Ultrastructure of the germinating Welwitschia mirabilis seed.
The structure and chemical composition of quiescent Welwitschia mirabilis (Hooker fil.) seeds and resultant changes occurring during the first 7 days of germination were studied. Research was focussed on the megagametophyte and its interrelationship with the non-vascularised outgrowth produced by the embryo. This outgrowth was given the name "feeder" by Bower in 1881 who postulated that it functions as an absorptive organ. However the possibility existed that it merely fulfilled the mechanical role of anchorage. Following hydration activation of embryonic collar cells precedes that of gametophyte cells whose rate of activation is governed by relative distance from the embryo. This sequence of activation is suggestive of a stimulatory factor diffusing from the embryo into the gametophyte. Starch, protein and lipid reserves in the collar and developing feeder are consumed within 36 to 48 h. As a consequence the rapidly developing seedling is probably largely dependent on nutrient material in the gametophyte until the plumule emerges, after approximately 5 to 6 days germination. Ventral feeder cells in contact with gametophyte tissue apparently act as transfer cells, developing numerous small wall projections invested with plasmalemma which result in a much greater absorptive surface area. The large numbers of mitochondria occurring in these cells might suggest active uptake of nutrients. At the 3- to 4-day-stage the feeder and gametophyte adhere firmly. While this adherance probably facilitates translocation of nutrients it could also have the secondary function of anchoring the feeder in the gametophyte, thus providing the emerging plumule with a firm base. The apparent root cap origin of ventral feeder cells might explain the positive geotropism of the feeder, as recent work inter alia by Wilkins and Wain (1975) has shown that root cap cells may be geoperceptive. Cytochemical methods used at the light and electron microscope level suggest that reserve material within protein bodies of the embryo and gametophyte might exist as a proteincarbohydrate complex and that globoid origin might be cytoplasmic. The immediate digestion of protein body reserves in the embryo and gametophyte interface zone argues the presence of pre-existing hydrolytic enzymes laid down within the protein bodies prior to quiescence. However the enzymes responsible for reserve breakdown in deep gametophyte tissue seem to be synthesised de nova. Protein hydrolysis precedes lipid digestion which possibly indicates that some of the resulting free amino acids might be used in the de nova synthesis of lipases. Lipid bodies, microbodies, mitochondria and amyloplasts encircled with ER seem to form a complex. Fatty acids resulting from lipase action in the lipid bodies (Ching 1968) are probably converted by microbodies (glyoxysomes) to succunate (Breidenbach and Beevers 1967) which is converted to sucrose by the action of mitochondria (Cooper and Beevers 1969a, b). Excess sucrose is probably converted to starch and stored in the amyloplasts. In 5 days the mean dry mass of the gametophyte decreases by approximately 47% during which time the total amount of lipid decreases by 76.5% and protein by 14%. Although some of the hydrolysed fatty acids and amino acids are no doubt utilised in the gametophyte it is suggested that the majority of fatty acids are probably converted to sugars· which, together with free amino acids (and possibly simple peptides) are transported to and absorbed by the embryo via the feeder whence they are utilised for seedling growth.