Abstract: Trochodendron, Tetracentron and Euptelea are considered as a small group of the primitive angiosperms. They are endemic to or mainly distributed in China. Their systematic position has long been debated. The purpose of the present work is to present palynological data both for systematic discussion and for the identification of fossil pollen. The three genera share a number of pollen characters, for example, being spheroid in shape, exine surface distinctly reticulate and lumina rather small and irregular in shape, etc. They are, however, obviously different in other respects of pollen morphology. The pollen grains of Euptelea are the largest of three (c. 30 μm in diameter), those of Tetracentron, the smallest (c. 15) and those of Trochodendron, intermediate (c. 20), 3-colpate in Trochodendron and Tetracentron, and those in Euptelea are 3-celpate but with transitional apertures in one species and 6-rugate, not 3-colpate, in the other. The ruga membranes are coarsely granular (rod-shaped elements under SEM). The lumina are the smallest in Euptelea. Muri in Tetracentron are distinctly striate. Thus, the establishment of three separate families is supported by pollen morphology. Since the pollen grains of Trochodendron and Tetracentron are 3-colpate and those of Magnoliaceae are 1-sulcate, it seems unreasonable to refer them to Magnoliaceae (or Magnoliales). If the rugae of pollen grains in Euptelea further shortened and their number increased, they would resemble those of Hamamelidaceae (such as Sycopsis). Pollen morphology, therefore, suggests that Euptelea is related to Hamamelidales, and supports Cronquist’s viewpoint. Hutchinson (1969) includes both Trochodendron and Euptelea in the same family (Trochodendraceae). However, these two genera are rather distinct in morphology and structure, the presence or absence of vessels and chromosome number, etc. Pollen morphology of two genera also disagrees with the Hutchinson’s viewpoint.

Key words: Trochodendron, Tetracentron, Euptelea, pollen morphology