We will use ISO to observe IR emission lines of [S III], [S IV],
 [Ne II], [Ne III], and Brackett-alpha in H II regions in two spiral
 galaxies and three irregular galaxies. The data will be used to determine
 the variation of the abundances of Sulfur and Neon over the metallicity
 range 0.1 to 2.0 times solar, and to compare the variation of these
 elements with that of Oxygen. Because of the weak temperature dependence
 of the IR lines, these measurements provide critical measurements of
 abundances in metal-rich H II regions, where electron temperatures are
 poorly known. Comparison of the IR lines of [S III] and [Ne III]
 with their optical counterparts will provide a measure of electron
 temperatures in the H II regions and of the amplitude of temperature
 inhomogeneities in the nebulae, with important consequences for
 elemental abundances inferred from H II region spectra.

 There is considerable debate whether Sulfur abundance gradients in
 spiral galaxies are shallower than Oxygen abundance gradients. If
 Sulfur gradients are indeed shallower than for Oxygen, it would imply
 either that the massive star IMF varies with metallicity or that the
 yield of S from massive stars varies with metallicity. We propose to
 measure the [S III] 18.7 and 33.6 micron lines, plus the [S IV]
 10.5 micron line, to determine Sulfur abundances in H II regions,
 providing an independent test of the trend for Sulfur. Neon is
 generally considered to vary in lockstep with Oxygen. However, a
 recent study of H II galaxies found Ne++/O++ increasing with metal-
 licity. This result was attributed to an incorrect ionization
 correction for Ne++, but could be suspect because of possible bias
 in the analysis. We propose to measure [Ne II] 12.8 microns and
 [Ne III] 15.6 microns in our H II region sample to determine
 independently Neon abundances and to test the validity of commonly-
 used ionization correction schemes for Neon.