This is the complete data set used in the manuscript titled "Variation in the role of the flag leaf in the reproductive effort of semi-arid rangeland bunchgrasses". It includes the responses of total propagule production, percentage of filled and unfilled seeds, and filled seed-specific mass in response to seed head shading and flag leaf removal in two semi-arid perennial bunchgrasses, crested wheatgrass (Agropyron cristatum) and squirreltail wild rye (Elymus elemoides). Seasonal volumetric soil moisture data is also provided.

Plants were randomly assigned into one of six flag leaf clipping/seed-head shading treatments: 1) flag leaves intact and seed-heads unshaded throughout the study (referred to hereafter as control plants), 2) flag leaves intact with post-anthesis seed-head shading, 3) pre-anthesis clipping flag without seed-head shading, 4) pre-anthesis clipping with post-anthesis seed-head shading, 5) post-anthesis clipping flag leaf without seed-head shading, and 6) post-anthesis clipping with seed-head shading. Flag leaves were clipped at the leaf collar sheathing the culm, and seed-head shading was provided by enclosing the post-anthesis seedhead in a mylar sheath. We tried to sample from 5 culms per plant, if at all possible, though culm breakage did occur; only culms that remained intact through the entire study period were used for our statistical analyses. As crested wheatgrass and squirreltail have distinct seasonal reproductive phenologies, these treatments were initiated at different dates throughout the study period, as noted in the supplementary metadata (S1). Starting June 2, weekly volumetric soil moisture (mol³ mol^-3) across 10 cm rooting depths adjacent to each sampled plant.

Entire seed heads were harvested after seed filling was completed in control plants of either species. Seed heads were disarticulated, and visually separated into filled and unfilled seed fractions on a light table. The number of propagules in each fraction was counted, weighed on a microbalance to the nearest 0.001 mg, then scanned at 800 dpi resolution on a flat-bed scanner calibrated for use with WinRhizo image analysis software (Regent Software, Quebec, Canada) to determine the projected surface area (cm²) of each fraction. These data provided the total propagule number (empty + filled seeds), the proportion of filled seeds (filled seed/total propagules), aggregate individual filled seed area (total projected area/number), seed mass (total mass/number), and seed-specific mass (total area/total mass; mg cm^-2).