Lime Requirements Of Kenya A C I D Soils

Abstract

Four buffer methods were evaluated using Ca(OH)2~titration to pH 6.5 and CaC03~incubation to pH 6.0 and 5.5 as reference methods and using 26 soils representing major agricultural acid soils in Kenya. The buffer methods selected were Shoemaker - McLean - Pratt single buffer method (SMP-SB), Shoemaker - McLean - Pratt double buffer (SMP-DB) , Mehlich I (for crops with high lime requirements) and Mehlich II (for crops with low lime requirements). Exchange acidity (Ace), residual acidity (Acr) and total acidity (Act) were also correlated with lime requirements measured by the four buffer methods to pH 6 . 5 , 6.0 and 5 . 5 . The offeet of lime on nodulation of beans and growth of maize and beans were also studied in the greenhouse using three selected soils. The lime lequirement values measured by the reference methods to pll 6 . 5 , 6.0 and 5 . 5 were well correlated with the values measured by the four buffer methods. The lime requirement values measured by the four buffer methods were also well correlated with amounts of exchangeable aluminium, exchange acidity, residual acidity, total acidity and organic carbon. The SMP-DB method gave the highest correlation coefficient values with the reference methods, the least variations from the ideal lines, and was (vii) particularly impressive at lower pH target ( 5 . 5 ). The sensitivity of the SMP-DB and SMP-SB methods was, however, comparable for soils of high and low lime requirements, while the sensitivity of Mehlich methods was only adequate for soils of low and intermediate lime requirements. All the soil acidity components were well correlated with lime requirements measured by the buffer and reference methods to pH 6 . 5 , 6.0 and 5 . 5 . None of the buffer methods gave a significantly higher correlation coefficient for any of the acidity components than the other methods. There was an improvement in correlation coefficients when a multiple regression using exchange acidity and residual acidity was used over simple regression using exchange acidity, residual acidity or total acidity alone. In the greenhouse experiment the highest dry matter yield was obtained at pH 5.9, 5.7 and 5.8 in National Agricultural Laboratory (NAL), Mariene and Gituarriba soils respectively, while the highest nodulation of beans was obtained at pH 5 . 9 , 5.7 and 5 . 5 in NAL, Mariene and Gituamba soils, respectively. There was a decline in dry matter yield of maize and beans and nodule production of beans at pH levels beyond these. At these pH levels, exchangeable aluminium and manganese were essentially zero. It was concluded that liming of these soils to achieve pH 5 . 5 to 6.0 would be ideal. Overliming of the soils to high pH levels is likely to cause (viii) a nutrient imbalance resulting in a decline in yields. Response of maize and beans to lime on NAL soil with virtually no exchangeable aluminium indicates that exchangeable aluminium would not be a good criterion for liming these soils. Since it might not be necessary to lime these soils to pH levels higher than 6 . 0 , the performance of SMP-DB method at lower pH targets would make it the best choice. Besides its sensitivity over the entire lime requirement range, the SMP-DB method has the added advantage over the SMP-SB method of not requiring established tables. Thus any pll target can be selected.