THALLUS OF SPHAEROCARPOS DONNELLII AUST. 
185 
The next division is variable, and indeed the whole subsequent history, 
up to the time when the young thallus has been fully formed, is very in- 
constant. Figures 4 to 9 indicate some of the methods by which the 
development occurs. The end cell resulting from the second transverse 
division may divide longitudinally (fig. 4), or other transverse walls may 
first be put in, and any of the resulting cells may divide longitudinally 
(figs. 5, 7). The result is usually the same in all cases — namely, the forma- 
tion of a small plate of cells (commonly six or eight) of about equal size 
arranged in pairs (figs. 6, 8). Occasionally, however, some of the longi- 
tudinal divisions may fail to occur, the result being a single cell at some point 
n the double row of cells in place of a pair of cells (figs. 7, 9, 11). The next 
divisions are also longitudinal, but at right angles to the preceding, the 
result being a series of several groups of four cells each, with, however, many 
departures from absolute regularity. Very frequently the groups consist 
of three cells instead of four. The arrangement of cells here is difficult to 
determine under the microscope, as dividing walls are often either in 
the plane of the slide or superimposed vertically one upon the other. 
Figures 13, 20, 21, and 23 represent the arrangement of the cells at this time. 
These figures are of older stages in the development of the thallus, but all 
the cells except those of the terminal group have remained undivided. 
The cells of the end group now begin to divide, mostly by longitudinal 
walls, so that a plate of small cells is formed at the end of the germ tube 
(figs. 10-16). These divisions are apparently more rapid than those 
preceding, since the cells do not attain so large a size before redividing. 
The divisions seem to be quite irregular, and I failed to find any cell in these 
stages which could be recognized as an apical cell. In figures 10 and 12 
there is shown a cell (a) cut off by a diagonal wall suggestive of the apical 
cell described for many liverworts; this appearance, however, is not the 
usual one. The typical appearance of the young thalli at this stage is best 
shown in figure 13. The groups of cells below the terminal plate undergo 
no further divisions, though rhizoids are often produced from them; the 
rest of the thallus develops entirely from the germinal disc. This disc is 
very conspicuous, though in certain views, when mounted and observed 
under the microscope, it is sometimes flattened out so as to suggest a flat 
plate growing in the same direction as that of the original tube (fig. 26), 
instead of being placed at right angles on the end of the latter. I have 
seen cases in which the direction of growth of the plate seemed really to be 
a continuation of that of the germ tube, but in all but one of these (fig. 24, 
PI. XII) the sporeling had become separated from the spore wall, and hence 
could not be distinguished with certainty from the not infrequent cases of 
regeneration from small shreds of tissue in the soil. Regenerative shoots 
take on many different appearances, and often seem to grow for a time by 
means of a single two-sided apical cell, which is regularly wedge-shaped. 
The only resemblance to this condition that I have found in sporelings is in 
