Abstract:Taxus yunnanensis (Taxus wallichiana var. wallichiana) is an endangered species on China's national protection list. This plant is located mainly in Yunnan, Sichuan, Chongqing, and Tibet. In nature, the seeds of T. yunnanensis need about 2 years to eliminate dormant factors for germination. To shorten the time of seed germination and improve the germination rate, many studies of seeds on biology, physiological ecology, and cultivation have been conducted. However, there are no consistent views and conclusions about how to improve germination rates. Thus, clarification of the embryonal germination characteristics and the water permeability of seed coats form the basis for understanding causes of seed dormancy. In order to explore the mechanisms of T. yunnanensis seed dormancy and germination, fresh seeds were used for morphological observation, water permeability tests, and tests on the effects of different seed sizes and endosperm quantities on the germination rate for in vitro embryos. Our results showed that T. yunnanensis seeds were obovate three-sided prisms, embryos were white and bottle shaped, and the endosperm was pale yellow and oily. One thousand seeds weighed 72.934 g. The relative weights of seed coat, kernel, and embryo accounted for 50.56%, 40.45%, and 8.99% of total seed weight, respectively. The volume ratio of an embryo versus a seed (E ∶ S) was 7.4%. The volume range for embryos and seeds was 0.156-1.012 and 51.658-109.649 mm3, respectively. Mantel tests showed that there was a significant positive correlation between the volume of embryos and that of seeds (P < 0.05, r=0.810). The water absorption rate of intact seeds was lower than that of cut seeds. However, the differences of water absorption rates for seeds corroded by acids for 40 and 20 min were reduced with the extension of soaking time. When the moisture content of a seed was saturated, there was no significant difference of water absorption rate among intact (23.7%), cut (28.7%), and acid-corroded (28.0%) seeds. The germination rate of in vitro embryos was markedly different for different seed sizes (P < 0.05). Specifically, the germination rate of in vitro embryos for superlarge seeds (98.051-109.649mm3), large seeds (86.453-98.050mm3), medium seeds (74.855-86.452 mm3), small seeds (63.257-74.854mm3), and supersmall seeds (51.658-63.256mm3) was 80%, 77.8%, 67.5%, 62.0%, and 44.0%, respectively. Moreover, the endosperm inhibits the germination of embryos, and the germination rate of embryos without endosperm (41.3%) was significantly higher than that for embryos with endosperm (P < 0.05). The respective germination rates for seeds with a portion of the endosperm and full endosperm was 9.7%-13.3%, and 0%. Finally, the freshness of in vitro embryos also affected their germination rates. In conclusion, the present study showed that sampling materials and their cultivation conditions should be consistent between studies. In addition, the seed dormancy of T. yunnanensis was moderate and caused by physiological dormancy. Many factors can contribute to physiological dormancy; this study confirmed that inhibitors within seeds are the most important causes of T. yunnanensis seed dormancy. Additional studies should focus on the changes in quantity of inhibitors over time and their effects on the germination rates for in vitro embryos.