During the period from 1963-1973, experiments involving highly enriched uranium units were performed at the Oak Ridge National Laboratory Critical Experiments Facility to determine various critical configurations of three-dimensional arrays. The experiments formed a four-part series and were reported by several different experimenters; the results of interest for this evaluation are those reported for the fourth experimentation, Critical Three-Dimensional Arrays of Neutron Interacting Units: Part IV, published and performed by D.W. Magnuson (Ref 1). Information is also available in the experimental logbook. a This set of experiments, performed November 13-17 of 1972, utilized subcritical metal units on a split table apparatus to determine critical configurations for 2×2×2 arrangements of highly enriched uranium reflected by concrete. Magnuson manipulated the configuration of several uranium cylinders and blocks within a concrete reflector. The different permutations utilized uranium cylinders of two different heights in various positions in the three dimensional array; certain cases also placed thin uranium blocks on top of the cylinders. The thickness of the surrounding concrete, as well as the inner dimensions of the concrete reflector, was also varied in certain cases. The variations resulted in fourteen different experimental permutations or configurations. All fourteen configurations were judged to be unacceptable for use as criticality benchmark experiments due to apparent large extrapolations to obtain experimental k eff values with no supporting data and large uncertainties in the concrete composition and density data. All experiments were initially evaluated; however only three configurations were evaluated in detail. Configurations 2, 4, 6 and 12 were not evaluated in detail because they are subcritical and Configurations 5, 7, 8, 9, and 10 were also were not evaluated in detail because they were extrapolated or corrected to values greater than beta effective (~0.007) above delayed critical, or prompt critical. The experiments evaluated in detail for this report were Configurations 1, 3, and 11. The experimental report also contains information for HEU-MET-FAST-056. Closely related work has been recorded in HEU-MET-FAST-053, which is an evaluation of a different series of three dimensional array experiments with four different moderator materials. HEU-MET-FAST-023 and HEU-MET-FAST-026 are also related because they utilize the same metal cylinders as these experiments (Ref. 1).