32 A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



jet is shattered against a solid, the charging current, for the condenser at zero 

 potential, is about 8 per cent, of the constant current; but it may also be 

 absent. 



The ratio of the ionization to the nucleation does not always appear as a 

 fixed quantity ; from which it follows that the mean charge per nucleus depends 

 on incidental conditions of freshness, the nature of the jet, its impact, etc. 

 Similarly, the ratio of positive to negative ionization does not seem to be a 

 fixed quantity, but to vary under the same conditions. Nuclei generated in a 

 small space are more highly charged because they can be more swiftly trans- 

 ferred to the condenser. 



Finally, the maximum nucleation for any jet is reached when as many 

 nuclei are produced per second as are lost in the same time. Unquestionably 

 the air current accompanying the action of a violent jet contributes to this loss, 

 by washing the air against the sides of the vessel and the surface of water. 

 Hence jets with a strong single direction, even if made up of filimentary jets, 

 produce few nuclei. 



Finally, the reason for the unique efficiency in the capillary oblique jet was 

 specially verified. Supposing that the high ionization relatively to the nuclea- 

 tion in this case is due to keeping the water level near the jet and expelling the 

 nuclei by an auxiliary air current from a small volume, I made the following 

 experiments in which the nuclei were discharged by a rising surface of water by 

 aid of the Mariotte flask. Table 13 shows that on successive half minutes the 

 currents ds/dt increase rapidly, as was supposed. Moreover, when referred to 

 an efflux of 2 liters per minute, the amperes are now actually of the low order 

 corresponding to the nucleation of the jet. 



The effect of different volumes is also seen from the table, which proves that 

 proportionality is roughly admissible. This also follows necessarily from the 

 equation of the phenomenon given elsewhere (cf. 18), and has-been carefully 

 verified for phosphorus. 



15. Spontaneous time loss of nuclei. The following table (14) shows the 

 spontaneous loss of nuclei in the lapse of time. The nuclei were produced in the 

 receiver in the usual way, and their number was then determined by the con- 

 densation produced after a stated interval. The approximate number or order 

 of the corona in my series is nevertheless somewhat difficult to determine, and 

 the number of nuclei estimated therefrom not quite definite. As this number 

 is an exponent, arithmetical progression indicates geometric progression in the 

 number of nuclei. 



The radial jet, No. 2, shattering itself against the sides of the vessel is 

 strongest as a nuclei producer, and the large oblique jet, No. 3, considerably 

 below it in efficiency. The capillary oblique jet, No. 3, is remarkably efficient 

 relatively to the quantity of water used. The vertical large copper jet, No. 9, 

 used in tables 10 and n, is a very poor producer of nuclei, though using about 

 8 liters of water per minute and in spite of the turbulent churning of the pool 

 below. 



