There is no theoretical reason to think that the spin-flip component of the
Pomeron is zero. One can measure the spin-flip part using Coulomb-nuclear
interference (CNI). Perturbative QCD calculations show that the spin-flip c
omponent is sensitive to the smallest quark separation in the proton, while
the non-flip part probes the largest separation. According to HERA results
on the proton structure function at very low x, the energy dependence of t
he cross-section correlates with the size of the color dipole. Analyzing th
e data from HERA we predict that the ratio of the spin-flip to non-Rip ampl
itude grows with energy as r(s) proportional to (1/x)(0.1-0.2), violating R
egge factorization of the Pomeron.