168 Mr. G. H. K. Thwaitcs on the Diatomaceae; 



fig. A 2 a filament consisting of sporangial fnistules produced 

 from one sporangium. Fig. C is a filament of M. Borreri, con- 

 sisting partly of ordinary and partly of sporangial frustules. 

 M. nummuloides of Kiitzing I cannot help believing to be the 

 sporangial state of M. salina of the same author. 



In Aulacoseira crenulata the sporangium is spherical, with its 

 axis of elongation at right angles to that of the frustule from 

 which it originated. Around the young sporangium a consider- 

 able quantity of mucus is developed, by which the empty half- 

 frustules are for some time held attached. Fig. B 2 represents 

 filaments of Aulacoseira crenulata with sporangia ; and fig. B 3 

 sporangial frustules of the same species. 



Orthoseira Dickieii, n. sp. PL XII. fig. E 1-7 (f ). Filamentis 



brevibus ; cellulis l?evissimis. 



The filaments of this beautiful species consist generally each of 

 from two to four frustules, which are hyaline and perfectly smooth, 

 and each with its central cavity filled with a dark red-brown en- 

 dochrome. The sporangium of Orthoseira Dickieii is no less 

 beautiful than interesting : it is fusiform in shape and marked 

 with numerous annular constrictions, each with a coiTcsponding 

 internal septum or chamber, the origin of which can only be un- 

 derstood by paying attention to the early development of the 

 sporangium. In fig. E 3 is shown a filament of this species, the 

 terminal cells of which have each commenced to develope a spo- 

 rangium ; E 4 represents two such cells or young sporangia ; 

 andE 5 a mature sporangium. It will be observed that the for- 

 mation of the ring-like markings is progressive, and that they go 

 on increasing in number until the sporangium is fully developed. 

 At the commencement of the formation of the sporangium, the 

 endochrome, at the same time that it withdraws from the end of 

 the frustule, produces at its centre an additional ring of cell- 

 membrane ; and this process continuing to take place at certain 

 intervals — each new ring of cell-membrane exceeding in diameter 

 those previously formed — produces at length the structure re- 

 presented in E 5. Or it may be a more coiTcct explanation of 

 the process to say, that an entire new cell-membrane has been 

 developed by the young sporangium at the time each new ring 

 has been formed, and that thus have originated the several 

 chambers into which the ends of the sporangium are divided. 

 Fissiparous division of the sporangium subsequently takes place, 

 as shown in fig. E 6, and sporangial frustules are developed from 

 each half, E 7. 



Meloseira americana, Kiitz. Bacillarien, 55. tab. 30. fig. 69, is 

 evidently congeneric with this species ; differing from it princi- 

 pally in the ends of its frustules being striated. 



