Analyzing reaction norm variation in the field vs. greenhouse: Comparing studies of plasticity and its adaptive value in two species of Erodium

Despite a wide range of experiments characterizing patterns of selection on phenotypic plasticity in controlled environments there has been virtually no research assessing the extent to which these results reflect selection on plasticity expressed in natural populations. To test how well the patterns observed in controlled experiments match the patterns in field populations, we present two case studies in which we characterized the fitness consequences of plasticity both under controlled lath house conditions and in the field. We quantified selection on plasticity in response to soil nutrient variation in two annual plant species, Erodium cicutarium and Erodium brachycarpum. For both species, families collected from the same source populations were used in both field and lath house experiments. We ask whether the qualitative results obtained from field and controlled environment experiments are equivalent. In two cases we observed selection on the expression of plasticity by E. brachycarpum in the field while controlled environment experiments indicated that plasticity was selectively neutral. In three other cases we observed differences in the pattern of plasticity expressed in the controlled environment experiment relative to the field resulting in conflicting results regarding the form of trait expression favored by selection. Based on these results, we argue that the extent to which results from controlled environments can be accurately extrapolated to naturally occurring populations depends on whether treatments imposed in a controlled environment accurately mimic environmental variation in the field and induce plasticity in traits of interest. Ideally any controlled environment experiment characterizing plasticity would be paired with field survey data of environmental and phenotypic variation within naturally occurring populations.