Multivariate adaptation to climatic shifts may be limited by trait integration that causes genetic variation to be low in the direction of selection. However, strong episodes of selection induced by extreme climatic pressures may facilitate future population-wide responses if selection reduces trait integration and increases adaptive potential (i.e., evolvability). We explain this counter-intuitive framework for extreme climatic events in which directional selection leads to increased evolvability and exemplify its use in a case study. We tested this hypothesis in two populations of the lizard Anolis scriptus that experienced hurricane-induced selection on limb traits. We surveyed populations immediately before and after the hurricane as well as the offspring of post-hurricane survivors, allowing us to estimate both selection and response to selection on key functional traits: forelimb length, hindlimb length, and toepad area. Direct selection was parallel in both islands and strong in several limb traits. Even though overall limb integration did not change after the hurricane, both populations showed a non-significant tendency toward increased evolvability after the hurricane despite the direction of selection not being aligned with the axis of most variance (i.e., body size). The population with comparably lower between-limb integration showed a less constrained response to selection. Hurricane-induced selection, not aligned with the pattern of high trait correlations, likely conflicts with selection occurring during normal ecological conditions that favor functional coordination between limb traits, and would likely need to be very strong and more persistent to elicit a greater change in trait integration and evolvability. Future tests of this hypothesis should use G-matrices in a variety of wild organisms experiencing selection due to extreme climatic events. 

SVL was measured from the tip of the snout to the posterior edge of the anal scale (Herrel et al. 2008); femur length was measured from the axilla to the tip of the femur; humerus length from the base (detected by palpation) to the tip of the humerus; tibia length was measured from the femoro-tibial joint to the tibia-metatarsus joint; ulna length was measured from the humero-ulnar joint to the ulna-metacarpus joint; metatarsus length was measured from the proximal-most part of the metatarsus to the base of the longest toe; metacarpus length was measured from the proximal-most part of the metacarpus to the base of the longest toe; longest toe length was measured from the base to the tip of the toe for both the hand and the foot, not including the claw. All measurements were taken using digital calipers (MitutoyoCD-20DC, Japan; precision: 0.01 mm), and were taken on the left side of the specimens. Toepad surface area was quantified using a Moment Macro Lens attachment on an iPhone 7. In the field, we took a picture of the forefoot and hindfoot of each lizard. Later, CMD measured the surface area of toepads on digit III on the forelimb and digit IV on the hindlimb to the first scale after the toepad widens (see Donihue et al. 2020), using ImageJ (version 1.51a; W. Rasband, National Institutes of Health, Bethesda, MD). All measurements are in log scale.

main_data.rda: open using function 'load' in R programming environment.

water_data.csv: open using a Excel or a text editor (e.g., Notepad).

pine_data.csv: open using Excel or a text editor (e.g., Notepad).