The native state dynamics of the small globular serine protease inhibitor eglin c has been studied in a long 336 ns computer simulation in explicit solvent. We have elucidated the energy landscape explored during the course of the simulation by using Principal Component Analysis. We observe several basins in the energy landscape in which the system lingers for extended periods. Through an iterative process we have generated a tree-like hierarchy of states describing the observed dynamics. We observe a range of divergent contact types including salt bridges, hydrogen bonds, hydrophilic interactions, and hydrophobic interactions, pointing to the frustration between competing interactions. Additionally, we find evidence of competing water-mediated interactions. Divergence in water-mediated interactions may be found to supplement existing direct contacts, but they are also found to be independent of such changes. Water-mediated contacts facilitate interactions between residues of like charge as observed in the simulation. Our results provide insight into the complexity of the dynamic native state of a globular protein and directly probe the residual frustration in the native state.principal component analysis ͉ protein dynamics ͉ protein energy landscape ͉ tree-like hierarchy of states