( 383 } 
tt will be seen that the accordance between the calculated and 
the measured effects leaves nothing to be desired. 
We find in WaArLLeER’s work still another series of numbers ( Philo- 
sophical Transactions of the R. S. Vol. 188 p. 60, 1897). 
From this series the following Table may be calculated. 
TABLE IT, 
A. WALLER. Philos. Trans. Roy. Soc. 
1897 p. 60. 
————————————————————— 
A = 20.1779 
B= 0.117467 
2.4972 
a mtr 
R E calc. oad Q 
| 
5 5.140 5.5 | + 0.360 
10 11.820 44. — 0.820 
15 15.533 16.5 | + 0.967 
20 17.596 18. + 0.404 
30 19.380 19.5 | + 0.220 
50 20.102 20. — 0.102 
100 20.177 20. — 0.177 
In this series there occurs again one single value, that causes a 
great deviation. If the most probable values were calculated with 
rejection of this most probably erroneous observation, the mean 
observation-error was reduced from 0,705 to 0,164. 
