A statistical study of results from experimental plots of Marquis wheat grown annually (though not on the same soil) at five points in Saskatchewan and Alberta indicates that there was a significant negative correlation between the amount of rainfall during the growing season and the nitrogen content of wheat. The main effect of rainfall was exerted during May and June. The data do not justify the conclusion that the amount of rain falling in July or August, or the amount of pre-seasonal precipitation, modified the nitrogen content significantly. Mean maximum temperature for July or August failed to show a significant correlation with nitrogen content, but may not be a satisfactory measure of the temperature conditions actually experienced by the crop.It is suggested that the preponderating effect of early rainfall may be due to the fact that it stimulates tillering and vegetative development generally. The available nitrogen must thus be distributed amongst an increased number of culms, whilst at the same time the total leaf area devoted to the production of carbohydrates is augmented.