Abstract: The present paper deals with the embryological study and the systematic position of Amentotaxus argotaenia (Hance) Pilger. The material used was collected during 1980-1981 from Jin-fo Shan, 1400-1600 m, Sichuan Province, China. The species is dioecious. The male cone sheds its pollen during the period from the end of May to the middle of June. The pollen at mature stage is 2-celled. Pollen chamber appears obvious at the end of the nucellus. When pollen grains are dispersed, megaspore mother cell, which is situated deep in the nucellus, is in meiosis. The megaspore divides mitotically after pollination and the free nuclei of female gametophyte divide for the last time at the end of June. The wall formation takes place at the stage of 256 free nuclei. The development of archegonia takes place at the beginning of July and the fertilization occurs about July 20-23. The fertilized egg divides successively four times and results in a 16-nucleate proembryo. The young embryo is developing in August. It is interesting to note that the development of the young embryo is very slow. When the seed reaches the mature stage from June to July in the following year, the multicellular masses of the young embryos resulted from simple polyembryony remain immature within the female gametophyte. No cleavage polyembryony has been found. The subsequent embryogeny takes place after the seed has shed. Keng (1975) considers that Amentotaxus links the Taxaceae with Cephalotaxaceae. Our embryological data support Keng’s conclusion since they share (1) compound microstrobilus, (2) 2-celled pollen grains at shedding stage and (3) the rather long life cycle. Keng (1975) also mentions that Podocarpaceae may connect with Taxaceae through Phyllocladus. According to Keng the Podocarpaceae is related to Taxaceae to certain degree. It is obvious that the primitive spike-like male strobilus like the one in Cordaitales is obviously retained in Podocarpus spicatus and P. andinus of Podocarpaceae and Amentotaxus of Taxaceae. In addition, like in Amentotaxus there are 16 nuclei before wall formation in the proembryo of Podocarpus nivalis. These facts may well indicate that at least the Podocarpaceae and the Taxaceae were derived from a common stock. As far as the Taxaceae is concerned the authors tend to support the view of Koidzumi (1932) that Amentotaxus and Austrotaxus should be put in the same tribe since both possess the spike-like strobilus, the long life cycle and the seed maturation in the following year. They are probably rather primitive genera in the Taxaceae. The proembryogeny of Torreya is more or less specialized. It may be placed in a rather advanced tribe and the tribe Taxeae (including Taxus and Pseudotaxus)may be between the above two tribes. In conclusion, the Taxaceae is related to the Coniferales in certain respects and, as Keng (1975), Harri (1976) and Wang et al. (1979) have pointed out recently, placing of the Taxaceae in Coniferales is rather justifiable.

Key words: Amentotaxus, Embryogeny, phylogeny.