OF THE DIATOME^. 73 



-^-ith the fonnation of these inner cells in HimanticUum undaJaium, by 

 gatheiing specimens at short intei-vals. Dming great, part of the T\inter, the 

 filaments increase in bulk, by repeated division of the frustules, until they 

 form large masses, filling the ditches ; at length the inner cells make their 

 appearance, at first spaiingiy ; but as spring advances, it is difficult, in many 

 situations, to obtain a filament without them. I have found that when these 

 become abundant, the filaments cease to grow, and the entire mass soon breaks 

 up and disappears. The same thing happens in the other species of Himan- 

 tidium, and in Meridion. 



'' I do not find that the inner cell commences in the centre and pushes its 

 valves outwards, as stated by Prof. Smith. Were this the case, the internal 

 matter also would necessarily be pushed outwards by the advancing valves, 

 and thus condensed between them and the walls of the frustule. On the 

 contraiy, in the Himcuit'idium the internal matter, before nearly fluid, collects 

 within the new cell, becomes dense and more granular, and the new walls are 

 formed round it in the situation they are to occupy, leaving an empty space 

 between them and the walls of the frustule. 



*' The alteration and condensation of the colouring matter, and the ap- 

 pearance, or at least great increase of vesicles, have a strong resemblance to 

 what takes place previous to the formation of sporangia, the completion of 

 which, as in this case, usually preludes the death and disappearance of the 

 mass. 



" As in most acknowledged sporangia, the ceil thus formed always tends 

 to assume an oval or orbicular form. It, however, is very frequently, and 

 perhaps generally, divided in halves, as in the fission of the frustules, so that 

 the oval seems made up of two neighboimng frustules ; but this is not the 

 case, as may readily be ascertained by noticing the marginal puncta of the 

 original fnistule. 



" Do these newly- constituted cells ever continue to divide, as Prof. Smith 

 supposes ? I beheve not ; at least I have never seen a specimen in which 

 the semi-elliptic portions were separated by the interposition of other valves 

 resembhng either themselves or those of the ordinary fnistule. For my own 

 part, I have been unable to trace the species after the formation of these cells, 

 owing to the quickly succeeding disappearance of the mass. If, indeed, this 

 renewed division does occur, the resemblance to what takes place in the 

 sporangia of some species of Melosira would be increased. 



" Prof. Smith, in his most interesting and valuable account of the ' Eepro- 

 duction in the Diatomacese,' enimierates four modes in which sporangia are 

 formed. The third is thus defijied : — 



" ' The valves of a single fnLstule separate ; the contents, set free, rapidly 

 increase in bulk, and finally become condensed into a single sporangiimi.' 



" As far as regards the Melosira varians, the only one in this group which 

 I have had an opportunity of noticing, I beheve the process is essentially the 

 same as in the examples already described. The only difi'erence is, that the 

 new-formed cell being inflated, and much larger than the original fnistule, 

 the valves of the fnistule must necessarily be either ruptured or iDushed apart 

 by the increasing growth of the sporangium, and the latter alternative happens. 



" I have seen no specimen of Mr. Brightwell's Chcetoceros Wighamii, but 

 from his figures I beheve the goniothecia-hke bodies constitute another 

 example of the formation of internal cells. 



" I have said that I consider these interaal cells sporangia, and essentially 

 of the same nature as the inflated ones of Melosira varians. At the same 

 time we should not forget that Mr. Thwaites discovered the Himantidium 

 pectinale in a truly conjugated state, and that it is contrary to our experience 



