322 N. E. BROWN ON THE STRUCTURE OF DIATOMS. 



seen to be thickened in a beaded manner at short equal intervals. 

 Tn some cases a few of these rods are curved away from the 

 surface of the valve, and one of them in such manner that part 

 is seen in surface view and part seen in side view, and traceable 

 from one view to the other (see Mr. O'Donohoe's photograph, 

 op. cit., t. 14, fig. 2. where it or a similar rod is shown out of 

 focus). Now it is obvious that the bead-like swellings occur at 

 the points where, in the perfect grating, the transverse bars 

 crossed and were fused with those parallel to the raphe, and 

 that these transverse bars were either more easily dissolved or 

 lie at a slightly lower level than the longitudinal bars, and so 

 are more quickly attacked by the dissolving action. In all cases 

 the films are very closely appliel to the cover-glass, indicating 

 that its cooler surface has in some way retarded the dissolving 

 process, so that the parts of the diatom farthest from the cover - 

 glass were always dissolved first. That the longitudinal bars over- 

 lie the transverse bars seems to be probably the correct view, as 

 under certain conditions of illumination the longitudinal bars seem 

 to pass over the transverse ones, and is supported by the testimony 

 of the curved bar mentioned above as seen in side view. At 

 a magnification of 3,000 diameters it is clearly seen that the 

 edge of the bar facing the outside of the diatom is perfectly 

 even, while the edge facing the interior projects into little hemi- 

 spheres at the points where (in surface view) it is bead-like (fig. 2). 

 Also at the marginal part of the valve, where the longitudinal 

 bars are normally undeveloped and only the transverse bars 

 are evident, these latter become pressed nearer to the cover- 

 glass, and are not dissolved. 



The bars can be distinctly seen to be solid pieces of silex, which 

 go to form the strengthening grating and support the membrane 

 which covers the exterior of the diatom. At a certain focus the 

 beads or nodes at the crossing of the bars, owin2f to refraction 

 or diffraction, assume the appearance of black dots so familiar to 

 all microscopists, demonstrating conclusively that these black 

 dots are not pores, but shadows produced by some refractive 

 or diffractive property of the nodes of the grating-bars. 



From the movements T have seen diatoms perform it is evident 

 they must have some means of communication through the valves 

 with their surroundings, and finding that the black dots on this 

 diatom are certainly not pores, I sought for them in the membrane 



