Coordinates describing the position of a point in a triangle or simplex. The position of any point P inside, or on the boundary of, a given triangle of reference A1A2A3 can be specified by finding masses m1, m2, m3 such that P is the centre of mass of particles of masses m1, m2, m3 placed at A1, A2, A3 respectively. In this case P is said to have barycentric coordinates (m1, m2, m3). For an internal point of the triangle the masses are all positive. If k>0 then (km1, km2, km3) represents the same point, so barycentric coordinates are usually chosen so that m1+m2+m3=1. With this convention, the barycentric coordinates of A1 are (1, 0, 0), and those of the midpoint of A1A2 are (, , 0). Suppose a random trial can result in one of three possibilities with probabilities p1, p2, p3, such that p1+p2+p3=1, then allocation of these probabilities can be represented by a point inside, or on the boundary of, the triangle. The triangular display is sometimes referred to as a ternary diagram. The idea of barycentric coordinates can be generalized to the case of a tetrahedron in three dimensions or a simplex. See also areal coordinates.