AND LIFE-HISTORY OF DIATOMS. 429 



towards the surface of the nucleus. As the nuclei retreat from 

 the broad end towards the centre of the cell, they again turn 

 through an angle of 90°, so that their convex surfaces face the 

 new cell wall in each case. Nucleoli appear in the central space 

 which is now rapidly filling up, and the centrosome shows a 

 connection with the nucleus by radiating lines. No trace of the 

 polar ends of the spindle now remains, and Lauterborn suggests 

 that they may possibly have passed into the substance of the 

 centrosomes. 



Each daughter nucleus occupies, of course, one valve of the 

 parent frustule, but the commencing formation of a new valve 

 for each daughter cell may now be noticed, including the first 

 appearance of the transverse canals (Fig. 13). A little later the 

 nucleus and centrosome will exhibit a change of position, so that 

 the convexity of the now kidney-shaped nucleus will once more 

 be turned towards the broad end of the cell. 



The various changes thus briefly described are estimated to 

 occupy about five to five and a half hours from start to finish,, 

 but the unnatural conditions which prevail when the organism is 

 subjected to examination under a cover-glass must be taken into 

 consideration, and it is probable that under natural conditions 

 the work of division is performed in a much shorter time. 



It is by no means an easy matter to find dividing nuclei. 

 The beautiful series of figures of Surirella given by Lauterborn 

 represent the work of two summers. Weeks would sometimes 

 pass without any result, although hundreds of specimens might 

 be examined, but this apparent scarcity might possibly be due 

 to the time of day when observations were carried on. Dividing 

 cells were relatively most abundant in the morning hours, and 

 thus it may be that in Diatoms — as in Sjnrogijra and Ceratium — 

 division usually takes place during the night. 



Castracane's view — that division was the exception, not the 

 rule, and that Diatoms usually multiplied by a process of spore 

 formation — found no support in Lauterborn's work. Although 

 thousands of examples passed under his notice from first to last^ 

 not the slightest trace of any such process was ever detected,, 

 but some hundreds of dividing cells were observed and studied. 



