On the Great Hungarian Plain, the correlation between soil variables a
nd plant data was studied in a heterogeneous solonetzic grassland. A 2
1-m transect was sampled to identify the basic features of the soil-pl
ant correlations. The abrupt change from nonsaline meadow to salt-affe
cted grassland vegetation was explained by the sharp increase in the s
odium concentration of the groundwater. Two 50-m transects were sample
d, and the relationship between soil variables and plant data was used
to predict soil variables. The observed data of the 50-m long transec
ts were divided into two groups of plants, one used for the multiple r
egression equations and the other used as a check. When comparing the
efficiency of various groups of variables for estimation, plant cover
data was only half as precise as the soil data.