(26(2):138-146)EFFECTS OF SOIL COMPOSITION ON THE RELATIONSHIPS BETWEEN P-UPTAKE OF RICE PLANT AND SOIL TEST VALUES FOR PHOSPHORUS FERTILIZER REQUIREMENTS

從20處田間試驗所得無磷區與施磷區之水稻磷吸取量與7種方法所得磷測定值以及田間供試土壤特性因子，經用直線及曲線複迴歸統計分析結果，顯示水稻磷吸取量，在供試地區受土壤特性因子的影響甚大，而受單獨磷測定值的影響則甚小。但各法所得磷測定值與水稻磷吸取量的迥歸方程式中，若參與適當土壤特性因子如pH、Clay等，則在第一期作之下，其關係均可獲得顯著改善；而在第二期作之下，若用Al-Abbas-P(NaOH-Na2C2O4）參予CEC於其迥歸方程式中，或用N. Caro ina-P(HCI-H2SO4)參與Exchangeable Ca時，兩者對於水稻磷吸取量之校正，均獲得極顯著的改進。該兩種方法所用抽取液性質懸殊，更可顯示參與土壤特性因子對於水稻磷測定值校正之重要。 從上結果，推測舉行磷測定值田間校正時，應減少土壤特性因子的過分影響，宜參照本省農田詳測資料，劃分為若干區域，分別進行，由於同一區域內，個別農田間之土壤特性未必完全一致，仍應採用本文所述方式IV或V或其他更適合公式，予以校正，藉可提高磷測定值估計稻田磷肥需要量的可靠性。 This paper was a continuation of the previous one for the purpose of obtaining the relationships between P-uptake of rice plant and soil P-test values as affected by soil composition participating in the calibration equation. The P-uptake of rice plant including grain and straw was obtained from plots treated with and without phosphorus from twenty field trials scattered in Central Taiwan. The soil composition and 8011 P-test methods were all the same as given in the previous paper. The results indicated that P-uptake of rice plant under field condition was highly affected by soil composition but less by P-test value alone in the region. The relationships between P-uptake of rice plant and soil P-test values obtained by various procedures were generally improved, if allowance was made for the effects of soil composition, particularly with respect to pH and clay, in the case of the first crop. While in the second crop, the P-test calibration equation was remarkably improved either by adding the variable for CEC in Al-Abbas-P test value or by adding exchangeable Ca in North Carolina-P. As these two extractants (NaOH-Na2C2O4 vs HCI-H2S04) are extremely different in nature, it further confirms the importance of participating soil composition in the P-test calibration equation. From the effects of soil composition on the relations between P-uptake of rice plant and soil P-test values that have been demonstrated, it is suggested that to reduce the effects of variation in soil composition for soil testing, a region may be partitioned into areas of relatively uniform soil and developing soil tests for these more uniform areas. As the variation in soil composition from paddy to paddy is unavoidable, the P--test calibration equation for grain yield may be improved by taking the form corresponding to regression IV or V as given in the paper or other fitting model to allow for variation in soil composition