Floral structure and evolution of primitive angiosperms: Recent advances

Concepts of primitive angiosperm flowers have changed in recent years due to new studies on relic archaic groups, new paleobotanical finds and the addition of molecular biological techniques to the study of angiosperm systematics and evolution.Magnoliidae are still the hot group, but emphasis is now on small primitive flowers with few organs and also on the great lability of organ number. Of the extant groups, a potential basal position of the paleoherbs has been discussed by some authors. Although some paleoherbs have a simple gynoecium with a single orthotropous ovule, anatropous ovules may still be seen as plesiomorphic in angiosperms. Anatropy is not necessarily a consequence of the advent of closed carpels. It may also exhibit biological advantages under other circumstances as is the case in podocarps among gymnosperms. Valvate anthers have now been found in most larger subgroups of theMagnoliidae (recently also in paleoherbs) and in some Cretaceous fossils. Nevertheless, as seen from its systematic distribution, valvate dehiscence is not necessarily plesiomorphic for the angiosperms, but may be a facultative by-product of the thick connectives and comparatively undifferentiated anther shape inMagnoliidae and lowerHamamelididae. A perianth is relatively simple in extantMagnoliidae or even wanting in some families. In groups with naked flowers the perianth may have been easily lost because integration in the floral architecture was less pronounced than in more advanced angiosperm groups. Problems with the comparison of paleoherb flowers with those ofGnetales are discussed. The rapid growth of information from paleobotany and molecular systematics requires an especially open attitude towards the evaluation of various hypotheses on early flower evolution in the coming years.     

Fossil pollen records of the problematical primitive angiosperm familyLactoridaceae in Australia

Microfossils which matchLactoris (Lactoridaceae) pollen more closely than those of any other living angiosperm occur in Campanian to Paleogene sediments around the margins of Australia. These are referred to the fossil genus Lactoripollenites (Zavada & Benson 1987). A species belonging to the same genus occurs in older (Turonian-Santonian) deposits off southern Africa but Australian specimens represent not only the most southern, but also the youngest known (Oligocene) records to date. Our data support suggestions that theLactoridaceae were widespread across the Southern Hemisphere during the Late Cretaceous (Lammers & al. 1986,Zavada & Benson 1987). An homology between gymnosperm sacci and the saccus-like structures found in Lactoripollenites and some specimens ofLactoris pollen is contested, as is the use of (anasulcate) apertures to support the primitive position of the family.     

Structure and development of the ovule and seed in Aristolochiaceae, with particular reference to Saruma

All members of Aristolochiaceae have anatropous, bitegmic, crassinucellate ovules, which are endostomic except in Saruma and Asarum arifolium where ovules are amphistomic. The outer integument is two cell-layered and the inner integument is three cell-layered. The chalazal megaspore is the functional one. All these conditions appear to be plesiomorphic for the order Piperales, which consists of five families, Aristolochiaceae, Hydnoraceae, Lactoridaceae, Piperaceae and Saururaceae. The embryo sac in Aristolochiaceae is eight-nucleate and corresponds to the Polygonum type; a hypostase is frequently present in this family. The seed coat of Aristolochia s.l., Asarum, Saruma and some Thottea species consists primarily of a two cell-layered testa, and a three cell-layered tegmen. In some species the cells of the outer epidermis become radially elongated, forming reticulate wall thickenings. Cells of the inner layer of the testa have crystals and thickened inner walls. The three layers of the tegmen are tangentially elongated, and become cross fibres at maturity, as fibres of the outer and inner layers are parallel to the seed axis, whereas those of the middle layer are perpendicular to it. This type of seed coat anatomy is synapomorphic for Aristolochiaceae. In addition, the gross morphology of the seed and elaiosome histology are remarkably similar in Asarum and Saruma, thus supporting a sister-group relationship between them. Embryological and seed characters do not supply any synapomorphy that support a close relationship between Aristolochiaceae, Hydnoraceae and Lactoridaceae. Instead, some seed features such as the absence of seed appendages and the collapsed cells of endotesta may indicate a close relationship of Lactoris with Piperaceae plus Saururaceae, although this is the subject of further analysis.