We consider a new three-nucleon force generated by the exchange of one pion in the presence of a 2N correlation. The underlying irreducible diagram has been recently suggested by the authors as a possible candidate to explain the puzzle of the vector analyzing powers Ay and iT11 for nucleon-deuteron scattering. Herein, we have calculated the elastic neutron-deuteron differential cross section, Ay, iT11, T20, T21, and T22 below break-up threshold by accurately solving the AltGrassberger-Sandhas equations with realistic interactions. We have also studied how Ay evolves below 30 MeV. The results indicate that this new 3N F diagram provides one possible additional contribution, with the correct spin-isospin structure, for the explanation of the origin of this puzzle.PACS numbers: 24.70+s, 21.30.Cb, 25.10.+s, 25.40.Dn, 21.45.+v, and 13.75.Cs The A y puzzle (or, more appropriately, the puzzle of the vector analyzing powers) is probably the most famous of the open problems in 3N scattering at low energy. The problem with this observable has been observed quite early at the Tokio & Sendai Conference (Few-Body XI, 1986) [1], since at that time the first reliable Faddeev calculations with realistic 2N potentials were becoming available, thanks in particular to the employment of separable expansion methods which transform the 3N scattering equations of Alt, Grassberger, and Sandhas (AGS) [2] into an effective, multichannel Lippmann-Schwinger equation. This method of calculation has been pushed forward to obtain accurate results particularly by the Graz group [3]. Since then, various alternative methods of solution of the 3N scattering equation have been developed and tested [4], and outstanding progresses have been made in the computational techniques in order to: 1) include three-nucleon forces (3N F ) in the 3N scattering equations [4][5][6]; 2) treat explicitly the ∆ dynamics in the 3N system [7]; 3) provide a combined description of the 3N dynamics with Coulomb, realistic 2N , and phenomenological 3N forces [8].The puzzle was confirmed by these new approaches, and it turned out that the existing 2π-3N F [9-11] provided a too small effect for A y [4,5,8], and not always in the right direction. In absence of new 3N F 's that could explain the puzzle, and since the 3N A y is rather sensitive to the 3 P j N N phase shifts, it was concluded in Ref.[12] (see also references therein) that such phases and the associated N N potentials derived from modern phase-shift analysis must be modified at low energy. These modifications can be achieved without affecting appreciably the 2N data because the low-energy 2N observables cannot resolve the 3 P j phases uniquely due to the Fermi-Yang ambiguities. However, as has been argued in Ref.[13], it is not possible to increase the 3N A y with reasonable changes in the N N potential, hence additional 3N F 's of new structure have to be considered. Recently, an attempt has been made [14] using a purely phenomenological 3N F of spin-orbit type, constructed ad hoc to affect only the tripl...