The method developed by Fisher (5) for examining the influence of rainfall on the yield of wheat at Rothamsted has formed the basis for a number of investigations: Davis and Harrell (3) examined the effects of rainfall and temperature on corn yield at various locations in the corn belt; Hopkins (7) made a similar study of the influence of weather on wheat yield in Western Canada; Schumacher (8), the effect of climate on timber growth; while Cochran (2), Davis and Pallesen (4), Tippett (10), and others have extended the applications. These studies have been stimulated by the demand for more reliable methods of forecasting and estimating crop yields as well as by scientific interest in them.

The application of Fisher's device gives a regression curve which shows the effect on yield of a unit change in a given meteorological element at any time during the growing season. For example, suppose one were investigating the relationship between the amount and distribution of rainfall and the yield of corn. The ordinate of a point on the regression curve corresponding to any particular time represents the mean effect, in bushels per acre, of an inch of rain at that time. If data are available, curves for each of several plots of different soil type or plots receiving different fertilizer treatments can be calculated. An examination of the curves might reveal some important facts about the manner in which the crop's response to weather changes with soil type or fertilizer treatment.