56 BRITISH DESMIDIE^. 



of unequal breadth, but the form of its joints also varies, as more or 

 less of the angles is seen at the margin ; in short, as they are at one 

 time fully visible, and at length entirely disappear. 



In a transverse view the joints are circular or broadly elliptic, with 

 two minute opposite projections which are formed by the angles. 

 The endochrome is radiate ; its rays from four to seven. 



Conjugated specimens of both species have been observed, and 

 possess much interest from their being the representatives in this 

 family of analogous states in some of the Conjugatae. 



Didymoprium differs from Desmidmm in having only two angles ; 

 in a transverse view also the number of rays of the endochrome does 

 not depend on the number of angles. 



Kiitzing separated this genus from Desmidium and Hyalotheca, and 

 as this separation meets with the approbation of the Rev. M. J. 

 Berkeley, I am induced, in deference to their joint opinion rather 

 than from my own conviction of its necessity, to adopt it in the pre- 

 sent work. Mr. Hassall is averse to its adoption ; but surely he is 

 mistaken when he puts the differential characters of Didymoprium 

 and Desmidium on the same footing as the presence or absence of a 

 mucous sheath, and thinks that if Didymoprium is a good and vahd 

 genus, the absence of a mucous sheath would be sufficient to separate 

 from it Didymoprium Borreri. 



The absence or presence of a mucous sheath, or rather its com- 

 parative development, since traces of it may be found in all the fila- 

 mentous Desmidieae, can only be of specific value ; and Mr. Hassall 

 himself, in the analogous example of the Conjugatae, amongst which 

 species of Tyndaridea and Mougeotia have equally developed sheaths, 

 has only in one instance thought it necessary even to allude to its 

 occurrence. 



In addition to the characters already enumerated, I may observe 

 that in Desmidium a transverse view shows that the cell itself, irre- 

 spective of the bidentate projections, is angular, and that the endo- 

 chrome is divided into a number of rays corresponding with the 

 number of angles : neither of these circumstances occurs in Didymo- 

 prium. 



The filaments in this genus increase in length by the repeated divi- 

 sion of the joints exactly as in the other genera, the new portions 

 being formed between the original segments, which in other respects 

 remain unaltered j the teeth of the angles furnish a guide to the change 



