THE CONJUGATION OF THE DIATOM EHABDONEMA ARCUATUM. 133 



other, and thus form a zigzag chain (like Grammatophora) which 

 is frequently somewhat twisted. (See Fig. 3, x 200.) 

 The frustules themselves are noticeable for — 



1. Their small size. They are the smallest observed, — the 

 length of valves or annuli being usually but '00156 inch, though 

 occasionally reaching *002. 



2. Their delicate appearance. This is partly owing to the 

 divisions between the annuli being less marked than in ordinary 

 frustules, but also through the pale colour of the endochrome. 



3. The absence of one isthmus and the small size of the other. 



4. But chiefly the paucity and arrangement of the endochrome. 

 In all vegetating frustules there is much variety in this respect, no 

 regular figure being generally found. In these, however, there is 

 always a central pale circle or disc, this being the nucleus, and the 

 remaining endochrome forms wavy curved lines, radiating in a 

 roughly-stellate manner, with a few small granules. (See Fig. 

 4,x600.) 



These are undoubtedly the male frustules, as will presently be 

 seen. 



II. There are other filaments nearly as small, generally about 

 •00188 inch, but a few as wide as '00225. These may be dis- 

 tinguished from ordinary filaments by the large number of annuli 

 in a frustule. Near the middle of a frustule is a " hoop," but much 

 wider than the " hoops " of a self-dividing ordinary filament. The 

 endochrome is usually of a darker green, and the granules denser. 

 The nucleus is frequently obscured by the aggregating granules, 

 but when detected is found to be altered, and there is no distinct 

 central mass as in the smallest frustules described in I. 



These we shall find are the female frustules. 



III. We have, then, two clearly-distinguishable kinds of frustules 

 which are concerned in the act of conjugation, but the mode is so 

 curious, .and as I believe it has hitherto been unrecorded, I would 

 specially draw attention to it. 



Most frequently the conjugation takes place with a female 

 frustule at or near the free end of a filament. At the earliest 

 stage a number of male frustules have attached themselves by an 

 isthmus to each of the half-frustules divided by the wide hoop or 

 band. They do so at any part of the annuli indifferently, either on 

 their flatter sides or ends. The number varies : it is never less 

 than two, and sometimes as many as eight crowd round it. Of 21 

 I have counted, the average on each half-frustule was four. In 



