Longitudinal and cross-sectional selection on flowering traits in a self-incompatible annual

Net selection on a trait reflects the association of phenotype to fitness across an entire life cycle. This longitudinal selection estimate can be viewed as the summation of selection episodes across time, each characterized by a cross-sectional estimate. Selection may be consistent in direction and strength across episodes for some traits, fluctuating in others, and for some, concentrated in a single intense event. While selection on plant reproductive traits is predicted to be stronger through male fitness than female fitness, male fitness remains less studied. We investigated how selection on flowering traits in Brassica rapa varies across a season by measuring male reproductive fitness in four experimental populations. To estimate longitudinal and cross-sectional selection, we introduced plants at successive intervals within a single reproductive season. We genotyped over 3000 plants and calculated selection on flowering time, duration, and total flowers. Cross-sectional analyses revealed directional selection was common, but patterns were masked by longitudinal estimates. Manipulation of spatial arrangement within populations further revealed to significantly impact pollen movement. Our experiment demonstrates that breeding timing and spatial aggregation interact to create complex evolutionary dynamics.