Since the conference review on this same subject by Díaz (1989) both the average number of observed and analyzed H II regions per galaxy and the number of galaxies sampled have steadily risen. Also, a larger number of elements is now being studied. The following points seem solid: (1) Global metallicity in spirals is influenced by galaxy mass. Evidence in Vila-Costas & Edmunds (1992) also suggests that interstellar metallicity scales directly with total surface density, perhaps through a star formation rate which is a function of local density, and thus an abundance gradient merely traces matter distribution. (2) N/O behavior with metallicity clearly suggests a primary then secondary origin of N with chemical evolution. (3) Finally, constant ratios of S/O, Ne/O, and Ar/O are ubiquitous and appear to suggest primary nucleosynthesis in massive stars with metallicity-insensitive yields.
While significant progress has been made over the past decade, more questions remain to be addressed. Is the abundance scatter at a specific galactocentric distance real or observational? Why do collisional and recombination lines give abundances which often differ by as much as a factor of two? Do gradients really flatten in outer regions of disks? Finally, there is the problem of measuring good abundances in low excitation H II regions, where auroral line electron temperature determinations are difficult/impossible to obtain, preventing adequate probes of early spirals as well as the inner disks of late ones. Clearly, there is still plenty of work to do.
I am grateful to Dennis Zaritsky for sharing his abundance data, Reggie Dufour for his stimulating comments, and the University of Oklahoma for assisting with travel support.