A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 57 



determine the apertures of the coronas (so long as these are normal) by a suit- 

 able goniometer, or one may find the rate of subsidence of the cloud particles. 

 Both are approximate and limited in scope, as they fail in the cases of the higher 

 transient coronas. Two methods, furthermore, are available for measuring the 

 nucleation, n, or at least relations of n. Aitken's direct dust counter or a 

 similar apparatus may be applied (work ' with this end in view is given in Chaps. 

 VII, VIII), or the values of n may be made to decrease geometrically in the 

 way just specified until normal coronas are obtained, for which d follows from 

 aperture. For the last of these methods I have already published data; but 

 in the course of over a year's additional experimentation a number of new 

 developments have shown themselves which it is my purpose here to elucidate. 

 In the first place the method formerly used for determining ra gave results 

 much too small. These are corrected in the present paper. In the second 

 place, the coronas were supposed to be observed under adiabatic conditions of 

 temperature ; direct experiments in this paper show that the air temperatures 

 during which the coronas are observed are nearly isothermal. Moreover, the 

 new results prove that in addition to the systematic loss of nuclei by exhaustion, 

 as thus fully computed, there is an additional loss which has hitherto escaped me. 

 Each exhaustion, in fact, is accompanied by a definite loss of nuclei for which 

 reasons must be investigated ( 10). 



Finally, I have in this chapter used both electric- and mono-chromatic 

 light as a source, as well as the Welsbach mantel employed for practical pur- 

 poses. Naturally from the introduction of intense violets the coronas become 

 more complicated, but it is only in this way that their true nature may be 

 detected. 



2. Apparatus. The apparatus in which the present experiments were 

 made differs from the earlier forms merely in the employment of plate-glass 

 condensation chambers. A variety of forms were used, some bulky and nearly 

 cubical, like figure i (20 cm. deep, 25 cm. high, 35 cm. long), others (figure 2) 

 long and narrow (15 cm. deep, n cm. high, 55 cm. long). Practically an appa- 

 ratus 25 cm. deep, 10 cm. high, or less, and 60 cm. long would be most generally 

 suitable. They were all lined, except on the opposed plate-glass faces, with a 

 double layer of cotton on a copper frame. The chamber is to be mounted on 

 trunnions, E, t, so as to admit of easy rotation around a horizontal axis at 

 right angles to the line of vision. Holes, A, A', must be provided so that the 

 plate-glass may be cleaned within, with a probang. The chamber carries a 

 stopcock, F, leading to a cotton filter, and another, P, leading to the nucleator 

 (preferably phosphorus). The trunnions are wide and hollow, and exhaustion 

 is made through one of them, E, while the other, t, is either closed or may 

 serve for the admission of a thermometer. To produce a definite amount of 



1 Aitken's dust counter may be dispensed with, and the intensity of the nucleator deter- 

 mined by condensation in benzol vapor, in which the coronas are all normal. See Smithsonian 

 Contributions, 1. c., p. 55 et seq. 



