164 p. T. CLEVE, SYNOPSIS OF THE NAVICULOID DIATOMS. 



larity of the valves appears to me, just as the asymmetry, rather as a facies common to different 

 types, than a characteristic of allied forms. There are among the true naviculoid diatoms several 

 more or less arcuate forms with slightly or strongly dissimilar valves, although both have a 

 central nodule and a median line. For such forms, belonging to the sections Lineolatee and Micro- 

 stigmaticse of Navicula Grunow formed the genus Rhoiconeis, and Pbragallo has recently proved 

 that the valves of Gyrosigma compacfum are very dissimilar. In Bhoicosphenia, which belongs 

 to the true Achnanthese, we meet with a form, which in many characteristics, especially of the 

 cell-contents, appears to be nearly akin to Gomphonema. Many of the species of Cocconeis have a 

 loculiferous rim and are so nearly akin to the elliptical forms of Mastogloice, that it seems justifiable 

 to believe them to be M-astogloice, degenerated by their parasitical habits. It thus appears that the 

 flexure of the frustule is connected with a dissimilar development of the valves, and the differences 

 become more striking, when the frustules are stipitate or attached to algae or other objects. The 

 lower valve maintains its naviculoid characteristics, but the upper undergoes more or less com- 

 plete changes. From this it seems probable that the forms of Cocconeis and Achnanthes are Naviculm 

 degenerated by their manner of living. The original Naviculse probably belonged to very different 

 types and have undergone analogous changes. 1 think it evident that Cocconeis and Achnanthes 

 comprise forms which frequently have very little real affinity. The classification offers many 

 difficulties, but I believe that to this end the structure of the lower valve is of more importance 

 than that of the upper valve. A characteristic, which appears to me to be also of importance, is 

 the presence or absence of a marginal rim, which occurs in C. PlacentulcL, C. Scutellum and others, 

 and is probably of the same nature as the loculiferous plates in Mastogloia. Some forms are probably 

 akin to the section Mesoleice of Navicula, as C. minor to Navicula Eotaeana. Others, such as 

 C. reticulata seem to be connected with the section Punctatai. The curious A. danica is remarkable 

 for the highly inclined striae of the lower valve and resembles in this respect the section Hetero- 

 stichce of Navicula. If we consider also that Rhoicosphenia is a Gomphonema-like Achnanthes, we 

 feel that the family Achiianthese represents rather a facies belonging to widely different types 

 than a family of allied species. The interior of the cells and the formation of the auxospores 

 point to the same conclusion. There is a great resemblance in this respect between Cocconeis 

 Pediculus and Achnanthes lanceolata, both having a single chromatophore-plate along the upper 

 valve, but in Achnanthes brevipes the chromatophore-plates are two, placed along the walls of the 

 connecting zone, as in Navicula. 



It is possible that by a long continued degeneration, the lower valve of the frustule may 

 also lose its central nodule and median line, in which case the form would become a Rhaphoneis. 

 Some points of resemblance between Achnanthes and Rhabdonema (and its allied genera) seem to 

 exist. In Gephyria we meet with the arcuate and stipitate frustule of Achnanthes. The structure 

 of the upper valve of A. groenlandica and A. longipes is the same as in Rhabdonema. The 

 zone of both forms is striate, and rudiments of septa exist in A. groenlandica and A. baccata, 

 also in Rhoicosphenia curvata. There is certainly a great difference between Rhabdonema (and 

 allied genera) and Achnanthes in the cell-contents, but it may be observed that the chromato- 

 phore-plate of Cocconeis by its marginal incisions seems to have a tendency to become split up in 

 several patches. 



In a truly natural system the genera, belonging to the Achnanthese, would be distributed 

 in widely distant places, but I think it advisable, to avoid a too intricate synonymy, to keep 

 them provisionally together in a family, and to retain for the forms the old generic names of 

 Cocconeis and Achnanthes, under which they are usually known. For the new subgenera, which I 

 propose, I give the following key: 



^ / Valves asymmetrical . . 2, 



\ — symmetrical . . .... . . .... 3. 



I Asymmetrical to the longitudinal axis . . . Anorthoneis Grun. 



' I — — transverse Rhoicosphenia Gkun. 



