Periodic Seeding -56- 



"Just before the start of the periodic seedings, the corre- 

 lation coefficient CC(7) based on the seven average values for 

 the successive days of the week of the 28-day period amounted 

 to only 0.23, but in the next 28-day period the value of CC(7) 

 rose to 0.91. 



"Table 1 gives the average rainfall in inches per station 

 day during 140 days at 20 stations designed as Group A in the 

 Ohio Valley Basin, representative of an area of about 600,000 

 square miles „ The successive rows correspond to five succes- 

 sive 28-day periods. It will be noted that the average rainfall 

 on Monday was 0.272" , whereas on Saturday it was only 0.064" , 

 a ratio of 4.3:1. The next to the last column gives CC(28), the 

 periodic correlation coefficients for each 28 -day period, and 

 the last column gives the phases in the successive periods. 

 Taking the 35 separate values for the 4-week averages given 

 in the table, one gets CC(35) = 0.689 with a phase of 1.60 days. 

 This result is statistically highly significant, 



c er 



'These periodicities in rainfall were evident at almost 

 any set of stations in the northeastern part of the United States. 

 Table 2 gives the rainfall on successive Tuesdays and Saturdays 

 during a 12-week period during the winter of 1949-1950 at Euffalo, 

 Wilkes -Ear re, and Philadelphia. This periodicity is almost the 

 same as that found in the Ohio River Easin but with a one-day 

 phase lag. The striking contrast between the total rains on 

 Tuesdays and Saturdays runs parallel to the total number of 

 days on which rains of 0.1" or more occurred on Tuesdays and 

 on Saturdays. 



"Maps have been prepared giving for 24 successive 28-day 

 periods the distribution of correlation coefficients, CC(28), 

 among 17 subdivisions of the United States, these data being 

 based on daily weather reports of 24-hour rainfall at 160 sta- 

 tions. During the first five 28-day periods there were always 

 several adjacent subdivisions that showed high weekly perio- 

 dicities in rainfall. After May 1950, however, the periodicities 

 became somewhat sporadic, although highly significant perio- 

 dicities over large areas still occurred during more than half 

 of the periods after July 1950. Presumably the large amount of 

 commercial silver-iodide seeding in the western states (not done 

 with a weekly periodicity) masks the effects of the periodic 

 seedings in New Mexico. Ey a map, the areas were shown in 

 which known seeding operations have been carried on in 1951. 



