To produce an accurate model of a gaseous nebula we should take its most relevant properties into account. These properties include: the geometry, the temperature structure, the density structure, the velocity structure, the dust content, the chemical abundances (together with possible chemical inhomogeneities inside a given object), and the energy sources.
Precise models of individual nebulae permit to determine accurate abundances and the abundances permit to test models of stellar evolution, galactic chemical evolution, and the evolution of the Universe as a whole.
To have a good model of a given gaseous nebula a very good knowledge of the temperature structure is needed. The temperature structure is crucial for the determination of accurate chemical abundances. For the best observed objects usually we have a value of the average temperature To and of the mean square temperature fluctuation, t2. When t2 agrees with the value predicted by chemically homogeneous photoionization models we are confident of the derived chemical abundances. Often the observed values of t2 are higher than those predicted by the models and a source for the discrepancy should be sought. Two points should be made here: a) the observational errors present in the t2 determinations are high, but maybe lower than I expect because the overwhelming majority of the observational t2 values present in the literature are positive, b) as Daniel Péquignot mentioned during his talk t2 is just an empirical parameter that should be adjusted by the model, a larger observational value for t2 than that predicted by the model is only telling us that something is wrong with the model, but it is not telling us what is wrong nor which is the temperature structure. This review will be mainly centered on the relevance of the temperature structure in the determination of the chemical composition of different objects, another view of the role of the electron temperature in abundance determinations is presented by Stasinska (2001).