Frosts in New York 523 



night when the evening dew-point and the following minimum were the 

 same. On every night throughout the month the evening dew-point 

 was several degrees above freezing, so that in every case the effect of 

 the liberation of heat from condensation was operative, yet frost occurred 

 several times. On the 25th the dew point was 40°, yet in spite of the 

 heat released by the process of condensation the temperature fell to 25° 

 before morning. On the 30th the dew point at 6 p. m. was 49° and a 

 minimum of 21° was experienced during the night following — a fall of 

 28° from the evening dew-point, notwithstanding the continuous liberation 

 of latent heat. It cannot be doubted that heat was liberated in the 

 process of condensation and that it had a warming effect, but in the 

 cases cited the amount obviously was too small to have an appreciable 

 influence in checking the fall of temperature. 



Professor Willis I. Milham, of Williams College, Williamstown, Mass., 

 conducted a series of observations covering a period of ten years. He 

 found the condition of the atmosphere at Williamstown with respect 

 to moisture almost exactly the reverse of that found by Professor Cox 

 in the cranberry districts of Wisconsin. In the cranberry districts the 

 evening dew-point nearly always was much above the ensuing minimum 

 temperature, so that the effect of the liberation of heat nearly always was 

 operative. At Williamstown the evening dew-point was generally lower 

 than the minimum reached during the following night, and, in consequence, 

 there was no liberation of heat to retard the fall of temperature. Professor 

 Milham summarizes his conclusions as follows: 



1. The cool nights of spring, when a frost might be expected, are very 

 dry and the dew point lies so low that it plays practically no part in 

 determining the minimum temperature. 



2. The amount of the drop from the maximurn to the following mini- 

 mum is very far from constant, even if the characteristics of these nights 

 seem very nearly the same. 



An examination of observations for the months of April, May, Septem- 

 ber, and October for several years, made by cooperative observers in this 

 State, shows that practically the same conditions exist here as were found 

 by Professor Milham at Williamstown. In this State, however, the dew- 

 point observations were made late in the evening, at 9 p. m., but the dif- 

 ference in time is not considered important. If there is any advantage 

 it should be in favor of the later observations. In many cases examined, 

 the air at the time of observation was saturated or nearly so, so that the 

 liberation of heat began as soon as the temperature commenced to fall ; but 

 on clear nights when frost was probable, the air was often so dry that the dew 

 point was several degrees below the freezing point, and therefore no effect 

 from the liberation of heat could be expected. Moreover, the difference 



