KITE WORK OF THE WEATHER BUREAU 59 



The central fact of importance which the study of inversions de- 

 veloped was that they were most pronounced with the radiation of 

 relatively warm southeast to southwest winds, the marked cold at the 

 surface and the higher warm air presenting a marked contrast, which 

 was not so apparent when the upper air blew from a colder northerly 

 direction. 



Inversions were sometimes caused by cloud formation in the early 

 morning. Below the clouds there would be little or no temperature 

 change, Avhile above there would be an increase on account of the di- 

 rect action of the sun's rays. On October 1, 1898, a marked case of 

 this character occurred at Dodge City, there being a rise in tempera- 

 ture of 11.5 degrees within a few minutes after the kite emerged from 

 the upper surface of the clouds. 



The most remarkable instances of inversion were found at Duluth, 

 Minn. Nearly one-half of them occurred in the late morning and 

 earl}^ afternoon, during cloudy weather, and were due to the easterly 

 surface winds from Lake Superior, the warming effect of these winds 

 being sensible at times to the height of nearly 6,000 feet. During 

 these inversions the direction of the upper air current would be almost 

 or entirely diametrically opposite to that at the surface. On June 3, 

 1898, there was an inversion of 2 degrees at an elevation of 5,372 feet, 

 and on September 20, 1898, one of 1 degree at an elevation of 5,714 

 feet. 



The effect of the presence of clouds upon the temperature gradient 

 was quite uniform, though not at all times decided. In a great ma- 

 jorit}^ of instances there was a decrease in the rate of temperature 

 fall, frequently amounting to a complete arrest, and less frequent!}^ to 

 an inversion. After the kite emerged from the clouds the rate of 

 temi)erature change would be diminished. 



In a number of instances tlie clouds appeared to liave no effect 

 whatever, and in a few rare ones there was a fall in the tein})erature 

 as the kite came in contact with the clouds. 



As a rule, the temperature gradients were greater in clear than in 

 cloudy weather, although exceptions were quite numerous. 



The relative humidities at and above the surface of the earth dif- 

 fered but little, and, except at 2,000 and 8,000 feet, the upper air per- 

 centages were the lower. The mean results obtained from all the 

 observations were 60 per cent at the surface and 58 per cent above, a 

 difference of 2 per cent. There wore, however, some marked differ- 

 ences at individual stations, viz., Washington, 14 i)ercent; Omaha, 



