The need for tools capable of handling non-stationarities in the spectral content of the data was recognized as early as 1946. The Wigner-Ville Distribution (WD) has been extensively used since its introduction in 1948, but suffers from some associated problems (e.g., spectral cross-terms and requiring the use of analytic signals). An alternative Distribution is proposed, which has its origin in the definition proposed by Page of 'Instantaneous Power Spectrum' (IPS). Its characteristics are examined and, when pertinent, compared to the WD. It is shown to be less sensitive to the problems afflicting the WD, but provides less frequency resolution. The usefulness of a parametric (AR) version is investigated. Some typical test signals are examined, to demonstrate the performance and trade-offs of IPS and its parametric version.