experimental area of the horticulture sector of the Federal Institute of Triângulo Mineiro, Campus Uberaba, at the coordi nates 19°45'27" S; 47°55'36" W at an altitude of 764 m a.s.l. The experimental area was previously pre pared with two harrowings, followed by a soybean cultivation cycle and two years without crop activi ties before the experiment was installed in the winter of 2017.

The climate of the region is characterized as tropical rainy, with a cold dry winter and hot humid sum mer - Aw climate type, according to ^{Beck et al. (2018)}. The average annual precipitation and temperature are 1,600 mm and 22.6°C, respectively (^{Inmet, 2018}); however, during the evaluation period, rainfall of 20 mm (June/2017) was recorded; the total accumulated in the year was 1,995.3 mm, above the normal for the region (Fig. 1).

Figure 1 Climate variables obtained from the Meteorological Database for Teaching and Research (^{Inmet, 2018}). 

The soil of the experimental area was classified as a dystrophic red latosol (^{Santos et al., 2018}), medium texture, presenting the following values in the 0-0.2 m layer four months prior to the liming: 210, 680 and 110 g kg^-1 of clay, sand and silt, respectively, pH (CaCl2) 6.5; 20.9 mg dm^-3 of P (Mehlich-1); 161 mg dm^-3 of K⁺; 2.9 cmol_c dm^-3 of Ca²⁺; 1.5 cmol_c dm^-3 of Mg²⁺; 2 cmol_c dm^-3 of H+Al and 18 g kg^-1 of organic matter and base saturation (V) of 71%.

The sowing of the green-leaf lettuce (Vanda cultivar, Sakata^®) occurred in plastic trays (200 cells), grown under greenhouse conditions prior to field transplant.

The seedlings were transplanted when they present ed 4 to 6 final leaves that were completely expanded (7 to 10 cm height), approximately 25-30 days after sowing (DAS), and spaced 0.25 m between plants and between rows, in 4 planting rows, with 10 plants per row (40 plants per plot) in plots 1 m apart. The experiment plots were mechanically built and were 0.3 m tall, with 2.5x1 m (2.5 m²). In each plot, only 12 plants from the two central rows were evaluated.

The experiment design used the randomized block design in a 6x3 factorial scheme, with six doses of fertilizers (0, 25, 50, 100, 150 and 200% of the rec ommended fertilization for lettuce crop) and three fertilizers sources [cattle manure (CaM) and chiken (hens) manure (ChM), decomposed, on a wet basis and applied 100% at planting at the doses: CaM - 0, 12.5, 25, 50, 75, 100 Mg ha^-1; ChM - 0, 5, 10, 20, 30, 40 Mg ha^-1; mineral fertilization (MF) varying the N levels: 0, 37.5, 75, 150, 225, 300 kg ha^-1 of N as urea, plus 400 kg ha^-1 of P2O5 (superphosphate, 20% P2O5) and 60 kg ha^-1 of K₂O (KCl, 60% K₂O), as recommend ed in ^{Ribeiro et al. (1999)} - urea and KCl were fractioned: 20% at planting, 20, 20 and 40% in successive side-dressing fertilizations; all treatments had 4 rep lications (n = 72). The organic fertilizers and minerals were distributed and incorporated in the plots 5 and 3 d before seedling transplant, respectively.

The lettuce plants were irrigated daily (drippers), with a flow rate of 1.8 L h^-1 spaced every 0.5 m, for approximately 40 min, maintaining the soil close to field capacity. Infestant plants (weeds) were manu ally controlled.

The harvest occurred 67 DAS and 42 days after trans plant (DAT), when the plants presented commercial standards without evidence of flowering and with the maximum vegetative development. The plants were cut just below the basal leaves, close to the soil level, and brought to the laboratory to conduct the evaluations.

After the lettuce harvesting, the number of leaves (NL) was quantified, the plant height (PH) was mea sured with a graded ruler and the stem diameter (SD) were evaluated with caliper; the shoot fresh mass (SFM), shoot dry mass (SDM) and yield (P) were re corded using a digital scale with a maximum capacity of 20 kg. The harvested plants were dried at 65 °C with a forced air circulation oven for 72 h to deter mine the dry mass (SDM).

