60 Generating Economic Cycles 



saved, and beyond that a continually decreasing 

 amount, though at fifteen years still nearly one-half 

 is saved. . . . Thus while for eight, nine, and ten- 

 year periods the process saves about 80 per cent of the 

 coefficient, it falls to 60 per cent on the one side for 

 six years, and to 45 per cent on the other for sixteen 

 years." ^ 



With this large proportion of the amplitudes of pos- 

 sible cycles between six and sixteen years in length 

 preserved by the process of curve smoothing, one would 

 suppose that the author would have been eager to 

 know whether there was any regularity in his data. But 

 Professor Poynting's caution led him to stop at the 

 most interesting phase of his investigation. The 

 Chairman of the meeting of the Statistical Society 

 at which the paper was read. Sir Rawson W. Rawson, 

 was at pains specifically to point out that "Professor 

 Poynting did not suggest that there were periods, or 

 cycles of prices or of anything similar. He had merely 

 adopted a ten year's period for forming an average, 

 in order to establish a curve for the examination and 

 comparison of periodicities of every kind, and did not 

 suggest that there were periods, or cycles of prices, 

 or anything else." ^ 



We now ask the definite question: Is there evidence 

 that cycles in the yield of wheat recurred during the 

 interval from 1760 to 1875? 



By the application of the method of the periodogram 



1 Journal of the Royal Statistical Society, March, 1884, pp. 46, 47. 



2 Ihid., p. 68. 



