We present multifrequency VLA continuum and spectral line observations of the three H II regions G31.28 + 0.06, G33.92 + 0.11, and G34.25 + 0.14. Our observations emphasize the importance of resolution effects in any scheme of classifying H II regions on the basis of observed morphology. The continuum images of G31.28 + 0.06 at 20 and 3.6 cm show a cometary morphology, whereas the combination of resolution and sensitivity of the 6 cm image shows that G31.28 + 0.06 is actually a shell source with the continuum emission from one part of the shell dominating the emission from the rest of the shell. Hydroxyl and water masers are projected off the edge of the brightest part of the shell. The high resolution (5".5) 20 cm continuum emission from G33.92 + 0.11 presented here shows a shell source whose continuum emission is dominated by one side of the shell. This H II region was originally classified as a core-halo source by Wood & Churchwell [ApJS, 69, 831 (1989)] and appears to have cometary morphology in our 6 and 3.6 cm images. Our 20 cm image of the G34.3 + 0.2 region details the continuum emission from two H II regions, G34.25 + 0.14, and the well studied "cometary" H II region G34.26 + 0.15. Our image of G34.25 + 0.14 shows a ring of emission of diameter approximately 1'. This structure appears to be a shell which is 75% complete. While the source is too large to be considered an ultracompact H II region, the morphology of the emission has important ramifications for models of the nearby cometary H II region. Both G34.25 + 0.14 and G34.26 + 0.15 are embedded in a 1'-2' NH3 core whose temperature and density peak at a point between the two H II regions. The brightest parts of both G34.25 + 0.14 and G34.26 + 0.15 are in the direction of the ultracompact molecular core. This argues against a bow shock interpretation for the H II regions since the bow shock model, applied to these H II regions, would require the unlikely circumstance that both regions are moving into a preexisting ultracompact molecular core. We suggest that the observed properties of both regions can be explained by expansion in an anisotropic medium. Density gradients in the ambient molecular cloud would slow the expansion of the H II regions in the direction of highest density.