Four days after harvest, soil samples were collected in each plot in the 0-0.2 m soil layer for chemical analysis and determination of the residual effect of the fertilizers. The soil water pH, phosphorus (P) by Mehlich-1, potassium (K⁺), calcium (Ca²⁺), magne sium (Mg²⁺), potential acidity (H + Al), soil organic matter (SOM), sum of bases (Ca+Mg+K+Na) (SB), effective cation exchange capacity (CEC) (t = SB + Al³⁺), CEC at pH 7 (T = SB + (H+Al)), and base saturation (V = 100 SB/T) were recorded.

The results were submitted to analysis of variance us ing the F test for significance. When the results pre sented significant differences (P<0.05), the data were submitted to analysis of regression for the quantita tive factors (doses) and Tukey's test to the averages of qualitative factors (sources), both at 5% probabil ity, using the AGROESTAT statistical software.

RESULTS AND DISCUSSION

The lettuce morphological parameters when using the CaM and ChM fertilizers were significantly high er (P<0.05) than when mineral fertilizer was applied (urea). For the doses, 150 and 0% were significantly higher and lower than the other doses for all param eters, respectively, but there was no interaction be tween the sources and doses (Tab. 1).

Table 1 Agronomic evaluation of green-leaf lettuce under different fertilizers and doses. 

^ns = non significant; ** = significant at 5%, by the Tukey test (P<0.05).

ChM = chicken manure; CaM = cattle manure; NL = number of leaves; PH = plant height; SD = stem diameter; SFM = shoot fresh mass; SDM = shoot dry mass.

The use of CaM and ChM improved the soil chemi cal attributes after soil incorporation and increased all agronomic characteristics of the green-leaf lettuce. These results were due to the low natural fertility of Oxisols in the Cerrado biome. These soils generally present an accelerated rate of plant residue decom position, low levels of organic matter, Ca, Mg, P and, consequently, a low cationic exchange capacity; thus, the use of organic fertilizers can improve the attri butes of such soils.

^{Cunha et al. (2012)} and ^{Oliveira et al. (2014)} showed that the application of organic compounds improves soil aeration and water absorption, as well as the chemical, physical and biological soil attributes, gen erating a balance in the availability of nutrients to plants as a function of useful microorganisms, macro and micronutrients that are available, natural antibi otics and growth substances, such as the humic frac tions of organic matter.

The organic fertilization not only boosted productiv ity but also produced plants with a better nutritional quality (arising from the improvement of soil qual ity) than the plants grown exclusively with mineral fertilizers (^{Silva et al., 2011}). ^{Oliveira et al. (2014)} em phasized that leafy vegetables respond well to organ ic fertilization and that the constant use of mineral fertilizers causes reductions in a soil's biological ac tivity, which can affect the productive performance of crop cultures. ^{Vaz et al. (2019)} emphasized in their study that the commercial productivity of 14.5 Mg ha^-1 was observed at 157.4 kg of formulated 05-25-15 (NPK) and that doses above 200 kg ha^-1 reduced the number of total leaves and productivity.

The polynomial regression analysis demonstrated that the best fit to the averages of the analyzed vari ables was the quadratic model for NL, PH, SD, SFM, SDM and yield; the maximum values were 23.9, 17.4 and 1.6 cm, 182.4 and 11 g/plant and 29.5 Mg ha^-1, which corresponded to the doses 229.5, 181.5, 195, 213, 274.5 and 216 kg ha^-1 of N, which were equiva lent to 153, 121, 130, 142, 183 and 144% of the rec ommended dose, respectively (Fig. 2).

Figure 2 Effect of doses in the lettuce for number of leaves (A), plant height (B), stem diameter (C), shoot fresh mass (D), shoot dry mass (E), and yield (F), corresponding to the doses of 229.5 (153%), 181.5 (121%), 195.0 (130%), 213.0 (142%), 274.5 (183%) and 216.0 (144%) kg ha^-1 of N. 

According to ^{Filgueira (2013)}, increases in the NL and PH favor vegetative development and expand the photosynthetically active area of plants, raising the productive potential. This response was seen in the present study since, in the plots fertilized with CaM, the NL and HP increased by 43.7 and 25.9%, which caused an increase of 36.6% in the SD, 119.4% in the SFM, 188.23% in SDM and 118.25% in the let tuce productivity. In the plots fertilized with ChM, the values did not differ from the plots with CaM (P>0.05); the NL and PH increased by 37.5 and 20.9%, which caused an increase of 36.6% in SD, 103.6% in SFM, 176.5% in SDM and 103.2% in the productivity when compared to the plots with min eral fertilization.

