Does mantle convection operate as a single whole-mantle system, or as two separate layers above and below the mantle transition zone, and do deep, geochemically distinct reservoirs survive within it?

Seismic tomography and mantle convection modeling now favor whole-mantle flow, but geochemical evidence for isotopically distinct deep reservoirs, drawn from differences between mid-ocean-ridge and ocean-island basalts, is difficult to reconcile with a fully mixed, single-layer mantle. The geophysics and geochemistry communities have not settled which picture is correct.