Experiments in a Highly Rarefied Gas. 413 



the Gaede pump when the gas pressure was already 0*00001 

 mm., as indicated by the McLeod gauge, I feel sure that the 

 partial pressure of gas finally was under 0*000001 mm. 



Now the quartz apparatus is ready for the experiment. 

 It is brought on the optical bench, put in position, and the 

 condensation-chamber (1) submerged in liquid air. Then an 

 exposure is made. Thirty minutes give a photo as shown in 

 fig. 3 (PL VI.), ninety minutes in fig. 4 : in both cases very 

 distinct fringes. In fact, fig. 4 (PI. VI.) shows the fringes 

 of practically the same intensity as fig. 1 (PL VI.). 



No trace can be found of any diminution in number or 

 distinctness or intensity in fringes if produced in a gas of 

 very low pressure, compared with the appearance of fringes, 

 given by the same apparatus at atmospheric pressure. 



Conclusions. 



The experiments show that we have no reason to assume 

 that a subdivision of light units is taking place when these 

 units pass through matter. The resonance hypothesis is 

 therefore not able to reconcile the existence of interference 

 phenomena with the conception of " units " sketched above. 



But are we forced by this experimental result to assume 

 the hypothesis that the energy of the elementary radiator is 

 only spreading in spherical waves ? Not, if we succeed in 

 developing a conception of a unit which produces effects 

 similar to those of waves when falling upon an interference 

 arrangement. 



Such a conception seems, in fact, possible. 



According to J. J. Thomson we should not assume that a 

 line of force is more or less at rest and could not have rapid 

 sideways motion. Then, of course, a succession of oscil- 

 lations travelling along such a moving line is no longer 

 propagated in a straight line. Therefore, we can imagine 

 quite well that different parts of one and the same train of 

 oscillations are passing through both slits of the interference 

 arrangement used in our experiments. If all the lines of 

 force move across the interference apparatus at the same 

 rate, the parts of the trains which pass through the slits will 

 have a constant lack in phase. So we should get a pattern 

 much the same as we should expect for waves. 



With regard to the velocity of the sideways motion 

 of a line of force, we are practically not limited as this 

 velocity represents a change in direction and not a real 

 velocity. 



Phil. Mag. S. 6. Vol. 27. No. 159. March 1914. 2 F 



