Multisite phosphorylation is a signaling mechanism well known to give rise to multiple steady states, a property termed multistationarity. When phosphorylation occurs in a sequential and distributive manner, we obtain a family of networks indexed by the number of phosphorylation sites
. This work addresses the problem of understanding the parameter region where this family of networks displays multistationarity, by focusing on the projection of this region oπnto the set of kinetic parameters. The problem is phrased in the context of real algebraic geometry and reduced to studying whether a polynomial, defined as the determinant of a parametric matrix of size three, attains negative values over the positive orπthant. The coefficients of the polynomial are functions of the kinetic parameters. For any
, we provide sufficient conditions for the polynomial to be positive and, hence, preclude multistationarity, and also sufficient conditions for it to attain negative values and, hence, enable multistationarity. These conditions are derived by exploiting the structure of the polynomial and its Newton polytope and employing circuit polynomials. A relevant consequence of our results is that the sets of kinetic parameters that enable or preclude multistationarity are both connected for allπ.