1 891.] NEW-YORK MICROSCOPICAL SOCIETY. 65 



upon me by the continual work of nearly three years in a manner 

 which left me no room to come to any other conclusion. My state 

 of mind on beginning to work on a dry slide of P . formosum was 

 one of utter bewilderment at the variety of appearances presented 

 by the different valves, and my first impulse was not only to 

 throw up the work, but also the microscope, in disgust, as giving 

 no image of any structure which could be depended upon. After 

 a time, however, I began to classify the appearances by observing 

 that it was different sides of the same diatom presented to me, 

 and, given the same side, the same appearance. After this my 

 work became easy, and all my subsequent work a natural out- 

 growth from the labors preceding. 



In coming before you with a brand-new theory of diatom-struc- 

 ture — as far, at least, as the Pleurosigma are concerned — I know 

 the difificulties of the subject; how easy it is to mistake interfer- 

 ence phenomena for new structure; how even the assumption 

 by me of different layers would put me out of court on the laws 



of these phenomena alone, did I attempt to base my theory on 

 these appearances only. But I have made no such attempt, have 

 never assumed the truth of any appearance until I have examined 

 the structure causing it, by seeing it isolated from anything 

 above or below. This I have been fortunate enough to do in 

 numerous instances, and, after you have seen the results on the 

 screen, will leave it with confidence to your judgment to decide 

 whether I have been deceived in my conclusions or not. 



The ultimate structure of the outer layer of the outer, or con- 

 vex, side of P. formosum is seen in slide No. 9, and consists of a 

 long fibril subdivided into short bars (Fig. 3), but how joined 

 together I have not been able to discover. These fibrils are 

 placed side by side lengthways on the valve, and run from end to 

 end, and when perfect the appearance is as in Fig. 4, as plainly 

 shown around the nodule in slide No. 11. When the structure is 

 perfect, a white focal "bead" is formed in each of the larger 

 interspaces and the whole run diagonally across the diatom, as in 

 slides Nos. 5 and 13. On focussing inward you lose the bright 



