An excellent data set has been gathered for the November 24, 1996, Geospace Environment Modeling substorm interval that included an ~95 min interval of strongly southward interplanetary magnetic field. There were two expansion phase onsets and a pseudobreakup during this period. For both onsets the classical signatures of onset in the auroral ionosphere (ground magnetometer signatures of electrojet formation, ground observations of Pi2 pulsations, and auroral brightening) all occurred within less than a minute of each other, indicating consistency between these onset indicators and giving reliable identification of onset times. On the other hand, low-latitude positive bay observations that were most likely made near or within the longitude range of substorm onset did not provide accurate or consistent onset timing. Low-latitude Pi2s identified both onsets, though one was delayed by 1 min from the auroral zone onset. Our most important and unexpected result is that particle injection at synchronous orbit was observed to initiate ~2.5 min prior to both expansion onsets in the auroral ionosphere. We suggest that the early detection of substorm onset at synchronous orbit was at least in part due to the unusually low latitude (63¿ magnetic) of the auroral zone onsets, which maps closer to synchronous orbit than usual. Onsets more often occur at somewhat higher latitudes, which map to a few RE beyond synchronous orbit. As a result, particle injections at synchronous orbit are generally delayed with respect to the time of current wedge initiation within the plasma sheet. We do not currently know whether or not onset within the near-Earth plasma sheet generally occurs ~2.5 min prior to expansion onset in the auroral ionosphere. While we have no reason to believe that the timing we have observed is unique, other studies of low-latitude substorm onsets will be needed to ascertain the generality of this result. ¿ 2001 American Geophysical Union