NO. 10 SOLAR RADIATION AND WEATHER STUDIES ABBOT 57 



In the same journal " I have used Saville's data to indicate evidence 

 for a periodicity of 22| years. I now incline to prefer 23 years, and 

 have some reason to trace a periodicity of 46 years as well as one 

 of 23 years. Accordingly, I have rearranged the data, omitting deci- 

 mals of percentages, as given in table 10. 



These data are also shown graphically in figure 24. The first cycle 

 of 23 years is discordant. It is, indeed, almost the exact inversion 

 of cycle II. The latter is shown inverted by the dotted line on cycle I. 

 A similar, though less complete inversion occurs with cycle VII. For- 

 tunately, the cycles are in almost exact step with the important date, 

 January 181 9. This adds interest to these inversions, which, as we 

 have seen, are apt to occur at integralmultiples of 11^ years measured 

 from 1819. Noting the considerable similarity of cycles III, V, VII 

 as forming one group, and cycles II, IV, VI, VIII as forming another, 

 I have plotted in curve IX the mean of groups III, V, VII, omitting I. 

 In curve X, I have plotted the mean of groups II, IV, VI, VIII. Al- 

 though both curves IX and X agree in many particulars, and both 

 show a marked maximum at about the thirteenth year, they also tend 

 to show opposition in some minor features, of the type which I have 

 hitherto called, to give it a name, "right- and left-handedness." This 

 tendency is apparent even in the individual 23-year cycles, for they 

 show alternately the " left " and " right " tendency, corresponding to 

 a 46-year period superposed on one of 23 years. The range from the 

 first to the thirteenth year in the mean of group II, IV, VI, VIII is 

 4-18 percent, and in group III, V, VII, -I-9 percent. Having com- 

 pleted cycle VII in the year 1933, and assuming that the average 

 march shown by group III, V, VII will now take place, we may expect 

 nearly 10 percent more annual precipitation in Southern New England 

 about 1945-1946 than in 1934. Should group II, IV, VI, VIII prove 

 the more representative, then the precipitation about 1945-1946 would 

 be nearly 20 percent above that of 1934. 



19. A Lake Level Test of the 23-YEAR Hypothesis 



By courtesy of the United States War Department, Corps of Engi- 

 neers, a set of charts of the levels of the Great Lakes was obtained. 

 These charts were cut and pasted by the present author so as to present 

 23-year intervals superposed. These charts all began with the year 

 i860. R. E. Horton, hydraulic engineer, was good enough to send 

 me additional data covering nearly completely the 23-year period 1835 

 to 1859. This furnished valuable additional evidence. 



Figures 25 and 26 show these data on lake levels. Figure 25 gives 

 original data for Lake Ontario. Figure 26 gives the march of yearly 



"Quart. Journ. Roy. Meteor. Soc, vol. 61, pp. 90-9-2, I935- 



