The experiment in which a candle is burned inside an inverted vessel partially immersed in water has a history of more than 2,200 years, but even nowadays it is common that students and teachers relate the change in volume of the enclosed air to its oxygen content. Contrary to what many people think, Lavoisier concluded that any change in volume in this experiment is negligible; moreover, the explanation relating oxygen consumption in the air with its change in volume is known to be wrong. In this work we briefly review the history behind the candle experiment and its relationship with some typical erroneous explanations. One of the key factors behind Lavoisier's success was the use of experiments carefully designed to test different hypotheses. Following these steps, we performed several closed volume experiments where the candle wick was replaced by a capillary stainless steel cylinder supported and heated by a nichrome filament connected to an external power supply. Our recorded experiments are displayed as web pages, designed with the purpose that the reader can easily visualize and analyze modern versions of Lavoisier's experiments. These experiments clearly show an initial phase of complete combustion, followed by a phase of incomplete combustion with elemental carbon or soot rising to the top of the vessel, and a final phase where the hot artificial wick only evaporates a white steam of wax that cannot ignite because no oxygen is left in the closed atmosphere. After either a complete or incomplete combustion of the oxygen, our experiments show that the final gas volume is nearly equal to the initial air volume.