DIATOMACEAE. 19 



cell casts off its two valves and begins to spread itself out and increase hi size, being 

 sometimes surrounded by a gelatinous envelope, sometimes not. At first naked, it 

 soon appears encased in a thin non-silicious membrane, the perizonitnn. When the 

 auxospore has reached its full size, it secretes two silicious valves one after the other 

 within the perizonium which then disappears, while the new diatom continues to 

 multiply by division into two till it reaches the minimum size, when the formation of 

 auxospores again takes place. This is the mode of proceeding in the case of a 

 considerable number of species, as Melosira varians, Cyclotella Kiitzingiana, Cocconeis 

 Pediculus. 



The second mode of formation of auxospores differs from the first only in this, that 

 the protoplasmic mass of a mother-cell divides into two naked daughter-cells, which 

 issue from the gaping valves of the mother- cell and develope each into an 

 auxospore. This type is given by Smith and Liiders for Rhabdonema arcuatum only. 



A third mode, common to very many diatoms, always shows two individual cells 

 combining to form auxospores but without the occurrence of a real conjugation. Two 

 cells lay themselves side by side and secrete a gelatinous substance which encloses the 

 pair of cells in a common and usually ellipsoidal envelope. Then the two cells within 

 the gelatinous envelope cast off their old valves, and so lie side by side as naked cells. 

 In some cases the formation of the gelatinous envelope does not begin till after the 

 valves are cast off. The two cells then lie naked, i.e. without cell-membranes, inside 

 the gelatinous envelope, in some cases closely approximated to one another, in others 

 on the contrary so separated from each other by tolerably thick layers of the jelly, that 

 no contact takes place between them. Then both elongate and grow parallel and 

 alongside of one another to the normal size of auxospores, and on their outer surface 

 sooner or later appears a membrane of cellulose, the perizonium. Then each cell forms 

 its two silicious valves and begins its usual course of development. Examples of this 

 type are Frustulia saxonica, Cocconema Cistula, and others. It must remain a 

 question in this case, whether the two cells exercise a material influence on one 

 another, and if this influence is due to the exchange of some substance in solution ; 

 if such an exchange takes place, it must be, as Schmitz remarks, when the two 

 cells lay themselves side by side, but it has not yet been directly observed. 



A fourth mode is exemplified according to the observations of Pfitzer and others in 

 the genera Himantidium, Surirella and Cymatopleura. Two individuals here co- 

 operate in the formation of one auxospore. The two cells, usually wrapped in a 

 common gelatinous envelope, cast off their old valves and unite into a single naked 

 mass of protoplasm which grows into an auxospore. In this case therefore the 

 auxospore is formed by the conjugation of two gametes, and must consequently be 

 called a zygospore. 



Lastly, the fifth mode is where two cells, surrounded by a common gelatinous 

 envelope, throw off their old valves and divide each of them transversely into two 

 naked daughter-cells. Each pair of these four daughter-cells, lying opposite to one 

 another, unite to form a single naked cell, which then grows into an auxospore. Thus 

 it happens, according to Schmitz, with Epithemia Zebra. Here too there is a conjuga- 

 tion ; the auxospores are zygospores ; only they do not, as in the Conjugatae to be 

 hereafter described, pass through a period of rest, but commence their further 

 development at once. 



Thus the formation of auxospores in the Diatomaceae varies in different species, and 

 indeed within the limits of a single genus, Cocconeis. The question is, which of the five 

 modes should be regarded as the original one. It is obvious that with our present 

 knowledge we cannot decide this point with certainty; but my opinion is that the 

 cases of apogamy mentioned above in other Thallophytes make it most probable in 

 this case also, that the forms of asexual formation of auxospores must be derived as 

 instances of apogamous suppression from the formation of zygospores. In the case of 

 the third type the cells lay themselves side by side and empty their contents into the 



C 2 



