ANIMAL NATURK OF DIATOMEjE. 415 



as well as in plants, the multiplication by reproduction 

 is accompanied frequently by alterations more or less 

 important in size and external form. When, again, the 

 multiplication is effected by simple division, both the 

 forms and dimensions remain constant. Nor will we here 

 enter upon the difficult question that relates to reproduc- 

 tion ; neither do we intend to define in what this differs 

 from simple division, though a division it certainly is. In 

 the Diatomese the distinction is easy. In these a repro- 

 duction certainly takes place, since mixed among the 

 larger individuals of every species we often see some 

 smaller, some very small, and others of every intermediate 

 size. But their enormous abundance seems to proceed 

 from division {divisione) rather than doubling {dimezza- 

 mento). Now, in doubling {dimezzamento), the forms 

 and dimensions (of the secondary surfaces) continue per- 

 fectly equal. Hence that wonderful uniformity in myriads 

 of individuals which present themselves to our observation, 

 all of which, perhaps, were derived by successive partition 

 from a single one. Hence the natural tendency that 

 must be felt by every observer to distinguish more spe- 

 cies when, among these individuals mathematically equal 

 to one another, he sees some rather different in form and 

 dimension ; or when he sees other thousands of indi- 

 viduals differing only by slight conditions from the 

 former, but all precisely equal one to another. If, again, 

 he then happen to meet with forms mingled among 

 them, different but in degree, and which successive ob- 

 servation proves to belong to the same species, he reflects 

 upon the difference that prevails among them, as well in 

 proportion as in dimension, and he easily believes, again, 

 that these differences are really greater than those fre- 

 quently proposed to distinguish species. Kiitzing, for 

 example, gives us four figures of Navicula viridula, 

 (PI. iii, fig. 44 ; 1, 3, 3, 5, 6 : PI. iv, fig. 10, 15 : PI. xxx, 

 fig. 37.) These he properly refers to the same species, 

 although the proportion between the breadth and length, 

 the degree of convergence of the sides, and the ventral 



