During 1999–2003, 293 samples of preinoculated and custom-inoculated lucerne, subterranean clover, white clover, red clover and miscellaneous species (mainly other clovers) were sourced from commercial outlets and assessed for numbers of rhizobia, seed pellet pH and toxicity, and nodulation in a ‘grow-out’ test. Average rhizobial counts were 8400/seed for preinoculated lucerne, 1380/seed for subterranean clover and <100/seed for white and red clovers and for the miscellaneous species. These counts compared poorly with the average counts of 35 100/seed, 13 800/seed and 10 000/seed for freshly-inoculated lucerne, subterranean clover and white clover, respectively. Thus, overall pass rates of the preinoculated seed were reasonable for lucerne (73%), marginal for subterranean clover (32%) and very low for white clover (3%), red clover (4%) and the miscellaneous species (0%). The ‘grow-out’ tests for nodulation were positively correlated with rhizobial numbers on seed, confirming the use of plate counting of rhizobia to assess quality of pre- and custom-inoculated seed. Many of the seed pellets were toxic to the 2 clover rhizobial strains tested, although the toxicity did not affect numbers of rhizobia on the seed. In light of these results and other data on rhizobial survival on seed, we suggest the current Australian standards for rhizobial numbers on pasture legume seed at the time of sale of 500/seed (very small-seeded legumes with seed numbers >750 000/kg) and 1000/seed (other larger-seeded species, seed numbers <750 000/kg) remain in place. We recommend shelf lives be restricted to 6 months for preinoculated lucerne and the annual medics, to 6 weeks for preinoculated subterranean clover, and to 2 weeks for white clover, red clover and other miscellaneous species. In the long-term, new products and procedures will hopefully enhance the numbers and survival of rhizobia on seed such that the needs of both manufacturers and customers are satisfied.