400 DESCEiPTioN OF IPoli/gastrica. 



buted to this aspect of the Diatomaccous frustules, as its foiin greatly 

 depends upon the stage to Avhich self-division has arrived, and may 

 vary from linear to oblong, or from very narrow to very broadly 

 lanceolate, in the same individnals. In p. 20, f. 13, is an illus- 

 tration of multiplication by self-di\dsion in PJenrosigma. That by 

 this mode Naviculece multiply to a surprising extent, is evident, from 

 the circumstance of so great numbers being found together nearly all 

 of exactly the same size ; but it is also certain that they, like some 

 other Diatomece, have a specific mode of reproduction, since we often 

 find fi-ustulcs in various stages of growth, as is evident fi'om the 

 diversities of their size, (P. 19, f. 7 and 9 ; p. 20, f. 1 to 3, 9 to 12), 

 and from the greater delicacy of the striae in individuals of the same 

 species J circumstances which are incompatible with the process of 

 self-division, where the half new fi'ustulcs must of necessity be pre- 

 cisely counterparts of the old. The mode in which the genninative 

 power is renewed, when exhausted by self-division, will probably 

 be found to be a process analogous to that of conjugation in the 



DesmidietB, and in some of the Biaiomece The figures of Pleuro- 



sigma, in p. 19 and 20, are di-a\\Ti by the camera lucida to a scale of 

 400 diameters." 



In Ehrenberg's an^angements, Naincida gives name, and is the 

 type of the section Nmicidacea, of the great family BaciUaria, although, 

 the characters of many of the forms introduced are far removed from 

 those of the genus Narmda, and any affinity with them scarcely 

 traceable. 



The only connecting link between the genus Navicula and many of 

 the so-called Karicuhicea, is the siliccoiis nature of the lorica ; 

 however, Navimla has close affinities Avith a large number of other 

 genera ; and each of its sub-divisions has also its own special 

 relations: what these severally are is pointed out under the head of 

 each genus. Still it is veiy difficult, oftentimes, to assign to its proper 

 genus each navicula-shapcd lorica which may be met "v\dth, especially 

 when but one or two at a time occur, and then only probablj^ one 

 surface presented to view ; or when the right portion is to be deter- 

 mined by the fact of its attachment or non-attachment, and only 

 detached frustulcs arc to be had, either from accident or li-om their 

 occurrence, in a fossil state. 



