GROWTH OF CELLS. 143 



always has its convex surface directed toward the original 

 cavity of the cell, and this lamella belongs to the cell 

 formed in the branch. The cavity of the parent-cell at 

 first appears to be closed in merely by the secondary-cell, 

 subsequently it becomes closed in by its own piece of 

 membrane. In this kind of cell-formation, therefore, as 

 I believe, not two, but only one cell originates in the first 

 instance. 



VI. -THE GROWTH OF CELLS. 



The cell is very varied in form at the moment of its 

 origin. Cells produced by free cell-formation are always 

 globular or ellipsoidal. Cells originating in numbers in 

 a parent-Bell, by parietal cell-formation, have that form 

 which is produced by the division of the parent-cell by 

 straight or curved surfaces. The variations of form under 

 which these newly-formed cells appear are innumerable. 

 The cell not unfrequently originates in a mathematically 

 regular form, such as a sphere, ellipsoid, hemisphere, a 

 quarter, segment, or section of a sphere, a cone, truncated 

 cone, cylinder, semi-cylinder, a quarter, section, or seg- 

 ment of a cylinder, tetrahedron, cube, table, prismatic 

 column, &c. &c. ; but in most cases the form of the nascent 

 cell is either intermediate between these mathematical 

 figures, or exhibits all possible irregular figures, with 

 plane (more rarely curved) surfaces. The faces and angles 

 of the nascent cell are either all curved or all straight, or 

 mixed. The number of faces by which the young cell is 

 bounded varies, according to my investigations, from 1 to 

 about 32, the number of plane angles from to about 90, 

 the number of solid angles from to about 60. 



From these facts it is evident how little the existing 

 theory of cells is applicable, setting out from the spherical 

 form, and deducing all the others from this as subsequent. 

 In general I hold it as a rule that the cell retains the 

 shape in which it originates. The shape is certainly 

 altered in many cases ; but taking a number as the test, 



