142 r. T. CLKVIO, SYNOPSIS OF THE NAVICULOID DIATOMS. 



4. Libcllas. Zone comijlex or with longitudinal divisiony. Central nodule dilated into a 

 stauros, or not dilated. 



5. Miciostigma. Zone simple. Central nodule not transversely dilated. 



This division of the whole group cannot be completely carried ont at present, as the con- 

 necting zones of many species have not been observed. I consecjuently class all the forms now in 

 two groups, viz. those with transvei^sely dilated central nodules in Htauroneis, and those without 

 such dilated central nodules in Microstigma; at the same time indicating as far as can yet be 

 made out to which of the five groups named above each species belongs. 



The genus Sfauroneis (Ehr. 184;')) has always been regarded as distinguished from Xavicula 

 by the transversely dilated central nodule; but the difference between a dilatation of the nodule 

 itself, or stauros, and a more transverse extension of the central area, or fascia, has not hitherto 

 been strictly carried out, and several forms with transverse areas only liave been placed in 

 Stauroneis. These I now remove to Pinnularia or Navicula. 



Still, if we include in Stanroneis all the naviculoid forms which have transversely dilated 

 central nodules, we shall not have a natural genns, as it will comprise species of Trachvneis, 

 Mastoneis, Psendoamphiprora, and other groups. All these 1 also remove to their respective groups 

 and coniine Stauroneis to such forms as possess the structure described above as belonging to the 

 Microstigmatica\ 



Stauroneis has affinities with Amphora, particulary with the group Psammamphora, in which 

 the structure is identical, and in which occur species with and without a stauros, which may be 

 regarded as asymmetrical forms of Stauroneis and Microstigma. 



In all the true species of Stauroneis the strict' are radiate at the ends. This is the case also 

 in the group Plcurostanron, which is nearly related to Stauroneis and passes over into it by gradual 

 transitions. 



Many s])ecies of both these groups are met with in fresh water in all countries, arctic or 

 tropical, a few inhabit brackish waters, but there is scarcely any undoubted marine species. 



The fresh water species of Schizosfauroii are closel}^ allied to those of Pleurostauron, but 

 I am not sufficiently ac(|uainted with the marine species of the former to be able to speak as to 

 their affinities. 



The division Lihellus comprises forms with and without stauros. They are all marine and 

 some of them live enclosed in gelatinous tubes, for which reason they have been classed in the 

 very unnatural genus Schizonema. As there are free forms, so closely connected with those 

 living in gelatinous tubes, that they cannot be specifically distinguished, there is no reason to 

 retain the genus Schizonema. As early as 1873 1 proposed the generic name Libellus for Naviculte 

 with complex zone, but this view was not accepted until recently by De Toni, who placed in this 

 genus N. aponina. Whether this be admissible or not 1 cannot say, as I have not examined this 

 species sufficiently; but as it is figured in Van Heurcks Synopsis as having subsidiary' longitudinal 

 lines it may belong to Caloneis. 



The species of Libellus are ]io doubt closely connected with those of the di^•ision Microstigma, 

 in which are some forms the zone of which has longitudinal rows of short stria^ (A. uuldandica 

 and N. Garkeana). Still closer allied are the forms of the section Oxyamphora among the asym- 

 metrical naviculoid diatoms. The same structure exists in these Amphorae as in Libellus; the 

 zone is similar and there are in Oxyamphora species with, and without stauros, exactly as in 

 Jjibellus. Most forms of Libellus have the terminations of the median line at some distance from 

 the ends of the valve. 



The division Microst'Kjma comprises forms without stauros, and with a simple connecting 

 zone, which however, as stated above, has in some species longitudinal rows of short striae. I 

 have enclosed in this di\'ision several forms, classed in different genera by authors, as Scolio- 

 2}leura fumida, and lihoiconeis Garkcana, the former having a sigmoid median line, the latter 

 arcuate frustules. On the sigmoid median line alone no natural genus can be founded, forms with 



