Nutrient concentrations in trifoliate orange as affected by lime and gypsum

Use of trifoliate orange [Poncirus trifoliata (L.) Raf.] as a rootstock has intensified in recent years in Brazil. Objectives of this study were to evaluate the effects of lime and agricultural gypsum on concentration of nutrients in trifoliate orange. Seedlings of trifoliate orange were grown in PVC pipe columns, presenting 15 cm in diameter and 35 cm in length. The columns were sectioned in two rings: the upper ring, 15 cm high, and the lower ring, 20 cm high. The factorial scheme (2×4)+1 was used, being two liming treatments and four agricultural gypsum doses (carried out only in the soil of the upper ring), and an additional treatment (with liming carried out in the soil of both upper and lower ring). Liming increased Mg and S concentrations in roots of the superficial soil layer (0-15 cm). Ca concentration was higher in roots of both superficial (0-15 cm) and subsuperficial (15-35 cm) layers. Gypsum without liming resulted in higher N, K, and Mn concentrations and lower Mg concentration in roots of the soil subsurface layer. Ca and S concentrations in root of the soil superficial layer were higher with gypsum. In plant shoot, the concentrations of K, Ca, S and Cu were higher with liming, and concentrations of Ca and P were higher and lower, respectively, with gypsum application.

irrigation, under greenhouse conditions of the CEDETEG Campus of the State University of Mid-West (UNICENTRO), Guarapuava, Paraná State, Brazil.
The PVC pipe columns, 15 cm in diameter and 35 cm in length, were composed of two rings: upper ring, 15 cm high (surface layer, 0-15 cm deep), and lower ring, 20 cm high (subsurface layer, 15-35 cm deep). To avoid contact between roots and inner wall of the PVC pipes, a coating of kaolin on the inner face of the PVC pipes was used before filling them with the soil. The upper and lower rings of each soil column were joined with "silver tape" adhesive. Both planting and topdressing fertilizations were carried out only in the soil of the upper ring, in order to provide macronutrients (except Ca and S, which were provided only with liming and gypsum dose treatments) and micronutrients recommended (Malavolta, Vitti, & Oliveira, 1997), followed by the planting seedlings.
The experimental design was of randomized blocks (DBC), with four replicates, in a factorial scheme (2×4)+1, with two liming treatments (without and with applications of calcium and magnesium carbonates, carried out only in soil of upper ring), combined with four gypsum dose treatments (0.0, 2.5, 5.0 and 10.0 g kg -1 of agricultural gypsum, also applied only in soil of upper ring). An additional treatment (with applications of calcium and magnesium carbonates, carried out in soil of both rings, upper and lower) was carried out to evaluate the potential of agricultural gypsum, applied on soil surface, in improving the conditions of soil subsurface, comparing with the soil subsurface corrected with liming. Each experimental plot consisted of a PVC pipe column with two plants of trifoliate orange.
Based on the results of the chemical analysis of the soil, in treatments that received liming, the application of calcium (MgCO3) and magnesium (MgCO3) carbonates was made aiming to raise the saturation of bases to 70% and to adjust Ca/Mg ratio to 4/1, with a dose of 3.84 and 1.87 g kg -1 of CaCO3 and MgCO3, respectively (both reagent p.a.). In treatments with agricultural gypsum, the doses were 0.0, 2.5, 5.0 and 10.0 g kg -1 (dry weight basis), Research, Society and Development, v. 9, n. 10, e7449109096, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i10.9096 6 simulating the approximate doses of 0, 5, 10 and 20 Mg ha -1 gypsum in the field. The agricultural gypsum was obtained from local commerce of Guarapuava, Paraná State, Brasil.
Plant shoot and roots in upper and lower rings were harvested in each separately, on April 10, 2018, at 153 DAT. Roots were previously washed in running water and deionized water to separate the soil particles, dried on paper towels and, subsequently, in an oven with forced air circulation at 58 °C for four days. The data obtained were submitted to analysis of variance (ANOVA, P ≤ 0.05), considering the design in DBC in a scheme (2×4)+1, using software R version 3.5.1. (R Development Core Team 2016). In ANOVA, degrees of freedom were adjusted according to treatments of liming, gypsum doses and interaction between liming × gypsum doses, and a contrast between additional treatment and factorial (2×4) experiment average. When there was a significant effect of gypsum dose treatments, linear and quadratic regression analysis was performed (P ≤ 0.05). The normality of data was tested by Shapiro-Wilk test (P ≤ 0.05). It may be seen in Figure 1 that nitrogen (N) ( Figure 1A) and phosphorus (P) ( Figure   1B) concentrations of the root had similar behaviors. In roots collected from 0-15 cm depth, N and P concentrations were not significantly influenced by the liming and gypsum dose treatments, presenting an average of 22.76 and 2.91 g kg -1 of N and P, respectively. In this soil layer there was also no significant difference between additional treatment and factorial experiment average. However, in roots collected from 15-35 depth, N (P ≤ 0.001) and P (P ≤ 0.05) concentrations were affected by the interaction between liming × gypsum doses. Without liming, N and P concentrations responded to gypsum doses in a quadratic form, decreasing at the lowest gypsum doses (0 and 5 g kg -1 ) but with a strong increase (126 and 82%, respectively) of values between the two highest gypsum doses (5 and 10 g kg -1 ). However, with liming, there was no significant variation in N and P concentrations of the root with gypsum doses. Research, Society and Development, v. 9, n. 10, e7449109096, 2020 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v9i10.9096  Figures 1A, 1B, 1C, 1D, 1E, 1F, 1G, 1H and 1I corresponding to root concentrations of N, P, K, Ca, Mg, S, Mn, Zn and Cu, respectively. *** , ** , * and ns = P ≤ 0.001, P ≤ 0.01, P ≤ 0.05 e P > 0.05, respectively, for linear and quadratic regression equations for gypsum dose response. this case, without liming, it was found that Mg concentration was high with highest gypsum dose.

