
Think of it the other way round, starting from the travel-time curve rather than \(\tau(p)\) directly. If \(v(z)\) increased monotonically everywhere, \(dp/dx\) would never change sign and the \(x(p)\)/\(t(p)\) curves would be single-valued and smooth, so \(\tau(p)\) would trace one unbroken curve. A break in the plotted \(\tau-p\) curve means that over some depth interval \(dv/dz\) must have gone negative — i.e. a low-velocity zone — because that is the only mechanism that removes a range of turning depths (and hence a contiguous range of \(p\)) from the set of rays that actually return to the surface as first arrivals (this range instead becomes a shadow zone, filled only by weaker diffracted/head-wave energy).
Reconstructing the depth profile implied by the two branches of the figure: the upper branch corresponds to rays turning above the LVZ (velocity increasing with depth down to the top of the LVZ), the missing/discontinuous region corresponds to the LVZ (velocity decreasing), and the lower branch corresponds to rays turning below the LVZ, where velocity resumes its normal increase with depth. Hence the depth-velocity trend is increase → decrease → increase, matching option (C), consistent with classical Earth models (e.g., the asthenospheric low-velocity zone beneath the lithosphere) that produce exactly this kind of triplicated \(\tau-p\)/travel-time behaviour.