E. M. NELSON ON NAVICULA RHOMBOIDES AND ALLIED FORMS. 97 



Let US take these four Naviculae we have been considering and 

 see what can be done with them. Obviously their lengths may be 

 expressed in terms of their breadths: this will give at once 

 one numerical value, but not of much use by itself, for there may 

 be many forms whose lengths, are, say, five breadths. But this 

 number may be farther narrowed down by taking the fineness of 

 the transverse striation into account. This was done, and an 

 interesting relation was discovered, viz. that when the fineness of 

 the transverse striation was multiplied by the length-breadth 

 ratio the constant quantity 35 was obtained. As we are dealing 

 with only a numerical relation, the figures have been cleared of 

 decimals by multiplying by 1,000, and only the number of 

 striae in the one-ten-thousandth of an inch has been used. 



For example, the English rhomhoides has 7*2 transverse striae 

 in the one ten-thousandth of an inch, and its length is about 4*8 

 or 4-9 times its breadth. Now, 4-85 x 7'2 = 35. 



For the Cherryfield rhomhoides we have 5*85 x 6 = 35 



For the Frustulia saxonica 4*3 x 8*2 = 35 



And for the Navicida crassinervis 3*8 x 9 "2 = 35. 



If then we divide the number of the transverse striae by 

 the leno^th-breadth ratio we shall find a number that will be the 

 numerical index for the variety. 



The index for Cherryfield rhomhoides is 10*3 

 English 14-7 



,, Frustidia saxonica 19 



,, ,, Navicida crassinervis 24 



On Moller's typen-platte slide are two diatoms ; it is required 

 to know what varieties they are. The first has a ratio of 5'6 and 

 60,000 transverse striae; the dividend is 10'7 : it is therefore 

 a Cherryfield rhomhoides. In the catalogue belonging to the slide 

 it is entered as " a fossil from America." The second has a ratio 

 of 4 '65 and 81,000 transverse striae ; its index is therefore 17*5 : 

 it is a Frustulia saxonica. The catalogue calls it "a N. crassinervis 

 from llolstein." It is totally unlike a crassinervis, for it has the 

 rhombic curve and is 0*0036 in length. We have therefore by 

 this simple means run these two varieties to earth.* 



* See remarks by J. E. Ingpen (at that time Hon. 8ec. Q.M.C.), 

 M. M. J., vol. 17, p. 221 (1877). 



