stereochemical features of the desired natural products and thereby perhaps simplify its total synthesis.3 However, all attempts at carrying out such a cycloaddition with 2 using Fisher carbene complexes4 3 [Z ) C(OMe)dCr(CO)5], which would have given the exo adducts, or the simple acid chloride 3 (Z ) COCl), which would have given more of the endo adducts, unfortunately gave only starting material, presumably due to the steric hindrance of these dienophiles. We proposed using R-methyl allenecarboxylates as the dienophiles in order to decrease this steric hindrance with a subsequent reduction of the exocyclic double bond to give compounds such as 4. We report herein the thermal cycloaddition of silyloxydienes 2 (X ) OSiR3) and allenecarboxylates which ultimately afford the formal exo Diels-Alder adducts 4. Heating a solution of E-4-methyl-2-silyloxybutadiene 55 with ethyl 2-methylbuta-2,3-dienoate 6 in toluene at 120 °C (sealed tube) for 14 h afforded, after chromatographic separation, a mixture of three compounds, the endo Diels-Alder product 7n (18%), the exo Diels-Alder product 7x (28%), and the [2 + 2] cycloadduct 8 (22%) (Scheme 1). The isolation of [2 + 2] cycloadducts from dienes and allenes was precedented in our own work and that of others.6 However, we have now shown that thermolysis of the purified cyclobutane 8 in toluene at 120 °C for 4 d gave only the exo cycloadduct 7x. Thus heating the diene 5 and dienophile 6 in toluene for 4 d gave only the two [4 + 2] cycloadducts 7n and 7x in 14 and 58% yields, respectively. Although we had observed the thermal rearrangement of a [2 + 2] cycloadduct into a [4 + 2] cycloadduct before,6ab this is the first case where the stereochemical course of this process could be determined. The mechanism of the rearrangement of simple 3-ethenylmethylenecyclobutanes I has been proposed to involve an initial bond homolysis to give a bis(allylic) diradical II which then recombines at the termini to give the 4-methylenecyclohexene III.7 Indeed it was shown that the divinyl compound I (R ) CHdCH2) had a lower activation energy for this reaction (Ea ) 27.3 kcal/mol) than did the parent I (R ) H) (Ea ) 35.7 kcal/mol).7a We believe that the substitution in our case may cause the reaction to proceed through a zwitterionic rather than a diradical pathway although the latter can certainly not be ruled out.