BINAET SUBDIVISION OF DESMIDIACE^E. 263 



constriction is seen across the middle of the primordial utricle of 

 each segment ; but there is still only a single chamber, which is that 

 belonging to one of the extremities of the original entire frond. 

 The globular circulation, for some hours previously to subdivision, 

 and for a few hours afterwards, runs quite round the obtuse end a 

 of the endochrome ; but gradually a chamber is formed, like that 

 at the opposite extremity, by a separation between the cellulose 

 coat and the primordial utricle ; whilst, at the same time, the obtuse 

 form becomes changed to a more elongated and contracted shape. 

 Thus, in five or six hours after the separation, the aspect of each 

 extremity becomes the same, and each half resembles the perfect 

 frond in whose self-division it originated ; and the globular circu- 

 lation witb in the newly-formed chamber comes into connection with 

 the general circulation, some of the free particles which are moving 

 over the surface of the primordial utricle being drawn into its 

 vortex and tossed-about in its eddies. 



203. The process is seen to be performed after nearly the same 

 method in Siaurastrum (Fig. Ill, D, e) ; the division taking-place 

 across the central constriction, and each half gradually acquiring the 

 symmetry of the original. — In such forms as Cosmariian, however, 

 in which the cell consists of two lobes united together by a narrow 

 isthmus (Fig. 113), the division takes place after a different method ; 

 for when the two halves of the outer wall separate at the sutural 

 line, a semiglobular protrusion of the Endochrome is put forth 

 from each half ; these protrusions are separated from one another 

 and from the two halves of the original cell (which their inter- 

 position carries apart) by a narrow neck ; and they progressively 

 increase until they assume the appearance of the half-segments of 

 the original cell. In this state, therefore, the plant consists of a 

 row of four segments, lying end to end, the two old ones 

 forming the extremes, and the two new ones (which do not 

 usually acquire the full size or the characteristic markings of 

 the original before the division occurs) occupying the inter- 

 mediate place. At last the central fission becomes complete, and 

 two bipartite fronds are formed, each having one old and one young 

 segment ; the young segment, however, soon acquires the full size 

 and characteristic aspect of the old one ; and the same process, 

 the whole of which may take-place within twenty-four hours, is 

 repeated ere long.* The same general plan is followed in Micras- 

 terias denticulata (Fig. 112) ; but as the small hyaline hemisphere, 

 put-forth in the first instance from each frustule (a), enlarges with 

 the no wing-in of the endochrome, it undergoes progressive sub- 

 division at its edges, first into three lobes (b), then into five (c), 

 then into seven (d), then into thirteen (e), and finally at the time 



* See the observations of Mrs. Herbert Thomas on Oosmarium marga- 

 ritiferum, in " Transact, of Microsc. Society," n.s., Vol. iii., 1855, pp. 33-36. 

 — Several varieties in the mode of subdivision are described in this short 

 record of long-continued observations, as of occasional occurrence. 



