The annual habit in plants is often accompanied by specialized physiological mechanisms which permit seeds to remain dormant for a few years to decades. This results in the establishment of a seed pool from which germinating individuals are drawn. The seed pool causes an overlapping of generations in a population which would otherwise be subdivided into discrete generations. Others have examined seed pools per se as evolutionary strategies, but we are examining the constraints a seed pool imposes on the evolution of loci controlling seedling and adult characters not directly related to seed dormancy. Since a seed pool introduces genetic constraints for all loci simultaneously, there is not one evolutionary impact of seed pools upon an annual population, but many. For those loci that yield constant fitness effects, the presence of a seed pool retards the annual rate of allele frequency change, but does not alter the final equilibrium. However, for those loci interacting with variable features of the environment in producing their fitness effects, the seed pool can alter the ultimate genetic outcome. In a cyclical environment, the seed pool can dampen fitness cycles with a short period (with respect to the extent of the seed pool over past years) and accentuate long cycles. More importantly, the seed pool may serve as an evolutionary filter that causes only a few years from the cycle to have any true evolutionary impact and to effectively eliminate the selective impact of other years. This evolutionary filter also operates for those loci interacting with environmental elements that fluctuate randomly from year to year. Hence, seed pools can greatly reduce the fitness uncertainty generated by cyclical or random environments and free the plant population from having to respond genetically to the fitness conditions realized in every year. Moreover, the seed pool allows the absolute quality of a given year's environment to influence allele frequency changes and not just the relative fitnesses of the genotypes at that locus. Finally, we examined the selectively analogous situation of seedling and juvenile selection in perennial plants lacking a seed pool. In this case, the standing crop replaces the seed pool as the agent causing an overlapping of generations. Most of the conclusions about the evolutionary effects of seed pools in annuals carry over to the evolutionary effects of standing crops on seedling and juvenile selection in perennials. However, the evolutionary filter caused by a standing crop is less effective in filtering random seedling and juvenile fitness fluctuations in perennials than the filter caused by a seed pool with respect to random fitness fluctuations in annuals.