OXYGEN. 13 



amply account, however, for the differences between the older and the 

 later determinations of the constant under discussion. 



Leaving out of account all measurements of the relative densities of 

 hydrogen and oxygen, to be considered separately later, the next de- 

 termination to be noted is that published by J. Thomsen in 1870.* 

 Unfortunately this chemist has not published the details of his work, 

 but only the end results. Partly by the oxidation of hydrogen over 

 heated copper oxide, and partly by its direct union with oxygen, Thom- 

 sen finds that at the latitude of Copenhagen, and at sea level, one litre of 

 dry hydrogen at and 760 mm. pressure will form .8041 gramme of 

 water. According to Regnault, at this latitude, level, temperature, and 

 pressure, a litre of hydrogen weighs .08954 gramme. From these data 

 O = 15.9605. It will be seen at once that Thomson's work depends in 

 great part upon that of Regnault, and is therefore subject to the correc- 

 tions recently applied by Crafts and others to the latter. These cor- 

 rections, which will be discussed further on, reduce the value of O from 

 15.9605 to 15.91. In order to combine this value with others, it is neces- 

 sary to assign it weight arbitrarily, and as Thomsen made eight experi- 

 ments, which are said to be concordant, it may be fair, to rank his 

 determination with that of Erdmann and Marchand, and to assume for 

 it the same probable error. The value 15.91, .0113 will therefore be 

 taken as the outcome of Thomsen's research. 



In 1887 Cooke and Richards published the results of their elaborate 

 investigation.! These chemists weighed hydrogen, burned it over copper 

 oxide, and weighed the water produced. The copper oxide was prepared 

 from absolutely pure electrolytic copper, and the hydrogen was obtained 

 from three distinct sources, as follows : First, from pure zinc and hydro- 

 chloric acid ; second, by electrolysis, in a generator containing dilute 

 hydrochloric acid and zinc-mercury amalgam ; third, by the action of 

 caustic potash solution upon sheet aluminum. The gas was dried ancl 

 purified by passage through a system of tubes and towers containing 

 potash, calcium chloride, glass beads drenched with sulphuric acid, and 

 phosphorus pentoxide. No impurity could be discovered in it, and even 

 nitrogen was sought for spectroscopically without being found. 



The hydrogen was weighed in a glass globe holding nearly five litres 

 and weighing 570.5 grammes, which was counterpoised by a second globe 

 of exactly the same external volume. Before filling, the globe was ex- 

 hausted to within 1 mm. of mercury and weighed. It was then filled 

 with hydrogen and weighed' again. The difference between the two 

 weights gives the weight of hydrogen taken. 



In burning, the hydrogen was swept from the globe into the combus- 

 tion furnace by means of a stream of air which had previously been 

 passed over hot reduced copper and hot cupric oxide, then through potash 



*Berichte d. Deutsch. Chem. Gesell., 1870, s. 928. 

 fProc. Amer. Acad., 23, 149. Am. Chem. Journ., 10, 81. 



