ABSTRACT.The expected volume of the convex hull of n random points chosen independently and uniformly on the d-dimensional torus is determined.In the 1860's J. J. Sylvester raised the problem of determining the expected area V$(C) of the convex hull of three points chosen independently and uniformly at random from a given plane convex body C of area one. For some special plane convex bodies the problem was solved by Woolhouse [16], Crofton [8] and Deltheil In 1917 Blaschke [3,4] succeeded in proving that, among all plane convex bodies C of area one, Vs(C) attains its minimum if C is an ellipse. In 1974 Groemer [11,12] generalized this statement to d-dimensional convex bodies C of volume one and an arbitrary number n of points. The value of this minimum is known for d + 1 points in a d-dimensional ellipsoid (Kingman [13]) and for an arbitrary number of points in dimensions d = 2 and d = 3 (cf. [6]).An obvious question is to ask for the expected volume of the convex hull of random points chosen on a compact metric manifold. A subset C of a compact metric manifold is called convex (cf. Bangert [2], Walter [14] ) if for any two points x, y G C all geodesic segments (i.e. all curves of minimal length on the manifold joining x and y) are completely contained in C. The convex hull of a set is the smallest convex set (on the manifold) containing it.In the case of the d-dimensional sphere S^ = {x G Ed+1: \\x\\ = 1} (Ed+1 denotes (d+ l)-dimensional Euclidean space), the metric is defined by the minimal Euclidean length of all curves on S^ joining two points. A convex set on S^ is either contained in a hemisphere or is the sphere itself. If n points are contained