The development of the megaspore wall in Ginkgo biloba has been traced from its inception to its completion. The foundation for the wall is constructed on the free surface of the coenocytic megaspore and consists of a fibrillar matrix forming a system of radially-aligned, hexagonal chambers. A finely granular material begins to accumulate in the chambers, appearing first in the proximal region and then extending radially to form a series of interconnecting columns. Concurrent with the development of the columns a continuous granular basement is deposited against or just within the undersurface of the fibrillar matrix. As development proceeds, the character of the columns and basement layer changes; they become electron-opaque and amorphous. The fact that at this stage they also become resistant to acetolysis is taken as evidence for the changes in appearance being visible consequences of the introduction of the polymer sporopollenin into the wall. Three substrata are then formed beneath the patterned layer. The first is polysaccharide-rich and the material for its construction possibly originates in the dictyosomes of the megaspore cytoplasm. The second substratum is acetolysis-proof and is formed when electron-opaque globules are extruded across the plasma membrane from large, membrane-bound vesicles of unknown origin. The third substratum, like the first, contains polysaccharides which arrive in vesicles, again, possibly of dictyosomal origin. The third substratum develops in contiguity with the walls which partition the megaspore cytoplasm. The structure of the megaspore wall in several cycad genera has also been examined. Although these differ among themselves and from Ginkgo in a number of structural details, the principal features of construction are very much the same. The enhancement in electron contrast exhibited by the fibrillar matrix in all the genera after fixation in glutaraldehyde solutions containing cationic stains and tannic acid suggests that it is an acid polysaccharide-protein complex or glycoprotein. With these fixation procedures, membrane-like lamellae, discernible only in the sporopollenin component of the patterned layer after conventional techniques, can be detected in the matrix material. On this evidence it seems likely that the lamellae are primary structures involved in the formation of the sculptural elements. In the ovules of both the cycads and Ginkgo the nucellus tissue degenerates progressively as the megaspore and its wall develop so that a thick covering of cellular detritus accumulates against the surface of the megagametophyte. This detritus contains sporopollenin and lipids. The characteristics of development and structure of the megaspore wall in zooidogamous gymnosperms are compared with the now well-documented features of the pollen grain wall. The similarities are far more marked than the distinctions; indeed, the resemblances are such that it is possible to adopt pollen wall terminology to define the various constituent parts of the megaspore wall. In an attempt to explain the presence of such an elaborate wall in a situation where its utility is not immediately apparent, a conjecture is advanced which, leaning heavily on the evidence from recent work on the function of the pollen wall in flowering plants, predicts a role in the control of breeding behaviour.