Supported by results obtained with semi-classical quantization techniques, and with a quantum mechanical "square-root Klein-Gordon" operator, it is argued that Positronium (Ps) may exhibit a proper quantum-mechanical ground state whose energy level lies $\approx 2m c^2$ below its "hydrogenic (pseudo-) ground state" energy, where $m$ is the empirical rest mass of the electron. While the familiar hydrogenic pseudo-ground state of Ps is caused by the Coulomb attraction of electron and anti-electron, modified by small spin-spin and radiative QED corrections, the proper ground state is caused by the magnetic attraction between electron and anti-electron, which dominates over the electric one at short distances. This finding suggests that the familiar "annihilation" of electron and anti-electron is, in reality, simply yet another transition between two atomic energy levels, with the energy difference radiated off in form of photons --- except that the energy difference is huge: about 1 MeV instead of the few eV in a hydrogenic transition. In their proper ground state configuration the two particles would be so close that they would electromagnetically neutralize each other for most practical purposes, thus giving the appearance of an annihilation. Once in such a tightly bound state such pairs would hardly interact with normal matter and not be noticeable --- except through their gravitational effects in bulk! If the existence of such a low-energy ground state is confirmed it would imply that a significant part of the mysterious "dark matter" in the universe may consist of such matter-antimatter bound states.

The original Jeans dispersion relation and instability criterion are derived by a mathematically well-defined limiting procedure. The procedure highlights Jeans' physical reasoning and vindicates the (in)famous ``Jeans swindle.'' A second, independent procedure is stated which yields the same result.