COSMAEIUM. 127 



by tlie subcylindrical cells (in vertical view circular), by the 

 slight median constriction, and by the absence of a basal 

 inflation of the semicells. Circular vertical views are, how- 

 ever, met with along with every possible grade of constriction 

 (cfr. C. annulatum, C. snltnryidiim, C. moniliforme, C. glalm- 

 latiiin, and C. Welwitschii) , and species are known, which, 

 although but faintly constricted, yet possess an elliptical 

 vertical view (cfr. C. anceps, C. Oocystiduni). It would be 

 unwise to fix an arbitrary depth of constriction, and to include 

 all those species more constricted than the demarcation limit 

 in the genus Cosmarium and those less constricted in the 

 genus Dysphinctium. Yet, there is no other way of clearly 

 defining Nao-eli's o-enus. Those authors who most strongly 



O O O / 



uphold the retention of Dysphinctium as a g-enus are not 

 agreed on the species which should be included in it. One 

 also finds Cosmarium speciosum referred to Dysphinctium, and 

 its nearest relative, C. subspeciosum, retained in the geuus 

 Cosmarium. Considering all these facts, we have thought it 

 best not to adopt a genus which is based upon indefinite 

 characters and the limitations of which could not be clearly 

 defined. Moreover, opinion would always remain divided 

 with regard to the inclusion of many species in Dysphinctium 

 or their retention in Cosmarium. 



The arrangement of the numerous species of Cosmarium in 

 the present volume is a purely artificial one, but one which 

 we think will best meet the requirements of the student. 

 We have divided the genus into two large groups, a section 

 with smooth cell-walls and a section with rough cell-walls. 

 This arrangement will be of great assistance to the student 

 in running down a species, and at the same time it allows of 

 the grouping together of species with close affinities. Of 

 course, the terms " smooth " and " rough " are only relative, 

 and there are a few species which might equally well be 

 placed in either section. The primary subdivisions of the 

 sections are in all cases based upon the external form of the 

 semicells as seen in front view. 



Division I. Cell-wall smooth, punctulate, or scrobicu- 

 late, outline never granulate, margins sometimes un- 

 dulate or crenate. 



Division II. Cell-wall granulate, verrucose, or papillate. 



As the genus is so very large we have divided the synopsis 

 of species into two parts, the first part being placed before 

 the descriptions of the smooth species and the second part 

 immediately before the descriptions of the rough species. 



