96 E. M. NELSON ON NAVICULA RHOMBOIDES AND ALLIED FORMS. 



Xavicula crassinervis, which, although known almost as early a& 

 the rhomhoides, is put last on the list because its striation is so 

 fine that it was not resolved until much later. It is a very small 

 form, about 0*0016 inch long, its ratio being about 3 "8, with a 

 striation about 88,000 per inch. Its shape in outline is elliptical 

 without a trace of any rhombic angle ; it also resembles Frustulia 

 in the uniformity of the size of its valves. When once seen it 

 cannot possibly be mistaken for any of the others. There is^ 

 however, another diatom in the same gathering as the English 

 rhomboides. It can be distinguished by having no rhombic angle, 

 the sides having much flatter curves, and at the ends the sides 

 come in with a shoulder. The end of the raphe is Romanesque, 

 like the American Cherryfield rhomboides ; its transverse striae 

 count 72 in the thousandth of an inch. This variety has 

 received an independent name, Schizonema virididum. The 

 shajDe of the exterior outline of a diatom more often determines- 

 its species than any other feature it possesses. 



You now have before you a particular description of the four 

 varieties of rhomboides, and it is for you to say if they are 

 merely variations of one form growing one from another, or 

 whether they are distinct varieties breeding true. Personally I 

 think they are distinct varieties, but not being a diatom-species 

 expert I decline to make any definite statement about this point. 

 Standing quite outside this branch of diatomic work I can only 

 express astonishment that a species such as rhomboides should 

 have been taken out of the genus Navicida and made a 

 Vanheurckia. If a rhomhoides is not a Navicula^ no other diatom 

 can be; and if this precedent, originated by M. de Brebisson in 

 1867, is to be followed we may expect to find other varieties 

 treated in the same manner: thus Navicida lyra may be named 

 Sollittiana lyra) Navicida serians, Ilarrisonii serians, and so on to 

 complete the confusion. 



In conclusion the following suggestion is put forward for your 

 consideration. So long as these minute variations are expressed 

 in words it is almost impossible to estimate from the many 

 different peculiarities or features which a diatom may possess the 

 resultant of the whole of them ; if only, however, we could give a 

 numerical value to these minute difierences a resultant number 

 could easily be found that would almost stand for the numerical 

 index of the variety. 



