Polycyclic structures containing seven-membered carbocycles are important synthetic targets because they constitute the key structural core of many classes of bioactive natural products. [1] Among the different strategies for their assembly, those involving the transition metal-catalyzed cycloaddition of readily available acyclic substrates is particularly attractive. [2] In this context, we have demonstrated that alkylidenecyclopropanes (ACPs) can work as 3C-atom components in several metal catalyzed cycloadditions that afford seven membered-containing polycyclic structures, [3] including an intramolecular (3 + 2 + 2) annulation promoted by Pd(0) catalysts (Scheme 1 route a). [3b] Despite the utility of this latter method, which generally proceeds with good yields and complete diastereoselectivity to give syn fused (3 + 2 + 2) cycloadducts, its synthetic potential is somewhat compromised by the competitive formation of cyclopentanic side products resulting from a formal (3 + 2) annulation. [4] In a related work, Evans and coworkers demonstrated in 2008 that Rh catalysts derived from [Rh(COD)Cl] 2 and P(OPh) 3 promote intermolecular (3 + 2 + 2) cycloadditions between alkenylidenecyclopropanes and activated alkynes, to afford 5,7-fused bicyclic systems (Scheme 1, route b). [5,6] Curiously, although this method uses alkynes as external cycloaddition components, intramolecular (3 + 2) cycloadducts were not isolated. Stimulated by this observation, and as part of our program on the discovery of new transition metal-catalyzed cycloadditions, [7] we analyzed the viability of a Rh-catalyzed intramolecular (3 + 2 + 2) cycloaddition of dienyne precursors of type 1. [3b] Herein, we report the implementation of such cycloaddition, a reaction that affords synthetically relevant 5,7,5-fused tricyclic systems of type 2 with moderate to good yields, high diastereoselectivities and total chemoselectivity (Scheme 1, bottom). Moreover, and in contrast to the Pd-catalyzed (3 + 2 + 2) annulations, [3b] the current methodology allows the use of di-or even tri-substituted alkenes. Also interestingly, an appropriate selection of the Rh catalyst allowed in particular cases a divergent access to the syn or anti cycloadducts 2 and 2'. We also provide preliminary DFT calculations that qualitatively support the experimental results and shed light on the mechanism and on the differences between the Pd-and the Rh-catalyzed versions of these cycloaddition reactions.[ [8] Importantly, (3 + 2) cycloadducts (i.e. 3a) were not detected in the 1 H-NMR spectra of the crude reaction mixture. [9] Therefore, we further analyzed other ligands and Rh sources in order to improve the yield and diastereoselectivity of the cycloaddition. Curiously, the use of a bulkier phosphite such as L1, instead of (PhO) 3 P, allowed to invert the diastereoselectivity (2a : 2a' > 20 : 1), although the yield of the process was lower (entry 2). As exemplified in entries 3-7, the combination of [Rh(COD)Cl] 2 with other phosphorousbased ligands was not particularly suc...