Navicula serians, N. rhomboides, and Pinnularia gibba. 167 



established by actual observation, since it seems just as natural 

 to me that the smaller or conical part should supply itself with 

 « larger portion for a base, and that the larger or basal part 

 should supply itself with a smaller portion for its cone, as that 

 the hemispheres of a divided spherical cell of Melosira should 

 make up their deficiencies as they do respectively without under- 

 going the least diminution in their new halves; if anything, I 

 think now that the frustules should increase in size, as Ralfs's 

 figure of ' Fragilaria virescens (Annals, 1843, vol. xii. pi. 2. fig. 6 a) 

 would appear to show. 



My view, that one of the frustules in conjugation was always 

 smaller than the other as a matter of course, and not an acci- 

 -dental occurrence, should also undergo modification, and has 

 therefore been rightly opposed by the late Prof. Smith; but, 

 although this excellent diatomist has stated that it is alto- 

 gether " irreconcileable " with the conjugation of the stalked 

 varieties, where it takes place between the halves of the frus- 

 tule which has just undergone division, and a matter of chance 

 rather than one "of course" among the other tribes, still the 

 entire separation of the halves of the free frustule on dupli- 

 cative division makes the chances of their coming together 

 again for conjugation so very remote that it can rarely occur. 

 Hence, although with the conjugations of the free Diatomese it 

 does not follow as a " matter of course " that one frustule should 

 be smaller than the other, it will be inferred, from what has been 

 above stated respecting the great diversity of size which must 

 exist among individuals of the same species, that with them there 

 is very little likelihood of two of exactly the same length meeting 

 again for this purpose ; so that it probably seldom happens that 

 there is not some disproportion between them, although this does 

 not generally amount to very much, and is not a necessary part of 

 the process, as I had formerly supposed when I thought the object 

 of conjugation might be chiefly to restore the size of the frustule, 

 already diminished by repeated duplicative division, rather than 

 (as I now think, and shall presently state that it is chiefly) to 

 raise it to a size fit for impregnative generation. 



As regards the diversity of size generally among individuals 

 of the same species of Diatomese, it now seems to me that we 

 must seek for the explanation of this chiefly, if not entirely, in 

 the different sizes of the conjugations, together with the varying 

 sizes of the conjugating frustules themselves, commencing from 

 the embryonal frustules upwards, until the maximum size of the 

 species is attained, or, to use other words, in the amount of 

 protoplasm, &c, furnished by these frustules; and therefore 

 that this conjugating, however much it may be a modification, 

 is not the process of impregnative generation in the Diatomese, 



