ANTHER-CELLS POLLEN-GRAINS. 397 



tents of each of these cells secrete a layer of cellulose, which does 

 not adhere to the wall of the parent-cell to form a layer of secon- 

 dary deposit, but lies free against it, so that a new free cell is 

 formed within each old one, nearly filling it. The walls of the 

 old cells then dissolve, so that the free cells become free, no 

 longer in their parent-cells, but in a cavity which is to constitute 

 the pollen- chamber or loculus of the anther. These free cells are 

 the i parent-cells of the pollen' of authors. A new phenomenon 

 soon occurs in these. These parent-cells divide into four by or- 

 dinary cell-division ; either by one or two successive partings, by 

 septa at right angles to each other, but both perpendicular to an 

 imaginary axis (as when an orange is quartered) ; or by simulta- 

 neously formed septa, which cut off portions in such a manner, 

 that the new cells stand in the position of cannon-balls piled into 

 a pyramid (tetrahedrally). These new cells are the ' special parent- 

 cells of the pollen ;' and in each of these the entire protoplasmic 

 contents secrete a series of layers, which, in the ordinary course, 

 by the solution of the primary walls of the special parent-cells 

 upon which they were applied, become the walls of free cells, 

 which constitute the simple ordinary pollen-cells. These subse- 

 quently increase in size, and their outer coat assumes its charac- 

 teristic form and appearance, while free in the chamber of the 

 anther." 1 This history bears a very close parallel with that of 

 the development of the spores within the "theca" of the Mosses 

 ( 217); and it is not a little curious that the layer of cells which 

 lines the pollen-chambers, should exhibit, in a considerable pro- 

 portion of plants, a strong resemblance in structure, though not 

 in form, to the elaters of the Marchantia (Fig. 132). For they 

 have in their interior a fibrous deposit; which sometimes forms 

 a continuous spiral (like that in Fig. 157), as in Narcissus and 

 Hyoscyamus ; but is often broken up, as it were, into rings, as m 

 the Iris and Hyacinth; in many instances, forms an irregular 

 network, as in the Violet and Saxifrage ; in other cases, again,, 

 forms a set of interrupted arches, the fibres being deficient on? 

 one. side, as in the yellow Water-lily, Bryony, Primrose, &c. ; 

 whilst a very peculiar stellate aspect is often given to these cells, 

 by the convergence of the interrupted fibres towards one point 

 of the cell-wall, as in the Cactus, Geranium, Madder, and many 

 other well-known plants. Various intermediate modifications 

 exist ; and the particular form presented, often varies in different 

 parts of the wall of one and the same anther. It seems probable 

 that, as in Hepaticae, the elasticity of these spiral cells may have 

 some share in the opening of the pollen-chambers and the dis- 

 persion of the pollen-grains. 



253. The form of the Pollen-grains seems to depend in part 

 upon the mode of division of the cavity of the parent-cell into 

 quarters ; generally speaking it approaches the spheroidal, but 

 it is sometimes elliptical, and sometimes tetrahedral. It varies 



1 " Micrographic Dictionary," p. 516. 



