M. Hess on the Scientific Labours of Richter. 93 



viating greatly from the true number, which is 31*9. Having 

 a false idea of the composition of arsenious acid, he neces- 

 sarily deduces a false composition for arsenic acid. Now 

 this is the way he obtains a number so far from the truth : 

 he converts a given weight of regulus of arsenic into arsenic 

 acid, and then into arseniate of lead. But instead of drawing 

 a conclusion from the weight of this latter, he first tries to 

 determine the quantity of arsenic acid which the precipitate 

 should contain, and for that purpose sets out from the arseniate 

 of magnesia, which must necessarily compromise all the re- 

 sults ; for in order to determine the composition of that salt, he 

 saturates a solution of arsenic acid by the carbonate of magne- 

 sia, a salt whose composition is not always constant. Then he 

 determines the quantity of arsenic acid from a table of density 

 previously constructed. In this then consists Richter' s greatest 

 fault, I will even say the only one which he has committed, but 

 from which several others originate : he did not yet quite ap- 

 preciate the difference which exists between a direct and an in- 

 direct method. This is the true source of all his errors. To 

 make amends for this, each time that he makes a direct expe- 

 riment, he approaches very nearly to the truth ; for example, 

 if he wished to know of how much oxygen and cobalt the oxide 

 of this metal is formed, he determines this quantity in a direct 

 manner, and finds for 100 parts of cobalt 26*5 of oxygen, which 

 does not widely deviate from 27, which is the real number. 

 But Richter distrusts himself. He tells us (vol. ix. 1798, pre- 

 amble) that he cannot easily manipulate ; that he was never 

 able to finish an analysis without losing something at the end 

 of all the operations ; and that he never dared to undertake 

 an investigation if there was any question of stcechiometrical 

 determinations, with so small a quantity as 100 grains, but that 

 he needed 500. This is perhaps the reason why Richter at- 

 tached great value to the tables of density, whether for acids 

 or for salts. A considerable part of his time too was employed 

 in making them. At the end of each acid he gives a table in- 

 dicating the acid contents in a solution at different degrees of 

 density. He does the same for the salts which are most used. 

 Richter was also much employed at different times in con- 

 structing areometers and alcoholometers ; we still use many 

 instruments which bear his name. 



It is not only when Richter treats of general questions that 

 he merits all our attention ; he often captivates it by questions 

 which are quite special. A few examples will suffice in order 

 to appreciate him. We have seen that he confirms the re- 

 searches of Klaproth on the composition of the salts of stron- 

 tian, but, he says, my conclusions are not just unless the salt 



