We report a linear dependence of the phonon splitting ∆ω on the non-dominant exchange coupling constant J nd in the antiferromagnetic transition-metal monoxides MnO, FeO, CoO, NiO, and in the frustrated antiferromagnetic oxide spinels CdCr2O4, MgCr2O4, and ZnCr2O4. It directly confirms the theoretical prediction of an exchange induced splitting of the zone-centre optical phonon for the monoxides and explains the magnitude and the change of sign of the phonon splitting on changing the sign of the non-dominant exchange also in the frustrated oxide spinels. The experimentally found linear relationh∆ω = βJ nd S 2 with slope β = 3.7 describes the splitting for both systems and agrees with the observations in the antiferromagnets KCoF3 and KNiF3 with perovskite structure and negligible next-nearest neighbour coupling. The common behavior found for very different classes of cubic antiferromagnets suggests a universal dependence of the exchange-induced phonon splitting at the antiferromagnetic transition on the non-dominant exchange coupling.The interplay of magnetism and the underlying crystal lattice is a topical issue of condensed-matter physics. This spin-phonon coupling can relieve frustration via a spin-driven Jahn-Teller effect in frustrated magnets [1,2], lead to novel excitations such as electromagnons in multiferroics [3,4], and can even bear the potential for future applications via magneto-dielectric effects [5]. For transition-metal monoxides (TMMOs) a magnetisminduced anisotropy in the lattice response was predicted theoretically [6]. This approach has been extended to other material classes such as Cr based spinels, which are hallmark systems for highly frustrated magnets [7][8][9][10], where spin-phonon coupling leads to a splitting of characteristic phonon modes [11][12][13][14][15].TMMOs are both textbook examples for antiferromagnets governed by superexchange in a cubic rock-salt lattice and benchmark materials for the understanding of strongly correlated electronic systems [16,17]. The magnetic structure of the TMMOs consists of ferromagnetic planes coupled antiferromagnetically, e.g., along [1 1 1] as depicted in Fig. 1(a). The antiferromagnetic 180• nextnearest neighbor (nnn) exchange J 2 is supposed to be the driving force of the magnetic ordering [18,19], leaving the nearest-neighbor (nn) exchange J 1 frustrated, since it cannot satisfy all its pairwise interactions [see Fig. 1(c)]. In Fig.1(d) , and find a linear slope of k B T N /J 2 S(S + 1) ∼ 3 (solid line) close to the expected relation in mean-field approximation (dashed line). In a pioneering paper Massidda et al. [6] showed that even for purely cubic TMMOs the antiferromagnetic order is accompanied by a Borneffective-charge redistribution from spherical to cylindrical with the antiferromagnetic axis being the symmetry axis, e.g. [1 1 1]. Consequently, the cubic zone-centre optical phonon is predicted to split into two phonon modes with eigenfrequencies ω and ω ⊥ for light polarized parallel and perpendicular to the cylindrical axis, resp...