Abstract A suite of dose–response bioassays with white mustard ( Sinapis alba L) and sugar beet ( Beta vulgaris L) in the greenhouse and with three herbicides was used to analyse how the fluorescence induction curves (Kautsky curves) were affected by the herbicides. Bentazone, a photosystem II (PSII) inhibitor, completely blocked the normal fluorescence decay after the P‐step. In contrast, fluorescence decay was still obvious for flurochloridone, a PDS inhibitor, and glyphosate, an EPSP inhibitor, which indicated that PSII inhibition was incomplete. From the numerous parameters that can be derived from OJIP‐steps of the Kautsky curve the relative changes at the J‐step [ F vj = ( F m − F j )/ F m ] was selected to be a common response parameter for the herbicides and yielded consistent dose–response relationships. Four hours after treatment, the response F vj on the doses of bentazone and flurochloridone could be measured. For glyphosate, the changes of the Kautsky curve could similarly be detected 4 h after treatment in sugar beet, but only after 24 hs in S alba . The best prediction of biomass in relation to F vj was found for bentazone. The experiments were conducted between May and August 2002 and showed that the ambient temperature and solar radiation in the greenhouse could affect dose–response relationships. If the Kautsky curve parameters should be used to predict the outcome of herbicide screening experiments in the greenhouse, where ambient radiation and temperature can only partly be controlled, it is imperative that the chosen fluorescence parameters can be used to predict accurately the resulting biomass used in classical bioassays. Copyright © 2003 Society of Chemical Industry