﻿32 Mr. Carl Barus on 



boiling-point, and the absence of colour in this case is more the 

 result of insufficient supersaturation than of absence of dust. 

 Now let the compressed air be admitted, so that the tube A A 

 is considerably cooled. If the steam-pressure is high enough, 

 and the gas in sufficient quantity, the field first clouds over 

 and then turns quite opaque. Thus at a pressure of 50 cm., 

 and a temperature registered at t of 47°, such a result 

 (opaqueness) was obtained. At 30 cm. steam-pressure the 

 field turned cloudy, but not opaque ; at 15 cm. no clouding 

 was even apparent. Colours were not observed in any case. 



In view of the insufficient air-supply obtained in case of the 

 filter G, I replaced it by the form H, differing from G in 

 having a drum much larger in diameter (length 25 cm., 

 diameter 15 cm.), and special sieves rr and s s. Similar 

 parts are similarly lettered. Under like conditions, such 

 a filter should supply nine or more times as much air as the 

 other (G). 



The absorbent cotton inserted between the sieves ss and rr 

 (the former and the lid being removed, and the latter \_r r] 

 soldered in place) was carefully laid in layers parallel to r r, 

 quite filling the width of the drum. When full to the top, the 

 cotton was compressed by the second sieve ss, which thereafter 

 was also soldered in place, the lid with the inlet pipe o' w! 

 soldered on, and the space between lid and s s filled with loose 

 cotton to catch dirt. The compressed charge between s s and 

 r r mainly acts as the filter. When not in use it should be 

 heated, so as to be thoroughly dry. 



In the first experiments with the filter H, compressed air 

 (issuing as usual from a narrow pipe, and at a tank-pressure 

 of about 15 atmospheres) was used as before. Spreading 

 through the filter, the air enters the colour-tube at a much 

 reduced velocity. As the air in the tank is gradually freed 

 from dust by the subsidence of the latter, there is here an 

 additional means of purification. But the method is much too 

 lavish to be practical, even if it be conceded that the filter is 

 actually efficient. The following data were thus obtained for 

 the margin of the opaque field at temperatures comparable 

 with the values in the charts. Colours were not observed : — 



Temperature . 



. 20° 21° 



22° 



23° 



33° 



Centigrade 



Pressure . . . 



. 41*5 42 



43 



44 



58 



Cm. Hg. 



In the set of data for 20° to 23° the filter was not so well 

 packed as at 33°. Observations were discontinued for want 

 of gas. I have inserted these data in the chart, fig. 3, at 

 F and F'. When taken together they suggest a locus of 

 the same nature as fig. 4, i. e. implying less curvature in 



