INTRODUCTION

In Mexico, mango is grown either under irrigation (84 207 ha) or rainfed (117 256 ha) conditions. The most important cultivars are: ‘Ataulfo’ (62 136 ha), ‘Manila’ (33 573 ha), ‘Tommy Atkins’ (24 907 ha), ‘Haden’ (23 356 ha), ‘Kent’ (23 994 ha), and ‘Keitt’ (10 411 ha) (^{SIAP, 2017}). For each of these cultivars, the production area under rainfed condition is 38%, 17%, 10%, 4%, 9%, and 9%, respectively.

Besides irrigation management, mango orchards in Mexico are established in a range of soil types. In the Pacific mango producing region, Vertisol, Litosol, Cambisol, Luvisol, and Regosol soils predominate, while in the Yucatan peninsula, Rendzina and Solonchak soils are typical soils (^{IUSS, 2007}).

Leaf mineral concentrations are affected by several factors. In Nayarit, Mexico, the evolution of foliar macro and micronutrients varied among Ataulfo, Kent and Tommy Atkins cultivars, and it was influenced by leaf age, vegetative flush, and tree phenology (^{Castro-López et al., 2012}). In an organic mango orchard in Veracruz, Mexico, leaf nutrient concentrations were also affected by cultivar and tree phenology. Before flowering, N concentration was higher in ‘Tommy Atkins’, Ca in ‘Ataulfo’ and Zn, Fe, and Mn in ‘Manila Cotaxtla 2’; and previous to harvest, ‘Tommy Atkins’ had greater concentrations of K and Zn, ‘Ataulfo’ in Ca, Mg and Cu, and ‘Manila Cotaxtla 2’ showed higher concentration of N (^{Peralta-Antonio et al., 2015}).

Nutrient concentrations in mango fruit have been addressed in other studies. In mango ‘Bocado’ from three locations of the Cojedes State in Venezuela, different concentrations in the mesocarp were reported for K (20.2, 23.9 and 20 g 100 g^-1), Ca (0.93, 1.05 and 1.17 g 100 g^-1) and Zn (0.05, 0.20 and 0.28 mg 100 g^-1) (^{Milagros-Garrido et al., 2013}). In Veracruz, México, the ‘Manila’ mango grown on a Vertisol soil showed differences for nutrient concentrations among fruit tissues. The highest concentration of Mg, Fe and Mn was in the skin, K in the mesocarp, Ca in the endocarp, and N, P, Cu and Zn in the seed (^{Guzmán-Estrada et al., 1997}).

Mango harvest implies the removal of nutrients from the soil. If these nutrients are not returned to the soil, fertility and productivity of the orchard gradually diminish. For efficient management of mango nutrition, it is important to know the amount of minerals removed from the soil by fruit harvest.

In Mexico, there are few studies on nutrient removal by mango harvest. In Apatzingan, Michoacan, the following amounts of nutrients were removed from the soil by fruit of ‘Haden’ and ‘Tommy Atkins’, respectively: N (1.03 and 1.11 kg t^-1), P (0.22 and 0.24 kg t^-1), K (2.14 and 1.88 kg t^-1), Ca (0.31 and 0.21 kg t^-1), Mg (0.14 and 0.15 kg t^-1), S (0.33 and 0.28 kg t^-1), Fe (3.8 and 3.5 g t^-1), Cu (1.1 and 1.0 g t^-1), Mn (4.8 and 3.2 g t^-1), Zn (2.8 and 2.0 g t^-1), and B (1.5 and 1.6 g t^-1) (^{Mellado-Vázquez et al., 2012}). In Cotaxtla, Veracruz, ‘Manila’ mango fruit harvest removed from the soil the following amounts of nutrients: N 1.2 kg t^‑1, P 0.2 kg t^-1, K 2.0 kg t^-1, Ca 0.2 kg t^-1, Mg 0.2 kg t^-1, Fe 5.4 g t^-1, Cu 1.4 g t^-1, Mn 0.4 g t^-1, and Zn 2.1 g t^-1 (^{Guzmán-Estrada et al., 1997}). Similar amounts of N, P, K, Ca, Mg, and S were removed by cv. Zihuaman in China (^{Xiuchong et al., 2001}), ‘Kensington Pride’ in Australia (^{Huett and Dirou, 2000}), and ‘Tommy Atkins’ and ‘Keitt’ in Costa Rica (^{Fallas et al., 2010}).

Given the economic importance of mango in Mexico, it is necessary to fertilize efficiently to maximize prof its and produce high quality fruit for the export market. The objective of this research was to survey mineral concentration in fruit tissues and calculate nutrient removal by fruit harvest of the most important mango cultivars from several production regions of Mexico.

MATERIALS AND METHODS

Orchard characteristics. In 2008, 15 to 20-year old healthy trees from commercial orchards of four mango cultivars were selected. Cultivars and locations included ‘Ataulfo’ in the states of Chiapas, Nayarit, and Oaxaca; ‘Kent’ in the states of Nayarit and Sinaloa; and ‘Tommy Atkins’ in Nayarit and Campeche. In Nayarit, three orchards per cultivar were used, and in the remaining states, two orchards per cultivar. In each orchard, three bearing trees at least 150 kg of fruit were randomly selected. The cultivar, location, type of climate (^{García-Amaro, 1988}), water management and soil type (^{INEGI, 2004}; ^{IUSS, 2007}) for each of the 17 mango orchards are listed on Table 1.

