Abstract
It is found that the E1 contribution to the C12(α,γ)16O cross section at the low energies of interest in astrophysics is poorly determined by K-matrix fits to the available data when the background term contains 1 subthreshold echo pole.