42 



CLIMATIC CYCLES AND TREE GROWTH 



The multiple plot from which the periodogram was made was itself 

 laborious to make, and a year's thought was put on the problem of producing 

 it automatically (1913-14). Such diagram must consist of duplicate solid 

 plots of the data, placed one below the other at equal distances and with 

 equal horizontal off-sets from each to the next. A solid plot illuminated 

 (later called cycleplot) may be seen in Plate 14, general view of the 

 cy olograph. The first bit of progress was obtained by putting on an 

 opaque background a transparent solid plot inclined at 45° slant in a window 

 of a well-darkened room. Then a camera was mounted on a spectrometer 

 plate and rotated a small angle each time between exposures. This gave a 

 fixed multiple plot as reproduced in Plate 13A, and incidentally showed a 

 20-plus year cycle in Arizona trees. It was recognized at once that this 



8 



10 



22 



24 



12 14 16 18 20 



Years 



Fig. 21 — Schuster's periodogram of sunspot numbers. Note: See also Stumpff's 

 periodogram computed by Schuster's method, reproduced in figure 55. 



told a story that could not be found in the periodogram, for it gave not only 



periods but their changes as time passed, -f-, and for a time it was called the 



dt 



"differential pattern." 1 This was still not a satisfactory solution and the 

 transparent plot was the only part that survived to become an essential part 

 of the final instrument. 



The cyclogram is a chrono-periodogram or a periodogram differentiated 

 with respect to time. The cyclogram was produced mechanically in Decem- 

 ber 1914 as follows. If a transparent plot is photographed with a camera 

 having a positive cylindrical lens with its axis perpendicular to the direction 

 of the plot, the resulting image consists of a bank of parallel vertical bands, 

 each band receiving its light from a corresponding maximum in the curve 



1 See Astroph. Jour., April 1915, p. 173. 



