Abstract
The determination of the heavy element abundances from giant extragalactic H II regions is based on collisionally excited lines. We argue that in the presence of temperature variations the abundances determined are lower limits to the real heavy element abundances. To determine the real abundances it is necessary to take into account the temperature variations present in these nebulae. We discuss the relevance of obtaining accurate line intensities of recombination lines of H, He, C, and O to determine the chemical composition of extragalactic H II regions. We suggest that Pagel's method to derive the O/H ratio should be calibrated by using recombination lines instead of photoionization models or abundances derived from collisionally excited lines.
