DECOMPOSITIONS BY TERNARY ARRANGEMENTS. 31 



ted by boiling from the water of a spring, which flows from a sandy valley, and also 

 from the dews which fall on a neighbouring hill, but too remote to be affected by the 

 exhalations of dwellings, I found the proportion, when care was taken in the analysis, 

 to be uniformly 33^ per cent., or as 1 to 2 by volume. This gas, thus extricated, is 

 isomeric with protoxide of nitrogen, with the particular exception that, in the protoxide, 

 the two volumes of nitrogen are compressed into half their bulk. 



94. In a quart jar, which was filled with spring water and inverted into a tin cap- 

 sule, I collected all the aeriform matter that could be disengaged from the water by 

 means of a fire placed beneath the jar and its tin. This gas, from many prior trials, I 

 knew to contain 33 i per cent of oxygen. When all the gas was collected that could 

 be extracted, at a temperature long continued close upon the boiling point, the arrange- 

 ment was suffered to cool, and kept undisturbed for four days ; at the close of that time, 

 considerably more than three fourths of the gas disengaged was reabsorbed; the residue, 

 on analysis, contained 5-25 per cent of oxygen only. A portion of the water in the 

 jar was now submitted to a boiling temperature in a small close vessel, and the gas col- 

 lected was analyzed. It contained, instead of 33 j, rather more than 47 per cent, of 

 oxygen. There cannot, therefore, be any doubt that oxygen may be obtained from the 

 atmosphere, in a pure and undiluted state, by the action of a tissue, or a binary, and 

 also a ternary arrangement. 



95. DECOMPOSITIONS BY TERNARY ARRANGEMENTS. After this consideration of the 

 case, in which two elements are employed, we are prepared to understand how ternary 

 arrangements effect decompositions. Let b (Jig. 13, pi. 1) be a compound gas, which 

 is placed above a barrier, a, of such a nature that one of the elements of b shall pass 

 more rapidly through it, or, in other words, be more readily absorbed by it than the 

 other. Also, let the other substance, c, which is on the opposite side of the barrier, be 

 of a kind capable of removing the quicker passing element of b from the under surface 

 of a, as fast as it arrives there. It is immaterial how this removal be accomplished, 

 whether by chemically uniting with it, or by mechanical action ; the quick passing el- 

 ement, finding at its approach to the under surface of the barrier a ready exit, continu- 

 ally passes off, and its place is supplied by fresh portions from above, so that, in the 

 lapse of time, only the less absorbahle element will be found in b. 



96. The general conditions, therefore, of chemical decomposition by ternary arrange- 

 ments are, that one element of the compound to be decomposed shall pass more 

 easilv through the barrier or bounding tissue than the others, and, on its arrival at the 

 opposite side of the barrier, it shall be rapidly removed. 



97. Reasoning upon this principle, I succeeded, nearly two years ago, in effecting 

 decompositions in this manner, which have some important physiological applications. 

 Havins taken a tube, one of the ends of which was expanded into a trumpet-shape, and 

 closed with a thin serous membrane peritoneum stripped from the liver which was 

 tightly tied on with a waxed thread, while it was wet, I poured through the orifice, 

 which was open, a strong but clear solution of litmus in water. The tube thus situated 

 was placed in a wine-glass containing strong alcohol, and the level of the liquid, inside 

 and outside, made to coincide. The conditions for decomposition were thus fulfill- 



