20 INTRODUCTION. 



In all these respects they differ from the Diatomaceae. In 

 the latter each frustule consists of three pieces, one central 

 and ring-like or continuous all round, and the others lateral. 

 The division is completed by the formation of new portions 

 within the enlarged central piece, which then falls off, or 

 else by a new septum arising at the centre ; but I believe 

 that, in every case, the separation commences internally 

 before it extends to the covering*. Their coverings, with 

 very few exceptions, are siliceous, withstand the actions of 

 fire and acids, and may be broken but not bent ; the frustules 

 are often rectangular in form, are never warted, and scarcely 

 ever spinous. Their internal matter is usually brown when 

 recent ; and, although some species are greenish, or become 

 green after they have been gathered, none are of a truly 

 herbaceous colour. Their vesicles bear some resemblance to 

 those in the Desmidiese, but they are of a yellower colour, 

 and no starch has been detected in them. The Diatoraacese 

 do not conjugate!, and in general they very soon give out an 

 offensive odour. 



I have preferred treating of these topics first and at some 

 length, because I wish the reader to enter upon the subject 

 unprejudiced, so that after a fair examination of the con- 

 flicting arguments he may be ready to surrender his judge- 

 ment to that which is supported by facts of greater weight. 



The first reason advanced by Ehrenberg in support of the 

 animal nature of the Desmidiese is, that they possess a 

 " voluntary motion ;" but I must protest against his use of 

 the term voluntary as prejudging the matter in dispute and as- 

 suming more than in the present state of knowledge can be 

 ascertained ; more indeed than he has attempted to prove. 



That the Desmidieee move must be admitted ; for this fact 

 has been noticed by too many accurate observers to permit 



* For detailed descriptions of the mode in which cells are multiplied by 

 division in the Diatomaceae, see the Transactions of the Botanical Society 

 of Edinburgh, vol. ii. p. 37. 



I Since the above was in type Mr. Thwaites has detected four species of 

 Diatomaceae in a conjugated state ; the si)orangia are elongated and in pairs, 

 and the internal matter is similar to that of the frustules. 



