PHENOMENA TYPICALLY SHOWN BY ACTINOCYCLUS RALFSII. 19 



3 and 4). The same processes occur ; but as the path of the rays is 

 obviously shorter, the half-wave-length difference must be for light 

 of some smaller wave-length. Consequently it is some other colour, 

 for instance yellow-green, that gets stopped out, and the tint 

 of the diatom will be purplish at this spot. And I have verified 

 that, as the diatom is thicker or thinner in different parts, 

 so are colours of longer or shorter wave-length extinguished ; 

 and of course their place is taken by light of just the com- 

 plementary colour, the merging of one colour into another being, 

 as mentioned, in accordance w^ith the Newtonian series. 



Now for some further proofs of the correctness of the above. 

 So far, Ave have assumed that the light impinges on the diatoms 

 almost perpendicularly ; if it impinged on them obliquely, as 

 shown in Figs. 5 and 6, the relative differences between the 

 length of passages in balsam and in silex will change and 

 consec^uently the colours ought to change — and so they do. 





Again, if instead of using a comparatively narrow cone of 

 light, we use a very wide one as compared -with the objective, 

 we get for each obliquity white light passed minus some different 

 colour. Tf we were to sum up the light for, say, seven different 

 slants, it w^ould be : White minus violet ; w^hite minus indigo ; 

 white minus blue ; white minus green ; white minus yellow ; 

 white minus orange ; white minus red ; which is like saying 

 7 w4iite minus 1 white, and this equals white : in other words, 

 with a suitably large cone, the colour would disappear and the 

 diatoms would appear white — and so they do. 



This, again, is the explanation why with dark ground illumina- 

 tion these diatoms appear mostly white, but we can readily make 

 them assume colour on a dark ground by narrowing the annular 

 cone of light passing round the dark spot, as is easily verified. 



The above explains the main colour-effects, but there are 

 also some subsidiary ones that deserve notice. If the iris 



