Dray and Streubel proposed a definition of angular momentum in general relativity based on `Bondi-Metzner-Sachs (BMS) charges'. I show here that the natural definition of center of mass in this program has an infinite-dimensional ambiguity. (This seems not to have been noticed before because previous work has tacitly carried over a special-relativistic assumption.) A related point is that the natural definition of spin in this context is translation-, but not supertranslation-, invariant.
Categories: gr-qc