The aim of this work is to study the possibility of using the ESTRO mini-phantom for transmission measurements of primary kerma in water at a point free in air. We discuss in-air measurements in general, with special attention given to in-air equivalent measurements using a water equivalent mini-phantom. The study includes four different photon energies (4, 6, 10 and 18 MV), where scoring of dose and primary kerma inside a mini-phantom in narrow beam geometry is performed with the Monte Carlo code EGS4. The results reveal that relative measurements (i.e. with and without a water absorber present) at 10 cm depth in a mini-phantom do not represent the variation of primary kerma in water at a point free in air (deviations as large as 7% at 4 MV are observed). Minimum deviations are obtained at depths somewhat larger than the depth of dose maximum. The observed deviations are due to a considerable beam hardening in the water absorber, which changes the amount of attenuation and scatter inside the mini-phantom.