10 Messrs. Richardson, Nicol, and Parnell on the 
explained, covering a range of temperature from 576° C. to 
1136° C., and of pressure from 76:0 to 0°2 cms. of mercury. 
The results of all the measurements which were made, using 
this method, are given in Tables 1. and II. 
TasueE IT. 
Tempera- | Pressure, | : | Tempera- | Pressure, | ‘ 
ture ems. of | ture ems. of | Q. 
(Cent.). | mercury. (Cent.). | mercury. | 
PRs ie) ay | | es po ae 
reve 76°56 | 1:841x10-%)|) 995 +} 148 | 6382 x10 
856» |... 7656 |589 x10-8/|., 1136), | 148 | 54 KOR 
995 720... 12488 X1007)) 6 TUF) ly ASS) FG ee 
1136 | 714  |3218x10-7}| 9856 | 1438 |264 10-8 
G17. |) 87-7. | 1171x1078) | 995 | 143 1.651 sede 
856 | 377 |409 x107°}| 1136 | 143 | 1-47 x10-7| 
Oe 1:084x10—"| |) 717" |, | Tl) | bee sel 
1136 | 377 = | 2:356x10~" e561 Tl | i6e4x to 
646°.) 27-4 1 498° x1077 11" 995 i 439 x107 4 
vary 27-4 CUO IOC Ll) Iles. fae | ek 985 x10>" 
787 274 4 189x108)! G7 296 13:47 x10—° 
856 27-4 «+1341 x10-8)| ~ 856 2°06 | 1:16 x10" | 
926 274 | 572 1078) | OS |. 8406.) Te ee 
995 274 |937 x10-8,| ~ 1136 | 2:96 | 5-91 x107° 
1065 M4} P88 RWW T | 9OR 1 > IS ea 
1136 974  |206 x10-*|| 1136 1120/1 $4) 
856 148 .(252'x10-"|| 1136 02 °|1:28 x1075 
With the exception of those corresponding to pressures 
nearly atmospheric, all the values of @ have been corrected, 
in the manner already described, for the small variation of pres- 
sure caused by the loss of hydrogen during any one experiment. 
Thus for the four readings where the pressure is stated in the 
above table to be 37°7 ems. the actual pressures were 37°7, 37°6, 
37°4, and 36°7 cms. respectively, whilst the actually reduced 
values of Q were smaller than those tabulated in the ratio 
of the square root of the actual pressures to the square root 
of 37-7. 
By inspection of the numbers in the preceding tables, it is 
easy to see that the rate of diffusion increases very rapidly 
with the temperature of the metal, and also that it falls off 
with decreasing pressure much less rapidly than if it were 
proportional to the pressure. For constant increments of 
temperature, the ratio of the values of Q for the initial and 
final temperature falls off as the temperature increases. Thus 
