It is established that noncovalent complexes can be maintained both during and after electrospray and that assemblies of increasing size and complexity often lead to broadened peaks in mass spectra. This broadening arises from the tendency of large protein assemblies to form adducts with salts and is compounded when complexes are isolated directly from cells, without the full protein complement. To investigate the origins of this broadening in mass spectral peaks and to develop the optimal method for analyzing mass spectra of large protein complexes, we have carried out a systematic investigation of a series of noncovalent complexes representing a range of different sizes and architectures. We establish a positive correlation between peak width and the increased mass observed and show that this correlation is independent of the instrumental parameters employed. Using this relationship we show that we can determine masses of both 30S subunits and intact 2.3 MDa 70S ribosomes from Thermus thermophilus. The masses of both particles are consistent with multiple populations of ribosomes. To identify these various populations we combine simulated mass spectra of ribosomes, with and without the full protein complement, and estimate the extent of adducts from our study of known complexes. The results allow us to determine the contribution of the different subpopulations to the overall mass spectrum. We confirm the existence of these subpopulations using tandem mass spectrometry of intact 30S subunits. Overall, the results show that, rather than uniform particles, gas-phase ribosomes consist of a number of discrete populations. More generally, the results establish a rigorous procedure for accurate mass measurement and spectral analysis of heterogeneous macromolecular assemblies.