330 C. Bar us — Effect of Temperature and of Moisture 



for a period of 2-5 seconds (longer at the lower temperatures), 

 and then suddenly opening it again. The puffs are at first 

 vaguely recognized at about 8°, or even below. They become 

 more marked as temperature rises. They are still marked at 

 even 12°, when the fainter colors are beginning to be perma- 

 nent. They show a maximum degree of darkness depending 

 on temperature, beyond which they do not increase even if the 

 cock is closed indefinitely. 



From a theoretical point of view this result is noteworthy. 

 Below the reaction temperature (say 12*5°), what may be called 

 the vapor pressure of the reaction is a definite quantity, but 

 decreases with temperature at an enormously rapid rate; above 

 the reaction temperature, the vapor pressure is relatively con- 

 stant as temperature increases. All this recalls the well 

 known analogy appropriated by the physical chemist. 



Now if we suppose the nuclear velocity to be a relatively 

 constant quantity, within a short range of temperature, while 

 the emanating activity decreases, the density of the emanation 

 formed within the ionizer will clearly diminish as temperature 

 decreases below the reaction temperature, however long the 

 air is in contact with the ionizer. Hence the color of the puffs 

 should gradually become fainter with decreasing temperature, 

 as actually observed. If JV nuclei are produced per super- 

 ficial square centimeter of phosphorus, if k be the corresponding 

 average nuclear velocity and n the number present per cubic 

 centimeter, N — kn. Thus n depends on the ratio of i\^and Jc. 

 The vapor pressure analogy suggested is not wholly tenable 

 inasmuch as nuclei are actually absorbed at the walls of the 

 vessel (tube ab, figure 1), so that JV vanishes with n in the 

 lapse of time. Since k is of the order of one unit, A^and n 

 may be regarded as about of the same order, roughly speaking. 

 The number of particles generated per square centimeter of 

 the phosphorus will not greatly differ from the number present 

 per cubic centimeter of the emanation. 



6. Above the reaction temperature, if the rate of production, 

 iV, were regarded as relatively constant, the means of comput- 

 ing the increase of speed of the nucleus with rising temperature 

 would be at hand. If m be the mass of the nucleus and mtf/V 

 varies as absolute temperature, hj\ = \/(^ a + 273)/(^4- 273). 

 Turning now to the chart, figure 3, let £ 2 = 30° and t 1 = 20°. 

 Then Jcjh 1 = 1*02, whereas the chart gives l\Jk A — 1'25. 

 These two results being out of keeping with each other, the 

 thermal variation of h is insignificant compared with the cor- 

 responding decrease of JV. 



7. Data for low temperatures. — After finishing the electri- 

 cal investigation presently to be discussed, it seemed desirable 



