Discontinuity in the Pheyiomena of Radiation 55 



already indicated in the ripple-tank photograph the wave length 

 of the light must be comparable with the distance separating 

 the successive spaces of the grating for the experiment to succeed. 

 Fourtnately the skill of such physicists as the late Professor 

 Kowland, of Baltimore, has placed at our disposal gratings with 

 as many as 14,000 lines to the inch, and also photographic copies 

 of these. Such fineness of ruling is absolutely necessary for 

 accurate work on light. The wave length of any homogeneous 

 light is so small compared with our usual standards of measure- 

 ment that we find it convenient to introduce two new units of 

 length into use, viz., the millionth part of a metre, i.e., the 

 thousandth of a millimetre, which is called a micron, and the 

 thousandth part of this length called a millimicron. Rowland's 

 gratings have their spaces separated by distances of the order 

 2 to 3 microns, the separation being known from the mechanical 

 means by which the ruling was carried out. The results of 

 measurements with such gratings and other apparatus designed 

 for similar purposes is to prove that when we analyse the white 

 light of the sun or electric arc into a spectrum band on a screen 

 by prism or grating the light which illuminates the extreme red 

 end of the band has a wave-length of -8 micron approximately, 

 while the light illuminating the other, the violet, end is 

 about -4 micron in wave length. To indicate the precision of 

 measurement possible in spectrometric work, it is known that 

 the yellow light emitted from incandescent sodium vapour is 

 composed of two homogeneous qualities, one having a wave length 

 •5890 micron and the other -5896 micron, a difference in wave 

 length of but -6 millimicron. This excessive shortness of wave 

 length carries with it as its natural counterpart an extreme 

 rapidity of vibration on the part of the mechanism inside the 

 atoms of the radiating, luminous material, and also on the part 

 of the ethereal medium through which the light is transmitted. 

 These frequencies can be calculated very easily from the known 

 speed of light through the ether, which is 300 millions of metres 

 per second. For extreme violet light it turns out to be about 



