An experiment was conducted in a critical assembly to determine water channel power peaking in a heterogeneous, highly enriched, water-moderated reactor subassembly. In addition to horizontal profiles of power density in a plane below a bank of control rods, data were obtained on the interacting effects of the control rod channel and bottom reflector power peaks. Experimental techniques are discussed. A computation for a comparable multiregion cell was made using a few-group diffusion theory digital computer code. Comparison of the results showed that variations in fast neutron spectrum and in manner of homogenization of materials within cell regions had no appreciable effects upon the power density peak. The choice of thermal neutron spectrum is important. Agreement between all experimental values and calculations with a Maxwellian spectrum was excellent. The comparison between experiment and calculations with a hardened thermal neutron spectrum was relatively poor. The results of this investigation indicate that great care must be exercised in interpreting experimental data on power distribution, and that two-dimensional diffusion theory calculations of power density are substantially verified. The use of U235 monitor foils is recommended in order to obtain a true fission power distribution. It was found that variations in the water gap width of the order of 10% had negligible effect upon the horizontal peak-to-average power ratio in the fuel. The use of metallic control rod extensions was found to decrease the horizontal peak-to-average ratio substantially, the magnitude of the decrease verifying analytical predictions.