Potassium relationships in selected eutrophic hutton soils of irrigation areas
Abstract
The relationship between red leaf disease and soil fertility
status was studied in the glasshouse. For red leaf disease
affected plants grown on Mangano soils, no effective fertilizer
treatment could be found. However, yield could be
predicted more accurately by soil parameters other than the
traditional ammonium acetate extractable potassium or exchangeable potassium. These parameters were the ratio (Ca+Mg)/K in the soil and exchangeable potassium percentage.
These initial findings were tested in fifteen field trials
on Mangano soils with variable clay contents and exchangeable
potassium contents at Vaalharts.
A large matrix of data resulted from the field investigation.
However, only statistical procedures relating soil
and leaf-K parameters to yield and leaf-K were investigated.
The data were statistically analised by means of analysis of
variance, correlation and multiple regression on an IBM 900
computer. For each soil and leaf potassium parameter relationships
were established with yield and leaf-K content for
individual trials, pooled data and grouped data.
Attention is drawn to the fact that ammonium acetate extractable potassium and exchangeable potassium are not reliable
predictors of cotton yield for this soil/crop relationship.
These two parameters should rather be considered
together with the ratio (Ca + Mg)/K in the soil, exchangeable
potassium percentage and leaf-K content when cotton yield is
to be predicted on Mangano soils at Vaalharts.
The initial results from one pot experiment and fifteen
field trials were largely substantiated by two additional
pot experiments. In order to find an explanation for the lack
of response to applied potassium, potassium fixation and K
fractionation were also investigated. All the soils tested
fixed appreciable quantities of K. Total K amounted to
approximately one percent.
Recommendations are made to clarify the complex nature of
potassium reactions in the soil/crop relationship.