170 THE FLOATINa-MATTER OF THE AIR. 



descence was doubled, ten minutes being allowed in- 

 stead of five, while the wire was raised to the higliest 

 possible degree of incandescence. The infusions era- 

 ployed were turnip and cucumber, a group of three 

 tubes being charged with each. After the air had been 

 calcined, the infusions were boiled for five minutes in 

 an oil-bath. With this mode of treatment not a single 

 failure occurred in 1875, turnip-infusion being among 

 the number of liquids thus treated. This year two days 

 sufficed to render every one of the six tubes turbid with 

 organisms and to cover the infusion with a heavy scum. 



I, however, had occasion to doubt the closeness of 

 these shades. Tlie wax intended to seal the junction 

 of the tin collar with the plate of wood had cracked and 

 yielded here and there, and the entry of contamination 

 through such cracks was possible. Six new shades were 

 therefore mounted and surrounded by collars which were 

 imbedded in white lead and firmly screwed down to 

 the plate of wood. The height of the collar, which 

 measured the depth of the filtering layer of wool, was 

 much greater tlian it had ever been last year. As 

 before, the period of incandescence was ten minutes, 

 during which the platinum spiral was brought as close 

 as possible to its point of fusion. 



Each of these six shades covered a group of three 

 test-tubes. Two such groups were charged with turnip, 

 two with cucumber, and two with artichoke-infusion. 

 The infusions, as usual, were boiled for five minutes 

 after calcination. They were all brilliant wlieu pre- 

 pared ; but in two days every one of them had become 

 turbid, and had covered itself with a fatty scum. Tins 

 gradually augmented until it reached in some of the 

 tubes a thickness of lialf an inch. The weight of the 

 scum caused it in some cases to bag downwards, form- 

 ing a kind of inverted cone, the apex of which was 



