The firmest evidence supporting theories1,2 which ascribe the voltage oscillations (narrow-band noise) to phase-slippage at the ends of the sample come from experiments performed with the sample in a thermal gradient3,4. In this short note we will review three classes of experiments which clarify different aspects of the problem. All experiments are done on high purity (RRR > 100) NbSe3 in two-probe configuration, except otherwise indicated. Where appropriate we will stress experimental details and difficulties. See K. Maki, these proceedings, for theoretical details.