Results and Discussion
In roots of the 15-35 cm depth, there was no significant effect of liming on Mg concentration of the root. Gypsum doses resulted in a decrease (P ≤ 0.01) of Mg concentration, especially between 5 and 10 g kg -1 of gypsum (the linear equation was statistically significant, but with a low determination coefficient). In additional treatment, there was a significant increase in Mg concentration of the root of both soil depths, compared to factorial group average.
In general, the effect of liming on Mg concentration of the roots was small. A more pronounced effect was expected, due to addition of MgCO3 to soil of the upper ring with liming. In a study by Fochesato, Souza, Schäfer and Maciel (2006), it was found that trifoliate orange naturally has a lower potential for Mg 2+  In roots of the 15-35 cm depth there was also interaction (P ≤ 0.01) of the treatments on S concentration. As in soil surface layer, gypsum was more efficient in raising the S concentration when liming was not carried out, with an increase of 84% between doses 0 and  Fig. 3A, 3B, 3C, 3D, 3E, 3F, 3G, 3H and 3I corresponding to root concentrations of N, P, K, Ca, Mg, S, Fe, Zn and Cu, respectively. ** , * and ns = P ≤ 0.01, P ≤ 0.05 e P > 0.05, respectively, for linear and quadratic regression equations for gypsum dose response.
It may be seen in Figure 2 that nitrogen (N) concentration of the shoot (Figure 2A) was not significantly affected by the liming and gypsum dose treatments, exhibiting about 30.03 g kg -1 of N in shoot. For phosphorus (P) concentration of the shoot ( Figure 2B) there was no significant effect of liming, but the two highest gypsum doses (5 and 10 g kg -1 ) resulted in 11% decreased in P concentration (P ≤ 0.05), with values varying from 2.24 to 2.0 g kg -1 . For both N and P concentrations, there was no significant difference between