\begin{abstract} We classify central configurations of positive masses on a circle centered at their center of mass. For every number of bodies and every potential $r^{-\alpha}$ with $\alpha>0$, the only such configurations are equal-mass regular polygons. The same conclusion holds for the logarithmic potential. The proof first separates the equality of the masses from the angular geometry. Comparisons obtained from the force equations then reduce mass equality to inequalities for positive kernels and finite probability laws. Different exponent ranges share these comparisons, while requiring different estimates to complete the argument. \end{abstract} \begin{center}\small The proof and its exposition were developed with AI agents. A companion Lean development is separate from the proof below; no formal theorem is used as a premise. This manuscript has not received an independent human review. \end{center} \section{Introduction} Albouy, Cabral and Santos asked whether a central configuration on a circle centered at its center of mass must be an equal-mass regular polygon, throughout the attractive power-law range \cite[Problem~12]{ACS}. Their question asks for a symmetry conclusion from the force equations and the geometric constraint, without prescribing the masses. We prove this classification for the full range, including its logarithmic endpoint. Let $m_1,\ldots,m_n$ be positive masses at distinct positions $z_1,\ldots,z_n\in\R^2$. We consider the attractive force equations \begin{equation}\label{eq:force} \sum_{j\ne i}m_j\frac{z_j-z_i}{|z_j-z_i|^{\alpha+2}}=-\lambda z_i, \qquad 1\le i\le n. \end{equation} For $\alpha>0$, these are the central-configuration equations for the potential $\sum_{i0\quad(1\le i\le n),\qquad \sum_i m_i z_i=0. \end{equation} The conclusion determines both the masses and the geometry. \begin{theorem}\label{thm:main} Let $n\ge2$ and $\alpha\ge0$. If positive masses at distinct positions satisfy \eqref{eq:force} and \eqref{eq:circle}, then all masses are equal and the positions are the vertices of a regular $n$-gon. Conversely, every equal-mass regular $n$-gon satisfies these equations for some $\lambda>0$. \end{theorem} We use the physical potential exponent $\alpha$ throughout. In the convention where the pair force has distance exponent $2a$, the corresponding parameter is $a=-1-\alpha/2$. The logarithmic case is an endpoint of the force family, not the constant potential obtained by substituting zero in $r^{-\alpha}$. Several parts of this picture were known. Cors, Hall and Roberts proved uniqueness for a prescribed cyclic mass order and obtained the equal-mass and logarithmic classifications \cite{CHR}. Hampton proved local isolation of the regular polygon in the Newtonian problem \cite{Hampton}. Wang obtained a sufficient inequality in the exponent and the number of bodies; in particular it gives the Newtonian classification for $n\leq6$ \cite{Wang}. Tang and Zhu excluded several classes of unequal mass words for power-law potentials \cite{TangZhu}. The argument here removes the restrictions on the number of bodies, the positive masses and the nonnegative potential exponent. The first reduction separates the two parts of the conclusion. For a fixed cyclic ordering of the masses, the angular energy is strictly convex after rotations have been removed. Equal masses therefore imply a regular polygon. We include this short convexity proof in Section~\ref{sec:geometry}. The main task is to prove equality of the masses from the full force equations. Section~\ref{sec:geometry} also settles the logarithmic case and configurations with at most four bodies. For positive exponents below two we use consecutive force rows to constrain nearby masses. The last interval below two combines these bounds with the remainders of a trial angular comparison. For exponents at least two we bound the mass ratio from the force equations and compare two bounds for the common radial value. Between two and four the comparison uses a probability law on the masses, with the diagonal term removed explicitly. Above four a quartic estimate controls the trial energy for the full range of larger exponents. The table locates the proof in each range. \begin{center} \begin{tabular}{@{}ll@{}} \toprule Range & Argument proving mass equality\\ \midrule $\alpha=0$, or $2\leq n\leq4$ & Section~\ref{sec:geometry}\\ $0<\alpha\leq1$ & Section~\ref{sec:low}\\ $1<\alpha\leq9/5$ & Section~\ref{sec:middle}; Appendices~\ref{app:middleendpoint}--\ref{app:middlefar}\\ $9/5<\alpha<2$ & Sections~\ref{sec:pressure-floor}--\ref{sec:pressure}; Appendices~\ref{app:pressure-scalar}--\ref{app:signedgap}\\ $2\leq\alpha\leq4$ & Sections~\ref{sec:source-low}--\ref{sec:native-low}; Appendix~\ref{app:escort}\\ $\alpha\geq4$ & Sections~\ref{sec:source-high}--\ref{sq:section}\\ \bottomrule \end{tabular} \end{center} Section~\ref{sec:completion} combines the ranges and states explicitly how the hypotheses of each comparison follow from the original equations. All decimal constants used in inequalities denote exact rational numbers.