210 Rate of Recombination of Ions in Gases. 
The discordance seems to me to be very serious, and 
requires explanation before we can accept either the theory 
or the conclusions. 
The following table was prepared from the numbers on 
. 298 of the paper, from which Mr. McClung showed that 
Na Was independent of pressure. 
narsiaes 
_ a ” 2 hd 2 2 
| Pressure in cae for ¥ fou = x 10-2. 2 | a 
| Atmospheres. 1 * 
Ak veeestes 12] 344 146 o44 425 
A/D savannas 193 ae’, 124 26°5 467 
BL seseeeee 67'9 166 ao | 6% 278 
ei ie ae 1032 35°5 42°6 14:2 300 
ee eee 70°4 15:2 495 | 162 326 
Be Bee 96°0 26°5 4671 13:2 348° 
D Lexan 128°6 38°2 165 38°2 / 433 
Br atee eka 148°7 49°4 147 | 82:0 eee 
E. | ree Cite 110 29°3 121 29:3 ) 415 
i a die 93°5 20°7 174 | 41° ee 
facil bes te 70 23:5 490 | 25:5 3/0) rae 
en th? Daas 43°5 - 707 14° | 210 
GL teereee 619 22°7 38'5 22°7. |. (aie 
| 125 29'9 51 71:5 40'8 Bas: 
If « is independent of pressure and Q is proportional to the 
pressure, then N? is proportional to the pressure. 
It would be untair to compare values of N?/P from different 
sets, say A and F’, as the intensity of the rays might be very 
different. This, however, does not apply to the values of 
N?/Q, which must be independent of the intensity. Nor 
does it apply to observations of any one set, e.g. I’, where 
the intensity must be approximately constant if the numbers 
are to mean anything at all. 
Was there, then, any difference in sensibility of the electro- 
meter in A, B, F, and G? 
The agreement between the values of N?/P or Q/P in A, B, 
© or D, although not very good, may be accepted as approxi- 
mately proved. But in H, I’, or G there cannot be said to be 
substantial agreement. 
The questions which naturally arise are: (1) Is the pro- 
posed theory correct? For the experiments cannot be held to 
support it. And (2) Do the measurements actually give the 
values of N, N,, and Q? 
In connexion with the second question, I have considerable 
