[22 BOTANICAL GAZETTE [FEBRUARY 
Wall formation in the cells open and growing toward the center, 
which is filled with sap, is carried on in the usual way at the peripheral 
end of each new cell, by the formation of a nuclear plate on the spindle 
of the division by which the new nucleus is formed (jig. 32). 
The mode of formation of the side walls of these same cells, how- 
ever, is not so easily determined. Curious double spindles (jig. 33) 
and extra fibers radiating from nuclei in the last stages of karyokine- 
sis (fig. 32) indicate some such procedure as this: When a nucleus 
divides, lengthening the centripetal row, fibers radiate from it, not 
only to the sister nucleus, but radially as well, connecting with similar 
fibers radiating from other nuclei, and upon the resulting spindles are 
formed the plates which later develop into walls.. This mode of wall 
formation has long been known in the endosperm of many angio- 
sperms. In some cases are to be seen similar double spindles which 
lack nuclei at the two ends, the single nucleus concerned being in the 
middle. Further work is necessary to settle this point. 
Centripetal growth continues through July and August, being most 
active at the base and filling the central cavity with tissue by the last 
week in August. In shape the prothallium is almost ellipsoidal, 
being slightly flattened on two opposite sides (figs. 34, 35). Asa 
result of the shape and the equal growth from the sides, the closure 
of the tissue is in a plane parallel with the broader sides of the gamet- 
ophyte. Only in sections cut at right angles to the flattened sides 
does the closure appear as a line (fig. 34). Upon approaching each 
other the open, centripetally growing cells of opposite sides do not 
unite and form a common end wall, as they are said to do in many 
other forms. Instead, each cell forms an independent end wall, 
separate from that of the other cells (fig. 36). |The resulting body is 
not a solid mass, but a tissue which may be easily opened at the 
middle. Later these separate end walls, lying as they do against each 
other, may so unite as to appear and really form a common wall, but 
such is not the case in the oldest prothallia examined. 
Growth.—Simultaneous with centripetal growth, the whole pro- 
thallium increases in bulk by growth and division of cells, the inner 
ones enlarging greatly and tone: at the periphery continuing meris- 
tematic (figs. 25, 38). 
Archegonium initials appear very early; the two-celled neck has 
