524 



Bulletin 316 



between the dew point at 9 p. m. and the ensuing minimum varied greatly. 

 At Alfred, N. Y., in April, 1892, the difference ranged from 0° to 21°; on 

 fourteen days the 9 p. m. dew-point was higher than the following minimum 

 and on fifteen days it was lower. The following table of days when the 

 minimum dropped to freezing or below at Alfred in April, 1892, will serve 

 to illustrate the variation and is considered typical of the conditions found 

 elsewhere : 



It is noticed that on the first three dates the ensuing minimum was much 

 below the evening dew-point, and, although the liberation of heat must 

 have been continuous, the fall of temperature apparently was checked 

 but little if any. In the two remaining instances the minimum did not 

 fall to the evening dew-point by a margin of 8 , which means that no heat 

 was liberated and therefore no check was offered to the fall of tempera- 

 ture. In three of the five cases the cause that was expected to check the 

 fall of temperature was operative, but apparently ineffectual, while in 

 two cases it was not in operation even. 



From the foregoing, it is apparent that this method of determining the 

 probability and degree of frost is by no means so reliable or so general 

 in its application as has been supposed heretofore. 



For the farmer who is prepared to make practical use of a frost warning, 

 the forecasts issued by the Weather Bureau should receive first considera- 

 tion because they may be obtained early in the day, before it is possible 

 to obtain any reliable indications from local observations as to the prob- 

 ability of frost. But when the warnings issued by the Weather Bureau 

 cannot be obtained and the farmer must rely on himself, there are no 

 instrumental readings that will take the place of a careful observation 

 of the condition of the sky, the direction and force of the wind, and the 

 trend of the temperature. 



