PHENOMENA TYPICALLY SHOWN BY ACTII^OCTCLUS RALFSIl. 21 



forations per unit area will appear of a paler tint than one 

 where the perforations are more closely set. 



All the peculiar colour effects of A. Ralfsii have now been 

 explained, excepting two, which puzzled me exceedingly a 

 long while. One of them was, that when the aperture of 

 the objective much exceeded '45 N.A. the diatom should appear 

 colourless. Surely the colour that had been extinguished in 

 the central or dioptric beam could not be added again by 

 admission of light which the object itself had thrown up to the 

 other parts of the objective. And the other thing was, why 

 should the diffraction spectra assume " fancy " tints instead of 

 the orthodox spectral colours ? 



It seemed at first that these two phenomena must hang 

 together in some mutual connection, but that turned out to be 

 a false scent, for the cause of the compound tints was found 

 to be due to the unequal thickness of the diatoms in different 

 parts. The consideration that some parts of the diatom would 

 more or less refract the dioptric beam afforded sufficient ex- 

 planation why the slant of refracted beams should vary accord- 

 ingly, so that it might well happen that differently coloured 

 portions of the diffraction fan proceeding from the differently 

 situated points on the object might overlap, and then of course 

 compound tints would be formed. 



This point settled, the question of the diatoms being colourless 

 with objectives of large N.A. became still more perplexing. At 

 last the mystery was cleared up, and by the very selfsame reason 

 that caused the interference effect with the dioptric beam ; for 

 when it was remembered that the diatom was not an ojyaque 

 plate studded with holes, and that in the general effect of the 

 diffraction, light w^hich passed through the silex, and consequently 

 was some fraction of a wave-length ahead of that passing through 

 the balsam, would participate, — as soon as that was realised the 

 whole difficulty vanished. 



It became at once evident that for the different slants of the 

 diffracted rays the light of any one colour would be enfeebled 

 more or less. It must not be thought that the light must 

 necessarily be of full intensity or completely extinguished be- 

 cause I have postulated this for the sake of illustration. 

 There may be a quarter-phase retardation, or some other fraction, 

 which would diminish the light without stopping it out altogether. 



