Interior of the Moon R. C. Weber Abstract A variety of geophysical measurements made from Earth, from spacecraft in orbit around the Moon, and by astronauts on the lunar surface allow us to probe beyond the lunar surface to learn about its interior. Similarly to the Earth, the Moon is thought to consist of a distinct crust, mantle, and core. The crust is globally asymmetric in thickness, the mantle is largely homogeneous, and the core is probably layered, with evidence for molten material. This chapter will review a range of methods used to infer the Moon’s internal structure, and briefly discuss the implications for the Moon’s formation and evolution. Key Words: Moon, lunar interior, lunar formation, lunar geophysics, core structure 1. Introduction Understanding the internal structure of the Moon is of key importance to deciphering its early history. The current consensus is that the Moon formed following the collision of a Mars-sized body with the Earth about 4.5 billion years ago. The rocky mantle of the impactor spun out to form the Moon, while the core of the impactor fell into the growing Earth. This model explains the high spin of the Earth–Moon system, the low density of the Moon relative to the Earth, and the Moon’s depletion of lighter volatile elements, which were likely ejected during the impact event. The model also provides a source of energy to melt the early Moon. The geochemical and petrological evidence strongly supports the theory that the molten Moon floated an anorthositic crust about 4.45 billion years ago. This forms the present high-albedo highland crust. As the initially hot, molten Moon cooled, the mantle likely crystallized into a sequence of mineral zones by about 4.4 billion years ago. Heavier elements sank to form a small metallic core. Following the formation of the crust, major impacts on the surface produced many craters and multi-ring basins, probably during a spike or “cataclysm” around 3.9 – 4.0 billion years ago. The oldest basin observed is the South Pole – Aitken Basin and the youngest is the Orientale Basin, which formed 3.85 billion years ago. Beginning about 4.3 billion years ago, and peaking between 3.8 and 3.2 billion years ago, partial melting occurred in the lunar interior, and basaltic lavas flooded the low-lying basins on the surface. This occurred mostly on the nearside, where the crust is thinner, resulting in the low- albedo lunar mare. Major volcanic activity ceased around 3.0 billion years ago, although minor activity may have continued until 1.0 – 1.3 billion years ago. The Moon has suffered only a few major impacts since that time (forming, for example, the young rayed craters such as Copernicus and Tycho). An overview of the commonly-accepted model of the Moon’s present day internal structure, illustrating the crust, mantle, and core layers, is shown in Figure 1, and includes the probable depths of the interfaces between the core layers. 2. Bulk lunar properties The mass of the Moon, determined from the orbital periods of various spacecraft using Kepler’s third law, is 7.35 × 1022 kg, which is 1/81 of the mass of the Earth. Although the Galilean satellites of Jupiter (Io, Europa, Ganymede, and Callisto), and Saturn’s moon Titan are comparable in mass, the Moon/Earth ratio is the largest satellite-to-parent mass ratio in the solar system (Table 1). The lunar radius is 1738 ± 0.1 km, or 27% of the Earth’s radius. This radius is intermediate between that of Europa (radius = 1561 km) and Io (radius = 1818 km). The Moon is much smaller than Ganymede (radius = 2634 km), which is the largest satellite in the solar system. The lunar mean density is 3.344 ± 0.003 g/cm3, while the Earth has a much higher mean density of 5.52 g/cm3. The lunar density is also intermediate between that of Europa (density = 3.014 g/cm3) and Io (density = 3.529 g/cm3). Most of the other known satellites in the solar system are ice-rock mixtures and so are much less dense. The Moon’s moments of inertia (MOI) are related to its second-degree gravitational harmonics, which have been measured to high precision by orbiting spacecraft. Current estimates indicate that the Moon’s mean MOI is 0.3931 ± 0.0002, very close to that of a uniform density sphere, which has a moment of inertia of 0.4. This requires a slight density increase toward the center of the Moon, in addition to the presence of a low-density crust. In comparison, the mean MOI value for the Earth, with its dense metallic core that constitutes 32.5% of the Earth’s total mass, is 0.3315. The mass of the Moon is distributed in a nonsymmetrical manner, with the center of mass (CM) lying 1.8 km closer to the Earth than the geometrical center of figure (CF) (Figure 2). This offset is due to the presence of a thicker farside crust. This is a major factor in placing the Moon into synchronous orbit with the Earth, such that the Moon always presents the same face to the Earth. The gravitational influence of the Earth (and to a lesser extent, the Sun) on the Moon’s asymmetric mass distribution resulted in torques that slowed down the rotation of the early Moon, until it became tidally locked. However, the lunar longitudinal and latitudinal librations in combination allow a total of 57% of the Moon’s surface to be visible at different times in the orbital cycle. Various explanations have been advanced to account for the offset of the Moon’s center of mass from its center of figure. Dense mare basalts erupted from the lunar interior cover about 17% of the lunar