shows that in the neighborhood of the magnetic axis of a Tokamak-type configuration, ballooning modes are stabilized simultaneously with localized modes. This result is surprising since in the case of magnetic surfaces with circular cross section the localized criterion does not yield 5 a critical j8. Indeed, the calculations of Adam and Mercier 7 have recently been completed 8 by taking into account second-order corrections in the curvature and it was shown that the stability limit is practically given by I/2TT<1. Hence, it seems that a critical /3 may perhaps be found only for perturbations with small values of m and a radial extension on the scale of the plasma radius.It is a pleasure tc thank Dr. C. Mercier for profitable suggestions and discussions.