Abstract: In this paper, the ³²P radio-tracer technique was applied to study the ways of phosphorus transfer between poplar (Populus euramericana cv.‘I-214') and black locust (Robinia pseudoacacia). Afive compartment root box (18cm×18cm×26cm) was used for testing the existence of the hyphal links between the roots of two tree species when inoculated with vesicular-arbuscular (VA) mycorrhizal fungus (Glomus mosseae). Populus I-214 (donor) and Robinia pseudoacacia (receiver) were grown in two terminal compartments, separated by a 2 cm root-free soil layer. The root compartments were lined with bags of nylon mesh (38μm) that allowed the passage of hyphae but not roots. The top soil of a mixed stand of poplar and black locust, autoclaved at 121℃ for one hour, was used for growing seedlings for testing. In5 compartment root box, mycorrhizal root colonization of poplar was 34%, in which VAmycorrhizal fungus was inoculated, whereas 26% mycorrhizal root colonization was observed in black locust, the other terminal compartment, 20 weeks after planting. No root colonization was observed in non-inoculated plant pairs. This indicated that the mycorrhizal root colonization of black locust was caused by hyphal spreading from the poplar. Test of tracer isotope of ³²P showed that the radioactivity of the treatment significantly higher than that of the control (P<0.05), 14 days from the tracer applied, to 27 days after, when VAmycorrhizal fungus was inoculated in poplar root. Furthermore, mycorrhizal interconnections between the roots of poplar and black locust seedlings was observed in situ by binocular in root box. All these experiments showed that the hyphal links was formed between the roots of two species of trees inoculated by VAmycorrhizal fungus.Four treatments were designed according to if there were two nets (mesh 38μm), 2 cm apart, between the poplar and black locust, and if the soil in root box was pasteurized. Most significant differences of radioactivity among four treatments appeared 44 days after feeding³²P, the radioactivity of the day was applied to estimating the contribution of the various possible transfer ways to the total amount of nutrient transfer.Level of ³²Pradioactivity was found to be significantly (P<0.05) higher in leaves of the treatment of "no separated and soil non-pasteurized"(17.1 pulse·g^-1·s^-1) than in leaves of "net separated and soil non-pasteurized" ( 5.3 pulse·g^-1·s^-1),and also significantly higher in leaves of "no net separated and soil pasteurized"( 11.5 pulse·g·s^-1) than in leaves of "net separated and soil pasteurized" (2.3 pulse·g^-1·s^-1), and very significantly (P<0.01) higher in leaves of "no net separated and soil non-pasteurized" than in leaves of "net separated and soil pasteurized",whereas the levels of ³²P radioactivity were not significantly different between the other treatments.The results showed that root contact and root exudations were the main ways of phosphorus transfer between the two species and the amount of phosphorus transfer through these two ways accounted for 62% of the total.The activity of the microorganisms including VAmycorrhizal fungi and the interaction between the microorganisms and root contact and root exudations made up 38% of total amount of phosphorus. The effect of mycorrhizal hyphal links in the direct nutrient transfer between poplar and black locust through separate mesh (38μm) was little.