Rapid, parallel evolution of field mustard (Brassica rapa) under experimental drought

Climate change is driving evolutionary and plastic responses in populations, but predicting these responses remains challenging. Studies that combine experimental evolution with ancestor-descendant comparisons allow assessment of the causes, parallelism, and adaptive nature of evolutionary responses, although such studies remain rare, particularly in a climate change context. Here, we created experimental populations of Brassica rapa derived from the same natural population and exposed these replicated populations to experimental drought or watered conditions for four generations. We then grew ancestors and descendants concurrently, following the resurrection approach. Experimental populations under drought showed rapid evolution of earlier flowering time and increased specific leaf area, consistent with a drought escape strategy and observations in natural populations. Evolutionary shifts followed the direction of selection and increased fitness under drought, indicative of adaptive evolution. Evolution to drought also occurred largely in parallel among replicate populations. Further, traits showed phenotypic plasticity to drought, but the direction and effect size of plasticity varied. Our results demonstrate parallel evolution to experimental drought, suggesting that evolution to strong, consistent selection may be predictable. Broadly, our study demonstrates the utility of combining experimental evolution with the resurrection approach to investigate responses to climate change.

Methods
We combined experimental evolution with the resurrection approach to explore adaptation to experimental drought in Brassica rapa (field mustard). 

We founded 24 experimental populations from lines which originated from a natural population and randomly assigned each experimental population to a watering regime. We assigned 8 populations for storage as ancestors, 8 to receive four generations of experimental drought, and 8 to receive four generations of experimental well-watered conditions. During these four generations of experimental evolution, drought regime populations received a draw-drown drought treatment aimed to mimic Mediterranean drought while well-watered regime populations were watered every other day. We collected data (flowering time, survival, and seed mass) on a subset of individuals each generation to explore trends in fitness and selection during experimental evolution.  

After experimental evolution, we grew all experimental populations (ancestors included) under unstressed (well-watered) conditions to reduce maternal and storage condition effects across the different regimes. 

We then grew two seeds (siblings) from 35 individuals per population while applying a watering treatment with one sibling from each individual under each treatment in what was the test generation. This design allowed us to assess evolutionary responses (ancestor-descendant regime comparisons), plasticity (watered-drought treatment comparisons), and consistency of responses (parallelism among replicates) to experimental drought.

Detailed methods, statistical analysis, and results are discussed in the associated manuscript published in Evolution.