NO. 4 sun's variation and temperatures — ABBOT 7 



have just been spoken of do occur. A cursory inspection did not in- 

 dicate any law governing these phase shiftings. They did not seem 

 to be correlated with any suggested variable, such, for instance, as 

 the sunspot cycle. I reserve the matter, though, for later consideration. 



4. GRAPHICAL ILLUSTRATIONS 



To show these results more clearly, I give in figure 3 all the plots 

 similar to figure 2, derived from Washington temperature records of 

 October of all years, 1910 to 1945. It will be obvious to all who ex- 

 amine these results for the years 1910, 1912, 1914, 1915, 1917, 1921, 

 1922, 1924, 1926, 1928, 1929, 1930, 1931, 1932, 1934, 1935, 1936, 

 1937, 1938, 1939, 1940, 1941, 1942, 1943, and 1944 that, though in- 

 dividually the curves represent means of only four or five recurrences 

 of the period in each of those years, and cannot be expected to be 

 smooth, they are perfect enough to indicate it clearly as a real and 

 important phenomenon of meteorology. It is indeed surprising, in 

 view of the opinion which has so long prevailed (namely, that the 

 details of changes in weather are attributable almost wholly to ter- 

 restrial influences) that this solar influence should stand out so 

 strongly against all terrestrial competition. 



The displacements of phase, above referred to, from year to year, 

 but the usual approximate constancy of phase in any month within a 

 given year, may be realized by looking down any vertical column of 

 figure 3. Compare, for instance, in the first column the results for 

 the years 1910, 1914, 1922, 1926, 1930, 1934, 1938, and 1942. The 

 shiftings of phase are obvious. And yet see how the curves for 1914, 

 1918, 1922, 1926, and 1942 covering 28 years altogether, find them- 

 selves in almost identical phases, verifying the accuracy of the average 

 period, 6.6456 days. Results of other years could be added to this list 

 of curves in similar phase from other columns of figure 3, as, for in- 

 stance, 191 1, 1931, 1935; 1924, 1936; 1913, 1917, 1921, 1929, and 

 1937; in all of which years the phases agree closely with the group 

 beginning with 1914, cited above from column i. 



The amplitudes in figure 3 are of especial interest. They range 

 from less than 2° F. in 1933 to 12° F. in 1915 and 1943. As I shall 

 show later, this great range of amplitudes in the temperature effects 

 of this periodic solar variation does not seem to be associated with 

 appreciable change of amplitudes in the solar-constant variations. 

 This tends to indicate that the temperature effect is not a direct heat 

 phenomenon. 



