48 DOUGLAS HOUGHTON CAMPBELL ON THE 



lar to the surface of the capsule, but directed in vai-ioiis directions (fig. 21). Walls are 

 never fonned in the parietal cells parallel to the surface of the capsule, so that the wall 

 of the complete sporangiiim is a single layer of cells. 



The tai^ctal cells very soon become also divided by walls perpendicular to their bi'oader 

 faces (figs- 21, 22), so that at this stage the central cell of the archesporium is sur- 

 rounded by a double layer of cells. Somewhat later, about the time that the central cell 

 becomes di\'ided for the second time, the taj)etal cells become further di^'ided by walls 

 parallel to the outer walls of the central cell, so that thetapetumis now composed of two 

 laj'ers of cells (fig. 23). These walls are not always perfectly parallel to the outer wall 

 of the capsule, but frequently inclmed more or less. 



Wliile these changes have been taldng place, the sjJOi'angium has mcreased greatly in 

 size, and become somewhat flattened laterally, this being much more strongly marked in 

 the later ones which often do not fully develop. 



A surf^ice view of the young sporangium shows that the arrangement of the parietal 

 cells is not the same in all parts. A row of cells running around the longest arc of the 

 capsule has divided more rapidly by transverse Avails than any of the other cells, result- 

 ing in the formation of a row of transversely elongated cells running all round the cap- 

 sule (figs. 22, 24). As the sporangium increases in size these continue to divide in the 

 same way until there are between thirtj' and forty of them. These are the cells of the 

 annulns. The other parietal cells divide more slowly and by longitudinal as well as by 

 transverse walls. 



About the time that the annulus can first be distinguished, and just before the tapetum 

 becomes two-layered, the innei* cell of the archesporium begins to divide to form the 

 spore mother-cells. The direction in which the first division takes place is not always 

 the same, being in some cases nearly vertical, in others horizontal, but in either case more 

 or less inclined (fig. 22) . Following this a second wall is formed, nearly at right angles 

 to the first, thus dividing each of the cells already formed into two (fig. 23). Almost 

 simultaneously the two layers of the tapetxnn are completed. 



The central cells, almost from the first, are distinguished from the outer cells by their 

 more dense and highly refractive contents. They now undergo further division until the 

 complete number, twelve, of the spore mother-cells is attained. The latter ai'e colorless, 

 but contain numerous fine granules, and each has a large and pretty evident nucleus. 

 As soon as the full numbei" is complete, the walls of the tapetal cells are absorbed and 

 the mass of mother-cells, Avhich are still connected, floats freely in the cavity of the spo- 

 rangium which it does not nearly fill (fig. 24) . The annulus has acquired nearly its full 

 number of cells, and the whole sporangium its complete form, though still consider- 

 ably smaller than when mature. 



The spore mother-cells now become sepai-ate by the partial dissolution of their walls, 

 and becoming rounded otf float about in the cavity of the sporangium which they nearly 

 fill, having increased very considerably in size after separating. Each spore mother- 

 cell is at this stage globular, with a large, but not usually very distinct nucleus which 

 is sometimes central in position, sometimes eccentrically placed. The protoplasm is 

 somewhat less granular than before, areas of considerable size being frequently quite 

 free from granules. 



The first trace of divisiou in tlie niotlier-cells is a cliange in the nucleus. This be- 



