406 Annals New York Academy of Sciences 



chary, 1952, figs. 17 and 18), and C. venlricosa Ag. (Desikachary, 1952, figs. 19 

 and 20, Helmcke and Krieger, 1953, plate 75) (c/., Xavicula cryptocephala 

 Klitz., Helmcke and Krieger, 1953, plate 69, N. digitoradiata (Greg.) A. 

 Schmidt, Helmcke et al., 1961, plate 292 and 293, N. radiosa Klitz., Helmcke 

 and Krieger, 1954, plate 172, and A^. viridula (Kiitz.) Kiitz., Helmcke and 

 Krieger, 1953, plate 73). Cymbella delicatnla Klitz. (Helmcke and Krieger, 

 1954, plate 180) and C. mexicana (Ehrcnb.) Cleve (Okuno, 1956, plate 21, 

 fig. 2), however, each have an ultrastructure which is different from that of 

 these species and from each other's. Electron micrographs of other species of 

 the genus have been published but none give adequate pictures of the ultra- 

 structure. 



Maslogloia. The ultrastructure of M. braunii Grun. (Helmcke and Krieger, 

 1953, plates 57 and 58, 1954, plate 159) and M. smitliii Thwaites ex. W. Sm. 

 (Helmcke and Krieger, 1954, plate 160) is similar and resembles that of the 

 only two species of Navicula subgen. Lyraneis Freng. of which adequate electron 

 micrographs are available, viz.: N. forcipata Grev. (Helmcke et al., 1961, plate 

 291) and N. pygmaea Klitz. (Helmcke and Krieger, 1953, plate 71). Maslogloia 

 angulata Lewis (Okuno, 1957, plate 7, fig. 2) and M. fimbriata Cleve (Okuno, 

 1953, plate 1, fig. 3) resemble each other in their ultrastructure, but this is 

 quite different from that of M. braunii and M. smithii. 



Discussion 



The principles of taxonomy have recently been much discussed, and from this 

 discussion it has emerged that the amount of overall similarity is the only basis 

 for a satisfactory taxonomic classification (Cain, 1962, Sneath, 1962). To 

 accord overriding importance to a particular character, or to characters derived 

 from a particular structure, even if there are a priori grounds for considering 

 these of particular importance, results in an artificial and unsatisfactory system. 

 Almost without exception, however, diatoms have been classified solely on 

 the basis of the symmetry and structure of their siliceous frustule as seen under 

 the light microscope: although this provides comparatively few characters, 

 some of these, in particular symmetry, have been treated as having an im- 

 portance overriding that of the others. This concentration of attention on the 

 frustule has not been based upon any a priori reasoning but purely on con- 

 venience; in both fossil and recent material the valves are always present and 

 recognizable, and provide sufficient information for identification at the specific 

 level. 



The current classification of the Naviculaceae rests on such a basis. The 

 species are separated into genera on the common possession of a single charac- 

 ter, or a combination of only two or three, all drawn from the structure of the 

 frustule. Some of the genera so characterized are probably natural groups, 

 e.g., Diploneis Ehrenb., Neidium Pfitz., and Pinnularia; others contain very 

 diverse elements, e.g., Amphora and probably Maslogloia Thwaites ex. W. Sm. 

 and Slauroneis. The species that do not possess any characteristic that has 

 been seized on as a mark of generic distinction are left in the very large genus, 

 Navicula, a hotchpotch of species of diverse affinity. The little that we already 

 know of the ultrastructure of the Naviculaceae shows that it provides a series 



