The phenomenon of interaction between air-spaced subcritical units of fissile materials has been investigated experimentally, and more recently using Monte Carlo calculations. Studies of array interactions using Monte Carlo codes indicate that certain techniques of array analysis which proceed from the extension of available experimental data may yield nonconservative results. These areas of apparent nonconservatism cannot be verified from studies of experimental data and may be subject to further question in view of inconsistencies observed among some of the Monte Carlo calculated arrays. Studies of experimental and calculated array data show the differing conclusions which may be developed and point out the need for future work.