In breeding population varieties, selection based on per se performance of spaced plants is less effective than selection for performance of their offspring would be. "is holds especially true for complex traits like biomass production. Nevertheless, many breeders use ratings of biomass from spaced plants as a selection criterion. Based on historical data from the Agroscope forage crop breeding programme, we analysed correlations between per se performance of spaced plants in a polycross and performance of their half-sib progeny as assessed in seeded rows. Correlations between plant per se and offspring vigour (rP,O), showed high variation among the different polycrosses analysed. Values of rP,O ranged from positive to negative, with mean values being close to zero for all species. A significance at the 5% level could not be attested for most of the parent-offspring comparisons, partly attributed to a low number of plants per polycross. "e inconsistent correlation patterns are likely caused by data character (scoring data), comparison levels (single plants vs swards), different plant materials and the nature of the trait itself (vigour as an omnium-gatherum of plant height and amount of leaves and stems). Breeders should not rely too much on selecting spaced plants based on their vigour