Vessel Structure of Gnetum and the Origin of Angiosperms

Abstract

Vessel structure of Gnetum gnemon L. and G. montanum Mgf. was investigated by means of light microscopy (brightfield, phase contrast and Nomarski interference optics), and scanning and transmission electron microscopy. The gyres of the first order framework are compound, consisting of three elementary strands (subunits). Pits ranging from slit-shaped to circular are formed either between the elementary strands of one gyre or between adjacent gyres. (Compound gyres are also figured for Alnus and lens-shaped pits in annular-helical elements are demonstrated for Juglans and Salix). In the two species of Gnetum studied, scalariform pits occur in addition to the well known circular pits. Scalariform perforation plates are present besides scalaroid, foraminate, and simple perforation plates as well as intermediates between the above types. These findings invalidate a common argument against the gnetalean origin of angiosperms, namely that pits and perforations of Gnetum are totally unlike those of primitive angiosperms and that therefore Gnetum must be ruled out as a potential ancestor of angiosperms. Variation in vessel structure of Gnetum is so great that it encompasses the typically circular pits of the coniferopsids as well as patterns of pits and perforation plates found in angiosperms. Some photographs of angiospermous taxa are interspersed with those of Gnetum to indicate the striking similarities. The determination of the patterns and shapes of pits and perforations is discussed in terms of four parameters: 1) the ratio of the width of the cell face to the distance between the gyres; 2) the types and distribution of the second order framework; 3) the stretching and distortion of gyres and/or second order framework; and in the case of perforations, 4) the pattern of wall and bar breakdown. Since the first parameter may change continuously, a continuum between circular pits or perforations and scalariform ones may occur. Such a continuum actually exists in Gnetum as well as in angiosperms. Patterns due to the other three of the above parameters are also formed in similar ways in both Gnetum and angiosperms. These similarities may, of course, be interpreted as the result of parallel evolution. However, when one considers the large number of angiospermous features of Gnetum, one is led to ask whether Gnetum may not have been close to the ancestral stock of all or at least some taxa of angiosperms.