ME. ~W. CEOOKES OH THE ATOMIC WEIGHT OF THALLIUM. 
279 
remarkable sentence was written Professor Stas had commenced bis classical researches 
on the atomic weights; and in 1861 he gave to the world the results of ten years’ expe- 
riments, which had been conducted with a care and perseverance never surpassed in the 
history of experimental investigation. These researches of Professor Stas, and others 
which he has since made public, constitute a standard of excellence which chemists who 
are engaged on the important task of the determination of atomic weights should strive 
to attain. They are, in my opinion, the most noteworthy chemical memoirs that have 
ever been written : not only have they determined in the most definite and unassailable 
manner atomic weights about which scarcely any two chemists have agreed since, the 
time of Berzelius, but they have raised the standard of accuracy in all chemical labo- 
ratories, and have set an example which, if followed, cannot fail to exert an important 
influence on the progress of chemical science. 
It has been with these researches before me that I have endeavoured to determine in 
a manner which should approach them in accuracy the atomic weight of thallium. 
In the determination of an atomic weight analysis is inferior to synthesis ; and espe- 
cially is this the case when the number sought is amongst the highest known. The 
method followed should be one in which as few chemical elements as possible are 
employed, so as to reduce to a minimum the errors arising from inaccuracy in the deter- 
mination of their atomic weights, — which errors, whilst they might on the one hand 
balance each other, on the other might accumulate in the same direction, and become 
a total error of exceeding magnitude in the atomic weight of the metal under investi- 
gation. The method adopted should also be one in which there is the greatest possible 
difference of weight between the substance taken for the starting-point and the one 
ultimately obtained ; for the greater the amount of this difference, other things being 
equal, the less likely are the unavoidable errors incidental to the method, and which 
may be looked upon as constant, to injuriously affect the atomic weight obtained. For 
these reasons processes in which a weighed quantity of the metal itself is taken and 
converted into one of its salts seemed likely to afford the best results ; and this accord- 
ingly is the principal method which I have adopted. 
Every substance employed in such a determination is liable to introduce errors pro- 
portionate to its own want of purity. The most extraordinary pains have therefore 
been taken to secure the absolute purity both of the thallium employed and of the 
agents used to act upon it. The glass and other apparatus have been specially constructed 
for these researches, and the balances and weights have been of an accuracy never before 
surpassed in any research. Whilst nearly every other branch of manipulative chemistry 
has advanced to an accuracy vieing with astronomical observation, the operation of 
weighing, as almost universally carried out, is attended with grave imperfections. For 
ordinary analytical work, and perhaps even for more refined and accurate researches, the 
errors attending the ordinary process of weighing are unimportant ; but in determining an 
equivalent so high as that of thallium no precaution whatever which can either reduce 
an error to a minimum or eliminate it altogether should be neglected. I am anxious to 
2 p 2 
