5IO Journal of Agricultural Research voi. xviii, no. io 



Applying this rule and then checking the temperatures (not' the nor- 

 mals) as they really occur, the writers found that the actual values 

 differed from the theoretical value by from i° to 15° F. with a mean 

 departure of 7°. A very few errors were more than that. The chances in 

 the arid West are i in 6 that the error will be less than 2°, 2 in 5 that the 

 error will be 5°, i in 4 that it will be 10°, and i in 7 that it will be 15°. 



SUMMARY 



This paper gives an approximate solution of the problem of the deter- 

 mination of the normal temperature at a certain place at an assigned 

 hour of the day on a particular day of the year. 



An equation that shows the seasonal change in temperature is presented 

 and gives the mean daily temperature in terms of the time of year. 

 Another equation gives the percentage of the mean temperature that 

 the temperature of a particular hour of the day is in terms of the time 

 of day. The product of the results of the solution of the equations gives 

 the temperature sought. The equation is of general application inas- 

 much as the first term is the mean annual temperature and the value for 

 the location considered is to be inserted in the equation before using it. 



An arithmetical solution is also presented. The hourly temperatures, 

 expressed as percentages of the mean daily temperatures, are given. 

 The mean monthly temperature for each of the 12 months must be 

 known for the location considered. The mean daily temperature 

 changes approximately one-third of 1° F. a day, except from December 

 15 to January 15 and from July 15 to August 15, when the mean tempera- 

 ture for any day is approximately equal to the mean temperature for 

 the month With this information, the mean daily temperature is 

 readily calculated, and by multiplying this value by the percentage 

 found in the table for the particular hour considered, the desired 

 probable temperature is obtained. 



In the arid West the mean error of the method of determining the 

 normal temperature is very small, but the mean error in predicting the 

 actual temperature is 7° F,, with 60 per cent of the errors less than this 

 amount. These errors are due largely to the abnormal tempeVatures 

 produced by rain and snowstorms. 



