H II regions are essential tools for the study of the chemical composition and chemical evolution of the Universe, specially in the extragalactic domain. It is necessary to be confident that our traditional techniques for deriving chemical abundances in ionized nebulae (based on the analysis of intensity ratios of collissionally excited lines, CELs) provide the real values.The main aim of this thesis is to derive the abundances of several ions (O++, C++ , O+ and Ne++ ) from the intensity of recombination lines (RLs), which are much more fainter than CELs, although well measured in sufficiently deep spectra. Previous results for bright Galactic and extragalactic H II regions show that abundances obtained from RLs are sistematically larger than those obtained from CELs (factors as high as 2-3). This problem (known as "the abundance discrepancy problem") may be related to the so-called temperature fluctuations suggested to be present in nebulae. In this thesis we present a detailed analysis of spectrophotometric data that have been taken in bright zones of eight Galactic H II regions. The data have been acquired with the Ultraviolet Visual Echelle Spectrograph (UVES) at the Very Large Telescope VLT) Kueyen Telescope in Cerro Paranal Observatory (Chile), during two runs on 2002 and 2003. For each object we have covered the region from 3100 to 10400 angstroms, with an effective spectral resolution R ~8800. We have detected hundreds of emission lines on each region (more than 2600 in the whole sample), which is the most detailed set of spectral emission lines ever obtained for a group of Galactic or extragalactic H II regions. Thanks to the huge amount of spectral information we have derived the physical conditions (electron temperature and density) of the gas from many different diagnostics involving several emission line ratios or continuum to emission line ratios. We have derived chemical abundances from CELs for a large number of ions of different elements. We have derived the abundances of O++ and C++ and, in some cases, the abundances of O+ and Ne++ , from faint RLs. We have obtained remarkable consistent estimations of the t2 parameter (mean square temperature fluctuation) using different methods: a) by comparing electron temperatures obtained from H I Balmer and Paschen discontinuities with those derived from CELs ratios; b) by comparing O++ (and in some cases C++ , O+ or Ne++ ) ionic abundances obtained from optical RLs and CELs and c) from the He I recombination spectra. The mean t2 for each object has been used to derive the ionic abundances in the presence of temperature fluctuations. These data allow us to derive the gas-phase C and O abundance gradients of the Galactic disk. These gradients are of paramount importance for chemical evolution models of the Galactic disk and the solar vicinity. This is the rst time the C gradient is derived from such a large number of H II regions distributed in such a wide range of Galactocentric distances. The Solar vicinity O and C abundances derived from these gradients agree very well with the expected values taking into account the most recent O and C abundances in the Sun and the chemical evolution of the solar vicinity since the Sun was formed. Finally, we have developed a global analysis of the sample. In this study, we have compared dierent temperature and density diagnostics, and have developed a comparative analysis of two different scenarios proposed to explain the abundance discrepancies in planetary nebulae and H II regions: the presence of temperature fluctuations or the presence of chemical inhomogeneities that are colder and denser than the "normal" chemical composition ionized gas. In this thesis we have found that there are signicative dierences among the results obtained in planetary nebulae and H II regions, and that the results found in H II regions are, in principle, consistent with the predictions of the temperature fluctuations scenario.