The Earth’s continental crust makes up only about one half of one percent of Earth’s mass, but it contains a large portion of its total budget of many chemical elements such as potassium and phosphorus, which are critical to soil fertility, or uranium and thorium, which produce much of Earth’s interior heat. In making the crust, these elements have been extracted via melts and volcanism from the mantle, the 3000 km thick layer of rocks beneath the crust. Therefore, part of the mantle is now depleted in these important chemical elements. But what portion of the mantle was involved in making the continents? In the past, geochemists thought that only its uppermost third was involved, leaving the lower two-thirds of the mantle essentially untouched. The measure used for this estimate is the difference between crust and mantle in the accumulation of the isotope neodymium-143, the decay product of the radioactive isotope samarium-147. This difference is caused by the crust and the depleted mantle having different samarium/neodymium ratios. This calculation yields a depleted mantle fraction of 30%. However, when we use an alternative measure for the same calculation, namely the concentration ratio of niobium to uranium (Nb/U), which differs by a factor of 8 between depleted mantle and continents, we find the depleted mantle fraction to be greater than 60%. Both of these results cannot simultaneously be correct. We therefore need a more complex Earth model, one that involves (at least) one additional “reservoir” with crust-like Sm/Nd but mantle-like Nb/U. We model this as a primordial mafic crust, which was buried and may now be hidden at the base of the mantle because of its high density. The bottom line is that nearly the entire mantle, not just its uppermost layer, was involved in making our continents. A buried ancient ocean crust might well explain the large density/temperature anomalies recently discovered at the base of the mantle by seismologists.