296 PHENOMENA, ATOMS, AND MOLECULES 



As a result of this work, which I shall briefly describe in the following 

 pages, I have gradually become convinced that molecules and atoms of 

 all kinds show Httle or no tendency to be reflected from a solid surface 

 against which they strike. In the case of metal vapors condensing on the 

 corresponding metals, the evidence seems to indicate that not over one atom 

 per thousand is reflected, if indeed, there is any reflection whatever. How- 

 ever, with molecules of the so-called permanent gases striking surfaces 

 covered with adsorbed films of gas molecules, there is some evidence that 

 a certain amount of reflection may occur. When hydrogen molecules strike 

 a hydrogen covered surface at high temperatures it is possible that as many 

 as 8i per cent, of the molecules are reflected. There is reason to believe 

 that this represents the maximum possible degree of reflection and that 

 with gases other than hydrogen and helium, the amount of reflection is 

 always less than 50 per cent., no matter what may be the nature of the 

 solid surface. 



I. EXPERIMENTAL DATA 



Evaporation and Condensation. — In connection with an experimental 

 determination of the rate of evaporation of tungsten filaments in a high 

 vacuum ^ the following relation between the vapor pressure p and the rate 

 of evaporation in (in grams per sq. cm. per second) was derived from 

 the principles of the kinetic theory 



m 



X 2tRI 



tute of Electrical Engineers at New York in 1894, described this phenomena as follows 

 (Trans. Amer. Inst. Elect. Eng., 11, 142, 1894) : "In the old Edison lamps the filament 

 was copper-plated to the platinum wires. When a break in the filament occurred near 

 the junction, the arc vaporized the copper and covered the bulb with a coating of 

 metallic copper, except that a line of clean glass was often left on the side opposite 

 the break, the line being the shadow of the unbroken leg of the filament. . . . The 

 plain old-fashioned unpretentious vapor tension that bursts our steam boilers is all- 

 sufficient to account for this rectilinear projection across the lamp bulb when there 

 is nothing in the way." In the very active discussion which followed. Professor Elihu 

 Thomson said "I am confident that I have seen it (this phenomenon) in series lamps 

 when the carbon broke. In a good vacuum the carbon vapor condensing would give 

 rise to particles which would move in rectilinear paths. The shadow effect is dependent 

 on a local evolution of vapor, with a part of the wire or filament between the point of 

 evolution and the opposite glass surface. I am also confident that I have seen platinum 

 shadows." In the further discussion, J. W. Howell said, "The molecules of carbon set 

 free by evaporation, fly from the filament as Professor Anthony says, in straight lines. 

 They are projected from every part of the filament and . . . cast shadows on the 

 globes. The reason why they do not cast shadows in most lamps is that the filament 

 is not all in one plane." Mr. Howell then exhibited to those present, both carbon and 

 platinum lamps which showed distinct shadows. 



^ Langmuir, Phys. Rev., 2, 329, 1913, and Phys. Zeitsch., 14, 1273, 1913. 



