High persistence and seed yield are important targets in white clover
breeding but their assessment is expensive. The efficiency of indirect
selection for these traits performed on characters recorded during a
medium-term dry matter (DM) yield evaluation was verified for large-le
aved populations in separate experiments for medium-term (three-year)
and long-term (fourth and fifth year) DM yield (Trial 1) and seed yiel
d (Trial 2). The best results were provided by stolen density observed
at the end of the third cropping year for prediction of long-term yie
ld, and by the product of number of heads times number of florets per
head assessed prior to mowing in Trial 1 for prediction of seed yield
in Trial 2. Indirect selection was about as efficient as direct select
ion in these cases. Predictions of persistence from three-year or thir
d year DM yield, and seed yield from single seed yield components in T
rial 1 were less accurate.