Twelve perennial ryegrass (Lolium perenne L.) varieties of different ploidy and maturity classifications were compared under a frequent cutting management in their second harvest year, equivalent to the simulated rotational grazing system employed in UK testing protocols. Varietal differences in canopy structure (proportion of lamina, green leaf mass, sward surface height, extended tiller height, bulk density) and in herbage nutritive value factors (water-soluble carbohydrate content and proportion of linoleic and α-linolenic fatty acids) were assessed and their importance evaluated with reference to total herbage production. Significant variety variation (P<0·001) was recorded in the annual means of all the canopy structure characteristics. Significant differences associated with ploidy were also recorded, with tetraploid varieties having significantly higher values than diploids in most plant characters, indicating better intake characteristics for these grasses. Temporal patterns of variation associated with maturity were also observed in several characters, thus making it impossible to designate a single assessment that would be representative of the annual ranking of varieties. Water-soluble carbohydrate concentration differed significantly (P<0·001) between varieties and although the tetraploids tended to have high contents, the highest value of all was recorded in a diploid variety, which had been selectively bred for this trait. The varieties did not differ in total lipid content but there were significant differences in the proportion of linoleic acid between varieties (P<0·001) while the proportion of α-linolenic acid differed between varieties (P<0·001), ploidy (P<0·001) and maturity (P<0·05) classes.

Overall evaluation of the extensive variety variation highlighted the need for better quantification of animal responses to differences of these magnitudes, before the high workload of including them in routine variety testing protocols could be justified. Potential for breeding improvement in these factors was also indicated and the future prospects for their use in farmer decision support systems was considered.