DISCONTINUOUS PERIOD IN CYCLOGRAM ANALYSIS 61 



and to indications of longer rotation periods (synodic) than 27 days. These 

 features of solar records are examined below by the cyclogram process. 



Significance of Conservation — A better understanding will be reached of 

 the analytical work of Bartels and others if we give some attention to the 

 physical meaning of conservation in certain solar problems. We can readily 

 see that daily values of sunspot data have a large conservation factor both 

 statistically and physically; for nearly all spots or groups would be visible 

 continuously for 10 or 12 days. Daily results then would amount to a run- 

 ning mean of 10 or 12. The monthly number, therefore, is a good figure for 

 general use. However, daily observations on spots within definite longitude 

 zones some 10° in width would be very desirable in order to secure closer 

 correlation in time with magnetic data. This need could be worked out in 

 the frequent photographs of the sun taken at many observatories. 



The magnetic character figure is believed to be more nearly related to 

 the solar activity near the sun's meridian, and therefore it perhaps involves 

 an equivalent running mean of only 1 to 4 days. Hence, in our accompanying 

 analyses of this solar record we can use 3-day means and get probably all 

 the "resolving power" which the data can yield. 



Bartels' estimates of quasi-persistence in the paper of 1935 were made 

 upon a similarity from rotation to rotation of the sun and temporary persist- 

 ence of the source areas of magnetic force in solar longitudes. His value of 

 v is 3.0 rotations, which means that the quasi-persistent localizing of mag- 

 netic source areas averages 3.0 rotations or between 80 and 90 days. 



But it is easy to see that at tinfes such as sunspot maxima we may be 

 averaging together outbursts of magnetic activity from many disconnected 

 localities in solar longitude, each of which complicates the problem of real 

 persistence in longitude. Spot groups, and probably magnetic source areas 

 also, cover at maximum a considerable range in solar latitude, and so may 

 exhibit a range in rotation periods because of the important fact that different 

 solar latitudes have different times of rotation. Hence, we have recourse 

 to the cyclogram process which helps to unscramble the various cycles near 

 any desired setting of the instrument. On the first attempt, covering data 

 near the 1923 sunspot minimum, we found so much more evidence of stability 

 in longitude than expected that we seemed to be dealing almost with a new 

 group of phenomena. Cyclograms of the magnetic character figure C are 

 given herewith in Plate 19B. As the necessary discussion is somewhat ex- 

 tended, further description will be found in the next chapter. We find in 

 the above discussion an illustration of how a search into the significance of 

 statistical conservation led to a better understanding of solar problems. 



STUMPFF'S PERIODOGRAPH 



Dr. Karl Stump ff's researches are presented in Analyse periodischer 

 Vorgdnge (Berlin, Gebruder Borntraeger, 1927 ; the method was first described 

 in Astronomische Nachrichten, BD. 223, Nr. 12, 1924). In this book there 

 will be found theory and description of the periodogram and various mechani- 



