We analyse spectral images across the full emission line in H_α of the discs of M100 and M101, extracting for each HII region its spectrum, and for the more luminous and extended, spectra over a range of circular apertures. The principal components of these spectra are presented in log-log plots of luminosity vs. turbulent line-width. These plots are apparently scatter diagrams, with no convincing linear fits, e.g. with slope ~ 4 as predicted for a population of virialized systems. Both plots do show, however, clean linear lower envelopes in turbulent width. We interpret these as the loci of HII regions in virial equilibrium. The emissions with greater widths (the great majority), show the effects of not yet fully damped kinematic inputs from stellar winds and supernovae. As expected, a greater fraction of the most luminous regions, which are the most massive, lie on or near the virial envelopes. The slopes of these envelopes are in fact 2.6, not 4, which is explicable if the most luminous regions are density bounded, so that a fraction of their ionizing luminosity, (increasing with luminosity above a critical value) is not registered in H_α. The regions on the envelopes are also those of greatest surface brightness, as predicted for density bounded regions. The masses of the regions on the envelopes, computed via the virial expression, agree well with masses estimated via their emission measures and assuming density bounding.