Plant Material

This research was conducted during 2009-2014 in two commercial rainfed (annual average summer rainfall 1300-1450 mm) ‘Ataúlfo’ mango orchards: “Las Palmas” in the Municipality of San Blas (21º 36’ 46.1” N, 105º 11’ 19.6” W; elevation 193 m), and “El Divisadero” in the Municipality of Compostela (21º 07’ 03.0” N, 105º 11’ 04.6” W; elevation 104 m). Mango trees were 11-year-old, established in an orchard frame 8 × 8 m, and grown in a subhumid warm climate (maximum and minimum annual average temperatures above 28 and 18 °C, respectively) (^{García, 1998}). Soil in orchards were a Humic Acrisol in Las Palmas and a Haplic Feozem in El Divisadero (^{INEGI, 1999}). At the beginning of the study, difference in tree vigor was verified by means of an analysis of variance. No significant differences between orchards or groups of trees comprising the treatments for height (5.8 to 6 m) or trunk cross-sectional area 20 cm above the graft site (609-688 cm²) were detected, so a completely randomized experimental design was used. In order to minimize the effect of initial tree crop load on subsequent harvests, in 2009, 60 trees of similar size and vigor that produced at least 80 kg tree^-1 were selected throughout each orchard. Treatments were randomly assigned among the 60 trees in such a way that three groups of 20 individuals each were formed, and each group was assigned a different fertilization treatment.

Leaf and Soil Analysis

Before treatments, leaf sampling for nutrient analysis was carried out in February 2009, and every year in February 2011, 2012, and 2013. In each treatment, 10 trees were randomly selected for sampling, and from each one, 30 healthy, mature summer vegetative flush leaves (leaf blade + petiole), situated in basipetal positions 6 and 7 of non-fruiting terminal shoots were collected according to ^{Salazar-García et al. (2011)}. Leaves were washed and dried in a digital forced-air oven (Thermo Scientific model WHO 180, Madison, WI, USA) at 70 °C, until constant weight. Subsequently, the samples were ground in a stainless steel mill (IKA Mod. MF 10 Wilmington, NC, USA) with a 40 caliber mesh and analyzed for N, P, K, Ca, Mg, S, Fe, Cu, Mn, Zn, and B in Fertilab laboratory (Celaya, Guanajuato, México) which is accredited by The North American Proficiency Testing Program of the Soil Science Society of America.

In January 2010, before the application of fertilization treatments, soil sampling was performed at 0-30 cm depth. In each orchard, three trees were randomly selected per treatment and from each tree a sample composed of four sub-samples from the tree’s shaded area was taken. Another soil sampling was collected in May 2013, after the fertilizations of summer 2010, 2011 and 2012. The following characteristics were determined in the aforementioned laboratory: texture, pH (1:2 water), organic matter, inorganic N, P-Bray, K, Ca, Mg, S, Na, Fe, Cu, Mn, Zn, and B.

Based on the soil analysis results, the hydrated lime [Ca(OH)₂] and/or calcium sulfate (gypsum; CaSO₄) requirements were determined (Table 1). The criteria for establishing the appropriate amount of lime were soil acidity (pH ≤ 5.5), calcium saturation point, cation exchange capacity, bulk density, soil sampling depth, tree’s shaded area and exchangeable aluminum. For gypsum, the same criteria as those used for lime, differing only in soil acidity (pH ≥ 5.5).

Fertilization Treatments

They were calculated independently for each orchard and considered: fruit nutrient’s needed to produce 20 Mg ha^-1 (^{Salazar-García et al., 2010}), available nutrients and other chemical characteristics from soil analyses, leaf nutrients concentration which were transformed to a nutrient balance index (^{Salazar-García et al., 2011}), fertilization efficiency for N, P, K, Ca, Mg, Fe, Mn, Zn, and B (40, 20, 60, 80, 60, 5, 5, 10, and 5%, respectively), tree’s shaded area, and periods of greatest root growth (July and September). During 2010, 2011, 2012 and 2013, three fertilization doses were applied: 1) “Normal” for a yield of 20 Mg ha^‑1; 2) “High”, 1.5 times the normal dose (Table 1); 3) Control, an annual application of 17 N, 7.4 P, 14.1 K as 17-17-17 commercial mixture at a dose of 3 kg tree^-1.

