THE CYCLONIC DISTRIBUTION OF RAINFALL. 7 



enough to secure a fairly representative average. In other words: 

 the number of times precipitation occurred orcloudiness preA'ailed 

 was noted when Davenport was located in any one of the desig- 

 nated tracts with reference to the centre of a low area, and also 

 the total number of times the opposite conditions prevailed. 

 From these two figures percentages were obtained showing the 

 comparative frequency of precipitation and cloudiness in each 

 tract. It will be seen that this is only a very simple method of 

 averaging weather conditions for different parts of an area of low 

 pressure. The results can be plotted on a chart. 



It was found that precipitation is most frequent at Davenport 

 when the station lies in the tract numbered eight, which is on the 

 \yest side of the central low. It was also found that precipitation 

 is infrequent in the region to the southeast of the centre, decreas- 

 ing very rapidly in that direction from the tract numbered eight. 

 From this distribution of pi"ecipitation it is evident that if fore- 

 casts were made on the supposition that precipitation is greatest 

 on the southeast side of the central low, a lai'ge percentage of the 

 predictions would announce the stormy weather ahead of time. 

 For it would often happen that the centre of the low would have 

 to move east some two or three hundred miles before it would 

 bring up that tract, where rains and snow are actually most fre- 

 quent. 



With the aid of some student friends the cyclonic conditions 

 were averaged in a like manner for some more stations, represent- 

 ing four other climatic regions in the United States. It was found 

 expedient to make use of data slightly different from those used 

 in the Davenport cj'clone. Thus we combined the observations 

 taken at Amarillo, Dodge City, Wichita, and Oklahoma during the 

 years 1894: — 1898, obtaining a chart \vhich presumably is charac- 

 teristic for the cyclonic conditions on the southwest plains. Other 

 charts combined into like averages the observations at Helena, 

 Miles City, Leander, and Boise City for 1899; those taken at nine 

 stations in the Upper Missiouri Valley in 1899; and those taken 

 at Detroit and at Buffalo' during the years 1900—1903. 



