EMBRYOGENESIS IN FLOWERING PLANTS 253 



that the fossil angiosperms— of which representatives of Fagaceae, 

 Moraceae, Menispermaceae, Magnohaceae and Lauraceae have been 

 observed in early Cretaceous strata — yield no information as to the 

 ancestry of the group beyond what is already provided by living 

 representatives. Along similar lines, Bailey and his colleagues (1942, 

 1949, 1951) have approached the general problem of angiosperm 

 phylogeny on the assumption that very few orders of flowering plants 

 have become completely extinct, and they have turned to the very 

 curious floral characters found in the Degeneriaceae and related 

 families for evidence of the early features of angiosperms. (The 

 Winteraceae, to which the Degeneriaceae are related, are placed by 

 Hutchinson (1926) in the Magnoliales, his first order in the Archi- 

 chlamydeae.) Here, our special interest lies in the embryogeny in 

 these primitive angiosperm species. At least one set of observations is 

 available from a study by Swamy (1949) of Degeneria vitiensis. Fig. 62. 

 In D. vitiensis the ovule is anatropous, with an outer and inner 

 integument, and an embryo sac with the normal organisation of eight 

 nuclei. There is double fertihsation and cellular endosperm formation. 

 The zygote divides only after about 300 endosperm cells have been 

 formed. The embryonic development is illustrated in Fig. 62. The 

 first division of the zygote is a transverse one, but thereafter the develop- 

 ment of both the terminal and basal segments does not follow an exact 

 pattern and may be variable and somewhat irregular. On further 

 development, the embryo becomes a bulky, mukicellular, club-shaped 

 structure, with no sharp distinction between the distal formative region 

 and the massive basal suspensor. At this stage, there is still no indica- 

 tion of tissue differentiation. The whole development is thus on the 

 massive eusporangiate basis found in primitive ferns and in gymno- 

 sperms. Later, the embryo usually forms three cotyledons (87 per cent) 

 and sometimes four (13 per cent); there is a bulky and somewhat 

 bulbous hypocotyl, a root and a root-cap. At this stage a procambial 

 tissue is also differentiated. The embryo is very small relative to the 

 endosperm. The consistently tricotyledonous condition (with occasional 

 tetracotyledonous seedlings) in Degeneria is a very remarkable feature : 

 no dicotyledonous embryos were observed in some 310 seeds. In 

 Magnolia grandiflora tricotyledons may be found as occasional abnor- 

 malities (Earle, 1938). The embryogenies in various genera of Magno- 

 haceae show a number of points of similarity with Degeneria, e.g. the 

 plasticity of the segmentation pattern, the organisation of an ovoid or 

 club-shaped, undiff'erentiated embryo, the development and persistence 

 of the massive suspensor, the swollen hypocotyl, the nearly triangular 

 shape of the cotyledon, and the relation of the small embryo to the 

 very large endosperm (Maneval, 1914; Earle, 1938; Swamy, 1949). 



