A model to predict wheat growth stage is briefly described. It is based on prediction of the number of emerged leaves and the final number of leaves on the main shoot, and the co-ordination between leaf emergence and apex development, including stem elongation. The input variables are daily maximum and minimum temperatures, date of sowing and site latitude, from which thermal time, vernalization and daylength are calculated.

Selected growth stages were predicted for six sites in each of three growing seasons. The differences between observations made by independent observers and predictions were mostly 7 days or less but in three site–season combinations the average difference was >10 days. Observer errors were implicated and examined, but it is concluded that the prediction scheme must also have been partly responsible for the discrepancies.