Available experimental material on phase-transitions of cubic compounds to type I- or type II-antiferromagnetism is briefly reviewed. Particular emphasis is paid to the question of whether they exhibit fluctuation-driven discontinuous transitions as claimed by a series of renormalization-group (RG) investigations for order-parameters withn≧4 degrees of freedom. It turns out that only less than one quarter of the known transitions are classified to be of first order. For most of these, the dominant spin-lattice couplings, i.e. exchangestriction of 3d-ions and magnetostriction and quadrupole-quadrupole interaction forf-ions, are known from independent experiments. Taking this into account, the discontinuous behaviour of the order parameter and related thermodynamic properties were fully explained neglecting critical fluctuations, i.e. by mean-field or randomphase approximations. In the remaining systems, the discontinuity predicted by the RG is not observed, for which several reasons may be responsible: (i) extremely small jumps of the order parameter that are below the limits of experimental resolution or are smeared out by imperfections, (ii) compounds with small spin-lattice interactions may exhibit an-K instead of a single-K structure postulated by the RG-work, and (iii) failure of the RG-criteria established at four dimensions for first-order transitions of readd=3 magnets.