Helium-3 sits flat at 1.1 × 10⁻⁵ across the Galactic disk, independent of metallicity and radius (Bania, Rood and Balser 2002), while stellar yield calculations demand low-mass stars enrich it steadily over ten billion years. The standard fix, thermohaline mixing destroying 3He in its own factories, must cancel production almost exactly: a primordial relic preserved by accident of opposing parameterized processes, with planetary nebula yields still contradicting the balance.
The model inherits the problem through its chemical evolution machinery: smooth gas cycling through well-mixed stellar generations whose yields are calculated, summed, and required to match the record. When the record refuses to move, the machinery must be retrofitted with a destruction term tuned to net zero, which explains the flatness by canceling it.
SCT keeps the BBN production standard (P40, P42: 3He/H near 1.0 × 10⁻⁵ from the same equilibrium network) and reframes the evolution side. The Galactic disk's gas is not a closed kettle of summed stellar yields: the deposition history (P25, P34) built the disk from thermalized material whose light-isotope baseline is the cascade's, continuously diluted by inflow along the collision-built web, so the interstellar 3He budget is dominated by the primordial baseline reservoir rather than by the net of stellar production and destruction. Flatness across radius and metallicity is then the signature of a baseline-dominated budget, the same structure that keeps deuterium's astration shallow, rather than a knife-edge cancellation inside red giants.
The stellar physics keeps its ordinary role as the secondary modulation: thermohaline mixing is real and bounded by the carbon isotope data, but it no longer carries the impossible duty of exact cancellation, because the reservoir arithmetic, baseline-rich inflow versus processed ejecta, sets the budget at first order. The honest flag: a quantitative SCT inflow-dilution model for the 3He gradient is outstanding work; the structural claim is that the flatness traces reservoir dominance, not tuned destruction.
This is the same baseline-reservoir reading behind the deuterium scatter and the metallicity floors. There is no need to invoke a cancellation calibrated to preserve an accident for ten billion years.
The reservoir reading is differential: 3He flatness should correlate with inflow-dominated regions and weaken where the gas is most stellar-processed and isolated. Resolved measurements finding the flat abundance equally in closed, heavily processed regions, with inflow tracers excluded, would restore the cancellation requirement and remove the baseline account. The shared BBN anchor carries its own kill: a primordial 3He inconsistent with the standard network at more than 2σ strains the cascade-termination timeline exactly as helium-4 would.