92 A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



While these inferences are essentially tmcertain and might even seem 

 attributable to the time coefificient, /3, neglected, the observations do not war- 

 rant svich exception. Thus at 6p = 22 cm. and St = /\, minutes, the total time 

 consumed was 36 minutes, whereas in the case of (5^ = 8.5 cm. it was only 32 

 minutes for 6t = 2 minutes, and only 21 minutes for 6t=i minute. The total 

 amount of time loss must therefore actually have been smaller in the latter cases. 



OTHER CAUSES OF CHANGE IN THE TYPES OF CORONAS. 



30. Thickness of cloud layer. — In passing from the cubical to the long 

 apparatus, the thickness of the cloud la3'er decreased from 20 to 16 cm. On 

 ti])ping the former vessel on its trunnions, and looking through the chamber 

 diagonally, the thickness of cloud layer could be increased to 30 cm. In all 

 these cases, the type of corona and its diameter showed no appreciable change. 

 If symmetr}^ about the center is maintained, the effect of the thickness of the 

 cloud layer seems thus to be absent. 



31. Obliquity of diffraction. — When the direction of the diffracted ray 

 differs by a small angle from the direction of the nonnal ray of light, the type 

 of corona does not change as to color of annuli. If, however, the angle of de- 

 viation due to diffraction is ver)' large, the corona may change in character. 

 Thus for great obliquity the orange-red corona was found to change into the 

 preceding green corona. In other words, a larger particle on very oblique 

 diffraction may produce the same corona as a smaller particle on less oblique 

 diffraction. 



32. Effect of wave length. — Since at Sp=i'j cm., a = 73\, and therefore 

 n = 6ms^/7ra^ = 6 ins^/7tV{'jT,)^=22S^/io"\% the equation for different colors 

 would be (denoting the colors by subscripts) n = 6'js^ = g']s';= io-js',= ii&Sy 

 = 15451= 22 i5b=36o5j, etc., while for other pressure differences the same 

 relations would be preserved. 



Testing this by the very full series in the second part of table 15, and re- 

 membering that the measurement is made to beyond the first ring, the fol- 

 lowing data may be deduced: 



If the w r g' corona is taken as correct, the w g b p corona will also be, as well 

 as the w oy (p) g corona. Btit in the latter case the narrow pui-plish ring which 

 intervenes before the green and the general difficulty of defining the mixed 

 colors of the second ring makes inferences of the present kind precarious. 

 Nevertheless it is probable that the rather sudden transition of green to blue in 

 the colors of the second ring is associated with the underlying cause of peri- 

 odicity. Examples of this kind might be multiplied. 



