G. Bar us — JS^ote on the Constants of Coronas. 327 



In table 1, s denotes the number of the partial exhaustions 

 each of volume ratio, y, and made in succession, t the current 

 time in minutes (the interval being about 3 min. to allow 

 for adjustments and for diffusion), s the chord of the angular 

 aperture, (^, at radius B, so that ^ = 60 sin <^. The eye and 

 source of light were at distances 85 and 250^"" from the inter- 

 vening condensation chamber and the former was focussed for 

 long distances. The pressure and temperature of the atmos- 

 phere were P and ^, and the fixed pressure decrement on 

 exhaustion uniformly hp =1^^^, nearly, so that the precipitate 

 per cub. cm. is m = 4:'f X 10~^g. Measurements were made to 

 the outer edge of the first ring. In the column marked 

 "coronas," the color of the annuli is specified from within 

 outward, using obvious abbreviations. The nucleation is 

 marked n' if computed from the aperture s, standardized with 

 lycopodium, n" if computed from s standardized by subsidence 

 measurements, iT if computed relatively as a geometric pro- 

 gression, n when the latter is reduced as suggested and the 

 absolute values corrected for time and exhaustion losses, etc. 

 The arbitrary initial nucleation is shown under n^, and corre- 

 sponds to 3 — 4:. The other coefiicients are yS, referring to 

 time losses, a referring to exhaustion losses, S referring to 

 subsidence losses. Though n is measured for the partially 

 exhausted receiver, a final correction {1/y) need not be added, 

 for the influx of filtered air leaves the nucleation undistui'bed. 

 The ratio n' /JSf^ 275 ^yiO^ ^""^ y if constructed in charts, 

 shows the wide departure from the constancy which would be 

 anticipated. Diameters of the fog particles are given under d^ 

 the equation referring to the method of computation. All 

 data will be fully discussed later. 



As to the nature of the color sequences of the first ring of 

 coronas corresponding to successively increasing sizes of parti- 

 cles, the clue may be obtained from extremely small particles 

 and excessively large (opalescent) coronas, using the electric 

 light as a source. In such a case the colors follow the order 

 of wave-length, and one sees w v, w b, w g, w y, w o, w r, w c, 

 with all intermediate color gradations. In succeeding series 

 this is repeated with more and more overlapping until after 

 two cycles have been passed all periodicity is lost (appreciably) 

 in the normal coronas. 



If the data of the table be mapped out graphically in terms 

 of ^, the periods of n and d are now sharply marked. In the 

 d curves for instance, cusps in the region of 6, 4, 3, 2 times 

 •00018'=" (large in comparison with wave-length) may be recog- 

 nized. 



Brown University, Providence, E. I. 



