he problem of distributing a large number of points uniformly over the surface of a tsphere has not only inspired mathematical researchers, it has the attracted attention ~f:/t/~176 ::t:u;hifi;l:: ::/i~:l :i~t;~ ogous problem is simply that of uniformly distributing points on the circumference of a disk, and equally spaced points provide an obvious answer. So we are faced with this question: What sets of points on the sphere imitate the role of the roots of unity on the unit circle? One way such points can be generated is via optimization with respect to a suitable criterion such as "generalized energy." Although there is a large and growing literature concerning such optimal spherical configurations of N points when N is "small," here we shall focus on this question from an asymptotic perspective (N--~ ~).The discovery of stable carbon-60 molecules (Kroto, et al., 1985)* with atoms arranged in a spherical (soccer ball) pattern has had a considerable influence on current scientific pursuits. The study of this C60 buckminsterfifllerene also has an elegant mathematical component, revealed by F.R.K. Chung, B. Kostant, and S. Sternberg [5]. Now the search is on for much larger stable carbon molecules! Although such molecules are not expected to have a strictly spherical structure (due to bonding constraints), the construction of large stable configurations of spherical points is of interest here, as an initial step in hypothesizing more complicated molecular net structures.In electrostatics, locating identical point charges on the sphere so that they are in equilibrium with respect to a Coulomb potential law is a challenging problem, sometimes referred to as the dual problem for stable molecules.Certainly, uniformly distributing many points on the sphere has important applications to the field of computation. Indeed, quadrature formulas rely on appropriately chosen sampled data-points in order to approximate area integrals by taking averages in these points. Another example arises in the study of computational complexity, where M. Shub and S. Smale [22] encountered the problem of determining spherical points that maximize the product of their mutual distances.These various points of view clearly lead to different extremal conditions imposed on the distribution of N points. Except for some special values of N (e.g., N = 2, 3, 6, 12, 24) these various conditions yield different optimal configurations. However, and this is the main theme of this article, the general pattern for optimal configurations is the same: points (for N large) appear to arrange themselves according to a hexagonal pattern that is slightly perturbed in order to fit on the sphere.To make this more precise, we introduce some notation. We denote by S 2 the unit sphere in the Euclidean space R3: S 2 = {x E R 3 : Jxt = 1}.Lebesgue (area) measure on S 2 is denoted by ~, so that *