One of the central puzzles in ecology is what determines the size of populations in nature. Strident debate between those claiming precedence of density-independent factors over density-dependent ones, and vice versa, has subsided with the realization that both may play a role in limiting population size of any species. Nevertheless, very few attempts have been
made to directly assess the effect of density on population growth in nature. This is particularly disturbing, because theoretical ecology 1s largely based on the density-dependent population growth routinely
documented in laboratory experiments.
Similarly, explanations of the existence of variation in life history among and within species, in particular the weedy vs. non-weedy habit, have invoked density-dependent selection. Yet such selection has never been demonstrated in nature, and thus its efficacy in maintaining genetic
variation is unknown. The experiments reported here were devised to determine the extent to which density limits growth of a population of Salvia lyrata L., an herbaceous perennial plant common in North Carolina grasslands, and whether density-dependent selection structures its genetic
The local density of Salvia was altered by sowing in seed at different densities, and by transplanting or removing established individuals. In most cases, these manipulations elicited weak or conflicting responses; however density-dependent mortality and stunting of seedlings was evident at unnaturally high densities of sowing. Thus density is rarely sufficient
to limit an individual's contribution to this population, but density effects can limit population size at extreme seedling densities.
In additional experiments, evidence of genetic variation in density response was sought by planting individuals of known genetic origin into arrangements of varying density. An experiment in protected conditions showed variation among families in the response of flowering probability to
density, but not in survival nor in the number of seeds produced. A field
experiment showed variation among families in the reponse of a size trait, number of leaves, to density. Given that survival and fecundity are size-dependent, as documented in observations of the natural population, this result suggests the potential for density-dependent selection in nature.