Abstract. We study the geometry of magnetic fluctuations in a CIR observed by Pioneer 10 at 5 AU between days 292 and 295 in 1973. We apply the methodology proposed by Bieber et al. to make a comparison of the relative importance of two geometric arrays of vector propagation of the magnetic field fluctuations: slab and two-dimensional (2D). We found that inside the studied CIR this model is not applicable due to the restrictions imposed on it. Our results are consistent with Alfvenic fluctuations propagating close to the radial direction, confirming Mavromichalaki et al.'s findings. A mixture of isotropic and magnetoacoustic waves in the region before the front shock would be consistent with our results, and a mixture of slab/2D and magnetoacoustic waves in a region after the reverse shock. We base the latter conclusions on the theoretical analysis made by Kunstmann. We discuss the reasons why the composite model can not be applied in the CIR studied although the fluctuations inside it are two dimensional.Key words. Solar physics · astrophysics and astronomy (magnetic fields) · Space plasma physics (turbulence; waves and instabilities)