surface, mostly on the nearside, but they are usually less than 1 or 2 km thick and constitute only about 1% of the total volume of the crust – insufficient by about an order of magnitude to account for the effect. It has also been suggested that the offset could arise if the lunar core is displaced from the center of mass. However, such a displacement would generate shear stresses that could not be supported by the hot, likely molten (or partially molten) deep interior. Another suggestion is that some form of density asymmetry developed in the mantle during crystallization of the magma ocean, with a greater thickness of lower density materials being concentrated within the farside mantle. However, it is unlikely that such density irregularities would survive stress relaxation in the hot interior, unless actively maintained by convection (for which there is no present-day evidence). The conventional explanation for the CM/CF offset is that the farside highland low-density crust is thicker, probably a consequence of an asymmetry developed during crystallization of the magma ocean. This explanation is supported by crustal thickness estimates derived from gravity mapping (see Section 3.2). The crust is massive enough and sufficiently irregular in thickness to account for the CM/CF offset. An equipotential surface is closer to the actual surface on the nearside. Magmas that originate at equal depths below the surface will thus have greater difficulty in reaching the surface on the farside, where the crust is thicker. This explains the scarcity of observed mare basalts on the farside. Lavas rise owing to the relative low density of the melt and do not possess sufficient hydrostatic head to reach the surface on the farside, except in craters in some very deep basins. 3. Methods used to probe the lunar interior The Apollo Lunar Surface Experiments Package (ALSEP), deployed across the lunar surface by the astronauts on Apollo missions 12, 14, 15, 16, and 17, gathered much data relevant to the lunar interior. Each ALSEP installation consisted of a set of geophysical instruments connected to a central base station. The base station acted as the command center for the entire package. It received commands and transmitted data to and from Earth, and distributed power to each experiment. The astronauts also gathered a wide collection of samples from the lunar surface that were returned to Earth for analysis. Instruments both onboard spacecraft in lunar orbit and Earth-bound also gather measurements that are useful for deciphering the Moon’s internal structure. These include gravity and magnetic field data measured from orbit and laser ranges originating from Earth. This section will review results of the active and passive seismic experiments and the heat flow experiment from ALSEP, analyses performed on samples gathered from the surface and shallow subsurface, and a variety of orbital and Earth-based measurements, and discuss interpretations of the lunar interior made from these data. 3.1 Apollo core samples The near-surface structure of the Moon was revealed by core samples taken by the Apollo astronauts (Figure 3). Core tubes were either 2 or 4 centimeters in diameter and were pounded into the surface with a hammer. The deepest core was nearly 3m at the Apollo 17 landing site, and a total of 24 cores were collected over all six Apollo surface sites. These cores revealed that the shallowest lunar layer, known as the regolith, is a complex array of overlapping ejecta blankets resulting from meteor bombardment on the lunar surface throughout the Moon’s history. This process is known as impact gardening, and results in a shallow layer of particles of varied size and texture (see section 4.1). 3.2 Gravity measurements The Moon’s internal structure can also be inferred through analyses of the lunar gravity field as measured from orbit (Figure 4). Variations in surface gravity across the Moon are caused by density heterogeneity in the subsurface, and these variations affect the position of orbiting spacecraft. First noticed during analysis of tracking data from NASA’s Lunar Orbiter program in the 1960’s, the Moon’s gravity field has been mapped in successively higher resolution by missions such as NASA’s Lunar Prospector in the 1990’s, the Japanese space agency’s SELENE orbiter (Selenological and Engineering Explorer) in the 2000’s, and NASA’s GRAIL mission (Gravity Recovery and Interior Laboratory) in the 2010’s. The GRAIL mission mapped the Moon’s gravity in unprecedented detail, resulting in the highest resolution gravity map of any body in the solar system, including Earth. The biggest features resolved in the lunar gravity field are known as mascons, or mass concentrations. They are associated with giant impact basins and are caused by the uplift of a central plug of dense mantle material during impact, followed by the much later addition of dense mare basalt. Smaller shallow features are also resolved in the gravity data, including tectonic structures, volcanic landforms, basin rings, complex crater central peaks and simple bowl-shaped craters. Young ray craters have negative gravity anomalies because of the mass deficit associated with excavation of the crater, combined with the low density of the fallback rubble. Craters less than 200 km in diameter have negative gravity anomalies for the same reason (e.g., Sinus Iridum has a negative anomaly of −90 mGal). Volcanic domes such as the Marius hills have positive anomalies (+65 mGal), indicating