The Structure of Diatoms. By Dr. J. E. L. FlogeJ. 513 



rations in the hands of the tyro. In one and the same section he 

 observes the most beautiful closed chambers and is fully convinced 

 that my description of the Pleurosigma chambers is accurate. A 

 few micro-millimetres further on he notices the magnificent pro- 

 jections of the wall, not inferior to those of the epidermis or the 

 vessels of higher plants, and with this he proves that I have erred 

 in all points. If he follows Pfitzer's advice with regard to breathing 

 on the sections and thence obtains a mass of fragments, he will know 

 neither what is outside nor what is inside the valve. For instance, 

 if he turns round fig. 5 he wiU then see, according to his taste and 

 intelligence, Pfitzer's furrows on the outer surface, and then he 

 confirms all Pfitzer's fairy-tale. Therefore I admonish every one 

 to use the utmost precaution in interpreting the images ! 



§ 2. Collodion Casts. — For the technical process I refer to my 

 former essay (6, pp. 489-90). These casts are of much greater 

 importance with Pinimlaria than with Fleuroslgma ; I can there- 

 fore seriously recommend sceptics to try my experiments. The 

 method is so easy in practice that even inexperienced manipulators, 

 unable to do the cutting, will in this way obtain a general view of 

 the details. It has been already observed that the chambers can 

 be injected from the opening, so that in pouring fluid collodion on 

 the inner side of the valves, it enters the opening and fills the 

 chambers. With the evaporation of the ether the mass contracts, 

 and after the collodion has hardened one sees the contents of the 

 long cyhndrical chamber shrivelled-up to a thread. The image 

 takes the shape of the letter T. The vertical line is the collodion 

 filling up the chamber-opening ; the horizontal is the collodion 

 which fills the space of the chamber. If, therefore, the Pinnularia 

 valve is taken away from the cast, these small T's stand in military 

 order in lines parallel to the midrib on the film. This T is 

 mostly so elastic that without breaking it can be pulled out of the 

 chamber. 



Fig. 7 is intended to bring clearly before the mind in a diagram- 

 matic form what I have described above for a small portion of the 

 cast ; it is impossible to draw it exactly, because the interpretation 

 chiefly depends on the alteration of the focus. With the lowering 

 of the tube the horizontal T threads appear before the surface of 

 the envelope is seen, and they disappear when the latter becomes 

 visible. It is, of course, desirable always to examine for oneself such 

 a cast. The central nodule leaves behind a pretty bold depression 

 of elliptic shape. Sometimes are seen two small flat cavities 

 adjoining each other, which also indicate the depression in the centre 

 of the nodule (as above stated). Except the broken line in the 

 middle, the collodion cast of the outer surface of a valve shows only 

 a perfectly plane surface; but this line increases in distinctness 

 near the central nodule and ends with a thickened point. This fact 



