1891.] MIOIIOSCOPICAL JOUIINAL. 5 



If a xvholc frustulc were formed inside each half then the size would 

 progressively diminish, but this is not the case, nor would it correspond 

 to the growth of the desmids, which is very similar. How lo7ig the 

 vitality of both of the original valves continues is of course an open ques- 

 tion, hut the two form one cell at the start, and the vitality of that cell is 

 etjually shared by both valves. In each new generation, then, by conju- 

 gation the two valves are of the same age. Furthermore, there is no 

 apparent reason why the two older valves should not maintain their vi- 

 tality for a considerable length of time, for the original semi-cells of 

 the desmids do, for in a filamentous desmid the halves of the original 

 cell are at the extreme ends^ and those two halves have been concerned 

 in the production of all the intermediate cells. At the start the two 

 halves were united. Tliey separated and each half formed a new half. 

 Then there were two cells instead of one, each having one-half of the 

 original cell. Each new cell divides again, that is to say, one of the 

 original halves forms again another half for itself, and the newer half 

 does the same, and the result is that the two original halves are at the 

 ends of the filament, while the two halves which were formed next are 

 at last in the middle of the filament. And how much vitality the two 

 original halves have we may gather from the fact that Wolle figures a 

 fragment of a series of adhering cellsof Mlcrasterias (fig. 8, pi. xxxviii) , 

 which he says must originally have numbered no less than sixty-four. 

 Nor is there any reason to suppose that even that is the limit for all 

 cases. My point therefore is that the larger valve of the original frus- 

 tule of the diatom can go on producing complementary valves again 

 and again which will be as large as the other half it originally had, and 

 so our frustule will keep the normal size for each individual of the 

 species that is present at the time division begins. Probably the diatoms 

 follow the same course as the desmids, multiplying by division, with 

 diminished vitality in both the original valves until the latter part of 

 the summer. 



I may be wrong, and, if so, am open to correction, but the impres- 

 sion Carpenter's description makes upon me is that the process of con- 

 jugation is necessary to the recovery of the normal size, because of the 

 progressive diminution which is incident to repeated binary multiplica- 

 tion in plants so co?istructed as the diatoms. For the above statement 

 occurs shortly after he has said " multiplication by binary subdivision 

 takes place among the diatomaceaj on the same general plan as in the 

 desmidiceaj, but with some modifications incident to the peculiarities of 

 the structure of the former group (p. 334) • At any rate my contention 

 is as to that idea, whether Carpenter implies it or not. I wish to say, 

 therefore, that, in my opinion, the reason for conjugation is primarily 

 for the recovery of vitality, and that while it also has to do with the 

 recovery of the typical form in the case of the diatoms, there is nothing 

 exceptional about it. For the sarnie is true of the desmids, and even of 

 the infusoria. In other words, multiplication by binary subdivision 

 tends to decrease in size any ivay^ in the long run, through loss of vitality, 

 and not merely in the diatoms, because one valve is smaller than the 

 other. Thus Wolle says of the desmids : "■ The frequent variations in 

 outline of form in certain species is calculated to mislead, and, perhaps, 

 suggest notions of variation of species, but close observation will soon 

 dispel such notions, and prove that they are merely temporary results of 



