“…Then we found, that in order, that we have (3) J <p eKp(e) (p) dv = $<f> ((J(e))'p) exp ( -/(<% p)) dv o ((J(e))' Stands for the transpose of J(e)} for an arbitrary choice of e e g and of φ e & (G), it is sufficient, but not necessary, that there exist a polarization which is also an ideal, for one (and hence for all) element of 0, or even that T be quasi-equivalent to a representation, induced by a one-dimensional character of a connected invariant subgroup of G (cf. Then we found, that in order, that we have (3) J <p eKp(e) (p) dv = $<f> ((J(e))'p) exp ( -/(<% p)) dv o ((J(e))' Stands for the transpose of J(e)} for an arbitrary choice of e e g and of φ e & (G), it is sufficient, but not necessary, that there exist a polarization which is also an ideal, for one (and hence for all) element of 0, or even that T be quasi-equivalent to a representation, induced by a one-dimensional character of a connected invariant subgroup of G (cf.…”