The Moon is not hollow. Decades of seismic data collected during and after the Apollo program reveal a layered interior with a solid crust, a thick rocky mantle, and a small metallic core, much like Earth’s structure in miniature. The persistent idea that the Moon might be hollow traces back to a real but widely misunderstood observation: when Apollo astronauts crashed spent rocket stages into the lunar surface, the resulting vibrations lasted far longer than anything geologists expected. That eerie, prolonged ringing sounded dramatic enough to fuel conspiracy theories, but the explanation turns out to be grounded in the Moon’s bone-dry, fractured outer layers rather than in any grand emptiness inside.
Why the Moon “Rang Like a Bell”
The story usually starts with Apollo 12 in November 1969. After the crew jettisoned the ascent stage of their lunar module, it struck the surface about 40 miles from the landing site. The seismometer they had installed recorded vibrations that built up gradually, peaked, and then slowly faded over roughly an hour. On Earth, a comparable impact would have produced a sharp signal lasting seconds, maybe a minute. NASA scientists at the time described the Moon as ringing “like a bell,” and the phrase stuck. Subsequent missions repeated the experiment with even larger impacts, crashing upper stages of Saturn V rockets into the Moon, and the pattern held: long, reverberating signals that seemed almost alien.
Taken out of context, that observation sounds like evidence for a hollow shell. If you tap a solid rock, the vibration dies quickly. Tap a hollow metal sphere, and it rings. The analogy is intuitive but misleading, because it ignores the actual physics of how seismic waves travel through the lunar near-surface. The real explanation is far more mundane and far better supported by data.
Seismic Scattering and the Dry, Fractured Crust
On Earth, water saturates cracks in rock and dampens seismic energy. The Moon has no liquid water, no atmosphere, and billions of years of meteorite bombardment have pulverized its outer layers into a thick blanket of broken rock called regolith and, beneath that, a deeper zone of fractured material called megaregolith. Seismic waves entering this dry, shattered material bounce off countless tiny interfaces between fragments. Instead of traveling in a clean pulse, the energy scatters in every direction, arriving at a detector gradually and taking a long time to dissipate. The scattering intensity on the Moon is roughly ten times higher than in the most heterogeneous regions on Earth.1Journal of Geophysical Research: Planets. Quantitative Evaluation of the Lunar Seismic Scattering and Comparison Between the Earth, Mars, and the Moon
High-resolution gravity mapping by the GRAIL mission confirmed the picture. The lunar farside, for example, has a thin surficial layer averaging about 4.7 kilometers deep with a very low density, consistent with highly porous megaregolith created by basin ejecta and impact fragmentation. Below that sits a denser, less porous anorthositic crust.2Journal of Geophysical Research: Planets. Shallow Crustal Structure of the Lunar Farside From High‐Resolution GRAIL Gravity The porosity of that upper layer decreases systematically with depth, meaning the deeper you go, the more intact the rock becomes. This gradient is exactly what you would expect from a solid body battered by impacts, not from a hollow shell.
So the prolonged ringing is really just energy bouncing around in dry rubble. It tells us the Moon’s outer layers are extraordinarily fractured and bone-dry, which is interesting in its own right, but it says nothing about a hollow interior.
What the Crust Actually Looks Like
Apollo missions placed seismometers at multiple landing sites, and those instruments recorded both natural moonquakes and artificial impacts for years. The data allowed geophysicists to build velocity profiles of the crust, identifying how fast seismic waves travel at different depths and using that to infer rock type and structure. At the Apollo 17 site in the Taurus-Littrow valley, for instance, the crust showed a stepwise increase in seismic velocity with depth. The surface basalt flows were about 1,200 meters thick, and the rock immediately beneath them had velocities around 4 kilometers per second, consistent with consolidated ignite bedrock rather than any void.3PubMed. Apollo 17 seismic profiling: probing the lunar crust
Estimating the total thickness of the crust has been an ongoing project. Combining Apollo seismic constraints with orbital gravity data from GRAIL, researchers have narrowed the range considerably. One widely cited analysis using GRAIL placed the average crustal thickness between 34 and 43 kilometers.4PubMed Central. The crust of the Moon as seen by GRAIL A more recent study using seven seismic anchor points and specific mantle-velocity assumptions refined this to about 43.6 kilometers on average.5Journal of Geophysical Research: Planets. Lateral Variations in Lunar Crustal Thickness Inferred From Apollo Seismic and GRAIL Gravity Data The numbers vary somewhat depending on the assumed densities and seismic velocities, but every estimate points to a solid crust tens of kilometers thick, with lateral variations that track the Moon’s geological history: thinner under the nearside mare basins, thicker on the farside highlands.
