Tlie Structure of Diatoms. By Dr. J. H. L. Flocjel 677 



the surface of the frustule such a second hue cannot yet be traced 

 at all ; on the contrary the transverse striae of both valves run to 

 the apparently common marginal line. This companion cell on 

 the right is in this respect slightly more developed, especially at 

 the one end (the lower in the figure). Not only is the small 

 nodule more distinct in the middle section, but it is also more 

 distant from the larger, and on the surface one can clearly see two 

 marginal hues not quite parallel. Fig. 34 gives the line of the 

 optical middle section again, but more highly magnified. Turning 

 now to another still younger specimen, fig. 35, we observe the two 

 marginal lines slightly further apart. The divergence of the lines 

 continues until the cell obtains the requisite breadth to divide anew, 

 fig. 37. The question arises with this divergence, what becomes of 

 the gii-dle-band and how is a new one formed ? With regard to the 

 old, one sees clearly the edges are extending over the other, of course 

 mostly only with an immersion objective, and it would easily be 

 overlooked if we did not know where to search for it from Pfitzer's 

 pioneer work. With regard to the processes accompanying the 

 formation and development of a new girdle-band we find but little 

 information in Pfitzer (19, p. 56). He states that the girdle-band 

 in Pinnularia is formed unusually late, only after the new valves 

 are complete, and then where it adheres to the valve ; that it is 

 seen almost at first in its definite thickness reaching slowly to its 

 normal breadth. According to Pfitzer (19, p. 9) the girdle-band 

 has an outer edge in organic connection with the valve, and an 

 inner free edge touching the other ring but not grown together. 



With Aclmanthes — and here the non-existing marginal line 

 of Pinnularia renders capital service — we can establish with all 

 desirable certainty that the girdle originates in these lines. One 

 has only to go backwards in the various stages of development in 

 order to establish the fact. At the point where the distance 

 between the old and the young line can be well observed, one sees 

 distinctly with an oblique position, fig. 36, that the cell-wall 

 between the two lines is of double thickness, hence at that point 

 there must be already a younger girdle-band. Of course, with the 

 extreme thinness of the two bands the duplicature cannot be 

 observed directly ; the line expressing one cleft extends, as far as I 

 see, up to the young marginal line, fig. 36, r I. But since this 

 was attached originally to the older line, the girdle-band without 

 doubt is so far a double membrane, as it has already been rendered 

 probable by its thickness. Following up backwards this very thin 

 narrow girdle-band, fig. 33, on the left, we find here its origin 

 in the depression of the marginal line. We further deduce 

 from these Aclmanthes images that, according to my exami- 

 nation, it is clearly evident the girdle-band from its commencement 

 ifl attached to the cell, since it fills up entirely the inner space 



