toward the Cold. 361 



though faintly. At 7 p.m. the dew formed a violet arc on the 

 side nearest the window. 



On the morning of the 22nd the arc was pale, and there was 

 a dense deposit of thick purple matter on the furthest side just 

 above the liquid ring. At 5 p.m., the sun being on the window, 

 the dark-coloured deposit was on the nearest side, but equally 

 low down. These are effects of capillarity evidently similar to 

 those so well described by Dr. Woods ; but I paid no particular 

 attention to them at the time, as my mind was occupied with 

 other phenomena. 



As the season advanced and became more settled, the pheno- 

 mena varied. A portion of the bisulphide was decomposed, and 

 another portion forced its way out of the bottle. During these 

 events the motions of the vapour went on with the regularity of 

 a clock. When the sun was on the window, a current apparently 

 rose on the nearest side and descended on the furthest, where the 

 deposit was made. When the sun was off the window, the cur- 

 rent set in in the opposite direction ; and the nearest side being 

 the colder, the deposit was transferred to it. These motions 

 were marked in an interesting manner at the sides of the bottle ; 

 at first by well-shaped parabolic spaces, marking the temperate 

 regions between the hot and the cold ones ; afterwards by lines 

 as straight as if ruled, diverging above and below from a centre 

 placed a little above the liquid. The lines above this centre were 

 several inches in length ; those below it about a third of an inch. 

 These lines, lastly, changed into a minute sea-weed pattern which 

 remained fixed for some weeks ; but I find by my note-book that 

 the bottle had become dry early in June. 



A similar experiment was also tried with phosphorus in bisul- 

 phide of carbon, &c. 



I also performed a number of experiments on the motion of 

 the vapour of water towards the light. During many years I 

 had often noticed, when preparing gases for the purpose of a 

 lecture, that the moisture condensed on the sides of the gas-jars 

 and bottles was of a different pattern with different gases. I spe- 

 culated a good deal as to why this should be so, but now deter- 

 mined to bring the fact under the ordeal of exact experiment. 

 Accordingly in May last I filled six quart stoppled bottles, No. 1 

 with oxygen, No. 2 with nitrogen, No. 3 with hydrogen, No. 4 

 with nitric oxide, No. 5 with carbonic oxide, and No. 6 with 

 carbonic acid. The gases were collected over water at the 

 pneumatic trough, and about half an ounce of water was left 

 in each bottle, which was tied over with soft leather wetted 

 in a solution of gum, which was also run round the stopple. 

 The bottles were placed on a table near a west window, and a 

 journal was kept of their proceedings, which were recorded twice 



