First observation of a transition between "parallel tracks" in the kHz QPO frequency vs. intensity diagram

Contrary to theoretical expectations, observations with the Rossi X-ray Timing Explorer (RXTE) show that in X-ray binaries timing properties are not uniquely correlated with X-ray luminosity. For instance, although the frequencies of the kilohertz quasi-periodic oscillations (kHz QPOs) correlate with X-ray flux on short (~few hours) time scales, on time scales longer than a day the QPO appears at more or less the same frequency, whereas the luminosity may be a factor of a few different. The result is a set of almost parallel tracks in a QPO frequency vs. X-ray flux plot. Despite the "parallel tracks" are a common phenomenon among kHz QPO sources, until now, after five years of observations with RXTE, not a single transition between two of these tracks had been seen. Here I present the first detection of such a transition, in 4U 1636-53.