A novel experimental approach based on a toroid Cavendish balance is used to evaluate the product of photon mass squared and the ambient cosmic magnetic vector potential A. The method is based on the energy density of the vector potential in the presence of photon mass, not on measurement of the magnetic field. The experiment discloses Am 2 g , 2 3 10 29 T m͞m 2 , with m 21 g as the characteristic length associated with photon mass. Consequently, if the ambient magnetic vector potential is A ഠ 10 12 T m due to cluster level fields, m 21 g . 2 3 10 10 m. If we conservatively use galactic fields prior to a reversal, then m 21 g . 1 3 10 9 m, a figure still superior to that derived from the Jovian magnetic field. [S0031-9007(98)05451-9] PACS numbers: 12.20.Fv, 14.70.Bh, 41.20.Jb, 98.80.Cq The photon mass is ordinarily assumed to be exactly zero. If there is any deviation from zero, it must be very small, since Maxwellian electromagnetism has been very well verified (in the classical domain). A nonzero photon mass would give rise to a wavelength dependence of the speed of light in free space, the possibility of longitudinal electromagnetic waves, a leakage of static electric signals into conductive enclosures, and a more rapid (exponential or Yukawa) falloff of magnetic dipole fields with distance [1,2] than the usual inverse cube dependence.Electromagnetism in the presence of nonzero photon mass is described by the Maxwell-Proca equations [1,3],in which E is the electric field, B is the magnetic field, r is the charge density, J is the current density, V is the scalar potential, A is the vector potential, c is the speed of light, and m 21 g h͞m g c is a characteristic length, the Compton wavelength of the photon, with m g as the photon mass. Maxwell's equations correspond to m g 0. The possibility of nonzero photon mass has been studied by Bass and Schrödinger [4], deBroglie and Vigier [5], Feynman [2], and others. Gauge invariance is lost [2] if m g . 0, since, in the Maxwell-Proca equations, the potentials themselves have physical significance, not just through their derivatives; the Lorentz gauge is required. Several experimental limits on the photon mass have been reported. Laboratory measurements of the speed of light at different frequencies [2] give m 21 g . 1.4 km (m g , 10 210 eV or 2 3 10 243 g), laboratory tests [2,6] of Coulomb's law give m 21 g . 3.1 3 10 7 m, measurements [7] of Earth's magnetic field give m 21 g . 1 3 10 8 m, more recently [8] m 21 g . 2.5 3 10 8 m, and measurements [9]of Jupiter's magnetic field give m 21 g . 5 3 10 8 m (corresponding to a photon mass m g , 6 3 10 216 eV or 8 3 10 249 g). More stringent limits based on inference from large-scale magnetic features in astronomical plasma objects have been reported as reviewed by Barrow and Burman [1], but such inferences are very indirect in view of the uncertainty regarding the mechanism of generation of such fields. Photon mass has been suggested by Georgi, Ginsparg, and Glashow [10] to affect cosmic background radiation. Photon m...