Small-scale genetic differentiation in mean flowering time, but not in plasticity, along a geothermal heating gradient

Genetic differentiation in traits is assumed to frequently occur in response to divergent natural selection. For example, developmental traits might respond to differences in climate. However, little is known about when and at which spatial scales environmental differences lead to genetic differentiation, and to what extent there is genetic differentiation also in trait plasticity. Using a crossing design and a greenhouse heating experiment, we investigated genetic differentiation in the thermal sensitivity of flowering time in a perennial herb along small-scale gradients in geothermal soil heating in Iceland. We found additive genetic variation in both flowering time and thermal plasticity of flowering time. Genetic differentiation in the median flowering date of individuals showed a counter-gradient pattern; flowering being earlier at higher greenhouse temperatures, while at a given temperature individuals originating from warmer soils flowered later than individuals from colder soils. We found no corresponding pattern for plasticity, suggesting that genetic differentiation in phenology in response to soil heating has occurred through changes in trait means rather than in plasticity. Findings such as these, identifying genetic trait differentiation along an environmental gradient are key to understanding how environmental variation can drive the process of local adaptation, and to predict responses to future environmental changes.

Methods
The study was conducted with plants originating from Hengill, a sub-arctic geothermal area in SW-Iceland, where. The study area (ca. 1 km2) shows large gradients of soil temperature (e.g. ranging from geothermal heating to > 20°C above ambient temperature over 10-20 meters), without significant differences in other abiotic factors. In this area, we selected individuals of Cerastium fontanum that represented the entire temperature gradient. Soil temperature was measured in the immediate vicinity of each individual and fruits were collected from flowering individuals. The seeds were then sown in a greenhouse at Stockholm University (SU), and seedlings were moved to a common garden at SU, and kept there for two years. We then used a North Carolina type II (NCII) crossing design to investigate genetic differences in flowering time and in thermal plasticity of flowering time among individuals originating from different temperature environments. Seeds from each of the crosses were shown in the greenhouse and seedlings were kept in the common garden until the next year, when a heating experiment was performed in the greenhouse, including two different treatments: a heated treatment mimicking soil heating and a control treatment with ambient temperatures. For each individual in the experiment, we recorded the total number of flowers produced and three measures of flowering phenology (first, median, and last flowering date).