Abstract
In the earliest discussion of this problem Nicholson (1) expressed the potential as a series of spheroidal harmonics with coefficients satisfying an infinite system of linear equations, and gave a formula for an explicit solution; but this formula appears to be meaningless and its derivation to contain serious errors. In the present paper, starting tentatively from Nicholson's infinite system of linear equations, a much simpler, though still implicit, specification of the potential is developed; this involves a Fredholm integral equation the existence and uniqueness of whose solution are deducible from standard theory. The specification so obtained for the potential is shown rigorously to satisfy the differential equation and boundary conditions of the electrostatic problem. The Neumann series of the integral equation is shown to converge to its solution, so that the potential, and other aspects of the field, can be explicitly formulated and thus computed. The errors in Nicholson's process of solving his system of equations are exhibited in detail, and it is concluded that attempts to carry through that process without error cannot lead to an explicit solution.