Salsola australis is an introduced weed of crop and pasture systems in the Western Australian broad acre cropping and pasture region (wheat-belt). This thesis investigated the classification, biology and ecology of the genus Salsola in southwest Australia, as well as modelling the effectiveness of possible weed control practices. Prior to this research, S. tragus was the only recognised species of the Salsola genus within Australia. However, genetic analysis revealed that four genetically distinct putative taxa of the genus Salsola were found in southwest Australia, none of which were S. tragus. The taxa that is the most prevalent agricultural weed was classified as S. australis, but the other three putative taxa could not be matched to recognised species. All four taxa were diploid (2n = 18), as opposed to tetraploid (2n = 36) S. tragus. Within the agricultural system of southwest Australia, S. australis plants established throughout the year, although the majority of seed production occurred in late summer and autumn. Total seed production (138-7734 seeds per plant) and seed viability (7.6-62.8%) of S. australis were lower than that reported for other agricultural weed species of the Salsola genus. Seed dispersal occurred when the senesced plants broke free of their root system to become mobile. Wind driven plants travelled and shed seed over distances of 1.6 to 1247.2 m. Movement of approximately half the plants was restricted to less than 100 m by entanglement with other S. australis plants within the stand. Some seed was retained on the senesced plants, but the germinability of this seed fell to less than 2% in the two month period following plant senescence (i.e. a decline of 79%). Once seed shed into the soil seed bank, anywhere from 32.3 to 80.7% of the viable seeds germinated in the year following seed production, with the rest remaining dormant or degrading. A model of the life cycle of S. australis based on the population ecology data indicated that the dormant seed bank had very little effect on annual seedling recruitment, but seed dispersal from neighbouring populations had a large impact on population growth rate. Therefore, the most successful weed control measures were those that restricted seed dispersal from neighbouring populations, or those that were applied to all populations in the region rather than to a single population. Weed control techniques applied to a single population, without reducing seed dispersal, could not reduce population size.