Calibration of crop responses to applied silicon (Si) serves as a basis for developing Si fertilizer recommendation guidelines. A greenhouse experiment was set up in a randomized complete block design with five replications, two sources of Si (wollastonite and slag) and four Si rates (0, 170, 340 and 680 kg ha⁻¹) to calibrate plant-available Si for growing rice in Louisiana soils. Silicon concentrations were determined in soils using seven different extraction procedures. Based on a quadratic model (p < 0.05), the estimated soil Si critical level using 0.01 M calcium chloride (CaCl₂) for Sharkey clay soil was 110 mg kg⁻¹ while for Crowley silt loam and Commerce silt loam, levels were 37 and 43 mg kg⁻¹, respectively. These results suggest that suitability of an extractant that gives the best estimate of plant-available Si could considerably depend on soil type and it is unlikely that there is a universal extractant for all soils.