1. Population differentiation was studied in the halophytic annuals Salicornia and Suaeda maritima at Stiffkey, Norfolk, England. 2. The salt marsh at Stiffkey shows considerable spatial and temporal heterogeneity. The gradient in height across the marsh results in extreme differences in tidal regime and edaphic conditions, while physiographic features such as salt pans, creeks and their associated levees, impose smaller-scale heterogeneity. Salicornia and Suaeda maritima occur throughout the marsh and exhibit striking variations in phenotype associated with particular microhabitats. 3. Monitoring of field populations of these annuals during successive phases of the life cycle showed that morphologically distinct populations differed in aspects of their life-histories. These differences included the control of dormancy and germination, the phenology of growth and development, and fecundity. 4. The genetic and environmental components of these differences were examined in reciprocal transplant experiments of both seed and seedling material and in uniform growth conditions in growth cabinets and in a glasshouse. Survival and many aspects of both vegetative and reproductive performance were measured throughout the life cycle. 5. Much of the variation in natural populations was attributable to differences in plant density. Plant size and fecundity were extremely plastic in their response to density. Phenological features were more stable and structural characters were most stable. Survival was density independent. 6. Environmental differences between sites also had profound effects on growth and survival of the plants. In particular, some aspect of hypersaline edaphic conditions and the presence of perennial vegetation reduced considerably the survival and growth of annual plants. 7. When the density dependent components of performance were removed, and plants from different parts of the marsh were grown under the same environmental conditions, significant differences remained between them in morphology and life-history. This applied both to plants grown on the same site in the field and to those grown under uniform conditions. These differences between populations were likely to be genetic. 8. Three sorts of explanation of the adaptive significance of genetic differences between populations were examined. First, the relative performance and survival of local and foreign populations planted on the same site were examined at all stages of the life cycle. These differences were expressed as relative selection coefficients. The relative success of populations was also measured over an entire generation by comparing numbers of seeds sown with those produced. This gives an estimate of the relative fitness of the local and foreign populations at each site. Second, correlations were examined between life history characteristics of the populations and features of their environment that were likely to act as selective agents. Third, parallel variation was examined between Salicornia and Suaeda maritima populations within the same zone of the marsh. 9. All three of these approaches were adopted to achieve a concensus of information. Each approach suffers different limitations and the interpretation of information obtained from each is discussed critically. 10. Selection usually favoured the survival and growth of local relative to that of foreign populations at all sites and at all stages in the life cycle. The highest selection coefficients were for survival of the lower marsh 1 t" ... upper morsn 1 . popu a lon ln competltlon wlth the~perennla vegetatlon during the growth phase (ca 0.7) and for seed production of upper marsh Suaeda maritima on the lower marsh (ca 0.9). 11. Measures of the intensity of selection at specific stages of the life cycle did not predict accurately the magnitude of selection over an entire generation. All populations planted as seeds on their native site were relatively fitter than foreign populations transplanted to same site. Relative fitness is likely to have been underestimated because the local and foreign populations at each site were not grown in competition with one another. 12. Chromosome numbers were counted in Salicornia plants typical of the populations recognized by their morphologies and life histories. Three tetraploid and three diploid Salicornia populations and three diploid Suaeda maritima populations were identified at stiffkey. 13. The breeding systems and possible genetic structures of Salicornia and Suaeda maritima populations are discussed in relation to population differentiation. salicornia populations appear to be largely inbreeding and consist of many predominantly homozygous lines. occasional outcrossing provides sufficient variability to allow divergence of the populations under selection. population differentiation in Suaeda maritima apears to be maintained by restriction of gene flow between populations by divergence in flowering times, as well as by intense selection.