412 ANIMAL NATURE OF DIATOMEiE. 



arrangement of at least two species {S. ? parvula, S. ? 

 Zenormandii,) out of the seven remains uncertain. 



29. GoMPHONEMA. — Corpuscula silicea a latere pri- 

 mario cuneata, basi affixi vel stipitata, stipite gelineo. 



As Cocconemce from Cynibellce, so Gomphonemce only 

 differ from Sphenellce by the stipes ; on which account 

 species are now referred to Gomphonema which formerly 

 belonged to Sphenella {G. oKvaceum). And with respect 

 to the whole thirty-three species of Gomphonema, it is 

 still doubtful whether they ought not rather to be placed 

 among the Sphenellce. Independently, therefore, of the 

 value which the presence of the stipes may have as a 

 generic character, it is important to consider it in an 

 organographical point of view. Kiitzing supposes the 

 Gomphonemm to be at first free, like Sphenellce, and that 

 afterwards they affix themselves by means of the (gela- 

 tinous) substance of the stipes, which in his opinion they 

 secrete from the inferior extremity. No direct observa- 

 tion confirms this hypothesis, and it is at least as just to 

 admit the other, that the Sphenellce are at first attached, 

 like the Gomphonemce, and afterwards become free. 

 Ehrenberg says, that the Gomphonemce can become free 

 and again adhere. The circumstance of a tubular cavity 

 through which this stipes runs, according to Kiitzing, 

 and the laceration that is produced in this tube, when in 

 the act of duplication the two new individuals separate 

 from each other, efiecting a dichotomy, if by any means 

 it could be reconciled with the idea of a simple secretion, 

 certainly agrees better with the supposition that the 

 stipes in Gomphonema, like that of Cocconema, may be 

 compared with the tube of Encyonema, and, like that, 

 capable of its own proper growth, and therefore endued 

 with life. But it remains to be proved that the stipes 

 can divide itself from above downwards, to produce the 

 dichotomy, as maintained by Kiitzing. There is little to 

 be remarked on the form of Gomphonema. The primary 

 surfaces are constantly cuneiform-truncate. In one only 



