302 J. RHEINBERG ON BLACK AND WHITE DOT PHENOMENON. 



am quite satisfied now that Mr. Nelson's view is the correct one, 

 and that the appearance of shifting arises from a curious optical 

 illusion to which I will refer a little later. 



Then Mr. Stokes' observation, that, since isolated objects (such 

 as bacteria) and edges also show the effects in question, they 

 cannot be due to bands or cones arising from periodic structui*e, 

 certainly merits attention, and makes it necessary to explain 

 that, according to my view, the bands or cones arising from a 



Fig. 1. 



single cylindrical perforation in a diatom are by themselves 

 competent to give two images, viz., one chief image, and one 

 above or below. Since, as above stated, a dark area is only 

 formed on one side of a partition, a vertical section of a 

 perforation would show us two more or less expanding bands of 

 darkness diverging from each other, and two similarly shaped 

 bands of extra brightness likewise diverging from each other. 

 Apparently it io where these bands cross respectively that we 

 get the maximum black and white dot effect. Fig. 1, a mere 

 diagram, gives a general indication of what occurs. I believe 



