Sensitivity studies of the response surface developed for a seven-variable, sparsely-noded set of RELAP calculations show (1) shape of input distribution less important than its mean and variance and (2) standard deviation less important than mean. A comparison of two models developed for the same problem shows the importance of careful choice of evaluation points. It is also shown that the fast Fourier transform is a useful tool for determining the distribution of a variable given by a quadratic response surface. Calculations have been completed for a series of two-dimensional wave propagation problems using the CSQ and WHAM codes. The validity of the Streeter-Wylie method has been demonstrated. Evaluation of the Babcock and Wilcox CRAFT 2 code has begun. The RELAP code has been significantly modified to improve editing and plotting, to increase heat slab capacity, to correct some errors and to evaluate new slip and heat transfer models. Several nodalizations of UHI plants have been performed which investigated the calculated effects of one and two downcomer sections, different slip and quench models, and different water pack assumptions. These various models have been shown to significantly affect the calculated blowdown behavior during a LOCA.