C. Barus — Apertures of Coronas. 341 



a given degree of snpersaturation at a given temperature. Full 

 allowance must, however, be made for the occurrence of peri- 

 odic variations of aperture in relation to the diameter of the 

 fog particles ; in other words, a given aperture is only of 

 value when qualified by the type of corona (whether of the 

 crimson or green order) to which the aperture belongs. Thus 

 it will not in any case be possible to dispense completely with 

 the color pattern. It is with the object of finding these correc- 

 tions systematically that I have recently begun a series of 

 experiments with a new form of plate glass apparatus and shall 

 then refer to other developments. Homogeneous light, though 

 in many respects desirable, gives effects so faint as to be use- 

 less in practice. 



With the above data I am able to make an independent esti- 

 mate of the number of particles in the saturated phosphorus 

 emanation. The number found for the first fog of the series 

 was (Phil. Mag. (6), iv, p. 25, 1902) »= 83,000; since ti'= 

 2*2 n, n'= 183,000 particles per cub. cm. JSTow the density 

 ratio before and after exhaustion is y, so that 1 — y is the volume 

 of saturated emanation added. As this has passed directly and 

 slowly over excess of phosphorus, it must be very nearly satu- 

 rated, becoming diluted on mixture with the dust free air of 

 the receiver. Hence if n particles per cub. cm. correspond to 

 saturation, n (l—y)= n' =183,000 ; or n = 10 6 . There must, 

 therefore, be at least a million nuclei per cub. cm. of the air 

 immediately in contact with a surface of phosphorus. The 

 value following from my electrometer work was^ =2xl0 6 . 

 The two methods are absolutely distinct but lead to data of the 

 same order. It is because of the general reasonableness of the 

 data which have followed from my simple hypothesis through- 

 out a very wide territory of observation that I have felt bound 

 to adhere to it, in spite of the more startling corpuscular expla- 

 nations which might be adduced. 



Brown University, Providence, E. I. 



