BRITISH DIATOMACE^. 23 



Were the frustules of Coscinodiscus eccentricus, for example, perma- 

 nently coherent after self-division, it would be difficult to separate 

 them in a generic point of view from those of Orthosira nivalis^ 

 W. Sm., which have the same cellular structure of the siliceous valve, 

 or from those of Melosira aurichalcea, Ralfs, or Orthosira sulcata, 

 Kiitz., which are furnished with a projecting fringe of silex, the 

 homologues of the spinous processes in C. eccentricus. It is true that 

 the Melosirea I have mentioned have valves much more convex, 

 occasionally even cylindrical ; but in Orthosira arenaria, W. Sm., this 

 difference disappears, and we have frustules as much compressed and 

 in consequence as narrow on the F, V. as in Coscinodiscus, and 

 differing principally in the firm and permanent coherence of their 

 surfaces. 



1 . Coscinodiscus minor, Ehr. Cells of V. circular, margin smooth, 

 v.s. 



Ehr. inKutz. Bacill. i. 12? 



Fresh water. Aberdeenshire, Dr. Dickie. 

 Plate III. 36. 



2. Coscinodiscus radiatus, Ehr. Cells distinctly hexagonal, ar- 

 ranged concentrically, largest in the centre of valve, margin smooth. 

 Diameter -0018" to •0055". v.v. 



Kutz. Bacill. i. 18. Prit. Anim. xiv. 39. 



Marine. Poole Bay, 1848. Stomach of Scallop, coast of Sussex, W. Sm. 

 Coast of Norfolk, Mr. jBr?!(//i^we//. Near Felkestone, Mr. Capron. (Peruvian 

 and African Guano.) 



Plate III. 37. 



3. Coscinodiscus eccentricus, Ehr. Cells arranged in eccentric 

 lines, margin spinous. Diameter -OOOG" to 0038". v.v. 



Kiitz. Bacill. i. 9. 



Marine or brackish water. Poole Bay, Aug. 1848. Near Lewes, July 

 1850. Stomach of Scallop in great numbers, Feb. 1851. 



Plate III. 38. 



