596 MICROSCOPIC FORMS OF VEGETABLE LIFE THALLOPHYTES 



each end-valve, so that the two valves are separated by a band, 

 which progressively increases in breadth by addition to the free 

 edges of the hoops, as is well seen in fig. 445, A.. In the newly 

 formed cell e, the two valves are in immediate apposition ; in d a 

 band intervenes ; in a this band has become much wider ; and in b 

 the increase has gone on until the original form of the cell is com- 

 pletely changed. At the same time the endochrome separates into 

 two halves ; the nucleus also subdivides in the manner formerly shown 

 (fig. 417, G, H, I) ; and the parietal utricle folds in, first forming 

 a mere constriction, then an hour-glass contraction^ and finally a 

 complete double partition, as in other instances. From each of its 



adjacent surfaces a new 

 siliceous valve is formed, as 

 shown at fig. 445, A, C, just 

 as a new cellulose wall is 

 generated in the subdivision 

 of other cells ; and this valve 

 is usually the exact counter- 

 part of the one to which it 

 is opposed, and forms with 

 it a complete cell, so that 

 the original frustule is re- 

 placed by two frustules, each 

 of which has one old and one 

 new valve, just as in Desmi- 

 diacece. Generally speaking, 

 the new valves are a little 

 smaller than their prede- 

 cessors ; so that, after re- 

 peated subdivisions (as in 

 chains of Isthmia), a diminu- 

 tion of diameter becomes 

 obvious. 1 But sometimes 

 the new valves are a little 

 PIG. U5.Biddulphia pulchella : A, chain of larger than their predeces- 

 cells in different states : a, full size ; 6, elon- gors ; SO that, in the fila- 

 gation preparatory to subdivision ; c, forma- ^p^,- <, npm - p fhprp TYI^V 

 tioii of two new cells ; d, e, young cells ; B, ment U s Species, tne > may 

 end view ; C, side view of a cell more highly be an increase sufficient to 

 magnified. occasion a gradual widening 



of the filament, although 



not perceptible except when two continuous frustules are com- 

 pared ; whilst, in the free forms, frustules of different sizes may 

 be met with, of which the larger are more numerous than the 

 smaller, the increase in number having taken place in geometrical 

 progression, whilst that of size was uniform. It is not always clear 

 what becomes of the hoop. In Melosira (fig. 444, A and B), and 

 perhaps in the filamentous species generally, the hoops appear to 

 keep the new frustules united together for some time. This is at 

 first the case also in Biddulpkia and Isthmia (fig. 457), in which the 



1 This could not be explained on the hypothesis of the rigidity of the walls within 

 which fission takes place. 



