282 PROFESSOR LANKESTER. 



siibstance which tends to take on the character of a cell wall. 

 The loculij as I have shown^ may be minute or relatively 

 large, few, numerous or absent ; and all these varieties occur 

 in one and the same form of plastid. 



That they contain sulphur is possible, as Professor Cohn 

 thinks he has shown; but that they consist of sulphur, cannot, 

 I think, be admitted. They appear to me rather to be 

 specially active points in the differentiating protopolasra of 

 which the Bacterium-plastid consists. 



The manner in which they are represented in Professor 

 Cohn's figures, is not at all like their real appearance when 

 examined by Hartnack's No. 10 immersion. This is more 

 nearly given in my fig. 12, of pi. xxiii, vol. xiii of this Journal, 

 and other figures on the same plate. 



The shape and minute structure of the plastids which build 

 up the Clathrocystis-like growths of Bacterium rubescens. 

 Professor Cohn says that the form of the plastids in what 

 he calls Clathrocystis roseo-persicina is spherical, oval or 

 angular. His figures ai'e for the most part not sufficiently 

 magnified to give an idea of the real structure of the consti- 

 tuent plastids. I must assert in opposition to Cohn that the 

 plastids in question are often elongated, and form ' myco- 

 derma-like ' and ' merismopedia-like ' tessellate aggregates 

 of oblong plastids. Further it is not true that the highly re- 

 friugent globules (erroneously termed ' dunkle Kornchen ') 

 by Cohn, and discussed above, are embedded in the cell-con- 

 tents. The plastid of this Clathrocystis- form may be per- 

 fectly homogeneous with no differentiation of cell-wall and 

 cell-contents, or you may have («) one, two, or more minute 

 highly-coloured ' loculi ' differentiated in the otherwise homo- 

 geneous plastid (these are Cohn^s ' dunkle Kornchen'), or you 

 may have {b), a great loculus highly coloured occupying the 

 whole of the medullary part of the plastid which may then 

 be said to consist of cell-wall (wall of the loculus) and cell- 

 contents (coloured, highly refringent substance filling the 

 loculus). These structural features are fully illustrated and 

 explained in my first paper. 



Hydro die ty on- like network. The condition of the Clathro- 

 cystis-phase of Bacterium rubescens, which Professor Cohn 

 compares to Hydrodictyon, is unfortunately chosen, for as 

 will be seen by comparing his figure with fig. 19, pi. xxiii of 

 vol. xiii. of this Journal — the condition which I have there 

 figured and have elsewhere compared to 'Hydrodictyon both 

 as to its actual form and mode of development — is very much 

 more closely in agreement with the beautiful green net-alga. 

 The figure which I have given represents only a small frag-i 



