P. T. Cleve: The Diatoms of Finland. 15 



ses. It is not advisable on that account to admit the latter ge- 

 nus. Frustulia forms a passage to Amphipleura, Berkeleya and 

 Brebissonia, if from the latter be separated Br. Weissflogii, which 

 as well as Navicula scopulorum seems to be more conveniantly 

 placed in the same group as Nav. rhombica. 



Schlzonema also has its frustules enclosed in gelatinous mas- 

 ses of different forms, but as it comprises species, which are very 

 nearly connected with Navicula? of the groups of Nav. radiosa, 

 Nav. directa and Nav. rhombica I do not see sufficient reason 

 to admit this genus. 



If the genera Pinnularia, Neidium, Anomceoneis and Fru- 

 stulia be admitted, the large number of forms remaining in Navi- 

 cula belong, with the exception of some smaller groups, to four 

 different groups viz: radiosa?, punctata?, formosa? and didyma?. 



The group radiosce is by some forms connected with Pin- 

 nularia, and by others with the group punctata?. It is very clo- 

 sely connected with Nav. directa, N. fortis and many forms of 

 Schizonema so that these forms scarcely can be separated from 

 the radiosa?. The stria? in all these forms are finely and trans- 

 versely lineate. Nav. cuspidata and allied forms are also (by 

 Nav. halophila and N. gregaria) connected with the group radiosa?. 



The small group of asperce seems to be allied to the radiosa?, 

 but the structure of these forms is so peculiar, that it seems 

 justifiable to place them in a separate genus, Trachyneis. 



The group punctatce is distinguished by the structure of the 

 valve, which is covered by rows of, usually, distinct puncta. This 

 group is connected with the radiosa? by some species (Nav. Pla- 

 centula). 



The large section of Nav. Lyra and Nav. Hennedyi, all 

 marine, seems to form a well defined group, but it has so many 

 transitions to the punctata? (Nav. pra?texta, Nav. carinifera, etc.) 

 that it does not seem desirable to place it in a separate genus, 

 though there might be some reason for combining the punctata? 

 and lyra? in a distinct genus. 



The group of formosce, to which the limosce may be united, 

 has scarcely any connection with the above named genera and 

 groups. The characteristic longitudinal lines, often indeed scar- 

 cely perceptible, remind one of Amphiprora (A. lepidoptera). The 