Fertilizers were distributed around the tree in a 50 cm wide band and 15-20 cm deep, initially at 1.5 m from the trunk and then progressively farther way from the trunk in each subsequent application. Every year, the fertilizer program was slightly modified according to the soil and leaf analyses.

Yield

Fruit yield was evaluated in each tree the first week of June 2011 to 2014. Fruit size (C) was classified according to regional mango packinghouses: C22 (196-220 g), C20 (221-250 g), C18 (251-283 g), C16 (284-315 g), C14 (316-365 g), and C12 (> 366 g).

Alternate Bearing Index

To determine treatment effects on bearing regularity, the alternate bearing index (ABI) was calculated using the equation suggested by ^{Pearce and Dobersek-Urbanc (1967)}:

ABI = (difference in yield between the previous and present year) / (sum of yields of both years) × 100. ABI ranges from 0 (no alternate bearing) to 1 (complete alternate bearing).

Data Analysis

For cumulative yield and fruit size (2011+2012+2013+2014), a 2 × 3 factorial arrangement (orchards × treatments) was used. In the case of leaf nutrient concentrations, the year (2011, 2012 and 2013) and fertilization treatments (Normal dose, High dose and Control) effect was analyzed. For fruit yield and size, 20 replicates (trees) per treatment and per orchard were considered. In the case of leaf nutrient concentrations, 10 samples per treatment per orchard were used. Analyses of variance were carried out with the ^{SAS software (2008)}. Means comparison was performed with the Waller-Duncan test (P ≤ 0.05).

Soil Characteristics

In 2013, three years after the start of the research, soil chemical properties varied among treatments. In El Divisadero orchard, the pH dropped from 4.7 (Control) to 4.3 and 4.2 (Normal and High doses, respectively). The same occurred in Las Palmas, where it dropped from 6.6 (Control) to 5.3 and 5.0 (Normal and High doses, respectively) (Table 2). The pH was more acid but within appropriate limits for mango (^{Castellanos et al., 2000}). Overall, fertilization increased the levels of most nutrients in the soil (K, Ca, Fe, Zn, Mn, and B), improving and/or maintaining its fertility (Table 2). In some cases, increases in the levels of some soil nutrients, due to fertilization, promoted higher concentration of some nutrients in leaves, specifically of Mg, Fe and B. However, it is not common to find a relationship between the concentration of these nutrients in the soil and leaves of fruit trees (^{Salazar-García, 2002}).

Leaf Elemental Concentrations

The average of leaf nutrient concentrations of all treatments varied with the year of evaluation. The highest concentrations of N, Mg, and Fe were recorded in 2011, a year after the start of the study, but for 2012, concentrations of K, S, Cu, Mn, Zn, and B were higher than those of 2011. In 2013, P concentrations were higher than in 2011 and 2012, and the Mn concentration showed no differences compared to 2012 but was higher than in 2011 (Table 3). In fruit trees grown under rainfed conditions, leaf nutrient concentrations usually show a slow change in response to soil fertilization. This was the case in other species, such as ‘Hass’ avocado in Nayarit, Mexico, where changes in leaf nutrient concentrations occurred after several years of fertilization (^{Salazar-García et al., 2008}; Cossio-Vargas et al., 2009¹; ^{Salazar-García et al., 2009}).

Fertilization treatments did not change leaf concentrations of N, P, K, Ca, S, Cu, Mn and Zn (Table 3). Mg showed its lowest concentration in the High dose treatment. For Fe, the High dose reflected the highest concentration of Fe (42.0 mg kg^-1) compared to the Control (33.9 mg kg^-1), although it was similar to the Normal dose (38.8 mg kg^-1). B concentration was higher with the Normal dose (24.7 mg kg^-1) than with the other two treatments.

