Scaling issues in the interpretation of dry matter yield differences among perennial ryegrass (Lolium perenne) cultivars


  • David F. Chapman DairyNZ
  • Gerald P. Cosgrove AgResearch
  • Barbara Kuhn-Sherlock DairyNZ
  • David R. Stevens AgResearch
  • Julia M. Lee DairyNZ
  • Laura Rossi DairyNZ



Perennial ryegrass breeding is estimated to be delivering rates of genetic gain in dry matter (DM) yield of 50–60 kg DM/ha per year of breeding effort. These estimates are based on DM yield data from tightly managed small-plot trials which are not necessarily representative of typical farm management. To assess the possible realised (on-farm) DMY gains from breeding, seasonal DMY data were collected contemporaneously at two ‘scales’ of evaluation for eight perennial ryegrass cultivars over a total of 14 seasons (comprising 5 ‘seasons’ in each of three years excluding the first winter) in each of four regions. The scales were denoted ‘small plot’ (SP, 10 m2, perennial ryegrass monocultures only) and ‘large plot’ (LP, 65–90 m2, subject to more intensive dairy cattle grazing and including all combinations of two nitrogen fertiliser rates with or without white clover as treatments). Relationships between DMY measured in SP versus LP were statistically significant (P<0.001) for all combinations of region, N fertiliser level and clover presence/absence, indicating good general agreement in cultivar performance trends. However, the slope of the relationship (range 0.492 to 1.171, mean 0.733) was significantly less than 1 in three-quarters of the region by treatment combinations. The slope was closely and inversely related to the size of the difference in total N supply between and SP (N from fertiliser only) and LP (where N was supplied as fertiliser at a lower rate but included biological fixation): the smaller the difference in N supply between the two ‘scales’, the greater the slope. Estimated realised DMY differences between cultivar ranking positions on New Zealand dairy farms under future projected N fertiliser rates were in the range 0.7–0.8 of those predicted from the SP scale of evaluation.


Download data is not yet available.






Vol 81 (2019)

Most read articles by the same author(s)