Knowledge of the structure of silicate glasses is one key to understanding their chemical and physical properties, as well as those of the high-temperature liquids from which they form. Because first-neighbour ordering into SiO4 tetrahedra is well known, much of the recent discussion about glass structure has concerned intermediate-range order. One of the most fundamental questions that can be posed is the distribution of silicate tetrahedra with n bridging oxygens (shared between two tetrahedra), labelled as 'Qn' (0 less than or equal to n less than or equal to 4). This distribution is probably important in determining the properties of melts and glasses, but considerable controversy exists concerning the identification and quantification of such species. Here I present new evidence from 29Si NMR (nuclear magnetic resonance) spectroscopy that clearly resolves one significant part of this dispute: at least in sodium silicate glasses, a variety of species is present that is greater than that required by composition alone.