MODE OF REGENERATION OF THE GLANDULAR EPITHELIUM. 451 



processes contain numerous nuclei, the nucleus of the columnar 

 cell still appears to be always perfect, spherical, sharply de- 

 fined, and without a trace of gemmation. Even with the 

 highest magnifying powers I have never observed any indica- 

 tion that a filament was given off from the nucleus which 

 could serve as a point of origin for the young nuclei. A few 

 processes even pass over the nucleus through the columnar cell, 



Fig. 95. 



Fig. 95. Multiplication of nuclei in the dilated and swollen processes 

 of the columnar cells. A, formation of small multipolar cells ; B appears 

 to be a dilated process of a columnar cell. Magnified 590 diameters. 



and their striae run parallel to its axis as far as to the free sur- 

 face directed towards the cavity of the salivary tube, so that it 

 scarcely appears to be possible that the nucleus originating 

 in the extremity of such a process could be derived from the 

 nucleus of the cylinder cell. The latter is almost always single, 

 rarely double. Very small and non-nucleated columnar cells, pos- 

 sessing processes that are filled with small nuclei, are also some- 

 times present (fig. 93). As on this ground I do not feel myself 

 justified in attributing the origin of the new nuclei developing in 

 the processes to that of the columnar cells, we must admit that 

 we have before us a case of free cell formation, if under this term 

 we understand that mode of cell increase in which the newly 

 developed nucleus originates independently in a cell, and is 

 not a morphological element proceeding from a division of a 

 previously existing nucleus. When we see the axis cylinder 

 and its fibrils to be directly continuous with the fibrils of the 



