SUMMARY Of two biotypes of Taraxacum officinale one (D) was shown previously in a glasshouse experiment to out-compete the other (A), but to set fewer seeds in unit time. The two biotypes have now been grown separately and in mixture in garden plots for four years; in certain plots they were subjected to artificial disturbance by defoliation or removal of the entire plant at midsummer for the first two years. After four years biotype D accounted for c. 85% of the plants and 91 % of the biomass in the mixtures on the undisturbed plots, but only 7-9% of the plants and 3-7% of the biomass on the disturbed plots.