"No satisfactory solution has been found as yet for the unexpectedly small vertical gradients [with respect to the observed ones] obtained by applying the divB=0 condition [to the solar photosphere magnetic field measurements]" (S. Solanki, 2003, Astron. Astrophys. Rev. 11, p. 184). This is always true today, when an observation review has just been published (H. Balthasar, 2018, Solar Phys. 293, 120). I propose to build an explanation in the frame of the vector law in magnetised media B=µ0(H+M), when the magnetisation M due to the plasma diamagnetism, itself due to the electrons or free charges spiralling about the magnetic field, is large enough. This requires larger than usual model free charges densities in the solar photosphere. However, it can be remarked that inside the Sun, like in the Solar Corona, the electron thermal velocity clearly surpasses their escape velocity. Their charge-dipole interaction with the neutral Hydrogen atoms in the photosphere slow down their escape. They accumulate there and become thus responsible for the non-negligible magnetisation. The last question is to determine if spectropolarimetry measures B or H: I will show that this is H, leading thus to the non-zero observed divergence revealing the presence of a non-negligible magnetisation due to the electrons escaping the Sun's interior.