Does genetic variation in controlled experiments predict phenology of wild plants?

Phenology and the timing of development are often under selection.
However, the relative contributions of genotype, environment, and prior
developmental transitions to variance in the phenology of wild plants is
largely unknown. Individual components of phenology (e.g., germination)
might be loosely related with the timing of maturation due to variation in
prior developmental transitions. Given widespread evidence that genetic
variation in life history is adaptive, we investigated to what degree
experimentally measured genetic variation in Arabidopsis phenology
predicts phenology of plants in the wild. As a proxy of phenology, we
obtained collection dates from nature of 227 naturally inbred Arabidopsis
thaliana accessions from across Eurasia. We compared this phenology in
nature with experimental data on the descendant inbred lines that we
synthesized from two new and 155 published controlled experiments. We
tested whether the genetic variation in flowering and germination timing
from experiments predicted the phenology of the same lines in nature. We
found that genetic variation in phenology from controlled experiments
significantly predicts day of collection from wild individuals, as a proxy
for date of flowering, across Eurasia. However, local variation in
collection dates within a region was not explained by genetic variance in
phenology in experiments, suggesting high plasticity across small-scale
environmental gradients or complex interactions between the timing of
different developmental transitions. While experiments have shown
phenology is under selection, understanding the subtle environmental and
stochastic effects on phenology may help to clarify the heritability and
evolution of phenological traits in nature.