378 ANIMAL NATURE OF DIATOMEiE. 



substance entirely lines the inner wall of the shell, it 

 becomes divided afterwards into minute portions, which 

 terminate by concentrating themselves in distinct globules. 

 I believe that these changes only occur after death. 

 Bailey succeeded in extracting a greenish resinous siib- 

 stance by means of alcohol. Only two well-defined 

 species inhabit the fresh water of all Europe ; two others 

 are doubtful. 



5. EuMERiDiON. — Individua aineiformia, prismatico- 

 trapezoidea (?), inflahellum velfasciam convolutam coalita, 

 demum stipitata. Stria tramversales pe^-vits, validcB. 



As Kiitzing has only seen dried specimens of the 

 Meridian constrictum of Ralfs, upon which he has con- 

 structed this genus, it seems that we may assert from the 

 exact description of this author, quoted at length by 

 Hassall, that there is no indication whatever of that 

 gelatinous pedicel like a cushion, upon which Kiitzing 

 asserts that the frustules are arranged in a fan-shape, 

 as in the Synedroe. Perhaps this erroneous impres- 

 sion may be founded on Ralfs' figure and descrip- 

 tion of the dried specimen, compared with the living 

 one. These are his graphic words :• — " As, however, 

 they are not arranged in a plane, as in Meridian 

 circulare, but stand nearly erect, somewhat like the 

 staves of a tub which is broader above than below, 

 when they are dry and fall down they necessarily 

 separate, and gaps are produced in the circular outline. 

 In the dried specimens I find some of the frustules 

 arranged in a circle, which however exhibits the gaps 

 already noticed, whilst others seem to be fasciculated." 

 Therefore there remain only the trapezoidal section from 

 the inclination of the lateral surfaces and the constriction 

 of these near the apex, which characters, though un- 

 doubtedly of value to discriminate species, are not suffi- 

 cient to establish a genus. 



The superfluous division of genera, whilst it disguises 

 the true affinities of objects, and fruitlessly complicates 



