512 Transactions of the Societtj. 



cause : — In transferring a frustule to the gum it may, of course, 

 be done according to method 2 without injury, but not always. 

 In further experimenting the mass, in hardening, may suffer un- 

 equal pressure on the valves, causing them to break, as also stated 

 by Pfitzer (19, p. 50), and this is very likely to occur in the mid- 

 rib. In this case they were injured before cutting. It is, how- 

 ever, more probable that injury occurs during the cutting. The 

 entire thickness of the silicified membrane adjoining the midrib 

 is about 8-10 times as large as that of the fine envelope. As soon 

 as the knife presses against this solid mass, the cap easily breaks 

 off, which appearance is also observed with thicker masses at the 

 central nodule crumbling out. In order to clear up this point I have 

 made another series of sections according to method 1, in which 

 injuries are more avoided. The best thereof show unmistakably 

 at the base of the furrow this fine capping envelope exactly as 

 represented in figs. 1-3. If the transverse section goes through 

 the central nodule, the image then becomes as in fig. 4. Hence 

 the nodule is also here a strong thickening of the midrib inwards ; 

 outwards it has no distinction beyond that the middle line is 

 broken and does not appear in the transverse section of the cleft. 

 The membrane is flat. Here it should be noted that Schumann 

 (23, p. 74) observed two focal images of this nodule in Finnularia 

 lata. I have seen them oftener in Finnularia major. But they 

 do not arise through a channel, as supposed by Schumann, but are 

 produced by a depression in the centre of the nodule. This may 

 often be seen without difficulty in a side view of the entire frustule. 

 Similarly, the end nodules are inward projections. 



I will now proceed to the description of the longitudinal sec- 

 tions. Their appearance must of necessity differ, depending upon 

 whether the section goes through the openings of the chambers, 

 or near the midrib, or near the edge of the frustule. Fig. 5 repre- 

 sents a section through the chamber- openings showing the siliceous 

 membrane with numerous long pegs projecting inwards in the 

 frustule ; these are the vertical partition-walls of the chambers. 



In fig. 6 we have a median longitudinal section ; the chamber- 

 spaces appear like beautiful squares slightly rounded off in the waU- 

 substance. The longitudinal section taken from the chamber- 

 opening towards the edge shows hardly any difference from that 

 last described. A vertical section along the midrib and through 

 the central nodule has hardly any importance. If a longitudinal 

 section goes a little obliquely towards the midrib, it shows in 

 places the image in fig. 5, and in other places that of fig. 6. 

 In all cases the outer limit of the cell-wall is perfectly straight 

 throughout ; nothing in the outline suggests furrows on the surface. 

 From this representation it may well be supposed that transverse 

 as well as longitudinal sections of Pinnularia are vexatious prepa- 



