Transition criteria and phase space structures in a three degree of freedom system with dissipation

The paper proves, in symplectic normal-form coordinates, that for a 3-DOF system near an index-1 saddle: (i) in the conservative case the transition boundary ∂Th equals the intersection of the energy surface with the stable manifold of the NHIM and is given by PX1 = −λ X1 with the energy entirely in the center modes (an S3 in the four center-momentum variables) so that ∂Th ≅ S3 × R; its projection to configuration space is a solid tube D2 × R, and the cone-of-velocity at a position inside the tube obeys θ = arcsin(−λX1/R) (eqs. (27), (28), (41), (43), (47) in the paper) . (ii) With linear damping, the stable manifold tilts to PX1 = ((kp+1)/(kp−1))λ X1, and ∂Th becomes a 4D ellipsoid (S4) in (X1, X2, X3, PX2, PX3); its projection is a solid ellipsoid of transition, and the cone-of-velocity satisfies θ = arcsin(λX1(kp+1)/(R(kp−1))) (eqs. (56)–(59)) . The paper explicitly states the topological types S3 × R (conservative) and S4 (dissipative) for ∂Th and gives the tubes/ellipsoids plus cones in position space . The candidate solution reproduces these statements with the same geometric argument (stable manifold of the NHIM vs. stable manifold of the saddle, in normal form) and the same formulas, including the line constraints PX1 = −λX1 and PX1 = ((kp+1)/(kp−1))λX1, the S3 × R vs. S4 topology, the tube vs. ellipsoid projection, and the conservative/dissipative cone-of-velocity relations. One minor derivational shortcut in the model relates the hyperbolic contribution to the aX1 scaling via S^2−1; although that intermediate identification omits the effect of the linear change of variables, the final aX1 used by the model matches the paper’s eq. (56). Overall, both are correct and follow substantially the same proof.