We consider the effect of a long range, flavor changing tensor interaction of possible gravitational origin. Neutrino mixing experiments provide the most sensitive probe to date for such forces---testing the equivalence principle at levels below $10^{-20}$. Here we justify and generalize a formalism for describing such effects. The constraints from neutrino mixing experiments on gravitationally induced mixing are calculated. Our detailed analysis of the atmospheric neutrino data confirms a remarkable result: the atmospheric neutrino data implies the same size force as does the solar neutrino data. Additional tests of this suggestive result are discussed.

In a recent paper I concluded that the neutrino background is important for calculating neutrino flavor evolution in supernovae, and including it weakens the connection between r-process nucleosynthesis and cosmologically relevant neutrino masses. Comments on this paper have been posted to this bulletin board; one by J. Cline and the other by Y.Z. Qian and G.M. Fuller. Here I briefly discuss these Comments and demonstrate that they are inaccurate and ill-considered. No corrections to my paper are necessary.

(The discussion is expanded, and a mistake in the crossing probability is corrected.)

The cosmological neutrino background will mediate long range forces between objects. For a background of temperature T, the potential decreases as 1/r^5 for r >> 1/T and as 1/r for r << 1/T. These forces have large spin-dependent components. If the neutrino background is nonrelativistic, the long range forces are enhanced by a factor of the inverse neutrino velocity. These long range forces may provide a method for observing the cosmological neutrino background.