Abstract

In the course of 1901, V. Crémieu published the results of some experiments carried out to test the magnetic effects of electric convection currents. According to Crémieu, his experiments had proved that convection currents had no magnetic effects and consequently they were not equivalent to conduction currents, that is they were not ‘real’ electric currents. These negative results conflicted with those of well-known experiments carried out by other researchers, in particular with Rowland's experiments, and with Maxwell's, Hertz's and Lorentz's theories, which was more shocking. The publication of Crémieu's experiments raised a controversy which involved directly or indirectly some leading French and British physicists, and led Rowland together with his student Pender to re-perform his experiments. Pender's results confirmed those preiously obtained by Rowland, in striking conflict with those of Crémieu. The importance of Crémieu's experiments was stressed by Poincaré in some of his publications, where he analysed and criticized various aspects of the question. However, he was involved more deeply: his correspondence with Crémieu shows that many of the experiments performed by Crémieu were conceived by Poincaré and carried out with his guidance. To put an end to the controversy, Poincaré obtained Kelvin's support to organize Crémieu's and Pender's joint experiments. The results of these experiments showed that the conflicting results obtained previously were due to an ‘anomalous’ behaviour of the dielectrics used by Crémieu in his apparatus