C. JBarus — Method for the Observation of Coronas. 281 



It is exceedingly difficult to account for this difference of 

 behavior. As the abscissas s — atynr/ftm) 1 '" 3 , or s a nf /3 , and the 

 ordinates, s (observed), are independent of each other, the 

 equality of the two values of s will in a measure check the 

 work apart from the constant a which determines n x . This is 

 actually the case for the lower series of coronas below s = 10, 

 which are used in practice. It is the observational value of 

 the aperture of the given coronas which varies. Yery proba- 

 bly mixed coronas are being observed. To this must be added 

 the subjective error or personal equation which enters into the 

 determination of contacts. Finally the tendency of a corona 

 to shrink at once after the formation of droplets, makes it diffi- 

 cult to catch the time at which coronas should be observed 

 soon enough. Under other circumstances there is even liable 

 to be an oscillation of the coronal aperture in the lapse of 

 time. All these difficulties are accentuated as the coronas 

 become larger, for here not only are the droplets more volatile, 

 but the coronas overlap, and there is an unlooked for tendency 

 for them to flatten at the point of contact. The dark rings 

 are liable to invade the bright. 



The green coronas in the series 1, 2, 3 show the following 

 average values, the diameter cl being given in cms. : 



Ser. Computed Observed Computed Observed 



s 3 



S 2 S3 



s 2 



10 6 d 3 



10 6 c7 2 



10 c d 3 



10 6 <2 2 



1 8 



16 8 



14 



400 



200 



400 



230 



2 8 



14 8 



15 



400 



230 



400 



210 



3 8 



13 8 



13 



400 



250 



400 



260 



Mean values are thus : 















o s = 8-0 





10 6 e£, = 



-400 









s 2 = 14-3 





10 6 < = 



= 230 







;reeing pretty well with 



my 



earlier data, where 









S 3 = 8-1 





10^ = 



= 398 









s„ = 14. 





10U = 



= 228 







The subscripts here show the c} 7 cle to which the corona 

 belongs, the 2d and 3d being producible here. The first cycle 

 is very rarely obtainable ; the fourth and fifth occur best in 

 experiments with atmospheric air, not filtered, the coronas 

 here being very crowded. 



Brown University, Providence, R. I. 



Am. Jour. Sci. — Fourth Series, Vol. XXIV, No. 141. — September, 1907. 

 19 