Total Cumulative Yield and Fruit Size

The highest cumulative yield was obtained in Las Palmas orchard (306.9 kg tree^-1), of which 89.6% was small-sized fruit (C22: 274.9 kg tree^‑1) (Table 4). In El Divisadero, the cumulative yield was lower (179.4 kg tree^-1), but 74.5% of this yield (133.6 kg tree^‑1) corresponded to the C22 size class. The lower yield (7.0 Mg ha^-1 year¹) in El Divisadero orchard was possibly influenced by the high production of “baby” mango (parthenocarpic fruit). ^{Salazar-García et al. (2016)} evaluated the presence of such fruit in this orchard, reaching 91% at fruit-set (March) and 63.7% at harvest (May). In Las Palmas orchard, these proportions were lower, 72.2% (fruit-set) and 40.1% (harvest).

Our results provide evidence that fertilization treatments modified fruit yield and size. The Normal dose produced the highest cumulative yield (292.8 kg tree^‑1), followed by the High dose (254.1 kg tree^‑1) and the Control (184.4 kg tree^-1) (Table 4). The highest fruit yield was obtained with the Normal dose treatment (73.2 kg tree^-1).

The average yield 2011-2014 for cv. Ataúlfo in the municipalities of Compostela and San Blas, where the experimental orchards were located, was 8.3 Mg ha^-1 (^{SIAP-SAGARPA, 2014}); this is equivalent to 53.2 kg tree^-1 (with a plantation of 156 trees ha^-1). For both orchards, the Normal dose produced 11.4 Mg ha^-1 (73.2 kg tree^-1), estimating a 38% increase.

Regardless the mango-producing region (China, Egypt, India, Mexico, Pakistan, Thailand, Taiwan, others), mango fertilization is commonly based on N, P and K (^{Suriyapananont, 1992}; ^{Sharma et al., 2000}; ^{Shu et al., 2000}; ^{Shakeel et al., 2001}; ^{Xiuchong et al., 2001}; ^{Vázquez-Valdivia et al., 2005}) and sometimes only K (^{Taha et al., 2014}). The applied doses of N, P and K per tree in these countries varied from 340 to 1500 g, 54.5 to 654 g and 249 to 623 g, respectively. For these same nutrients, doses that increased fruit yield in the present study were 509-608 g N, 21‑206 g P, and 132‑582 g K. The observed differences in relation to other producing regions support the use of the SSF approach to increase yield for mango in Nayarit. Besides, unlike the aforementioned reports, this study included applications of Ca, Mg, Fe, Mn, Zn and B. The addition of such nutrients could influence the results here reported; however, their specific effect was not the scope of this study.

In the fertilization treatments, the increased yield was accompanied by increased production of small fruit (C22) (Table 4). However, the sum of the intermediate (C20+C18+C16) and large (C14+C12) sizes was higher with the High (44.2 kg tree^-1) and Normal (43.6 kg tree^-1) doses, compared to the Control (28.8 kg tree^-1). For the 2011-2014 harvests, the ABI only showed differences between orchards (Table 4). Its values were 0.11 and 0.18 for Las Palmas and El Divisadero, respectively.

The SSF approach proved to increase fruit yield and size of ‘Ataúlfo’ mango. The highest total yield and C22 (196-220 g) and C20 (221-250 g) intermediate fruit sizes were obtained with the Normal fertilization dose. This treatment was based on the application per tree of 509-608 g N, 21-206 g P, 132-582 g K, 19‑234 g Mg, 6.5-18 g Fe, 6-46 g Mn, 2-6 g Zn and 3-13 g B, depending on the orchard. We found that the High fertilization dose was beyond tree needs, resulting in lower production composed of smaller fruit size, compared to the Normal dose. The results herein are now available for ‘Ataúlfo’ mango growers in these regions and the SSF approach used are worth to be tested in other producing regions.

Economic Analysis

The most profitable fertilization treatment was the Normal dose (Table 5). In addition to improving the fruit size, this treatment increased yield by 4.3 Mg ha^‑1 and had a net economic benefit of US$ 537.7 per hectare, each year, compared to the Control. As we can see, yield and fruit size showed a significant response in the short term, confirming that fertilization trials in fruit trees should be conducted over several seasons. The short-term response is important to mango growers to convince them that they do not have to wait several years to have an economic benefit from proper fertilization management.