INTRODUCTION. XXIX 



membranaceous envelope, in which the frustules are subsequently 

 imbedded, from those in which such secretion is altogether 

 absent, or is represented merely by a cushion or stipes, to which 

 the frustules are attached by a small portion of their surface ; 

 and I have placed the latter, as of simpler organization, in my 

 first tribe, arranging the genera belonging to it into sub-tribes, 

 depending upon the permanency or otherwise of the connecting- 

 membrane, another product of the self-dividing process. This 

 enables me to place apart those genera whose species present us 

 with frustules in which the union of the cells is dissolved almost 

 immediately upon the completion of self-division, as well as those 

 where a cushion or stipes still maintains a kind of indirect in- 

 dividuality in the divided frustules, from the genera in which 

 the cells cohere after gemmiparous increase, and by such co- 

 herence form filaments of various lengths and forms, allotting 

 the latter to sub-tribes which respectively present a compressed 

 filament, a zigzag chain, or a cylindrical thread. In the second 

 tribe, including those genera which have frondose forms, I find 

 characters for my sub-tribes in the nature of the frond and the 

 arrangement of the frustules. 



I do not propose this arrangement as free from exceptions or 

 even serious defects, but I have adopted it in preference to those 

 hitherto given, as bringing more frequently together forms 

 allied in structure and mode of growth, and as being at the 

 same time more strictly in accordance with the external physi- 

 ognomies of these organisms, and therefore more likely to be 

 apprehended by the inquirer entering upon the study of this 

 department of nature, A wider study of Diatomaceous forms 

 will doubtless lead to more accm^ate and more natural generali- 

 zations. 



Section VII. 



On Collecting and Preserving Specimens of the 

 DiatomaceyE. 



I have already described the various localities in which the 

 Diatomacese ordinarily abound. Supposing the observer to be 



