ZOOLOGY AND BOTANY, MICROSCOPY, ETC. 645 



It would appear, however, from observations made by otber writers 

 in tbe case of other plants (especially by M. Mellink on Agraphis 

 put nla) that there is a certain equivalence in the cells of the axile row, 

 and that one or other of them may develope into the embryo-sac, but only 

 one. Nevertheless, in some species the presence of two nuclei in the 

 penultimate cell is nearly constant. The occurrence, or otherwise, of 

 divisions in the subapical cell appears to depend on the greater or less 

 rapidity of the development of the embryo-sac. This tendency to 

 division is even continued in the embryo-sac itself, but is arrested and 

 reduced to the remarkable division of its primary nucleus into two 

 parts, which separate to the two extremities of the sac. The formation 

 out of this primary nucleus of eight distinct nuclei takes place in the 

 way described by Strasburger ; but their exact position varies according 

 to the form of the cavity. The upper nucleus divides into the oosphere, 

 the two synergida3, and a polar nucleus ; the lower one into the three 

 antipodals and a polar nucleus. The synergidse are more or less 

 closely attached to the summit of the embryo-sac ; in the Mimosefe 

 they often attain a considerable length. The oosphere is inserted 

 laterally, and descends below the synergidaa ; its appearance varies 

 considerably at the time of impregnation. The three antipodals are 

 attached to the base of the sac, and are generally clothed with a thin 

 membrane. In Phuseolus this is of considerable thickness ; but it 

 subsequently disappears, and the antipodals themselves have entirely 

 vanished at the time of impregnation. In the Banunculaceas and 

 Papaveracese, on the other hand, they survive this period. They are 

 most developed in the Mimoseae and Caesalpinieae. The fusion of the 

 two polar nuclei takes place at different parts of the embryo-sac. 

 It is completed before fertilization, except perhaps in the Vicieaa. 



The author regards all the cells formed in the embryo-sac of 

 Angiosperms as representing endosperm-cells analogous to those 

 formed in the embryo-sac of Gymnosperms. The oosphere forms by 

 itself a greatly reduced archegonium ; the synergidse are endosperm- 

 cells adapted for a special function. The endosperm formed after ferti- 

 lization by division of the secondary nucleus is the recommencement of 

 an interrupted development. 



With regard to the formation of the embryo, the first step is 

 invariably the appearance of a transverse wall in the fertilized cell ; 

 but from this point great diversities exist, the differentiation of the 

 subsequent product into embryo properly so called and suspensor not 

 being constant. The entire absence of a suspensor had been noticed 

 by Schacht, Treub, and others in a few isolated genera or species, 

 chiefly monocotyledonous. Guignard now establishes that it occurs 

 throughout the Himoseae and in some Hedysareaa. 



When there is a distinct suspensor, it is differentiated at very 

 different periods of development, according to the following six 

 types :— 



1. The suspensor may be rudimentary, and never composed of 

 more than three or four superposed cells (Soju, Ampliicarpcea, 

 TrifoUuiii). 



2. It may be composed of two pairs of cells placed crosswise, the 



