If I produce two different sets of data with ode113, that are based on the exact same inputs, but one is longer than the other (i.e. tf is larger for one set, we'll call it A. Both A and B have the same ti). If I compare the two plots with imagesc(A) (so that the vertical axis represents the indices themselves, which are the time steps of integration) everything is fine. If I try to append the time vector with: imagesc(xvalues,[0 (max(TA))]m A] for both A and B (and TA and TB respectively), I suddenly have a mismatch on the axes. This indicates to me, that the time vectors are inconsistent, but why? What's the point of them then? For a nonstiff solution (which ode113 is designed for) the difference is small (the same event is at TAe = 2220 and TBe = 2255 but in a different parameter point, where the solution terminates to stability rather quickly (stiff) the difference between TAe and TBe are on the order of thousands.