Seedling emergence strongly limits establishment of bunchgrass populations critical to restoring sagebrush steppe degraded by invasive annual grasses. Selecting for seed mass, an indicator of seed quality, could improve bunchgrass seedling emergence and establishment, but it is unknown how variation in soil moisture affects these outcomes. To address this, we sorted seeds of three important restoration species, crested wheatgrass (Agropyron cristatum), bluebunch wheatgrass (Pseudoroegneria spicata) and Sandberg bluegrass (Poa secunda) into large and small seed size-classes and determined individual seed mass, area and seed density (mg mm-2). Seeds were planted into pots with native soil under watering regimes equivalent to 0.5x (low), 1.5x (high) and average monthly precipitation from October to June under a rain-out shelter in intact sagebrush steppe.
Results/Conclusions
Germination decreased with lower water, and large size-class seeds in A. cristatum and P. secunda had improved germination, but not P. spicata. Pre-emergent seedling mass did not differ between watering treatments; P. secunda masses were constant over size-classes and time, while A. cristatum and P. spicata masses were greater in large size-class seedlings, and masses of both species declined over time. Emergence markedly reduced, if not eliminated, in low watering treatments. Large size-class A. cristatum and P. secunda had higher emergence, while P. spicata did not, with emergent seedling mass affected only by watering regime. Established plant tiller numbers and aboveground canopy area were identical between size-classes, and reduced with lower soil water. Agropyron cristatum root-shoot ratios (R:S) increased 3-fold in low water, and were greater in large size-classes. Small size-class P. secunda R:S reduced with lower soil water, while large size-class R:S was constant. Most seedling emergence and growth parameters in A. cristatum and P. secunda matched size-class seed density, while in P. spicata, seed mass, not density, related only to pre-emergent seedling mass. Thus, seed density more likely reflected available seedling energetic reserves in A. cristatum and P. secunda, but not for P. spicata. We believe these findings will help in developing selection criteria that recognize the differences and consequences of per seed energetic allocation, and these could improve restoration efforts in semi-arid rangeland systems.