Purcell proposes that the potential energy U of a localized distribution of electrostatic charge has physical significance only as the total energy of the whole configuration. Feynman, on the other hand, favors the idea that U be localized as field energy of density E2/8π however, he points out that the demonstration of E2/8π as localized energy remains ambiguous. The article shows that by calculating the work done by the Maxwell stresses during assembly from infinity of a charge distribution of density q(r), it may be demonstrated unambiguously how the Maxwell stresses create a flow of work that distributes U into localized field energy of density E2/8π. This result supports the Feynman point of view. PACS No.: 03.50De