The study of the production of lettuce (Vanda culti var) grown under protected environments and in an open field, with organic and mineral fertilization, us ing doses of 0, 1 and 2 times the recommendation for organic compost (3.6 kg m^-2 of NPK), mineral fertiliz er [0.02 kg m^-2 of N, via urea, 0.08 kg m^-2 of P (simple superphosphate), and 0.008 kg m^-2 K (KCl)], indicated that the open environment had the highest averages with the highest dose of organic fertilization for PH (15.6 cm), the recommended dose of mineral fertil ization for NL (8.6) and SD (1.13 cm) and a greater dose of mineral fertilization for SFM (96.8 g/plant) and productivity (36.0 Mg ha^-1) (^{Cavalheiro et al., 2015}). These values were lower than those found in the present study for NL (23.9), PH (17.4 cm), SD (1.6 cm), SFM (182.4 g/plant), and SDM (11.0 g/plant), except productivity (29.5 Mg ha^-1) (Tab. 2).

Table 2 Regressions of the effects of fertilizer doses on the plant variables. 

* = significant at 1% and ** = significant at 5%, by the Tukey test (P<0.05). NL = number of leaves; PH = plant height; SD = stem diameter; SFM = shoot fresh mass; SDM = shoot dry mass.

However, the greater organic fertilizer doses, which resulted in the greatest SFM, SDM and productivity, were 213 (142%), 274.5 (183%) and 216 (144%) kg ha^-1 of N, respectively; however, the technical and economic feasibility of such application must be evaluated for each situation (Tab. 2).

The evaluation of the mineral fertilization, ChM, CaM, earthworm humus and organic compost in the production of lettuce (Vera cultivar) indicated a high er production when using ChM (^{Abreu et al., 2010}), presenting a SFM and SDM of 542.3 g, significantly higher when compared to the 91.1 g obtained with CaM and the 233.1 g obtained with mineral fertiliza tion. ^{Peixoto Filho et al. (2013)} assessed the produc tivity of green-leaf lettuce (Cacheada cultivar) with doses of chicken, cattle and ovine manure in succes sive cultivations and observed that the ChM provid ed the best indices for the evaluated parameters: NL (15), SFM (141.4 g), SDM (5.5 g) and P (26.7 Mg ha^-1).

For the residual effect on soil chemical attributes for the type of fertilizer used, the CaM showed the high est levels of magnesium (Mg) (1.7 cmol_c dm^-3) and organic matter (21 g kg^-1). Significant interactions between the fertilizers and doses for soil water pH, P and K were detected (Tab. 3; Fig. 3). In general, it was observed that the ChM resulted in a linear increase in soil pH, P and K, which ranged from 6.5 to 6.8, 19.9 to 86.8 mg dm^-3 and from 0.2 to 0.5 cmol_c dm^-3, respectively; 5, 335 and 104% higher than the control (no fertilizer). The urea fertilizer linearly increased the content of K by 64%, which ranged from 0.4 to 0.6 cmol_c dm^-3 and linearly decreased the pH value by 7%, which ranged from 6.6 to 6.2.

Table 3 Chemical attributes of Oxisol cultivated with green-leaf lettuce under different fertilizers and doses. 

^; = non significant; * = significant at 1% and ** = significant at 5%, by the Tukey test (P<0.05). ChM = chicken manure; CaM = cattle manure.

Figure 3 Regressions of the effect of doses of urea, chicken and cattle manure fertilizers on soil chemical attributes. 

The increase in P availability, verified with the fertil ization with ChM and CaM, was due to the pres ence of this nutrient in the manure composition, to the soil pH and to the rise in the soil organic matter content. The soil pH near neutrality provided con ditions for P availability; in acidic conditions, there is a reaction of H₂PO₄ ^- with the ionic forms of iron (Fe) and aluminum (Al), forming compounds with low solubility; also, the presence of organic mat ter blocks the adsorption sites on Fe and Al oxides in the soil, decreasing the adsorption capacity of H2PO4^- (^{Novais and Smyth, 1999}).

The study of polynomial regression indicated that the soil pH in the soil fertilized with ChM and urea presented a linear adjustment, reaching values of 6.8 and 6.2, respectively, with the 200% dose. The soil P when fertilized with ChM and CaM present ed a linear and quadratic adjustment, respectively, reaching values of 86.8 mg dm^-3 with the 200% dose and 51.2 mg dm^-3 with the 129% dose. The soil K in the areas fertilized with urea and ChM presented a linear adjustment, reaching values of 0.6 and 0.5 cmol_c dm^-3 with the 200% fertilized doses, respec tively (Tab. 4).

