From model atmosphere analyses of the spectra of B-type stars in young clusters, we have found zero galactic abundance gradients for the solar neighbourhood (i.e. within 3 kpc of the sun). At larger galactocentic distances, we find considerable abundance variations over relatively small distance scales (i.e. 1 kpc). These results are significantly different from those deduced from H II region observations. However our discussion of these discrepancies has been compromised by the lack of data for both the B-type stars and their associated H II regions. We therefore plan to observe with ISO the H II regions directly associated with early-type stars that we have previously analysed viz. S 208, S 212 and S 285 for a spring and the first two only for an autumn launch. The ISO observations will be supplemented with our existing optical data for the H II regions plus new observations, which will be obtained using time recently awarded to us on the 4.2 m William Herschel Telescope. The nebular abundances will be compared with the B-type stellar values, which will allow us to investigate if the discrepancies discussed above are due to: (a) errors in (probably) the H II region analyses, (b) variations in their dust content, and/or (c) the stellar observations identifying abundance variations on a small spatial scale beyond those seen in the H II region results. As H II regions are used extensively to map abundances in external galaxies, it is important that these discrepancies are understood.