VOL. LXXVII.] PHILOSOPHICAL TRANSACTIONS. IQQ 



Exper. 1. My first object was to collect a sufficient quantity of the air 

 separated from water by silk, to determine its goodness by the test of nitrous air; 

 and to this end, having filled with clear spring water a globe of thin, white, and 

 very transparent glass, 4^- inches in diameter, with a cylindrical neck 4 of an inch 

 in diameter, and about 12 inches long, I introduced into it 30 grains of raw silk, 

 which had been previously washed in water, to free it of air ; and inverting the 

 globe under water, and placing its neck in a glass jar, containing a quantity of 

 the same water with which the globe was filled, I exposed it in a window to the 

 action of the sun's rays, and prepared myself to examine the progress of the 

 generation or production of the air. 



In 10 minutes an infinite number of exceedingly small air-bubbles began to 

 make their appearance on the surface of the silk ; and these bubbles continuing to 

 increase in number, and in size, at the end of about 2 hours the silk appearing 

 to be entirely covered with them, rose to the upper part of the globe. These 

 bubbles going on to increase in size, and running into each other, at length be- 

 gan to detach themselves from the silk, and to form a collection of air at the up- 

 per part of the globe ; but the measure of my eudiometer being rather large, it 

 was not till after the globe had been exposed in the sun near 4 days, that a suffi- 

 cient quantity of air was collected to make the experiment with nitrous air, in 

 order to ascertain its goodness by that test. 



Having at length collected a sufficient quantity of this air for that purpose, I 

 carefully removed it from the globe, and mixing with 1 measure of it 3 measures 

 of nitrous air, they were reduced to 1 .24 measures ; which shows, that it was 

 actually dephlogisticated air, and that of a considerable degree of purity. Com- 

 mon air, tried at the same time, 1 measure of it with 1 measure of nitrous air, 

 were reduced to 1 .08 measure. 



Having again exposed the globe with the same water and silk in the window, 

 where the sun shone the greatest part of the day, at the end of 3 days I had 

 collected 3^ cubic inches of air, which, proved with nitrous air, gave la -{• 

 3n = 1.18 ; that is, 1 measure of this air, added to 3 measures of nitrous air, 

 were reduced to 1.18 measure. A small wax-taper, which had been just blown 

 out, a small part only of the wick remaining red-hot, on being plunged into a 

 phial filled with this air, immediately took fire, and burnt with a very bright and 

 enlarged flame. The water in the globe appeared to have lost something of its 

 transparency, and had changed its colour to a very faint greenish cast, having 

 also acquired the odour or fragrance proper to raw silk. 



This experiment was repeated several times with fresh water (retaining the 

 same silk) and always with nearly the same result : with this difference however, 

 that when the sun shone very bright, the quantity of air produced was not only 



