Dr. Sutton will begin by giving a brief description of the MC21 Monte Carlo neutron and photon transport code jointly developed by the Knolls Atomic Power Laboratory and the Bettis Atomic Power Laboratory. He will then discuss two recent research projects undertaken using MC21. The first is a method for minimizing the largest statistical uncertainties in Monte Carlo reactor calculations. Unlike other methods for accomplishing this goal that rely on auxiliary adjoint calculations, this method is completely contained within the Monte Carlo calculation. The second topic discussed is an anomalous increase in Monte Carlo statistical uncertainties rear surfaces employing reflecting boundary conditions for iterated-fission-source calculations. A theory is presented that yields qualitative agreement with observation. It is found that the anomalous behavior is inherent in any Monte Carlo code.