Effectiveness of magnesium source and rate in the fertilization of banana grown on an Ultisol in Puerto Rico

A 33-month study was conducted to determine the response of banana grown on a clayey, mixed isohyperthermic Aquic Haplohumults to four sources and fiv e rates of magnesium. The sources of Mg were dolomitic limestone (10.4% Mg) , magnesium oxide (Fert-O-Mag, 51.5% Mg), and magnesium sulfate (Granusol, 45.0% Mg; Kieserite, 17.5% Mg). The Mg rates were 0, 45, 90, 135 and 180 kg /ha/crop. We obtained data of marketable bunch weight, Mg concentration in the third youngest leaf and the exchangeable Mg at two soil depths for two ratoon crops, R1 and R2. Results showed that sources of Mg had no significa nt effect on the three attributes studied. The rate x ratoon crop interacti on, however, was highly significant (P less than or equal to 0.01) for bunc h weight, leaf Mg concentration and the soil exchangeable Mg. Bunch weight, leaf Mg concentration and soil exchangeable Mg increased linearly with the amount of Mg applied in the first and second ratoons (R1 and R2). The high est bunch weight (28.4 kg) and the highest leaf Mg concentration (2 g/kg) w ere obtained with the application of 180 kg/ha in R2. This leaf Mg concentr ation, however, was considered sub-optimal for banana plants approaching th e flowering stage and grown on highly weathered soils, The highest applicat ion rate of 180 kg/ha also increased the soil exchangeable Mg to its highes t level of 1.6 cmol(+)/kg in R1 at a soil depth of 0 to 20 cm. Leaching and nutrient imbalance were factors that limited Mg availability. Although the banana response to Mg fertilization was linear, it was concluded that any further increase above the existing recommended rate of 55 kg/ha/year would increase production costs considerably unless a cheaper source of Mg is us ed.