321 
1890 - 91 .] Mr W. Dittmar on Composition of Water. 
I prefer to give at once the results of my recalculation of Dumas’ 
experiments, and apply the correction to the most “ probable value” 
as calculated by me. 
In his tabular statement of results Dumas gives, in the case of 
each of his nineteen experiments, two values for what he calls the 
“equivalent of hydrogen” (the term with him meaning the weight 
of hydrogen which combines with 10,000 parts of oxygen into 
water) — viz., firstly, the value as calculated from the uncorrected 
weights of water and oxygen ; and, secondly, the “ equivalent * as 
corrected for the air in the sulphuric acid ” (used for the evolution 
of the hydrogen from zinc). For reasons, which will be stated 
in the memoir, I have left these corrected values on one side, and 
recalculated and reduced only the “ equivalents hruts” 
Taking S as a symbol for the weight of oxygen consumed in a 
given experiment, and W for the uncorrected weight of water pro- 
duced, I formed the equations 
= 8 1 
W 2 -K 2 = 8 2 
W 3 -*S 3 = S 3 
and solved the nineteen equations in respect to Jc — firstly, in the way 
which reduces the algebraic sum of all the errors 8 to nil ; and, 
secondly, so as to reduce the sum of the squares of the errors 8 to 
its minimum. The first method gave &=1T25 43; the second 
gave &=1T25 47. 
The two values, as we see, are practically identical. Adopting 
the second, it may be read as stating that 1000 grammes of oxygen 
take up hydrogen to form a quantity of water whose apparent 
weight in air is 1 1 25*47 grammes. But, assuming the air to have 
the density corresponding to 15° and 760 mm. (which probably is not 
far removed from the air-density which actually prevailed during 
Dumas’ work), the air displaced by the water amounts to 1*38 
grammes ; hence we have, in reference to any given quantum of 
water, the following relative values for the weights of 
* As already pointed out by Lothar Meyer and Seubert, Dumas’ table of 
results includes quite a number of misprints. These, however, are all easily 
discovered, and set right without much fear of error. 
