Sporoderm and tapetum development in Eupomatia laurina (Eupomatiaceae). An interpretation.

For the first time, the developmental events in the course of exine structure establishment have been traced in detail with TEM in Eupomatia, with the addition of cytochemical tests. A new look at unfolding events is suggested using our recent hypothesis on self-assembling micellar mesophases. The process proved to be unusual and includes "ghost" stages. The first units observed in the periplasmic space are spherical ones (= normal spherical micelles). These accumulate, resulting in a granular layer up to middle tetrad stage. Sporopollenin precursor accumulation on these units makes the ectexine layer looking as homogenous at late tetrad stage. Simultaneously, the columns of globules are added in the periplasmic space, which reminds an attempt to form columellae; but, the process failed. Instead, a fimbrillate endexine layer of compressed globules appears. The latter augments via additional globules, appearing in the periplasmic space in the free microspore period. The endexine formation is double-stepped spatially and temporally. The second, lamellate endexine layer (laminate micelles) appears late in development, when the channeled intine-I is already established-a very unusual feature. Moreover, a "fenestrated" stage comes unexpectedly at vacuolate stage, when hitherto amorphous ectexine appears pierced by cavernae-the results of reversal of normal spherical micelles (constituents of ectexine) to reverse the ones that open their cores for the entrance of hydrophilic nutrients from tapetum and give them over to the microspore cytoplasm by exchanging their solubilizates.