ANIMAL NATURE OF DIATOMEiE. 375 



such as we have in the genus Himantidium, inasmuch as 

 the multiplication by bisection {dimezzamento) occurs 

 precisely in that direction. In such a case we should 

 have both in the genus Epithemia and the Iliviantidia, 

 incomplete division and lateral adhesion of individuals 

 in the family Eunotiese, and the first example of poly- 

 pariform association through adherence by one of the 

 primary surfaces, in this respect corresponding to 

 the Synedrce, with this difference only, that the latter 

 adhere by one extremity. Ehrenberg observed four in- 

 dividuals thus connected in the H. Westermanni. The 

 comparison is justified also by this, that in my Epithemia 

 there appears a kind of foot, which projects at the two 

 extremities, with truncated appendages. 



In all Epithemi(S the two principal surfaces present, at 

 each extremity, two small apertures, united to similar 

 apertures in the opposite extremity by fine longitudinal 

 lines. The transverse stria? of the lateral surfaces are 

 very evident, and in some species {E. ocellata, E. Argus,) 

 terminate on the dorsum in round apertures. They are 

 furnished with ciHa, according to Corda, in his Navicula 

 ciliata, which to me appears to be an Epithemia rather 

 than a Cocconema, as supposed by Ehrenberg {C.gihhum). 

 The internal substance appears to be uniformly extended 

 at the origin, divided subsequently into two lateral 

 masses; it is brownish-yellow, or green. A series of 

 minute oily drops occupies the median line. Many 

 species live in fresh water, many in the sea, and some are 

 found in a fossil state. Kiitzing enumerates twenty-one 

 species, to which we think two others [E. costata, E. 

 ciliata,) ought to be added. 



2. EuNOTiA. — "Lorica in sectione transversali trape- 

 zoidea, strice transversales tenuissiiJKe." 



Although the only character by which this genus is 

 distinguished from the preceding be the fineness of the 

 transverse striae on the lateral surfaces, yet the distinction 

 appears to be justified by the circumstance that the 



