A method for estimating the nitrogen-fixing capacity of a population of rhizobia resident in soil is presented. Legume test plants, growing under microbiologically-controlled conditions in test tubes packed with a vermiculite substrate moistened with a nitrogen-free plant nutrient solution, are inoculated directly with a suspension of the soil under examination. Rhizobia in the soil nodulate the test plants, and the amount of foliage dry matter produced in the 28 days after inoculation is regarded as an index of their effectiveness. An inoculum of at least 30, and preferably 100, rhizobia is needed to ensure that nitrogen fixation is not masked by delayed nodulation. The new method is tentatively described as the 'whole-soil inoculation' technique.Appraisals were made with Trifolium subterraneum L. and Rhizobium trifolii and with Medieago sativa L. and R. meliloti. Soil-borne pathogens did not interfere with plant growth. The whole-soil inoculation technique was less tedious and time-consuming than an alternative method which involved extracting representative isolates from the soil and testing their effectiveness individually, and appeared to give more realistic values for the nitrogen-fixing capacity of the soil as a whole. Used in association with a field experiment, the whole-soil inoculation technique confirmed microbiologically that there had been an agronomic response to surface application of inoculant to poorly-nodulated T. subterraneum pasture.It is submitted that this technique for determining the effectiveness of rhizobia in soil, combined with a plant-infection method for counting rhizobia, can be a reliable guide to the need for inoculation in the field.