164 



CLIMATIC CYCLES AND TREE GROWTH 



AN AUTOMATIC OPTICAL PERIODOGRAPH l 



VARIABLE STAR TESTS 



In all the previous work which had been done with the periodograph, a 

 continuous series of observations was used. The test of discontinuous ob- 

 servations was made on variable stars. In order to avoid prejudices and to 

 get a fair sample of variable stars, my assistant selected a certain number 

 from the table given in volume 37, in the Annals of the Harvard College 

 Observatory, page 202. They comprised the 10th, 20th, and 30th star, etc., 

 in the list. The accompanying table will give the results of the test and the 

 time taken to obtain each one. One must remember, of course, that this 

 was the first application of that instrument to these interrupted periods, and 

 one can say confidently that the average time to solve a variable star period 

 (after the plot is made) would be five or ten minutes. The columns in this 

 table marked "Curve Period" and "Computed Period" are from Harvard 

 Observatory Annals, volume 57, page 182. The "Curve Period" was ob- 

 tained directly from a plot. The "Computed Period" was obtained after 

 an application of the method of least squares. It is seen that the periodo- 

 graph results agree very well with the least square solution. The solution 

 of X Virginis, as seen in the table, was obtained after a very long study. The 

 period of 52 days seems applicable if the observations of J. D., '2815, are 

 omitted. 



The method by which the solution was obtained consisted in placing a 

 strip of slightly tinted yellow glass behind the minima of the curve in which 

 all the observations were plotted by their distance from a mean line. This 

 made a mixture of yellow and white spots on the differential pattern. When 

 this mixture of spots resolved itself into alternate yellow and white rows, 

 then a period was indicated. This could be done very readily. In the period- 

 ogram the alternate beads or corrugations come down in different colors. 

 Such a periodogram of R Arietis is shown in Plate 12b. 2 The marginal marks 



1 Presented at astronomical meetings, Pasadena, June 20. 1919, and Ann Arbor, 

 September 3, 1919. See abstract in Pub. Ast. Soc. Pac, Vol. XXXI, 188, June 1919. 

 Introduction omitted here. 



2 The reference to plates applies to this book but the photographs here are repro- 

 ductions from the original article. 



