20 J. RHEIXBERG ON THE ORIGIN OF CERTAIN COLOUR 



diaphragm above the objective be opened just, and only just, 

 sufficiently to resolve the perforations, and a high eyepiece 

 be used, it will be observed that surrounding each dot of the 

 main colour there is a ring less marked of another colour. 

 The cause of this in the case of light incident perpendicularly 

 is that the perforations are not truly cylindrical, as diagram- 

 matically figured, but that the silex is slightly convex or 

 irregular, so that some rays pass through a gieater thickness 

 of silex than others. This is best explained by the action of 

 rays impinging obliquely (Figs. 7 and 8) as with these the sub- 

 sidiary colour effects must occur whether the perforation is truly 

 cylindrical or not. On referring to Figs. 5 and 6 it will be 

 noticed that any rays passing between A and C in a parallel 

 direction must travel partly in silex and partly in the mounting- 

 medium. The rays B in Figs 7 and 8 illustrate this. They 

 meet the ray D, which started from the same point originally, 

 in P, and there interfere according to the difference of phase in 

 which they arrive. From the figures it is also apparent that 

 each successive ray as we pass from A to C would traverse more 

 of the medium and less of the silex, or vice versa, which means 

 that we should get a rainbow-coloured ring. In practice, 

 however, the rings ajDpear of only one tint, owing firstly to the 

 excessively narrow space to which the whole spectrum range is 

 confined, secondly to the lioles not being absolutely cylindrical, 

 thirdly because rays impinging at several angles at the same 

 time tend to counteract any pure colours,* and fourthly because 

 we usually get a perfectly dark annulus in connection with 

 them — which point I v/ill refer to later on. 



The total colour effect seen on Actinocyclus Ralfsii, viewed 

 mth a power not sufficient to resolve the perforations, is a blend 

 of the main colour effect with the subsidiary ones plus an 

 admixture of white light passing through the intercostal silex. 

 The sparser the perforations per unit area, the greater the 

 amount of intercostal silex, and the greater the amount of white 

 light passing through it. So that a diatom with fewer per- 



* It could be easily proved that the light impinging on the diatoms 

 at various angles at the same time would tend to annul the colours of 

 the rings to a far greater degree than it would affect the main colour 

 produced by rays traversing the whole thickness of silex and medium 

 respectively. 



