Salina wildrye (Leymus salinus [M.E. Jones] Á. Löve) is a perennial cool-season grass that potentially could become an important restoration species in the Colorado Plateau. However, its seed production has never been commercially viable due to sparse heading. We compared a 4x ssp. salmonis population, Lakeside C₃, to an 8x ssp. salinus population, 9043501, for seed production — related traits and measured the response of 9043501 to 2 cycles of selection for increased spike number over a 4-yr period at Millville, Utah. Seed yield of Lakeside and 9043501 was similar (P > 0.10) in 2013, but seed yield of 9043501 was 81% greater (P < 0.10) than Lakeside in 2014 and 191% greater (P < 0.01) in 2015. Lakeside spike number was 99% greater (P < 0.0001) than 9043501 in 2013, but they were similar (P > 0.10) in 2014 and 2015. Seeds per spike of 9043501 were 71% (P < 0.05), 80% (P < 0.05), and 209% (P < 0.01) greater than Lakeside in 2013, 2014, and 2015, respectively. Selection in 9043501 increased (P < 0.05) spike number by 4.3 spikes per plant (19.8%) per cycle of selection in the first seed-production yr (2013), but no change was seen in 2014 or 2015 (P > 0.10). Selection did not change (P > 0.10) seeds per spike or individual seed mass. Consequently, seed yield increased (P < 0.05) 0.32 g per plant per cycle (36.8%) in 2013, with no increase (P > 0.10) in 2014 or 2015. Dry matter per plant across the 4 yr increased (P < 0.01) 10.3 g per plant per cycle (9.3%), and canopy height increased (P < 0.01) 3.9 cm per cycle (6.6%) in 2013. AFLP DNA primers detected a 1.7% loss of genetic variation per cycle, presumably due to a combination of selection and genetic drift, but no plant traits were diminished as a result.