Five herbicides commonly used for suppression of weed growth in white clover seed crops were tested for toxicity against white clover (Trifolium repens), Rhizobium trifolii and the nitrogen fixing symbiosis formed between these two organisms. Trials were carried out on R.trifolii on solid and in liquid media to determine if growth of this bacterium was affected by the presence of the 5 herbicides. Paraquat and MCPB substantially inhibited bacterial growth on solid medium. Bentazone, fusilade and kerb caused very small zones of growth inhibition of R.trifolii on solid agar at high concentrations. None of the herbicides tested affected growth of R.trifolii in liquid culture. In vitro studies of herbicide toxicity toward white clover were carried out to identify interactions of herbicide activity with rhizobial inoculation and supplied nitrogen, and to attempt to identify the targets of herbicide activity. Nodules grown under in vitro conditions were excised and used for ultrastructural examination. Herbicides were applied to plants grown in vitro at two different stages of plant growth. White clover proved to be very sensitive to all herbicides at the early seedling stage. Three week old plants were more tolerant. Pot experiments in a glasshouse environment indicated the relevance of in vitro experiments of herbicide toxicity against plants and gave information on the effect of soil water levels on herbicide activity. Paraquat was extremely toxic to white clover both in vitro and in pot experiments. Nodulation is affected to some extent directly by this herbicide although dessication of foliage probably has some role in halting activity of the nitrogenise enzyme. MCPB caused severe deformation of root tissue both in vitro and in pot experiments. It must be either contaminated with the active form of this herbicide, MCPA, or is being broken down to the active form by bacterial or chemical action. Bentazone did not damage white clover or nodule activity in a consistent way in vitro. However this herbicide did have a deleterious effect on both plant weight and nodulation when applied to white clover grown in soil, particularly under conditions of low soil moisture. Fusilade showed a direct effect on the activity of nitrogenase in vitro. Fusilade also acted more severely against plants of higher nutritional status, and appeared to affect nodule activity directly rather than affecting nodules via damage to other plant parts. Kerb was very toxic to seedling white clover in vitro although older plants were not as susceptible and were stimulated by high concentration of kerb. In pots white clover was slightly inhibited by kerb at recommended concentration while 10 x this concentration did not cause any inhibition of nodulation or plant growth. Differences in results between in vitro and pot studies of toxicity of these herbicides to white clover appear to be due to the different application methods used. In vitro herbicides were applied to the whole plant while in pot experiments herbicides were foliarly applied, hence more uptake by roots would be expected. Pot experiments indicated that changes in nodulation generally reflected changes in plant growth rather than an independent activity of the herbicide on nodulation.