48 CLIMATIC CYCLES AND TREE GROWTH 



ment and during 1934 and 1935 was engaged in formal tests of the degree of 

 accuracy found in its use. His description of his work is found in the appen- 

 dix to which reference is made for details. The enormous amount of analyz- 

 ing that has been done upon a thousand or more cycleplots for many different 

 purposes offered excellent opportunity for comparisons of results. This 

 meant largely the repetition by him of analyses made originally by the writer, 

 together with many additional tests to bring out the comparisons. He found 

 that we might expect a difference of about one per cent in cycle length as the 

 mirror moved out or in. The "unknown scale" test mentioned above 

 showed differences in result of 1.9 to 2.5 per cent in cycle length; different 

 observers obtained results differing 2.5 to 3.0 per cent. Inversion of curves 

 might produce changed reading by 2.0 to 3.0 per cent. Discrepancies in read- 

 ings occurred almost altogether in very weak cycles recognized at the time of 

 observation as of small weight. Tests were made on synthetic plots prepared 

 in a manner totally unknown to the observer. The analyses of these were 

 satisfactory. I feel, therefore, that the cycle results obtained with the cyclo- 

 graph by Mr. Schulman and myself may be accepted with confidence. 



GENERAL CYCLOGRAM INTERPRETATION 



The cyclogram as previously described is a rectangular plot with a general 

 time scale usually placed horizontal and progressing from left to right, and 

 a differential time scale in the vertical. The analyzing lines showing in 

 the cyclogram are slightly inclined to the vertical, usually about 17°. It is 

 better to set them a little out of focus so as not to catch the eye, for a cycle 

 is indicated by an alignment of maximal images wholly or partly across the 

 cyclogram and not in the direction of the analyzing lines. A horizontal 

 thread mounted close to the analyzing plate shows in all cyclograms and 

 serves as a reference line for direction of rows of maxima both in calibrating 

 the instrument and for getting correct cycle settings. 



Our cyclograms in published prints have nearly always had the analyzing 

 lines inclined up to the left. The relation of this to cycle length is easily seen 

 by brief study of a cyclogram. A cylindrical pattern of the original data as 

 projected upon fixed analyzing lines has a definite and single time interval 

 across it from left to right whether taken in a horizontal direction or at any 

 inclination. In the horizontal line this time interval is divided by the 

 analyzing lines into a definite number of equal parts, equal to the number of 

 lines that intersect the pattern along the line of the thread. One can count 

 them if desired and divide their number into the time interval and find the 

 cycle represented by the horizontal line. This should check with the cycle 

 setting of the instrument. If the analyzing lines incline up to the left, then a 

 cycle alignment inclining down to the left will take in more of them (than 

 when horizontal) and be subdivided into smaller parts, and thus the cycle 

 length in this row is less than it would be if horizontal. In the same way, an 

 inclination of a cycle row up to the left will cover fewer of the analyzing lines, 

 and so an increased cycle length is indicated. 



