Jan., 1915.] Predicting AH iiimum Temperatures. 407 



It seems important, however, that some plan be devised 

 whereby a fruit man not in tcuch with the Weather Bureau and 

 its maps and forecasts can closely estimate the probable lowest 

 temperature at critical periods. 



In investigating this point in connection with our special fruit 

 service in Ohio, we have found that the prediction of the minimum 

 temperature from the time of the a^-crage afternoon median 

 temperature gives very close results. 



Figure 1 is a copy of the temperature record made b}^ a self 

 recording thermometer at Delaware, Ohio, from May 11 to May 

 17, 1914. This shows a rapid fall in temperature beginning at 

 about 10 a. m. of the 12th. This was due to a shift of the wind 

 to northwesterly and the small change that occurred in the tem- 

 perature from the afternoon of the 12th to the early morning of 

 the 14th was because of continued northerly winds and rainy 

 weather. 



Beginning on the 14th, however, and continuing through the 

 balance of this week and most of the following week, there was a 

 period of clear and comparatively still weather when an area of 

 high barometer pressure was centered over this district. Under 

 these conditions the temperature rises high during the daytime 

 under strong sunshine, and then falls quite low at night under 

 free radiation. It is under conditions of this kind that frosts 

 may be expected in the spring and fall. It will be noticed that 

 the rise in temperature is rapid in the early forenoon and that the 

 thennograph curve has a decidedly convex shape. 



The highest temperature will be reached at about 3 o'clock 

 in the afternoon. The temperature will fall slowly for two or 

 three hours, then there will be a rapid fall in the evening and 

 a slower fall until the lowest point just before sunrise. The 

 afternoon curve has a decided concave shape. There is a marked 

 similarity in the curves during these days when frosts threaten. 



This being true the question was raised whether the half way 

 point in the temperature fall from the maximum of one day to 

 the minimum of the next morning might not occur at about 

 the saine time each evening. 



A study of available records showed that in May the half 

 way temperature occurred at Delaware on an average at 7:36 

 p. m. and that the variations on either side of this time was less 

 than 20 minutes in either 1913 or 1914. 



For example the highest temperature at Delaware on May 14, 

 1914, as shown by the thermograph record in Figure 1, was 65. 

 The temperature at 7 :36 p. m., the average time of the median, was 

 51. vSubtracting this from the maximum leaves 14. If we take 

 14 from 51 then we shall have 37 as the predicted minimum 

 temperature during the coining night, by this method. The 

 lowest temperature that actually occun-ed was 36 or only 1 lower 

 than estimated. 



