CYCLOGRAM ANALYSIS 43 



and therefore its light intensity is proportional to the height of that maxi- 

 mum. This pattern was first called a "sweep" (Plate 13B, like the effect of 

 a broom on a sanded surface) and later it was called the cylindrical pattern. 

 If we cut across this cylindrical pattern in any direction with a straight line 

 — for example, with a narrow transparent line cut in opaque paper — we find 

 coming through this line a complete set of data, taking light intensities for 

 ordinates. If now we take a large number of such transparent lines parallel 

 to each other and equidistant, we shall find a perfect multiple plot coming 

 through. Each separate line contains the full set of data, as far as its length 

 permits, is parallel to its neighbors and presents equal offsets each to the 

 next, throughout the whole pattern. 



So the basic parts of the cyclograph were established as a cylindrical lens 

 in a camera and an analyzing plate made of equally spaced transparent lines 

 placed just in front of the photographic plate. In order to read the cycles 

 from a pattern so produced, it was necessary to measure the time scale of 

 the image and the effective spacing of the analyzing lines and the actual 

 angle between lines and bands. This was done by extra lenses and a small 

 filar position micrometer. It was a real task to get the cycle length. 



The trouble with any given cyclogram such as those produced in this 

 way is that only a small range of cycle lengths could be found in it and there- 

 fore many other photographs had to be made. The range can be increased 

 in several ways. Two of them have been tried and discarded; namely, rotat- 

 ing the analyzing plate and changing the spacing between its lines. The 

 former added very little range and made the computation of cycle length a 

 more difficult matter. We have called the latter a variable grating cyclo- 

 graph. It falls far short of the method finally adopted in flexibility and pro- 

 duces confusion between fundamentals and harmonics. 



The simple method finally adopted has all points in its favor. It pro- 

 duces results by changing the effective distance from the camera to the plot 

 that is being analyzed, thus altering the scale of the image that passes through 

 the analyzing plate. The change in scale becomes in effect a change in the 

 cycle from which residuals are automatically plotted as described above in 

 connection with the cyclogram plot. This change of scale is now done (since 

 1920) by altering the distance of a secondary mirror (see fig. 22) interposed 

 between the transparent curve and the camera lens. Motion of this mirror 

 by diminishing or increasing its distance from the camera changes in inverse 

 proportion the size of the cylindrical pattern that falls upon the lines of the 

 analyzing plate. The position of the mirror that by trial brings each definite 

 cycle length into horizontal position (compared to a stationary thread in 

 the field of view) is marked on its track and the instrument becomes in all 

 respects a direct reading cycloscope capable of instant adaptation over a 

 very large range. 



The general arrangement of the cyclograph is shown in figure 22. The 

 light from the illuminated curve, or cycleplot, is reflected in a movable mirror ; 

 it then returns to the camera which has in effect a positive cylindrical lens 



