A herbicide bioassay based on tissue cultures of Brassica napus L. was
evaluated with selected sulfonylurea herbicides. Data were analysed b
y fitting the results to a log-logistic dose-response model. Within an
experiment, the non-linear regression models were fitted simultaneous
ly to the individual dose-response curves. The results obtained showed
good response to even low concentrations of herbicide, with detection
limits in the range 0.008-0.69 nmol L-1 for chlorsulfuron and 0.02-0.
13 nmol L-1 for metsulfuron. The reproducibility of the assays, on the
basis of coefficient of variation of the ED50 values, was found to be
44% for chlorsulfuron and 48% for metsulfuron measurements. Assay of
herbicide dissolved in aqueous soil extract showed significant interfe
rence from this matrix on the response, requiring a five times dilutio
n of the extract to overcome this matrix effect.