86 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 94 



by symbols i and 2. Number 5 is a zigzag line. It represents an inde- 

 terminate form of curve not similar to those represented by i, 2, 3, 

 and 4. It is not intended to imply that curves i and 2 or 3 and 4 are 

 always similar in form as between representations of periodicities of 

 different stations or periodicities of different lengths. It is only im- 

 plied that all curves i and 2 within a single vertical column of the 

 same subfigure are approximately similar though inverted, and all 

 those represented as 3 and 4 within a single vertical column of the 

 same subfigure are approximately similar though inverted. 



Owing to local influences, it was not to be expected that complete 

 harmony would prevail throughout all the subfigures. But if the 

 changes of phase and form in terrestrial periodicities to which ex- 

 tended references have been made, are due to radical changes in the 

 solar radiation, it would naturally be expected that similar mutations 

 of phase and form would tend to occur in all terrestrial periodicities 

 and all stations at about the same time. 



Figure 38 seems to show that on the whole this expectation is fairly 

 supported by the facts. Though exceptions occur, there is a prevailing 

 tendency for inversions to occur in all periodicities and all stations 

 simultaneously. Thus, for illustration, at the years 1841, 1864, and 

 1 910, reversals or at least major modifications of form occurred in 

 nearly all cases, and this also frequently happened at the year 1887. 

 It is believed that the exceptions are neither more numerous nor more 

 radical than might fairly be attributed to local terrestrial influences 

 affecting conditions differently at these widely separated stations. 



If this conclusion is sound, modifications may well be expected from 

 the prediction I have ventured of solar variation for the years 1935, 

 1936. and 1937 as given in figure 7. For on that basis it is very 

 probable that a radical change in the phases or amplitudes of solar 

 variation, or in both, will have occurred about 1934, being 115 years 

 after 1819, and will greatly modify solar variation in subsequent years. 

 But yet this result might not occur, for at several epochs the terrestrial 

 periodicities appear to have continued stability for 23 years or even 

 longer, which might call for a similarly long-lived stability in the solar 

 variation, and no mutation of it in 1934. 



As for the third query, C, let us restrict our investigation to the in- 

 terval 1 920- 1 930, for it is only then that we have actual observations 

 of the amplitudes of the periodicities, both of the solar radiation and 

 the terrestrial temperature. In table 13 I give the amplitudes of the 

 periodicities expressed in percentages of the solar constant (1.94 

 calories per square centimeter per minute) and in percentages of 

 the absolute temperature of the earth, which I take as 290° Centigrade. 