The composition of that crust further rules out hollowness. The Moon’s primordial crust formed when a global ocean of magma crystallized and lighter minerals, primarily plagioclase feldspar, floated to the top. Experimental simulations of this crystallization process predict a primordial crust of anorthosite and quartz-gabbro that would have been 70 to 95 kilometers thick if all the plagioclase floated. The observed crust is only 35 to 50 kilometers, which researchers interpret as evidence that the interior underwent density-driven overturn and remelting, mixing layers rather than leaving hollow gaps.6Journal of Geophysical Research: Planets. Experimental Crystallization of the Lunar Magma Ocean, Initial Selenotherm and Density Stratification, and Implications for Crust Formation, Overturn and the Bulk Silicate Moon Composition
The Mantle and Core
Below the crust lies a thick silicate mantle that makes up the bulk of the Moon’s volume. Seismic waves from deep moonquakes, combined with gravitational and laser-ranging data, have allowed researchers to model this region with reasonable confidence. One key finding is a deep-seated zone above the core-mantle boundary where seismic wave speeds drop sharply. The S-wave velocity in this zone is roughly 2.9 kilometers per second, and the viscosity is very low, suggesting the material is partially molten. This zone extends more than 170 kilometers above the core.7Geophysical Research Letters. Internal structure of the Moon inferred from Apollo seismic data and selenodetic data from GRAIL and LLR The inferred density of this lowermost mantle suggests it is enriched in titanium oxide, which supports the idea that the Moon’s interior experienced gravitational overturn after the magma ocean solidified: dense, titanium-rich cumulates sank while lighter material rose.
At the very center sits a small metallic core. Reanalysis of Apollo seismic data using modern signal-processing techniques found evidence for both a solid inner core and a fluid outer core, with a partially molten boundary layer between the core and the overlying mantle. The core is roughly 60 percent liquid by volume.8PubMed. Seismic detection of the lunar core The core is small relative to the Moon’s total size, which is consistent with the Moon’s low overall density and weak magnetic field, but it is there. A hollow moon would have no core, no density gradient, and no way to produce the observed seismic wave reflections and refractions at depth.
Moonquakes as Structural Probes
The Apollo seismic network recorded thousands of moonquakes during its years of operation. They fall into several categories: deep moonquakes originating hundreds of kilometers below the surface, shallow moonquakes closer to the crust, thermal quakes caused by the extreme temperature swings between lunar day and night, and impacts from meteoroids. Each type provides different information about the interior.
Deep moonquakes tend to cluster at specific source regions and recur on a tidal cycle, triggered by the gravitational tug of Earth. Shallow moonquakes are rarer but more powerful and more interesting structurally. Recent reanalysis of the Apollo data has identified repeating shallow moonquakes that share a consistent fault-slip direction regardless of the tidal phase in which they occur. This is a meaningful distinction: deep moonquakes are closely tied to tides, but these shallow events are not, implying that tidal stress is not the dominant trigger for their fault-slip initiation. The repeating shallow moonquakes also show a relationship between seismic moment and slip-area size that resembles earthquakes on our own planet.9Geophysical Research Letters. Discovery of Repeating Shallow Moonquakes in the Apollo Lunar Seismic Data
All of these moonquake types are consistent with seismic waves traveling through a solid, layered body. Waves refract at boundaries between layers of different density and rigidity, they reflect off the core-mantle boundary, and they attenuate differently depending on the material’s temperature and composition. If the Moon were hollow, seismic waves from deep moonquakes would have nowhere to originate, since the deep sources sit hundreds of kilometers inside solid rock. And the wave patterns recorded at the surface would look nothing like what was observed.
