314 C. Barus — Behavior of the Phosphorus, etc. 



volts- Three condensers (K2, E = ll'T; K4, E = 6-5 ; K8, 

 P = 3*5 cm.) were treated in this way , admitting of electric 

 fields from 2 to 90 volts/cm. The observations were made as 

 above in sextuplets, and from these both the current, ds/dt, 

 (arbitrarily in scale parts of the electrometer), and the constant, 

 n x K, were computed. The results however, owing to the 

 variability of the phosphorus (whether due to this method of 

 applying electric fields or to incidental causes, I do not know), 

 are complicated, particularly in the case of weak fields. It 

 will suffice therefore to give a graphic digest (fig. 3) of the 

 data for K8, the smallest condenser selected, as this admits of 

 the greatest variation of field. The curvature of the line 

 ds/dt, shows that Ohm's law is not quite obeyed as the fields 

 grow stronger; i.e., the number of ions is not indefinitely 

 largef. Nevertheless the limit is as yet far off, showing that 

 but a small part of the ions convey current even in fields of 100 

 volt/cm. It happens moreover, that the phosphorus for these 

 experiments showed weak ionizing power. Usually the ioniza- 

 tion was 50 per cent stronger and the curves more nearly 

 straight. In case of K2 (P = 11.7 cm.), the line ds/dt, 

 observed up to 20 volt/cm was quite straight, the condenser 

 being the largest, admitting of best adjustment. 



Corresponding with the values of ds/dt, the curve nj£ shows 

 a downward slope and therefore a decreasing number of avail- 

 able ions (n^ as the fields increase in intensity from 20 to 100 

 volts per cm. The value of n 1 computed from figure 2 {$\K 

 = '00120, whence n 1 = 4:X 10 4 if ^is about 1 cm/sec. and e 

 about2xl0~ 19 coulombs) agrees very well with the value given 

 in Science (March, 1900, ?i =8X10 4 ) and obtained for plate 

 condensers under the same limitations. 



Brown University, 

 Providence, R. I. 



* Science, xi, p. 4. 1900. 



fin § 4 e is the average charge per particle. In reference to electrons a coeffi- 

 cient is thus implied; for all that I showed in my experiments with tubes is that 

 the number of ions conveying current is proportional to.the total number present. 



