514 DESMIDS AND DIATOMS. 
valves becoming impregnated with silex before they have 
become free of the connecting membrane. (Pl. 13, fig. 
31a.) In this way a complete septum is formed, the 
two new frustules often remaining enclosed within the 
connecting membrane, like the inner tubes in a telescope. 
These new cells, which in every way are the exact coun- 
terpart in size and structure of their parents, may remain 
thus enclosed, but more frequently are left free by the 
falling off of the connecting membrane. The new frus- 
tules thus formed, by the greater or less adhesion among 
themselves, which differs in different species, give rise to 
all the different forms of aggregation already alluded to. 
If adherent side by side, the effect is to produce a length- 
ened filament, either straight (Pl. 13, fig. 18), or spiral 
(fig. 12); if adherent only at corresponding corners, & 
star-shaped figure is produced (fig. 17); if at alternate 
corners, a zigzag chain (fig. 16). Frequently processes 
of greater or less size are developed at the corners, which 
serve as means of adhesion between the two (fig. 23). 
Lastly, the division may be complete, and the resulting 
cells remain aggregated in a spongy mass (fig. 19) 5 oF 
else, entirely free, swim slowly through the medium in 
which they live (figs. 9, 10, 28, 32, ete.). : 
Of the contents of the cells or “frustules,” their chem- 
ical composition, changes, and internal movements, it is 
unnecessary now to speak. Suffice it to say, that the ex- 
istence of the so-called stomachs, organs of locomotion, 
ete., which Ehrenberg asserted that he had detected in 
their interior, have been entirely disproved by the re- 
searches of later authors. The “endochrome,” as I have 
already said, is of a different color from that of most 
plants, and is even supposed by many to be of a different 
- chemical composition; but in most. other respects their 
