H II regions around OB associations have a thick-shell structure (see, e.g., the Carina and Rosette nebulae), and yet the standard “Hot Interstellar Bubble” model (e.g., Weaver et al. 1977) predicts thin H II shells around a large X-ray emitting volume, when associated with stellar winds. Observations suggest that strong dissipation must occur at the edge of the wind cavity: (i) expansion velocities there are much smaller than predicted by the standard model (e.g., Chu, 1983); (ii) in bubbles around WR stars, overabundances of N, He, etc., are seen, hence the need to cool these WR-produced elements down to observable temperatures (Kwitter, 1981). Also, two theoretical developments are important: (i) new stellar evolution models for massive stars, including mass loss and overshooting in convective cores (e.g., Doom, 1985); (ii) a non-linear theory for heat conduction with steep temperature gradients (Luciani et al. 1985).