18 J. RHEINBERG ON THE ORIGIN OF CERTAIN COLOUR 



they are composed of a clear, siliceous, glassy material, and any 

 lif'ht reaching them passes » through their substance as well 

 as through the perforations. And now see what happens 

 when light impinges from a single point (supposed to be a 

 good distance oft", so that the light may be taken as parallel) 

 on an A. Ralfsli. 



Suppose Fig. 1 to show a small part of a valve with a per- 

 foration in it, and let the dotted lines with arrowheads represent 

 light passing through the silex and through the perforation 

 respectively. As the valve is usually not of so dense a sub- 

 stance as the medium in which it is embedded, and as we 

 know that wave-lengths of light of any colour are longer 

 in a rarer medium than in a denser one, or, to put it another 

 way, more waves get squeezed into the same length of path 

 in the denser medium, it may happen that for some 

 particular wave-length the ray passing through the medium 

 has just got half a wave-length ahead of the one passing through 

 the silex. In the figure the transverse marks on the dotted 

 lines represent wave-lengths, and the retardation of half a 

 wave-length can be observed. This means that the crests of 

 one set of waves arrive at the same time as the hollows of the 

 waves of a neighbouring ray, and as the two rays have come 

 from the same far-off point, they are in a condition to interfere 

 and cancel one another completely according to the well-known 

 law. Expressed more shortly, in the language of optics, the 

 waves arrive with a half-phase difference and completely 

 interfere. 



Of course this complete interfei^ence has taken place for light 

 of one particular wave-length only. Suppose it to be yellow 

 light that has been stopped out in this way ; then the white 

 light in the vicinity of the perforation, being deprived of its 

 yellow component, appears blue. If the perforations in the 

 diatom were large, there would be a blue ring about their 

 edges, but being so small the inner edges of the ring overlap 

 and they apj^ear blue altogether. 



Fig. 2 shows how a similar result can come about if the 

 diatoms are mounted in some medium less dense than the silex. 

 Here the wave-lengths are shorter in the silex, and the diagram 

 shows an acceleration of a half-wave-length on emergence. 



Next consider a place where the diatom is thinner {vide Figs. 



