The present work measured the accuracy of VAX monitor unit calculations currently used for highly blocked photon fields at The University of Texas M. D. Anderson Cancer Center. The VAX program is an in-house computer program that calculates the setting for linear accelerator photons that is needed to deliver a prescribed dose to a muscle miniphantom located on the beam's central axis at a prescribed depth beneath the surface in water. Monitor unit settings were computed for the Varian 2100 with multileaf collimation using 6 MV and 18 MV photons at a depth of 10 cm with different blocking ratios. Concentric and lateral blocking methods were used. Measurements were taken with an ion chamber and then compared with the VAX calculations with off-axis factors and peak scatter factors added in that are not available in the program. The results showed less than 2% difference for the concentric fields containing no more than 80% blocking and an equivalent square of at least 3.5 cm for 6 MV and 5 cm for 18 MV X-rays. For the lateral fields with the central axis blocked, less than 2% difference was found when blocking was 80% or less and the equivalent square was no less than 3.5 cm for 6 MV X-rays. For 18 MV X-rays, the maximum blocking and minimum equivalent square were 60% and 5 cm for 2% accuracy. This confirmed that the VAX system currently in use is accurate unless a very large amount of blocking is present or the central axis is blocked in very small fields. Previous research done at dmax with similar blocking ratios produced similar results.