Why perform time-lapse seismic monitoring? Is it to verify the reservoir model? No! We should conduct time-lapse seismic surveys in order to find out what is incorrect in the reservoir model, in a way similar to the production history matching familiar to reservoir engineers as they look for improvements to the model. This being the case, it is difficult to determine in advance of monitoring just what it is we should be monitoring. Thus, surveys designed specifically to test one feature of a reservoir model may be missing other important features. In this paper, we present a set of very surprising results from the Teal South time-lapse multicomponent (4-D/4-C) study, in Eugene Island Block 354 in the Gulf of Mexico. We will show that time-lapse seismic observations have revealed that an undrilled reservoir near a producing reservoir is exhibiting time-lapse changes consistent with expansion of a free gas phase, and that this implies that oil is being lost through the spill point, never to be recovered, ev...