Real Voids That Actually Exist
While the Moon is emphatically not hollow, it does contain real voids, just not on a scale that anyone would confuse with hollowness. The most exciting of these are lava tubes: tunnel-like cavities left behind when flowing lava drained out from beneath a solidified crust. Because the Moon’s gravitational pull is about one-sixth of Earth’s, lava tubes on the Moon can be far larger than their terrestrial counterparts. Simulations and remote-sensing observations suggest that lunar lava tubes could be hundreds of meters wide and deep.10International Journal of Mining Science and Technology. A comprehensive review of lunar lava tube base construction and field research on a potential Earth test site
Researchers have identified numerous collapse pits on the lunar surface that may provide access to underground caves or tube systems. Several of these pits sit in maria regions, where basaltic lava flows stacked up over time. They expose the layering record of the top 20 to 100 meters of basalt. While only a handful of pits have geological contexts clearly suggestive of lava tubes, many are ambiguous, and some show significant overhangs that hint at present-day cave access beneath.11Journal of Geophysical Research: Planets. Lunar Pit Morphology: Implications for Exploration
Orbital radar data adds another piece to the puzzle. Subsurface echo candidates detected by lunar radar sounders can sometimes be explained by structures 50 to 600 meters wide with dielectric constants consistent with either voids or very high-porosity material (above 35 percent porosity). Some of these echoes align in patterns suggesting lava tubes, while others line up along the extension of surface graben features, suggesting gas voids at the tip of intrusive magma that formed those graben.12MDPI / Remote Sensing. Detection of Small-Scale Subsurface Echoes Using Lunar Radar Sounder and Surface Scattering Simulations with a DEM Generated Using a Generative Adversarial Network These are localized features meters to hundreds of meters across, not continent-sized caverns. They are scientifically fascinating and practically important for future exploration, since a lava tube could serve as natural shelter against radiation and micrometeorites, but they do not make the Moon hollow any more than a few caves make Earth hollow.
Where the Evidence Gets Thin
It is worth acknowledging that our seismic picture of the Moon is still surprisingly incomplete. The Apollo seismic network consisted of only four stations, all clustered on the nearside within a relatively small area. The entire farside, the polar regions, and most of the nearside went unmonitored. That limited geometry means the deep interior is constrained by only a handful of ray paths. The core was not conclusively detected until 2011, more than three decades after the Apollo instruments stopped transmitting, because the original data required modern reprocessing techniques to extract the faint core-reflected signals.
The Apollo stations also operated for a finite period, roughly from 1969 to 1977, before being shut down to save budget. So while the data are rich enough to rule out a hollow moon with high confidence, they leave real uncertainties about fine details: the exact radius of the inner versus outer core, the precise extent and composition of the partially molten layer above the core, the three-dimensional variations in crustal and mantle structure. These are not the kind of unknowns that leave room for a hollow interior. They are the kind that distinguish between, say, a core radius of 330 versus 380 kilometers, or a partially molten zone thickness of 170 versus 250 kilometers. The broad architecture, solid crust over solid mantle over metallic core, is settled.
Planned Lunar Seismic Missions
Several upcoming missions aim to fill these gaps. NASA’s Farside Seismic Suite, for example, was designed to place a broadband seismometer on the lunar farside for the first time, which would dramatically expand the geometry of seismic observations and improve resolution of the deep interior.13IEEE Xplore. The Farside Seismic Suite: A Novel Approach for Long-term Lunar Seismology Other proposals from space agencies and commercial landers envision placing networks of seismometers across the surface, essentially doing for the Moon what global seismic networks did for our understanding of Earth’s interior in the twentieth century.
A farside seismometer would be particularly valuable because it would sit on the opposite side of the Moon from the Apollo stations. Seismic waves from a quake on the nearside would have to travel through the entire body of the Moon, including the core, to reach a farside detector. That kind of whole-Moon ray path is exactly what is needed to pin down core size, state, and composition with much greater precision. The same data would help resolve the properties of the low-velocity zone at the base of the mantle, including whether it truly extends more than 170 kilometers above the core as current models suggest. Better seismic constraints also feed back into understanding of the Moon’s thermal evolution, magnetic history, and the giant-impact event that is thought to have formed it in the first place.
The Mars InSight mission demonstrated that even a single seismometer on another world can reveal far more about its interior than decades of orbital measurements alone. Applying that lesson to the Moon, where we already have a baseline of Apollo data to build on, could transform lunar geophysics from a field of educated estimates into one of precision measurements. The question is no longer whether the Moon is hollow; it is how precisely we can map the solid, layered, geologically interesting world that it actually is.