The aim of this study was to evaluate, by comparing simulation results with measurement results, the impact of the lack of scattering volume in experimental conditions of preclinical irradiation. First, a Monte Carlo model of a small animal irradiator, the Faxitron CP-160, was developed with GATE (Geant4 Application for Tomography Emission). To validate the model, simulated data were compared to depth dose and off-axis ratio profiles measured with a plane-parallel ionization chamber and Gafchromic® EBT films, respectively, in a solid water phantom. The AAPM TG-61 protocol was applied to measure the dose rate at the surface of a semi-infinite reference phantom. Then, the model was used to determine the dose distributions in three different phantom settings: a semi-infinite water phantom, a 2.8-cm-thick water phantom and a 2.8-cm-diameter cylindrical water phantom. The dose distributions measured and simulated with Monte Carlo methods in a semi-infinite water phantom were similar (<2%), thus validating our Monte Carlo model. The highest dose underestimation was observed between the reference and the cylindrical phantom (more than 15% difference for the entrance dose) and was due to the lack of lateral scatter and backscatter. The use of standard backscatter factors and AAPM TG-61 protocol may result in a significant underestimation of the dose absorbed by small irradiated phantoms, such as mice or cells, in preclinical studies. Background: For preclinical radiotherapy studies, radiobiologists were used to determine the irradiation time depending only on the source surface distance. This work aimed to demonstrate that scatter conditions have a large impact on dose rate. Measurements and Monte Carlo simulations were used.