Objectives The seed production of Festuca kryloviana Reverd. is often limited by water and nutrient supply. Hence, we tested the efficiency of combined application of low nitrogen and high phosphorous fertilizers rate and super absorbent polymers) on seed yield of F. kryloviana in an alpine region.

Methods This experiment was conducted in Haibei Prefecture, Qinghai Province. A three-year old perennial forage germplasm of the F. kryloviana was tested. Two super absorbent polymer rates (PM) (0 and 30 kg/hm²), two nitrogen (N) fertilizer rates (0 and 60 kg/hm²) and four phosphorus (P) fertilizer rates (0, 60, 75 and 90 kg/hm²) were combined into sixteen treatments making a 2 × 2 × 4 factorial arrangement in a randomized complete block design. The seed yield and traits were measured at harvest.

Results 1) The polymer alone did not affect the seed yield of F. kryloviana significantly, but its interaction with N or NP significantly did (P < 0.05). Application of 60 kg/hm² of N (N₆₀), with or without P and absorbent polymer, generally reduced seed yields by 16%–48%. Contrarily, P significantly increased the seed yield by 28%–49% (P < 0.05). Treatments P₇₅ and P₉₀ were similar (P>0.05), but both significantly (P < 0.05) obtained higher seed yield than P₀ and P₆₀. The highest seed yield (560 kg/hm²) and the lowest (260 kg/hm²) emerged under N₀PM₃₀P₉₀ and N₆₀PM₀P₀ treatments, respectively. 2) Treatments containing PM₃₀ generally aided increase in the seed weight per inflorescence with maximum (29%) found in N₀PM₃₀P₉₀. Treatments containing N₆₀ decreased the seed number, seed weight per inflorescence, 1000-kernel weight and numbers of reproductive shoots by range of 12%–18%. The maximum (50%) reduction of seed number per square meter was found under treatment of N₆₀PM₀P₀. The interaction of N and PM reduced the numbers of reproductive shoots and the seed number per square meter by 24% and 34%, respectively. With increase of P rate, the seed weight per inflorescence increased the most, followed by the seed number per inflorescence and the seed number per square meter. The index values of yield traits in P₇₅ and P₉₀ were higher than those in P₀ and P₆₀. 3) The correlation coefficients of the seed yields (seed number per square meter by 0.847; numbers of reproductive shoots by 0.822 and seed weight per inflorescence by 0.767) were high. According to the path analysis, effects of reproductive shoot number (0.637) and the seed weight per inflorescence (0.518) were direct and the most important. 4) With increase of P rates, the partial productivity of P fertilizer decreased significantly while the agronomic efficiency was not significantly different. N₆₀ significantly reduced the partial productivity and agronomic efficiency of P.

Conclusions The seed yield of F. kryloviana is greatly affected by nitrogen and phosphorus fertilizers, and less by super absorbent pdymer. Application of 60 kg/hm² of nitrogen decreases the seed yields while P fertilizer increases the productive branches, seed weight per inflorescence, and the seed yields. Therefore, application of low nitrogen (N₆₀) or super absorbent polymer alone are proved unproductive, but combined application of polymers (30 kg/hm²) and high phosphorus (P 75 kg/hm²) fertilizer rate promotes efficient seed production of F. kryloviana.