Beneath our feet lies a realm we will almost certainly never visit. The mantle — that vast shell of rock stretching from roughly 30 to 2,900 kilometres deep — makes up about 84 percent of Earth's volume. Yet everything we know about its chemistry comes from indirect evidence: seismic waves, laboratory experiments, and a handful of rocks that made the improbable journey to the surface.

Those rocks are ultramafic — dense, iron- and magnesium-rich assemblages dominated by the mineral olivine. Peridotites, the most common variety, are our closest thing to a hand sample of the deep Earth. They arrive as xenoliths ripped from the mantle wall by ascending magmas, or as massive slabs thrust onto continents during tectonic collisions.

Reading these rocks demands careful geochemical detective work. Their mineral proportions, trace-element signatures, and isotopic ratios preserve a layered history of melting, depletion, and chemical modification stretching back billions of years. Each ultramafic specimen is an archive of mantle processes — if you know how to decode it.

The most dramatic way mantle rock reaches the surface is inside a volcanic eruption. Certain magma types — particularly kimberlites and alkali basalts — ascend so rapidly from depth that they tear chunks of surrounding mantle wall rock and carry them upward. These fragments, called xenoliths, can range from fist-sized nodules to blocks weighing several tonnes. When geologists crack them open, they find assemblages of olivine, orthopyroxene, clinopyroxene, and an aluminium-bearing phase such as garnet or spinel — the classic mineralogy of peridotite.

What makes xenoliths extraordinary is their directness. Seismic tomography can tell us that the mantle transmits waves at certain velocities. High-pressure experiments can recreate mantle conditions in a lab. But a xenolith is an actual piece of the mantle, preserving its mineral chemistry at the pressure and temperature conditions where it last equilibrated. By analysing the partitioning of elements like calcium and aluminium between coexisting minerals, petrologists can estimate the depth and temperature from which each xenolith originated — a technique known as geothermobarometry.

Different volcanic settings sample different depths. Alkali basalts typically bring up spinel peridotites from the uppermost 60–80 kilometres. Kimberlites, which erupt explosively from depths exceeding 150 kilometres, deliver garnet peridotites that record conditions well within the lithospheric mantle and sometimes the asthenosphere. This depth stratification lets researchers build vertical profiles of mantle composition beneath specific regions — a kind of geological core sample assembled from scattered eruptions.

Xenolith suites from cratons — the ancient, stable cores of continents — have proven especially revealing. Peridotites recovered from beneath the Kaapvaal craton in southern Africa and the Slave craton in Canada show that these deep lithospheric roots are chemically distinct: strongly depleted in easily melted components, yet cold and buoyant enough to persist for three billion years or more. Without xenoliths, this picture of ancient, refractory keels stabilising the oldest continents would remain purely theoretical.

Takeaway

Xenoliths are the mantle's own dispatches to the surface — rare, direct samples that anchor our models of deep-Earth composition in physical reality rather than inference alone.

When mantle peridotite partially melts, the resulting liquid is enriched in certain elements — silicon, aluminium, calcium, sodium — while the solid residue becomes progressively depleted. The melt rises to form basaltic crust. What stays behind is a drier, more magnesian rock with less clinopyroxene and a higher ratio of olivine and orthopyroxene. By quantifying what's missing from an ultramafic rock, geochemists can reconstruct how much melt was extracted and under what conditions.

The key diagnostic tools are major-element chemistry and modal mineralogy. A fertile, unmelted mantle peridotite — sometimes called pyrolite in theoretical models — contains roughly 55–60 percent olivine, 25 percent orthopyroxene, 10–15 percent clinopyroxene, and a few percent of an aluminous phase. As melting proceeds, clinopyroxene is consumed first. A peridotite with almost no clinopyroxene has lost perhaps 20–25 percent of its mass as basaltic melt. These are called harzburgites, and they dominate the shallow oceanic mantle beneath mid-ocean ridges where melting is extensive.

Trace elements add further resolution. Elements like titanium, zirconium, and the rare earth elements are incompatible — they prefer liquid over solid and are efficiently stripped out during melting. A strongly depleted harzburgite will have vanishingly low concentrations of light rare earth elements compared to a fertile lherzolite. Plotting these depletion trends against experimental melting curves allows researchers to estimate not just the degree of melting, but the pressure range over which it occurred — crucial for understanding whether melting happened at a mid-ocean ridge, a subduction zone, or a mantle plume.

One of the most striking findings from this work is that the mantle is heterogeneous. Some regions have experienced multiple episodes of melt extraction over billions of years, leaving behind ultra-refractory residues. Others remain relatively fertile. This patchwork of depletion is not random — it maps onto tectonic history. The mantle beneath old cratons is severely depleted, while mantle beneath young ocean basins retains much of its original fertility. The rocks themselves carry a memory of every melting event they've survived.

Takeaway

A depleted peridotite is not a rock defined by what it contains, but by what has been taken from it — and reading those absences reveals the cumulative melting history of the mantle beneath our feet.

If partial melting strips the mantle of its easily fusible components, metasomatism is the process that puts some of them back. Fluids and small-fraction melts migrating through the mantle can infiltrate previously depleted peridotite, introducing elements that had been extracted long ago. The result is a rock whose bulk chemistry tells a more complicated story than simple depletion — one overprinted by later chemical enrichment.

Two styles of metasomatism are widely recognised. Modal metasomatism introduces entirely new minerals — amphibole, phlogopite, apatite, or exotic phases like titanate minerals — that would not normally crystallise in dry, depleted peridotite. Their presence is a clear flag that volatile-rich fluids have passed through. Cryptic metasomatism is subtler: the original mineral assemblage looks unchanged, but trace-element and isotopic analyses reveal enrichment in light rare earth elements, strontium, or other incompatible species that have been re-injected into existing mineral grains.

The sources of metasomatic agents are varied. Subduction zones deliver water-rich fluids and sediment-derived melts deep into the mantle wedge. Mantle plumes can generate small-fraction alkaline melts that percolate through the lithosphere. Even carbonatitic liquids — ultra-low-viscosity melts rich in carbon dioxide — have been implicated as potent metasomatic agents capable of travelling long distances through mantle rock along grain boundaries.

Why does this matter beyond academic interest? Because metasomatised mantle is more easily melted than its depleted counterpart. The introduction of water and other volatiles lowers the solidus temperature, meaning that regions of the mantle that should be too refractory to generate magma can become sources of volcanism once they've been chemically re-fertilised. This mechanism helps explain enigmatic volcanic activity far from plate boundaries — and it means that understanding a single peridotite's metasomatic history can illuminate the conditions that trigger magma generation hundreds of millions of years later.

Takeaway

Metasomatism reminds us that the mantle is not a static reservoir but a chemically dynamic system — depleted rocks can be re-enriched, and that re-enrichment can reset the conditions for future volcanism.

Ultramafic rocks are, in the most literal sense, messengers from a world we cannot reach. Each peridotite xenolith, each ophiolite slab, each depleted harzburgite records a chapter of mantle history — from primordial composition through episodes of melting and chemical modification.

The story they tell is one of a dynamic interior. The mantle is not uniform, not static, and not simple. It carries the chemical fingerprints of billions of years of tectonic recycling, melt extraction, and fluid infiltration — a palimpsest of planetary processes written in olivine and pyroxene.

Every time a kimberlite erupts or an ocean floor is obducted onto a continent, another page of that record becomes available. The challenge — and the reward — lies in reading it carefully.