The size and composition of pasture legume seedbanks were estimated from 2 surveys on a 460-km west-east transect of the wheatbelt of Western Australia. Survey 1 (in spring) sampled naturalised legumes, and survey 2 (in summer) measured the amount and botanical composition of legume seed from selected sites. Seedbanks were examined in greater detail on 2 farms in the higher rainfall part of the wheatbelt. Survey 2 revealed that mean seedbank size of the poorest 40% of sites (those with 5200 kg seed/ha) was 61 kg/ha, and that 72% of seeds were naturalised clovers. In contrast, the best 60% of sites (those with >200 kg seed/ha) averaged 533 kg seed/ha, of which only 35% was naturalised clover seed, the remainder in both surveys being mainly subterranean clover (Trifolium subterraneum). Mean seed bank size (kg/ha) varied from 359 (survey 2) to 587 (survey 1) and, in both surveys, was poorly correlated with long-term mean annual rainfall and a number of soil parameters. On the 2 farms, seedbank size ranged from 300 to 345 kg/ha (in spring) and from 650 to 740 kg/ha (in summer). Trifolium glomeratum (cluster clover) and subterranean clover were the most widespread species in both surveys. They were present at 35 and 30 of the 57 survey sites, respectively, and at both farms. Most of the subterranean clover collected was cv. Geraldton (22 of 30 sites), the next most frequent cultivar was Dwalganup (6 sites). The currently recommended cultivar, Dalkeith, was found at only 5 sites. Several other legumes including T. tomentosum (16 sites), T. suffocatum (8 sites), Medicago truncatula (7 sites), T. hirtum (4 sites), and M. minima (4 sites) were common, while M. littoralis, M. polymorpha, T. dubium, T. cernuum, T. cherleri, and T. carnpestre were found at single sites. With few exceptions, these are naturalised species and were characterised by flowering times about 20 days later than sown legume cultivars, and seed sizes < 1 mg. The value of these widespread annual legumes to agricultural productivity and sustainability needs to be quantified and their adaptation to wheatbelt farming systems assessed.