An efficient sampling scheme for monitoring flowering of Actinidia chinensis

Abstract

Flowering data from four vines of Actinidia chinensis (Planch.) var. chinensis ’Hort16A’ and three vines of A. chinensis ’37–3‐18A’ ('18A') were used to develop a robust and efficient scheme to monitor flowering. The frequency distribution of flowering for ‘Hort16A’ flowers was normal. Although the distribution of flowering for ‘18A’ terminal flowers was not normal, assuming normality produced estimates of the date and standard deviation of mid bloom within1 day of the true value. Three schemes for selecting shoots for monitoring flowering in A. chinensis vines were tested. The accuracy of flowering estimates was proportional to the number of shoots monitored and although totally random selection of shoots was slightly more accurate, stratified random and systematic sampling was more practical. Flowering can be monitored at 4‐day, rather than 2‐day intervals without affecting the accuracy of estimates. For the two cultivars we tested, monitoring flowering on 60 shoots selected using the stratified random and systematic scheme will result in estimates of mid bloom within ±1 day of the true value on 80–95% of occasions. This corresponds to a sampling intensity of c. 3–4% of the total number of terminal flowers on the vine, or 140–290 flowers/population.