Fertilization management. Orchard 1 (‘Ataulfo’ in Oaxaca): Each tree received 6 kg of 17-17-17 [nitrogen (N), phosphorus (P₂O₅), potassium (K₂O)], 2 kg urea (46% N) and three foliar sprays of 4% potassium nitrate (KNO₃); Orchard 2 (‘Ataulfo’ in Chiapas): 1.5 kg of 17‑17-17 per tree; Orchard 3 (‘Ataulfo’ in Chiapas): 1.5 kg of 17-17-17 per tree; Orchard 4 (‘Ataulfo’ in Oaxaca): 8 kg of 17-17-17 per tree and two foliar sprays of 3% KNO₃; Orchard 5 (‘Ataulfo’ in Nayarit): trees were not fertilized; Orchard 6 (‘Ataulfo’ in Nayarit): trees were not fertilized; Orchard 7 (‘Ataulfo’ in Nayarit): 0.9 kg 17-17-17 per tree; Orchard 8 (‘Kent’ in Sinaloa): trees received (kg tree^-1) 2.8 ammonium nitrate, 0.15 triple superphosphate, 0.36 potassium sulphate and 0.23 magnesium sulphate; Orchard 9 (‘Kent’ in Nayarit): trees were not fertilized; Orchard 10 (‘Kent’ in Nayarit): 0.9 kg 17-17-17 per tree; Orchard 12 (‘Kent’ in Sinaloa), Orchard 14 (‘Kent’ in Nayarit), Orchard 13 (‘Tommy Atkins’ in Campeche), Orchard 14 (‘Tommy Atkins’ in Campeche), and Orchard 15 (‘Tommy Atkins’ in Nayarit) were not fertilized; Orchard 16 (‘Tommy Atkins’ in Nayarit): 0.9 kg 17-17-17 per tree; Orchard 17 (‘Tommy Atkins’ in Nayarit): trees were not fertilized.

Fruit harvesting and processing. In 2009 and 2010, four fruit at physiological maturity were picked from each of the three experimental trees on each orchard. Sampling dates were: Chiapas and Oaxaca, April 2010; Campeche, May 2009; Nayarit, Jun 2009, and Sinaloa, July-August 2009. Harvest criterion was based on fruit size, shape, and color of fruit skin, cavity formation at the pedicel base, and visible lenticel size (^{Sergent, 1999}). At harvest, fresh fruit weight was recorded and fruit washed with tap water and double rinsed with distilled water. Each fruit was separated into skin, mesocarp, endocarp, and seed, and their fresh weight recorded. Tissues were cut in thin slices and dried in a forced air oven (Lab line 34887 Thomas Scientific, Madison, WI, USA) at 70 ºC, until constant weight. Samples were then ground in a Wiley mill (3383-L10, Thomas Scientific, Swedesboro, NJ, USA) with a 40 mesh.

Mineral analysis. A composite sample of skin, mesocarp, endocarp or seed tissues was made from each of the four fruit collected per tree. Three replicate samples per tissue (from four fruit per tree) per orchard were analyzed. Total N was determined by semi-microKjeldahl digestion, modified to include nitrates (NO₃) (^{Bremner, 1965}). P, Ca, Mg, S, Fe, Cu, Mn, and Zn were extracted by wet digestion with a mixture of HNO₃ and HClO₄ (^{Jones and Case, 1990}), and K was extracted in water (^{AOAC, 1990}). Except for P, minerals were determined by atomic absorption, using an ICE 3000 spectrometer (Thermo Scientific, Madison, Wisconsin, USA) (^{AOAC, 1990}). P was quantified by ascorbic acid method, and B was determined by calcination, using the spectrophotometric method of Azometin-H (^{Enríquez-Reyes, 1989}) using a Genesis 20 spectrophotometer (Thermo Scientif ic, Madison, WI, USA).

Nutrient removal. Nutrient removal per ton of fresh fruit was calculated by the formula, exemplified for nitrogen (N): , where NC_s= concentration of N in skin, DW_s= skin dry weight, NC_p= concentration of N in mesocarp, DW_p= mesocarp dry weight, NC_h= concentration of N in endocarp, DW_h = endocarp dry weight, NC_e= concentration of N in seed, DW_e= seed dry weight, F_t= number of fruit in one ton (obtained from the quotient 1000 kg divided by the fresh weight of the whole fruit; ^{Mellado-Vázquez et al., 2012}). The removal of macro- and micronutrients was expressed in kg×t^-1 and g×t^-1of fresh fruit, respectively.

Statistical analysis. A completely randomized design was used with three replications (trees) per orchard; each replication consisted of four fruit per tree. Analysis of variance were conducted using the GLM procedure (SAS for Windows V9). Mean separation was performed with the Waller-Duncan’s test (P ≤ 0.05).

RESULTS AND DISCUSSION

CONCLUSIONS

A survey of the fruit nutrient elements status of commercially important mango cultivars grown in several production regions of Mexico was made. Differences in the concentration of N, K, Mg, and Zn were found among cultivars and tissues. Concentration of P, S, Cu, and Mn in the skin, Ca, Cu, and Mn in the mesocarp, Ca, S, Mn, and B in endocarp, and S, Fe, and Mn in the seed were not affected by mango cultivar. Production region affected concentration of minerals in ‘Ataulfo’ fruit more than in ‘Tommy Atkins’ and ‘Kent’. Nutrient removal by mango fruit tissues was little affected in cvs. Ataulfo, Tommy Atkins and Kent. The regions with the greatest nutrient removal were Oaxaca, Campeche and Sinaloa for ‘Ataulfo’, ‘Tommy Atkins’ and ‘Kent’, respectively. This study provides baseline concentration in fruit tissues and nutrient removal values which can be used by growers, researchers, extension personnel, and mango consultants to implement best management practices for mango fertilization.