Loss and fragmentation of the native prairies in the Midwestern United States have resulted in isolated and smaller habitats and populations. The populations remaining in these prairies are expected to show a decline in the extent of genetic variation and an increase in genetic drift load (accumulation of deleterious recessive alleles due to genetic drift) in fitness-related traits. Using complementary greenhouse experiments, we tested whether these expected changes have occurred in the native annual prairie plant Chamaecrista fasciculata. In the first experiment, open pollinated C. fasciculata seeds from 12 prairie fragments representing a range in area of habitat were grown in competition with Schizachyrium scoparium to determine if there are changes in plant vigour with changes in fragment size and corresponding changes in population size. Plants from smaller prairie fragments exhibited a slight but significant decline in biomass, suggesting an increase in genetic drift load. In the second experiment, a formal genetic crossing design of four prairie fragment populations was used to estimate quantitative genetic diversity and genetic drift load. We did not find extensive quantitative genetic variation, but we did find a strong effect of genetic drift load on five traits in this experiment. Our overall conclusion is that a decline in relative-fitness traits in smaller prairie fragments is probably associated with fixation of deleterious alleles due to more isolated and smaller populations, i.e. genetic drift load. © 2012 The Linnean Society of London, Biological Journal of the Linnean Society, 2012, ••, ••–••.