A very subtle thing is in Wiki I found the energy-momentum tensor in GR is automatically symmetric.
So whether the relativistic energy-momentum tensor is always symmetric? Whether this implies some relativistic constrain?
So whether this is just some coincidence?