6 THE CYCLONIC DISTRIBUTION OF RAINFALL. 



of doing this, I made use of a simple devise, which I liave since had 

 the satisfaction of seeing- employed by others By marking off 

 eight radii in four concentric circles I plotted twenty-five areas in 

 a figure, which could he used to represent definite separate tracts 

 in a circular storm. The lengths of the radii of the successive cir- 

 cles had the ratios 1: 4: 7: 10 and were taken to represent the 

 same number of hundreds of miles in a composite cyclone two 

 thousand miles in diameter. The construction will be readily un- 

 derstood from the accompanying figures. The radii were drawn 

 at angles of 45°, but were not extented into the inner circle. The 

 figure was so oriented that the four points of the compass would 

 bisect four alternate octants. There were thus three tracts marked 

 off in each octant outside the smallest circle. With this represent- 

 ing the central region of a cyclone, the figure was used to delimit 

 twenty-five fixed areas inside its extent. Tract "l" thus covers a 

 central circle two hundred miles in diameter in the center of a cy- 

 clone. Tract "2" covers an area extending from one hundred to 

 four hundred miles away from the center to the north, and lying 

 between radii diverging 22}^° on either side. Tract "lO" covers 

 the area between the same two radii at a distance from four hun- 

 dred to seven hundred miles from the centre. Tract "18" lies at a 

 distance fi'om seven hundred to one thousand miles from the 

 centre, and so on, in the other octants. 



My method was then simply to take a sufiicient number of 

 observations on the weather at Davenport, when this station lay 

 in any one of the twenty -five corresponding tracts of an actual 

 cyclone, and to average these for each tract separately and thus 

 obtain for each separate percentages expressing frequency- of cer- 

 tain weather conditions, such as precipitation and cloudiness, re- 

 sultant wind directions, etc. I averaged these elements of the 

 weather, as observed at 8 a m., during a iieriod of about five 

 years, taking the data from the daily weather maps. There were 

 nearly a thousand observations in all. These were distributed 

 ainong the twenty-five tracts somewhat unequally, but it is be- 

 lieved that the number of obser\-ations in each tract was lars^e 