Table 4 Regression equations of the effect of doses of fertilizers on soil chemical attributes. 

* = significant at 1%. and ** = significant at 5%, by the Tukey test (P<0.05). ChM = chicken manure; CaM = cattle manure; pH = soil water pH; P = phos phorus (Mehlich-1); K = potassium.

^{Pimentel et al. (2009)} conducted two experiments on lettuce and carrot [the first experiment was fertilized with organic compound (70% of Napier grass, 10% of remnants of cultures, 20% of cattle manure) at 0, 4.2, 8.4 and 16.8 kg/parcel (dry basis); the second experi ment was fertilized with 1/3 of the previous organic compound plus various grasses at 0, 3.4, 6.9 and 13.7 kg parcel^-1 (dry basis)], and observed that the levels of P and K increased regardless of the organic compound used. The authors reported that the soil pH in both experiments was close to neutrality (6.7 and 6.6), as well as in the present study.

According to ^{Mantovani et al. (2005)}, green-leaf let tuce (cultivar Verônica) fertilized with 0, 30, 60, 90 and 120 Mg ha^-1 of urban waste compost presented linear increases in soil pH (CaCl₂), soil organic mat ter, P, K, Ca and Mg, with increases of 19, 28, 81, 27, 178 and 100%, respectively, as compared to the control treatment (highest dose of organic fertilizer). This authors also reported that this effect on soil pH can be attributed to the presence of soluble organic anions (R-COO- and R-O-) in organic waste, which can adsorb H⁺ from the soil solution with exchange reactions, especially with Ca²⁺ ions.

The correlation of the soil chemical characteristics and biomass production of the lettuce fertilized with organic compounds indicated that the addition of or ganic compounds increased the production of lettuce dry mass and the content of soil organic matter in the soil (^{Oliveira et al., 2014}). These authors also reported that the rise of soil organic matter provided greater P, K and Ca availability, in addition to a reduced soil potential acidity, favoring plant development and in creasing productivity, similar to what was observed in the present study.

After harvest, there was still a residual of 49.2 mg dm^-3 of P in the plots fertilized with ChM (Tab. 5), indicating that there was a greater availability of this nutrient in an Oxisol (high P adsorption), which is a common soil in the Cerrado.

Table 5 Soil pH, phosphorus and potassium levels according to fertilizer. 

Means followed by different letter indicate significant differences according to Tukey test (P≤0.05). ChM = chicken manure; CaM = cattle manure.

The microbiological quality and productivity of let tuce (Vera cultivar) under mineral and diverse organic fertilization treatments (control without fertiliza tion, chemical fertilization, chicken manure, cattle manure, earthworm humus, organic compost) indi cated that the fertilization with chicken manure pro vided the highest soil pH (7.1), K (0.46 cmol_c dm^-3), Ca (6.8 cmol_c dm^-3), Mg (1.1 cmol_c dm^-3), and CTC (11.2 cmol_c dm^-3), while the fertilization with cattle manure showed the highest values for soil organic matter (51.8 g kg^-1) and P (16 mg dm^-3), results that corroborate an improvement in soil chemical attri butes when cultivated with organic fertilizers.

^{Mantovani et al. (2014)} found that high doses of P resulted in an increased growth and production of green-leaf lettuce and observed that a dose of 350 mg dm^-3, equivalent to 800 kg ha^-1 of P₂O₅, was the most appropriate dose for clay soils (420 g kg^-1 of clay). ^{Abreu et al. (2010)} demonstrated that plants from treatments with organic composts showed a high foliar P and K content, higher than those observed in a mineral fertilization treatment. The authors also pointed out that N and P are the elements that most commonly limit lettuce production as a result of low availability in the soil.

CONCLUSIONS

The fertilization with cattle manure and chicken ma nure was more efficient than the mineral fertilization for the production of green-leaf lettuce, mainly be cause of P residual effects in the soil.

The chicken manure provided a higher soil pH, phos phorus and potassium, and the cattle manure provid ed a higher magnesium, organic carbon and organic matter content.

The 144% dose of organic fertilization (wet basis), corresponding to 72 Mg ha^-1 of cattle manure and 29 Mg ha^-1 of chicken manure, resulted in the highest yield for green-leaf lettuce.