It is frequently observed in Ziegler-Natta polymerization that the dependence of the overall polymerization rate on the catalyst/cocatalyst ratio has a more or less marked maximum (see, e.g., Porri and Giarusso 1 ). However, in lithium-initiated anionic polymerization, the addition of organometallic modifiers usually monotonically decreases the overall rate. As early as 1960, it was shown 2 that the addition of AlEt 3 or ZnEt 2 to n-BuLi gradually decreases the rate of styrene polymerization in benzene to full inactivity at Al/Li ≈ 1 and Zn/Li ≈ 10, respectively. This was easily explained by the formation of LiAlR 4 and LiZnR 3 "ate" complexes which are known to be inactive in polymerization of nonpolar monomers. 2,3 A similar monotonic dependence was observed later for butadiene polymerization in cyclohexane with the sec-BuLi-Bu 2 Mg initiator. 4 All of these findings did not encourage a search for new bimetallic anionic initiators for polymerization of nonpolar monomers.However, in the case of heavier alkali metals the situation seems to be different. True, a few earlier attempts to use a preformed stoichiometric NaMgR 3 complex in diene polymerization also were unsuccessful. From the data reported, 5 one can estimate that the firstorder rate constant [k′ ) -(1/M)(dM/dt)] for isoprene polymerization in cyclohexane at 20°C is on the order of 10 -6 s -1 , and at 50°C, it is on the order of 10 -5 s -1 in comparison to 3 × 10 -4 (30°C) for the uncomplexed organosodium initiator 6 (the initiator concentration was ca. 5 × 10 -3 M in all cases). According to a later publication, 7 only 16% yield of polybutadiene was achieved after "1 day or more" with the same initiator in hexane at 50°C.Much more interesting results were obtained with substoichiometric mixtures. It was observed recently that the addition of small amounts of various aluminum alkyls such as AlEt 3 , Al(i-Bu) 3 , HAl(i-Bu) 2 , and so on to the common disodium tetra-R-methylstyrene initiator greatly increased the rate of butadiene polymerization in hydrocarbon solvents up to a certain maximum value. 8 The Al/Na ratio at the maximum depended on the particular aluminum compound. The addition of an aluminum compound above the optimum ratio decreased the overall rate to the full cessation of the process.In all of these cases, small amounts of a polar solvent were introduced with the organosodium component; hence, the role of solvation effects and the solvation/ complexation competition could not be excluded. However, as was reported in the preliminary communication, 9 at least for one particular model system, namely, 1,1,3-triphenylpropylsodium-AlPh 3 , the dependence of the polymerization rate on the Al/Na ratio passed through a maximum even in a pure hydrocarbon medium (toluene) without any polar additives.Here, we reported the results of the investigation of butadiene polymerization with a RNa-R 2 Mg system in toluene. In this study the same R ) 2-ethylhexyl was used for both sodium and magnesium components. Because of the solubility of 2-ethylhexylsodium...