ON THE SPECTRAL SIDE OF ARTHUR’S TRACE FORMULA II