2 Transactions of the South African Philosophical Society. 



at this stage. A slightly older stage (Figs. Il.a and ll.h) shows the 

 megaspore mother-cell much more distinctly differentiated. It is 

 separated from the epidermal cells by three layers of cells, indicating 

 a probability that the archesporial cell * was hypodermal, and gave 

 rise to a row of four cells of which the lowest became the func- 

 tional mother-cell. [This would agree with Guignard's researches 

 on Cassia stipidacea and G. corymhosa referred to below.] 



The mother-cell seems to undergo a resting period while growth 

 of the other parts of the ovule proceeds, as the next stage (Fig. Ill.a) 

 shows the ovule distinctly anatropous, while Fig. lll.h (drawn from 

 the same section) shows the last stage of the first division of the 

 mother-cell. 



The second division of the mother-cell is shown in Fig. IV. The 

 considerable increase in the differentiation of the integuments since 

 the last stage (Fig. III.) indicates that this division also proceeds 

 very slowly. This stage (Fig. IV.) shows very clearly that a row of 

 four potential megaspores is developed. 



The development still proceeds very slowly up to the final 

 differentiation of the functional megaspore. This is shown in 

 Fig. N.a. 



The integuments have now reached practically their mature form, 

 and the diameter of the ovule is almost equal to the internal diameter 

 of the ovary. 



The peculiar arrangement of the integuments is well shown in this 

 Fig. (V.a). This arrangement is always found in the mature ovary, 

 namely, that the micropylar opening in the outer integument is not 

 immediately over that in the inner integument (as is usually the case 

 in other plants). It is probable, therefore, that the pollen tube 

 would not follow the sinuous course of the micropyle, but would 

 grow through the tissue of at least one integument.! 



Fig. N.h, drawn from the same section, indicates that it is 

 the third megaspore from the apex which develops, and not the 

 innermost, as is almost universally-, the case. The two apical 

 megaspores abort and in later stages are almost indistinguishable 

 from the wall of the embryo-sac, which rnerely appears somewhat 

 thickened at the apical end. ;,^ .,,.,.t.j fr.; .v.rf ..v.. 



The cases previously noted in which the innermost megaspore 

 does not become functional are very limited in number, and, among 

 the Dicotyledons, are almost all plants belonging to the natural 

 orders Eosaceae and Leguminosae. Since these two orders are very 

 closely allied, there would seem to be a tendency towards this pecu- 



*' Assuming that a single archesporial cell was originally present; > 

 f I hope to be able to investigate this point later. See note, p. 12. 



