Published online by Cambridge University Press: 27 March 2009
Twenty-five soils, including some subsoils, with widely differing properties were cropped with perennial ryegrass in the glasshouse, and measures of Mg availability in the soils were related to the Mg concentration in the plants.
No single measure satisfactorily characterized Mg availability. The ion activity ratio, √aMg/√aca+Mg when the soil neither gained nor lost Mg on equilibration with 0–02 N-CaC12, was more highly correlated (r = 0·;75***) with Mg concentration in the ryegrass than was the best of the capacity measures tested (CaCl2-extractable Mg, r = 0·68***), particularly if the potential buffering capacity (PBCMg) of the soils was also taken into account, in which case r = 0·83***.
Exchangeable Mg (r = 0·67***) was a poorer index of Mg availability than percentage Mg saturation (r = 0·73***) or exchangeable Mg expressed as a percentage of the total exchangeable bases (K, Na, Ca and Mg) in the soil (r = 0·81***). This latter quantity was better than the activity ratio as a predictor of Mg availability (r = 0·81*** compared with r = 0·75***).
A significant proportion of the variation in Mg concentration in the ryegrass could be attributed to the K content of the soils, and closer correlations were obtained when K was included in the composite activity ratio, √aMg/(√aca+Mg + B. aK) (r = 0·88***) or when a term for exchangeable K was included in regression analysis with √aMg/√aca+Mg and PBCMg (r = 0·93***).
The concentration of Mg in equilibrium soil solutions was generally a less satisfactory indicator of Mg availability than were the activity ratios. Soils derived from basaltic parent material had much higher contents of readily exchangeable Mg than the remainder: parent materials other than basalt had little influence on the availability of Mg in the soils. Exchangeable Mg and per cent Mg saturation were higher in gleyed soils than in freely drained soils derived from similar parent material.