Abstract
The key assumption is that of Leinaas and Myrheim and of Berry and Robbins, here specialized to spin zero: for n particles, the argument of the wave function should be the unordered multiplet {r 1,r 2,...,r n }. I also make use of the requirement that wave functions in the domain of the Hamiltonian must be continuous functions of the spatial variables. The new proof presented here has advantages of simplicity and transparency in comparison with earlier work based on the same two principles and it uses weaker assumptions, especially avoiding the use of rotations of the relative coordinate of identical particles