92 BULLETIN 100, UNITED STATES NATIONAL MUSEUM 



MASTOGLOIA SANSIBARICA A. Schmidt 



(Schmidt, Atlas, pi. 187, fig. 44.) 



MASTOGLOIA SERIATA Cleve and Grove 



(Le Diat., vol. 1, p. 66, pi. 10, fig. 6.) 



MASTOGLOIA SINUATA A. Schmidt 



(Schmidt, Atlas, pi. 186, fig. 11.) 



MASTOGLOIA SPLENDIDA (Gregory) Cleve 



(Van Heurck, Synopsis, pi. 28, figs. 1-2.) 



Cleve has rightly united Orthoneis with Mastogloia. 



MASTOGLOIA SQUAMOSA Brun 



(Schmidt, Atlas, pi. 188, fig. 19.) 



Cleve's attempt to unite this with Navicula ajfirmata Leuduger- 

 Fortmorel (Diat., Ceylan, pi. 2, fig. 22) can not be approved. 



Genus MELOSIRA Agardh 



MELOSIRA CORONARIA Mann 



(Mann, Diat., Alb. Voyages, p. 237, pi. 51, figs. 1-2.) 



MELOSIRA DURA, new species 



Plate 20, figs. 1, 2 



Valve circular, very convex inside of the narrow flat rim; marked 

 with fine watch-milling, over which are thickly scattered short, stout 

 but sharp pointed spines, increasing slightly in size and abundance 

 toward the margin; rim having a single row of beads. 



Diameter 0.041-0.044 mm. 



Spine-bearing and convex forms like the above are frequently 

 referred to the genus Stephanopyxis; as for example the quite similar 

 " Stephanopyxis robusta" Leuduger-Fortmorel (Diat. Malasie, p. 46, 

 pi. 4, fig. 6). But it is questionable if any form with a distinct rim 

 and radially arranged puncta can be referred to this genus, solely 

 because they are convex and spiny. They stand much closer to 

 the nearly related genus Melosira, many authentic species of which 

 are more or less spiny, as for instance, M. setosa Greville (Schmidt, 

 Atlas, pi. 182, figs. 42-46), M. Mspida Janisch (Schmidt, Atlas, pi. 

 182, figs. 54-55). For this and other reasons I think that the union 

 of the members of the genus Trochosira should be grouped with that 

 subgenus of Melosira, " Skeletonema," rather than united with Stepha- 

 nopyxis, as is done by Van Heurck. It is true that all these forms 

 are closely similar in their chain-forming method of growth and their 

 convex spiny valves. But if the chief purpose of taxonomy is to 

 classify and render available the objects of nature, rather than to 

 show generic relationships — the only view present possible with the 



