354 



KNOwij.nr,!-: 



Si TTKMBKR. 1012. 



sufficient diptli to be readily measured, when 

 seen in fragim-nts and in profile . . . The 

 floor of tile cells is also minutely punctate, beiii^,' 

 arran;,'ed in (|uincun.\. The minute puncta recpiire 

 careful illumination and a power of four liundrcd 

 diameters to render them quite distinct." 



We next come to 1872. In a brief coniinuiiica- 

 tion to The Lens — a short-lived American micro- 

 scopical journal, Dr. \\'ood\vard, the eminent 

 .\merican microscopist, says of the " Double 

 Markings of Tricerafiiiin," of which photogra[jiis 

 are enclosed : " When the rows of dots seen in the 

 broken frustule are examined bv immersion oii- 

 jectives of the hij^hest power and best quality, thev 

 appear with white liij;ht as garnet-red beads on a 

 greenish ground, approximating in size and appear- 

 ance to the beads of I^leiirosigma aiii>iilatiiiii 



The comparative thickness of the frustule, and the 

 marginal walls of the areolae, interfere considerably 

 with clear definition with these high powers; still, I 

 think an impartial examination of the figures 

 will dispel all doubts as to the real nature of the 

 markings." 



The above passage is interesting and instru(ti\c 

 in even more senses than Dr. \\'ood\\ard intended 

 at the time of writing. He speaks of immersion 

 objectives of the highest power and quality as being 

 em[)loyed to resolve the markings. Oil immersions 

 were not known then, but his own favourite glass 

 was a one-sixteenth iiicli water immersion of 

 Powell cS: Lealand with a X.A. of aiiout 1-21. 

 Figure J87 of T. favus was taken with a dry. 

 one-sixth inch of 0-85 N.A., and speaks elocjuentix 

 of what we owe to the new Jena optical glass in 

 producing cheap lenses. This, however, in passing. 



That the markings on the floor of T. favus and 

 allied species occupied a lower level than the walls 

 of the hexagons, was the commonly received opinion. 

 Indeed, it appears to be the common opinion now. 

 Messrs. Nelson and Karop, in a paper read before 

 the yuekett Club in 1886 : " On the Finer Structure 

 of certain Diatoms," say of T. favus: "The 

 coarse areolations are hexagonal in form and very 

 deep. At the bottom of these is a delicate perforated 

 membrane, the perforations being circular and 

 arranged for the most part in rows. Figure 3 

 shows a fracture, passing through the minute perfor- 

 ations, the resolution of which may be considered 

 one of the most crucial tests for the microscope of 

 the present day." The drawing in ([ucstion shows 

 the walls of the hexagons sharply outlined, with little 

 dots at the corners, which the present writer only 

 finds upon focusing lower down from that side. 



It was, in fact, a case of not being able to see the 

 forest for the trees. By the sheerest good luck the 

 writer hapi)ened upon one where most of the trees 

 had been felled, as it were, and thus was able to see 

 things in their proper relationship. This is the 

 view he now wishes to place before microscopists 

 through the pages of " Knowledgp:." The normal 

 appearance of this diatom is shown in Figure jX7 

 magnified six Inmdred diameters, yet to be well 



Seen it should be enlarged some three or four 

 times more. Another view would show the white 

 tiot instead: only, after all, a question of focus. 

 I'igure 388 is part of the same valve magnified 

 two thousand five hundred diameters, in which the 

 w hite dot is represented. 



Then, using the same dry lens, he came ujjon one 

 looking rather queer. Naturally, the desire was to 

 find out the cause of this divergence from the normal, 

 and a one-twelfth inch oil immersion was substituted, 

 when the appearance was as shown in F"igure 389, 

 magnified one thousand five hundred times, taken 

 from one corner of the same valve. It was here 

 simply a question of torn structure spread all over 

 the surface, from which it would have been quite easy 

 to produce another half-a-dozen s[)ecimens, had it 

 served any good purpose. Figures 390 and 391 , taken 

 on a larger scale from the same valve, and showing 

 the white and black dot aspect respectively, illustrate 

 his views of the normal structure of this side. The 

 reader can take whichever rendering he likes, though 

 one thinks a great deal too much is being made of 

 this white and black dot business. The climax of 

 astonishment is arrived at, certainly, when one hears 

 of ten optical level sections of a diatom photo- 

 graphed between the points of the two dots. These 

 refinements of manipulation seem as little relevant 

 to fact as was the ancient academic discussion : 

 " \\'hcthcr a thousand angels," or was it ten 

 thousand ? " could dance on the point of a needle." 

 It will lie seen in Figures 395 and 396 that 

 there is not the slightest difference, either in size 

 or contour, between the two aspects. 



The opinion already given, in the article in .\ugust 

 and September numbers of "Knowledgk" for last 

 year, was that the conventional appearance of 

 diatoms, at least in the finer forms, is not the real 

 structure. More recent researches of the writer 

 only tend to confirm that opinion, which the various 

 prints presented here should prove. The torn 

 structure of T. favus all takes on the same 

 character as the fibrils in P. fonnosum and of the 

 Pleurosigma generally. They assume the same 

 zig-zag, or wavy shape, either long or short, 

 as the case ma\' lie. In Pleurosii^iiia thev 

 stretch lengthwa\s upon the vaKe without inter- 

 ruption from end to end. Here in Triceratium. 

 and the discoid forms, they are found in short 

 lengths onl\-, springing from the structureless parts 

 covering the tojis of the hexagons immediately above 

 the enclosing walls. Loosely thrown together the\' 

 are easily torn, and to this we are indebted for the 

 knowledge of their structure. As if to remove the 

 last reasons for doubt, in some instances a single fibril 

 is found sticking out by itself, only connected at one 

 end to the structureless membrane. Figures 390 

 and 391 illustrate both the zig-zag formation of 

 the fibrils, and how they combine to produce the 

 appearance of the hexagons similar to those on one 

 side of the \alve of P. angulatum. It must be under- 

 stood tliat most of the structure is torn away here, 

 leaving all the more op[)ortunit}- to judge of the 



