406 
MR, W. CROOKES ON THE VISCOSITY 
the curve for the portion between 60 M and 0'02 M. On this scale one millionth of 
an atmosphere is represented by 5 millions, vertical height.'" 
The “ air” curve on diagram C shows the further diminution in the viscosity of air. 
At the foot of the diagram a straight diagonal line starts from the logarithmic decre¬ 
ment 0T124 at 1,000,000 M (one atmosphere) up to the logarithmic decrement 0'098 
at 1000 M, and a second line to the logarithmic decrement 0'0794 at 60 M. The slope 
of these lines is that of the condensed curves in diagrams A and B, and is greatly 
exaggerated, since in order to get the true slope proper to join on to the rest of the 
curve the lines should he respectively 200 metres and 5000 metres long instead of 
30 millims. 
661. The dotted curves on the left of diagrams B and C show the variation in the 
force of repulsion under the influence of radiation; the force gradually increases until 
it reaches a maximum between 25 and 40 M, when it rapidly sinks. This curve 
agrees very w T ell with similar curves given in former papers, representing the action of 
light on the radiometer! (334, 382, 481). 
The close agreement between the loss of viscosity and the increased action of radia¬ 
tion is very striking up to the 35 millionth, when the repulsion curve turns round and 
falls away as rapidly as the viscosity. 
662. On the right side of diagrams B and C I have given a third curve. The 
abscissae represent the mean free paths of the molecules at the various pressures used 
as ordinates. The mean free paths being very small, to make them comparable with 
the other dimensions adopted on diagram B, I have multiplied their lengths by 20 . 
Thus at 1000 M the mean free path is 0*1 millim., and at 100 M it is 1‘0 millim. In 
diagram B, drawn on a scale of 20 to 1 , these become respectively 2 - 0 and 20 millims., 
but in diagram C, I have given the actual lengths of the free paths of the molecules. 
The curves of increasing mean free path and diminishing viscosity closely agree. 
This agreement is more than a mere coincidence, and is likely to throw much light on 
the cause of viscosity of gases. 
RESISTANCE OF AIR TO THE PASSAGE OF AN INDUCTION SPARE. 
663. In describing the apparatus (fig. 1, par. 641) I have mentioned that it is 
furnished with a resistance tube, l. This tube is provided with aluminium poles, so 
that at any pressure the appearance presented by the induction spark when passed 
through the tube may be observed. Since the publication of my researches on the 
phenomena presented by the passage of the induction discharge through high vacua, 
the present results—which, although never published, precede by a year or two those 
* To give some idea of the high exhaustions at which accurate measurements can be taken, I may 
mention that the highest exhaustion on the table—0’02 M —bears the same proportion to the ordinary 
atmospheric pressure that 1 millim. does to 30 miles, or, converting it into time, that 1 second bears to 
20 months. 
! Phil. Trans., Part I., 1878: The Bakerian Lecture; Ditto. Part I., 1879. Proc. Roy. Soc., 
vol. xxv., p. 305. 
