514 Transactions of the Soeiety. 



implies that the cleft furrow is very deep, a fact which is also con- 

 firmed by the transverse sections. This is simply the consequence 

 of the gradual increase in thickness of the entire membrane, whilst 

 the closing envelope is probably of even thickness. Closer obser- 

 vation will show that the surface is not everywhere alike ; the area 

 along the- midrib, i. e. the portion free from chambers, appears quite 

 plane ; the other, on the contrary, a little granulated, and occasion- 

 ally one can even see a kind of glitter of chambers. In this case the 

 cast teaches more than the longitudinal section, since it seems to 

 display an unusually fine surface-difierence which is not brought to 

 the eye by the longitudinal section. Similar experiments were 

 made with Pleurosigma. As a matter of course, this condition of 

 surface has nothing to do with transverse striae oi Pinnularia. I am 

 still in doubt whether this image is not called forth by the difierent 

 evaporation processes above the chambers, therefore perhaps it may 

 not correspond to any real difierence between the valve surfaces. 

 It is true that where air-bubbles are in the collodion the surface 

 after hardening looks difierent from what it does when free of 

 bubbles. 



Thus a complete and exhaustive explanation is given of all 

 appearances of the surface image of a Pinnularia. In the imbed- 

 ding of the valves in balsam, chamber and opening are filled with 

 the strongly refractive substance and thus produce the coarse trans- 

 verse strise. Each stria is a chamber. 



§ 3. Staining Processes. — I cannot call these experiments more 

 than tentative ; they were intended, after I had recognized this 

 most interesting condition of the chambers, to provide preparations 

 in which the chamber-spaces alone should be filled with colour. 

 The wall-substance, as is well known, does not take the staining. 

 The experiments were made with solution of silver, picro-carmine, 

 and Prussian blue ; with the latter substance only I obtained pre- 

 parations which were partially serviceable. The valves were put 

 in aqueous solution of Prussian blue, poured ofi" after some time, 

 immersed in alcohol and constantly shaken to remove the blue 

 which had deposited in and upon the valves. They were then put 

 into balsam. In successful instances, not occurring frequently, the 

 chamber is seen blue in the colourless wall. I have not persevered 

 with these experiments. 



§ 4. Passing on now to the literature on Pinnularia sculpture, 

 Schumann as far back as 1867 was approximately correct in his 

 views. In his work (23, p. 78) he says that " in a fragment in 

 partially reversed position the channels were most raised at the 

 middle ; each channel seems to consist of two vertical walls, the 

 vault across being open towards the middle line." On pi. IV., 

 fig. 54, B, he exhibits such a fragment, from which one can imagine 

 what he means. The real state of affairs could not be discovered 



