W. HOFMEISTER ON THE DEVELOPMENT OF ZOSTERA. 243 



arranged with strict regularity, as is well known, according to 2, 

 in such a manner that single anthers alternate with single ovaries 

 in each of the two longitudinal rows of floral organs, and organs 

 of different sex constantly stand next together horizontally. 



The anther first appears as a longish papilla of cellular tissue, 

 the larger diameter of which is parallel to the longitudinal line 

 of the spadix (fig. 4 a). The two ends of the young anther soon 

 appear curved outward to some extent ; they rapidly increase in 

 magnitude, and become globular protuberances (fig. 3 a, a), which 

 gradually become spindle-shaped through continued rapid mul- 

 tiplication of the cells. Thus each anther consists of two appa- 

 rently independent, moderate-sized halves, which are connected 

 together by a comparatively long riband-shaped cellular body, 

 the altered connective, attached to the spadix by the narrow 

 edge*. 



At the time when the half of the anther is passing from the 

 globular into the ovate form, two parallel rows of cells lying in 

 its longitudinal axis, take on a different character from the sur- 

 rounding tissue (PL VI. figs. 5-7). The process of division ceases 

 in the former, while the three layers of cells of the latter continue 

 to multiply. The cells of these parallel rows are the primary 

 parent-cells of the pollen. Their form is pretty nearly that of a 

 cube (fig. 7) ; in their subsequent development they become 

 elongated into short prisms, in a direction inclined downwards 

 from the surface of the spadix. 



The pollen-cells are developed from these cells through re- 

 peated longitudinal division, by means partly of perpendicular, 

 partly of horizontal walls (figs. 6-12). The process of their 

 formation is thus very unlike that occurring in the great 

 majority of Phanerogamia. There is not the slightest indica- 

 tion of parent-cells becoming isolated, or of special parent- 

 cells. There occurs in the primary parent-cells a series of 

 " halvings " or bisections in only two directions, differing in 

 no respect from vegetative cell-multiplication ; the last genera- 

 tion, alone, of the daughter-cells become disconnected and so 

 form the pollen-cells. This development of the pollen of Zoster a 



* This part of the development of the anther, very completely treated and 

 correctly figured by Grbnland, loc. cit. 187, is mentioned here merely for the 

 sake of giving a connected account. 



16* 



