Ecological constraints on the evolution of plasticity in plants

Signal detection and response are fundamental to all aspects of phenotypic plasticity. This paper proposes a novel mechanism that may act as a general limit to the evolution of plasticity, based on how selection on signal detection and response is likely to interact with gene flow in a spatially autocorrelated environment. The factors promoting the evolution of plasticity are reviewed, highlighting the crucial role of information acquisition and developmental lags, and of selection in spatially and temporally structured habitats. Classic studies of the evolution of plasticity include those on shade avoidance, on morphological plasticity in clonal plants, and on selection in spatially structured model populations. Comparative studies indicate that, among clonal plants, extensive plasticity in growth form is favored in patchy environments, as expected. However, among woody lineages from Madagascar, plasticity in photosynthetic pathway (CAM vs. C3) appears to confer competitive success in areas of intermediate drought stress, rather than allowing individually plastic species to expand their ranges, as has often been argued. The extent of phenotypic plasticity cannot only determine species distributions, it can also affect the sign and magnitude of interactions between species. There appears to be some relationship between developmental plasticity and evolutionary lability: traits that show relatively few transitions within and among plant lineages (e.g., zygomorphy vs. actinomorphy, phyllotaxis, fleshy vs. capsular fruits) usually show no plasticity within individual plants; traits that show extensive plasticity within individuals or species (e.g., leaf size, flower number, plant height) generally also show extensive variation within and across lineages. Transaction and cybernetic costs, as well as long-lived leaves or roots, can limit the tempo of adaptive developmental responses, and create a hierarchy of responses at different temporal scales. Traits whose variation entails few transaction costs (e.g., stomatal conductance) are more likely to be shifted more frequently than those with higher costs of variation (e.g., leaf cross-sectional anatomy). The envelope of responses at the physiological and developmental time scales appears to be an important determinant of adaptive performance. However, adaptive plasticity can limit its own range of effectiveness as a consequence of energetic and competitive constraints, as seen in the allometry and zonation of emergent vs. floating aquatic plants. Plants' inherently low rate of energy capture (and, hence, developmental response and growth) and the high energetic costs of a central nervous system (CNS), may explain why they lack a brain and integrate environmental signals with a slow, hormone-based set of feedback loops rather than with a fast CNS. Finally, environmental spatial autocorrelations – especially those involving factors that determine optimal phenotype – can combine with gene flow and selection for reliance on the locally most informative signals to produce a fundamental limit on the extent of adaptive plasticity.