Correlations between visual biomass scores and forage yield in space planted red clover (Trifolium pratense L.) breeding nurseries

Red clover (Trifolium pratense L.) forage yield remains a prime breeding target for improved variety development. In a world of decreasing forage legume breeding resources, rapidly and cheaply phenotyping plants for the highly quantitative trait of forage yield is vital. Many red clover selection programs are based on space planted nurseries. The objectives of this study were to determine: (1) the accuracy of visual forage yield scores in predicting actual forage yield; (2) the nature of the relationship between visual scores and actual measurements; and (3) The repeatability of visual scores between different evaluators. Twenty-seven halfsib families were transplanted at two locations in three replicates of six plant plots. Individual plant fresh weights and visual scores, by two evaluators, for forage yield were taken. On an individual plant basis visual forage yield scores showed an exponential relationship with actual fresh weights. Individual plant visual scores were very accurate with a pseudo-R (2) of 0.79 observed for the exponential model. On an entry mean basis using a linear model, visual scores could explain 90% of the variation of actual fresh weights. Agreement among evaluators scoring the same plants was very high with coefficients of determination at 0.84 for individual plants and as high as 0.96 on an entry mean basis. This study suggests that visual scores of plants in space planted red clover breeding nurseries are basically as accurate as measuring actual yields and that plants can be consistently scored the same by different evaluators.