ME. Vi. CEOOIvES ON THE ATOMIC WEIGHT OF THALLIUM. 
329 
Let 
* = ■ 01 , 
-20=47619, 
\/w= 218 , 
t=lc\/w= 248, 
t=H 2 . 17 =-99795. 
(See Table I., at end of Professor De Morgan’s ‘ Essay on Probabilities’.) 
The result, =-99795, is so near to unity, the measure of certainty , that for every prac- 
tical purpose it may be considered certain that the truth is really comprised within the 
limits named. 
By Table II. in Professor De Morgan’s Essay, we can test the correctness of the pre- 
ceding deductions; for 
62 
-== the probable error, 
130 V w 1 ’ 
probable error 
Then K, answering to t in Table II., is the probability required. 
Hence 
where w= 47619, 
» \A>=218, 
62 
130 V iv 
- = • 0022 , 
-2_=^L = 4-G = t 
•0022 -0022 5 
to which corresponds in Table II. K = 99808, against 99795, as before. 
I may therefore conclude that, within the limits of error (as small as possible) of 
observation, the 
ATOMIC WEIGHT OF THALLIUM =2 03 ’64 2, 
Professor Stas has shown the hypothesis of Prout — that the atomic weights of the 
elements are severally multiples of the atomic weight of hydrogen — to be without the 
corroboration of experimental result. This view of the hypothesis is further borne out 
in the present investigation; for the number 203-G42 cannot, within the limit of what 
has been shown to be the probable error, by any liberty be made to follow the hypo- 
thesis. Without doubt, when the atomic weights of all the metals are redetermined 
according to the standard of recent scientific method, it will be found that there are 
more exceptions to the hypothesis than are commonly considered. Marignac gives, in 
his confirmatory discussion of Stas’s experiments and in his own results with calcium 
