ANIMAL NATURE OF DIATOMEiE. 383 



to Odontidia; and those with perfectly smooth frustules 

 {D. pectinate, D. vitreum, D. hyalinuin^ to TragillaricB. 

 Therefore there remains no other character to distinguish 

 them except the angular connection into flexuose chains ; 

 this character appears again in other genera {Tabellaria, 

 Grammatophora, JRhabdonema,) of distant sections. We 

 shall hereafter have occasion to resume this subject. We 

 will only observe now, that this condition is always 

 associated with another, that of a peduncle by which the 

 chains are affixed to submerged objects. Perhaps it 

 happens more fi-equently and more remarkably than in 

 Odontidia and Fragillarice that we find in Diatoma the 

 cuneate form of frustules, and it is surprising to see 

 the inconstancy of the forms which Kiitzing, with in- 

 comparable diligence, has been able to collect, describe, 

 and figure, referring them to difierent species. Three 

 species (D. mesoleptum, B. elongatum, I). Ehrenhergii) 

 differ greatly from the others, being contracted in 

 the centre of the principal surfaces, and two of them 

 {D. elongatum, B. Ehrenhergii^ still more because they 

 are incrassato-capitate at the extremities of their lateral 

 surfaces. (In respect to the B. Ehrenhergii, Kiitzing 

 cites as synonymous my Glaeonema Heuffleri, because in 

 the specimen I sent him he saw only the Biatoma, which 

 is parasitic on the Glceonema, which he did not notice, 

 and so he made me appear to confound a Biatoma 

 with an Hydrurea. In the arbitrary selection he makes 

 of some characters from external form to separate 

 genera, families, and ordei's, we cannot help feeling sur- 

 prise in finding united in the same genus species which 

 present differences so marked. The character of flexuose 

 concatenation, independently of organographic condition, 

 (which as we shall see is of little value, reappearing in 

 other groups,) ought to have much less systematic value 

 than those taken from the conditions of form, which may 

 be supposed to have a direct bearing upon internal 

 organisation. For my part I think it would be much 

 more natural to place the smooth species [B. pectinate. 



