12 H. MARSHALL WARD. 



proceeds rapidly ; but, since the objects now become of a more manage- 

 able size, I have been able, by actual sections through the perithecium 

 embedded in spermaceti or gum, or, better still, in elder pith, to 

 obtain some insight into the processes going on even in the centre of 

 the mass of cells. 



At stages just prior to the one last described, the central core of 

 thin walled cells — which it will be remembered has been derived from 

 continuous divisions of the cell A — is commencing to divide up by 

 septa in several directions (figs. 23, 24), while the outer layers sur- 

 rounding this — derived primitively from B, and, possibly, in part 

 from A — are divided more regularly by tangential walls, followed by 

 radial ones at right angles as the area enlarges. As the increasing- 

 small and delicate cells of the core become formed more rapidly, a 

 certain tendency at least to a regular arrangement can be recognised 

 in the later stages, as shown in such sections as figs. 28 and 29 and 

 fig. 27 : this regularity becomes interfered with by the mutual pres- 

 sure of the cells, and the outer ones, of which the walls are especially 

 soft and swollen, become flattened and pulled in the tangential direc- 

 tion, and only marked by the very granular yellowish protoplasm in 

 their diminishing cavities. In the central lower part of the core, 

 vertical sections at this, and slightly later stages, show that certain 

 cells, with very delicate outlines and finely granular refractive con- 

 tents, maintain then larger size and upright arrangement, and are by 

 these peculiarities well distinguished as a special group or tuft of cells 

 (see fig. 28, and fig. 31). In oblique (fig. 29) and horizontal (fig. 30) 

 sections passing through the lower third of the developing pei'ithecium, 

 they can also be readily distinguished by their special peculiarities, 

 and no question can be entertained as to their significance in the for- 

 mation of the essential parts of the fruit-body. This group of cells is 

 the forerunner of the young asci, and may be termed the Ascogonium. 



As development proceeds continuously, the outermost layers 

 acquiring thicker and more deeply coloured walls, the above named 

 group of upright cells become relatively larger, increasing slowly in 

 number by a few divisions, while the diffluent, compressed cells between 

 them and the outermost layers slowly give up their contents, and 

 become reduced to mere granular streaks embedded in a jelly-like 

 mass of swollen and fused cell-walls (see fig. 31). This process is 

 exactly comparable to what takes place in the developing embryo-sac 



