On the structure of the inclusive pomeron

The assumption of the exchange in the Regge-Mueller diagrams of a two component factorizable Pomeron of the general from P1 approximately s, P1 approximately s ((lnssup(epsilon)) (lnlnssup(g)), O<=epsilon<=2, γ>=0, leads to a contradiction with positivity if epsilon different from 0. Indeed studying the analytic structure of the generating function of the production cross sections, it is proved that sigmasub(el)(s) >sigma sub(tot)(s) unless epsilon=0