258 Profs. H. Rubens and E. W. Wood on 



board (0*4 mm.) • 4, smoke film, 1*8 rag. of carbon per 

 square cm. (determined with Nernst's micro-balance). 



The great transparency of the smoke film for the long- 

 waves is of interest as showing that the sensitiveness of our 

 radiometer was not much increased by smoking the metal 

 surface. We have increased the deflexions of the instrument 

 used in the present work by covering the thermo-couple with 

 a thin film of water glass, which is very opaque to tho 

 longer waves. The best material is probably a mixture of 

 water glass and finely divided carbon (soot deposited by flame), 

 which dries without cracking, and sticks firmly to glass or 

 metal surfaces. The transmission of fluids and vapours was 

 investigated by employing quartz cells of known thickness, or 

 large brass tubes closed by thin quartz plates. The fluid cells 

 are made by pressing the quartz plates against thin rings of 

 lead foil or glass, except in the case of the cell used for the 

 investigation of water which was found to be quite opaque 

 when contained in our thinnest cell (0*158 mm.) . To determine 

 the transmission of water we once more inserted the inter- 

 ferometer in the path of the rays, and, having separated the 

 plates a given distance, measured in sodium fringes, intro- 

 duced a drop of water between them. The water crawled in 

 by capillary attraction, and the sodium fringes were watched 

 throughout the process to make sure that no change in the 

 distance between the plates resulted. 



The tube used for the investigation of vapours was 20 cm. 

 in length and 6 cm, in width, an observation being made 

 first with the tube filled with air, and then with air saturated 

 with the vapour, obtained by allowing a slow stream of air 

 to bubble through two bottles filled with the fluid. The 

 absorption of water vapour w r e measured with a tube 40 cm. 

 long and 9 cm. in diameter, open at both ends, a current of 

 steam flowing constantly into it through a side tube placed 

 near the centre. The walls of the large tube we kept at a 

 temperature of 150°, by a coil of w 7 ire electrically heated. 

 This prevented the formation of clouds, condensation not 

 taking place until the superheated vapour had risen several 

 centimetres above the open ends of the tube. The results 

 are given in Tables IV. and V. The values observed with 

 the fluids are all a little too high since the introduction of a 

 fluid between the quartz plates more or less completelv 

 eliminates the reflexion from the two inner surfaces. A 

 knowledge of the refractive index of the fluid for the long- 

 waves would be necessary for the correction of this error. 

 Water and water vapour are still highly absorbing for 



