We present two methods to probe the energy landscape and motions of proteins in the context of molecular dynamics simulations of the helix-forming S-peptide of RNase A and the RNase A-3'-UMP enzyme-product complex. The first method uses the generalized ergodic measure to compute the rate of conformational space sampling. Using the dynamics of nonbonded forces as a means of probing the time scale for ergodicity to be obtained, we argue that even in a relatively short time (< 10 psec) several different conformational substrates are sampled. At longer times, barriers on the order of a few kcal/mol (1 cal = 4.184 J) are involved in the large-scale motion of proteins. We also present an approximate method for evaluating the distribution of barrier heights g(EB) using the instantaneous normal-mode spectra of a protein. For the S-peptide, we show that g(EB) is adequately represented by a Poisson distribution. By comparing with previous work on other systems, we suggest that the statistical characteristics of the energy landscape may be a "universal" feature of all proteins.