PHENOMENA TYPICALLY SHOWN BY ACTINOCYCLUS RALFSII. 15 



Another thing wanting examination was the shape of these 

 ■diatoms as a whole ; and it turned out that, whilst a number 

 were almost of even thickness, some appeared to be thin in the 

 •centre, with a bulging ring around it, as best seen in the photo 

 which I will hand round, and which was taken by dark-ground 

 illumination with oblique light.* 



Another experiment tried was to see if perchance the diatoms 

 polarised light in any way ; but, as expected, they did not do so. 



A fourth line of search was opened up by an examination of 

 the spectra at the back of the objective, and this showed the 

 remarkable fact that the diffracted beams did not appear as 

 ordinary spectrum colours, but that they comprised all sorts of 

 fancy tints — such as purple and chocolate ; and, still more 

 strange, the central or dioptric beam, which is ordinarily colour- 

 less, was distinctly coloured bluish in the case of the diatoms 

 which would turn blue when the iris was closed, yellowish in those 

 which would appear yellow, and so on. A further point was 

 noted — viz., that the colour present in the dioptric beam appeared 

 to be reduced or absent in the diffracted beams. This point has 

 already been recorded by Dr. Johnstone Stoney, who, writing of 

 A, Ralfsii in an appendix to "Modern Microscopy," by Messrs. 

 Cross and Cole, p. 109, says : "It will be seen that most of the 

 red is located in a ray of first lateral beams, with an equal defect 

 of red in the central beam. Hence the blue colour seen w^hen 

 the image is formed by the central beam only." Strangely 

 enough. Dr. Stoney only refers to this diatom in connection 

 with diffraction colours, to which, as we have already seen, its 

 distinctive phenomena are not due. 



Lastly, I carefully noted in what way the colours of the 

 diatoms themselves merged into one another, and the order was 

 invariably the same. It appeared that a brownish yellow would 

 merge into purple, purple into indigo, indigo into blue, blue into 

 pale green, pale green into green-yellow, green-yellow into a canary- 

 yellow ; and this furnished a strong clue, for on turning up a 

 table of Newton's colours of thin films the order of a part of the 

 series corresponded almost precisely with the changes mentioned. 



Of course it was not a question of thin films ; for, as already 



* Fig. 11 shows sections of A. Ralfsii, as near as I could ascertain from 

 ■examination of a number of specimens mounted on edge and in different 

 positions. 



