For Francis turbines, outlet cavitation is a major critical phenomenon in determining the prototype setting level. To predict the full-scale cavitation characteristics, it is technically interesting to provide reliable model cavitation tests. This can be done by a similar cavitation development on the model and on the prototype. In this paper, previous results obtained during systematic cavitation tests using nuclei injection and counting techniques are recalled. An alternative procedure to nuclei injection is then investigated with a high gas content in the water. The unique feature of the IMHEF test facility to perform cavitation tests either with a closed loop (low dissolved air content) or with an open loop (high dissolved air content), whatever the nuclei content, means that the difference between the influences of the dissolved gas content and of the cavitation nuclei content can be studied, independently. Moreover, the determination of the water active nuclei content leads to predicting the cavitation inception, provided the pressure distribution in the runner can be estimated. Based on the results of all our investigations guidelines for performing cavitation tests on Francis model are suggested.