Proteins undergo fluctuations under native conditions. Many lines of evidence are usually interpreted as implying that fluctuations are small excursions away from the native structure. By definition, fluctuations from the native conformation are small increases in free energy. But if protein folding energy landscapes are bumpy, such fluctuations could involve highly nonnative but compact ''misfolded'' structures, even while the excursions in energy are small. Using a model in which we can rigorously study fluctuations and rugged energy landscapes, we ask whether current experimental measures of structure, such as X-ray crystallographic Patterson maps and Debye)Waller . . factors or nuclear magnetic resonance NMR nuclear Overhauser effect NOE spectra, could detect bumpy landscapes. We find that even a substantial population of highly nonnative transients will generally be masked by the heavy averaging implicit in current experiments. This means that, in contrast to current interpretations, very nonnative or Misfold Fluctuations of native proteins may exist but are escaping experimental detection. A positive implication is that structure determination is robust to the presence of much conformational noise. Q 1999 John Wiley & Sons, Inc. Int J Quant Chem 75: 147)164, 1999