922 
which we conclude that our formula represents all points below 
200° within the errors of observation *). 
TABLE V. 
Ah Ww 1g W Tij W 
Jee 
| 
| 
| 
1 1.24875 55 14.1022 73 24.1347 
2 1.24883 56 14.6307 14 24.7118 
3 1.24917 57 15.1637 15 25.2900 
4 1.25008 58 15.7016 76 25.8691 
| 5 1.25197 59 16.2436 17 26.4489 
6 1.25534 60 16.7896 78 27.0294 
7 1.26079 61 17.3392 79 27.6104 
8 1.26898 62 17.8922 80 28.1918 
9 1.28064 63 18.4484 81 28.7738 
10 1.29657 64 19.0074 82 29.3561 
11 1.31758 65 19.5691 83 29.9386 
12 1.34454 66 20.1333 84 30.5213 
13 1.37831 67 20.6997 85 31.1041 
14 1.41978 68 21.2682 86 31.6871 
15 1.46981 69 21.8385 87 32.2701 
16 1.52926 70 22.4103 88 32.8531 
17 1.59896 11 22.9838 89 | 33.4360 
18 1.67968 12 23.5586 90 34.0187 
19 1.77219 
20 1.87715 
21 1.99519 
22 2.12687 
e 
A few words about the temperatures above 200°. There the 
deviations are seen to increase rapidly. If we had used the 
proportions W/W, instead of the resistances themselves as 
1) Yet it will be noticed that the deviations, at least these at liquid oxygen 
temperatures, are not distributed at random, but show a regular course. From 
this we draw the conclusion, interesting in itself, that the mutual agreement of 
these measurements is much better than would be indicated by an accuracy 
of 0°.02. Surely this only tells something about the purely accidental errors. 
