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



add something concerning the structure to be seen on it that 

 appears to have escaped the eyes of Mr. O'Donohoe. 



In the process of melting the realgar, either the great heat 

 required, or some chemical action set up by it, has acted upon 

 some specimens of P. balticum and completely dissolved part 

 of the shell, leaving only film-like strips flattened upon the 

 cover-glass, whilst others have been quite unaffected. One 

 specimen shows in a very clear manner the dissolving action in 

 progress, but arrested at the moment when a subcentral part 

 of the diatom had become fused into a structureless strand of 

 silica, connecting the two ends, which remain intact. These 

 films above mentioned, which Mr. O'Donohoe has photographed, 

 will prove, I think, to have an important bearing upon our 

 more complete understanding of diatom structure. 



If the outer surface of a perfect valve of P. balticum be 

 examined under a binocular, it will be seen that the sides curve 

 away from the raphe very much as the sides curve away from 

 the keel of a boat when turned bottom upwards, so that the 

 surface is nearly always oblique to the surface of the cover- 

 glass. From this cause I have found the structure of a perfect 

 specimen extremely difficult to understand, as a very slight 

 modification of the illumination or alteration of focus under high 

 powers, or the two combined, produce a number of different 

 appearances six or seven have been noticed all apparently 

 demonstrating true structure, so that it is practically impossible 

 to form an opinion as to which view, or views, represent the real 

 structure of the valve. Owing to this, I suppose, has arisen the 

 diverse views held of the structure by different authors. 0. Miiller, 

 for instance, in" the Deutschen Botanischen Gesellschaft for 1898, 

 Vol. XVI. p. 387, t. 26, fig. 8, regards the pores (by which 

 I understand he means the black dots) in the cell- wall as 

 perforations passing completely through the wall, which are 

 not perfectly tubular, but enlarged at their centre and contracted 

 to a minute opening on the internal and external surface of the 

 valve thus : 



porej 



Mr. T. F. Smith, however, in the Journ. Q. M. C, Ser. 2, 



