DATA OF VARYING PRESSURE. 17 



in the fog chamber on the barometer, on the ionization of the air, on 

 any form of external radiation, or on the temperature of the atmosphere, 

 can not be detected. All the variations may be referred to the temper- 

 ature of the fog chamber itself, as if it generates increasing numbers of 

 colloidal nuclei as its temperature increases. Since the colloidal nuclei 

 in dust-free moist air are to be associated (from my point of view)* 

 with the saturated vapor, and are only secondarily dependent upon the 

 air itself, the result so obtained is curious, as one would expect a decrease 

 of the colloidal nucleation with rise of temperature. Correction for the 

 increased water precipitated at higher temperatures merely accentuates 

 the difference. If r t is the low (absolute) temperature obtained by 

 sudden expansion adiabatically from r the ratio T X /T should be wholly 

 dependent upon the corresponding pressures; and yet, for the same 

 ratio, more nuclei are obtained as r is larger. This difference of be- 

 havior is maintained for larger and smaller ratios of r 1 /r, in like degree. 



12. Data. The results are given in tables 7 and 8, and refer to a 

 fog and vacuum chamber, the volume ratio of which is about v/V = o . 06, 

 combined with sufficiently wide piping (2 -inch bore) and an interposed 

 (2. 5-inch) stopcock. The former communicates with the filter, the 

 latter with 'the air-pump. At the same temperature the fog and vacuum 

 chambers are initially (before exhaustion) at pressures p and p' ' , finally 

 at pressure 3 , when in isothermal communication after exhaustion; 

 p 2 and p' 2 , respectively, would be the pressures at the given temperature 

 if the chambers could be isolated immediately after exhaustion and 

 before the precipitation of fog. P denotes the barometric pressure, and 

 p m the initial gage-reading within the fog chamber before exhaustion, 

 so that the drop of pressure is (apart from the moisture content, which 

 will be treated in turn below) dp = Pp m p 3 , and the drop of pressure 

 takes place from p = P p m adiabatically to p t , isothermally to p z if the 

 fog chamber were isolated as specified, or isothermally to p 3 when fog 

 and vacuum chambers are left in communication. 



For a given value of P the same drop of pressure dp may thus be 

 obtained in two ways either by giving a suitable value to p m , i. e., by 

 starting with a partially exhausted fog chamber and a vacuum chamber 

 at fixed exhaustion p f , which implies a nearly fixed p 3 ; or by keeping 

 p m constant (small, nearly zero), thus starting with the fog chamber 

 about at atmospheric pressure, and determining p' of the vacuum 

 chamber and therefore p a . 



Briefly, then, the condensational effects of a given difference dp when 

 lying between different pressures p and p a , are to be tested, and this is 

 best accomplished by constructing separate complete graphs for the 

 aperture 5/30 of the coronas, first by keeping p' and p 3 nearly constant 



*Am. Journ. Sci., xxn, p. 136, 1906. 



