On the Glass ijication of Diatoms. 65 



XIII. — On the Classification of J)iatoms. 

 By 0. Meresciikowsky *. 



The division of diatoms into three groups — Raphideai (or 

 l\a|)liidicge), Pseudoraphidese, and Anaraphidese or Crypto- 

 rapliidejB — generally accepted to-day is based on the presence 

 or absence of a true or a false raphe. 



The first group differs very distinctly in this respect from 

 the third — the Raphideai having always a true raphe and the 

 Anaraphidea3 having neither raphe nor pseudoraphe, forming 

 two natural and well-defined groups. 



This is not the case in the group of the Pseudoraphideai. 

 The name itself is but little applicable, a large number of forms 

 liaving no trace of a pseudoraphe. This is so in the genera 

 JS^itzschia, JViizscliiella, Hantzsckia, Eunotia, in a considerable 

 number of species of the genus Camjji/lucliscus, and in 

 several species of the genera Surirella, Synedra (5. affints, 

 S. laculus)^ Lianophora {L. araphidiea^ Mer.). Finally, the 

 genus Cliniacosphenia possesses two pseudoraphes and the 

 subgenus Aidissonia three. This character is therefore too 

 inconstant and variable to serve as a basis for classification. 



On the other hand, tiie group of Pseudoraphidese is com- 

 posed of forms very dissimilar with respect to the structure of 

 the valve. Between a Nitzschia or Surirellaon the one hand 

 and a Synedra or Fragilaria on the other there is nothing in 

 common^ and the first two genera have certainly a much 

 ntarer relationship with the Raphideaj than with the last two. 



The Pseudoraphideje thus form an extremely artificial 

 group. It seems as if all the torms which could not be 

 ranged either with the Baphideai or with the Anaraphideae 

 had been simply thrown pell-mell into this refuge. 



In a natural classification this group should disappear, ayd 

 its disa})pearance will necessitate a new division of all the 

 diatoujs. 



What is the most important character which should serve 

 as the base for a truly natural classification ? I believe that 

 it is the presence or absence of movement, which is merely 

 dependent on the presence or absence of a slit in the walls 

 of the frustule ; this character should be taken into con- 

 sideration before any other. Whether this slit be placed at 

 the centre of the valve, as in the majority of the Navicu- 

 loidea*,, or near one of its margins, as in the Nitzschioideaj 

 and the ISurirelloideai, is of secondary importance only, as is 



* Abstract of a paper iu ' Scripta Botauiea' Horti Uiiiveroitalid Petro- 

 pulitaua;, fasc. xviii. 



Ann. & Mag. N. Hist. Srr. 7. Vol. iv. 5 