support by a rigid lithosphere. The gravity signature of young, large, ringed basins, such as Mare Orientale, shows a “bull’s-eye” pattern with a central positive anomaly (+200 mGal) surrounded by a ring of negative anomalies (−100 mGal) with an outer positive anomaly collar (+30 to +50 mGal). In combination with topography data, the gravity field can also be used to infer the depth of the crust-mantle interface (known as the moho). The moho deflects in response to surface loads, and the resulting flexural signature contributes to the observed gravitational field. Crustal thickness largely correlates with topography, with the exception of the lunar mare regions. These areas of low elevation were resurfaced by high-density basaltic lava flows, resulting in more complex flexural signals. The average density of the highland crust calculated from GRAIL-derived crustal thickness estimates is 2250 kg/m3. The lunar highland crust is strong. High mountains such as the Apennines (7 km high), formed during the Imbrium collision 3.85 billion years ago, are uncompensated and are supported by a strong cool interior. The gravity data are consistent with an initially molten Moon that cooled quickly and became rigid enough to support loads such as the circular mountainous rings around the large, younger, ringed basins as well as the mascons. Even if some farside lunar basins do not show mascons, this may merely be a consequence of the greater thickness of the farside crust. The South Pole–Aitken Basin (the largest and oldest basin, age at least 4.1 billion years) is particularly significant in this respect. 3.3 Laser ranging Additional information about the interior of the Moon can also be inferred from data gathered by the ongoing Lunar Laser Ranging (LLR) Experiment. This experiment consists of Earth-based laser ranges to an array of retroreflectors emplaced on the lunar surface 30 years ago by both U.S. and Russian missions. A laser pulse is fired from the Earth to the Moon, where it bounces off a retroreflector and returns back to Earth (Figure 5). The round-trip travel time can be used to measure the Moon’s shape and position with accuracy better than 2 centimeters. The analysis of LLR data provides a wealth of information concerning the dynamics and internal structure of the Moon. The distances between the retroreflectors and the Earth change in part because of lunar rotation (physical librations) and tides. Values of the gravitational harmonics, the moments of inertia, the lunar Love number k (which measures the tidal change in the Moon’s moments of inertia and 2 gravity), and variations in the lunar physical librations are related to the Moon's composition, mass distribution, and internal dynamics. A range of internal structure models is compatible with the MOI values constrained by LLR. For example, a 60-km-thick lunar crust with density of 2.75 g/cm3, a constant-density lunar upper mantle, a lower mantle with a similar change in density relative to the upper mantle, and a variable-radius iron core with density of 7 g/cm3 produces an appropriate MOI. In this case the maximum core size is in the range of 220 to 350 km, and an increase in crustal density to 2.959 g/cm3 raises the maximum core size to 400 km, consistent with other estimates (see Section 4.4). All layers can be adjusted in thickness and density to produce a suite of plausible lunar structure models. For a perfectly rigid Moon, the mean direction of the lunar spin axis would be expected to precess with the Earth-Moon orbit plane. The LLR data show, however, that the true spin axis of the Moon is displaced from the expected direction. This is the result of ongoing active dissipation in the lunar interior, which has been proposed to be due in part to friction at the interface between the solid lower mantle and a fluid core. 3.4 Magnetic techniques At present, the Moon does not possess an internally generated magnetic field. However, samples returned by the Apollo astronauts from the lunar surface retain natural remanent magnetism. In addition, orbital estimates of surface magnetic field strength reveal regions of increased magnetic intensity (Figure 6), albeit with field strengths of only about 1/100th of the terrestrial field. Based on the age dating of Apollo samples, this magnetic signature suggests that between about 3.6 and 3.9 billion years ago, there was a planetary-wide magnetic field that has now vanished. The field appears to have been much weaker both before and after this period. Although taken from the lunar surface, these observed present-day remanent magnetic anomalies are relevant to the lunar internal structure since one interpretation is that the Moon once possessed a lunar dipole field of internal origin. The favored mechanism is that the field was produced by dynamo action in a liquid iron core, similarly to the way Earth’s magnetic field is generated. A core about 400 km in diameter could produce a field at the lunar surface with strength comparable to the observations. In early lunar history, this magnetization would be impressed into the cooling lunar crust. An alternative interpretation however suggests that the magnetic signature may not be internally generated in origin but rather results from shock magnetization in transient fields generated following the basin-forming impacts in lunar history. This theory is supported by the observation that the largest crustal magnetizations appear to be located at or near the antipodes of the largest impact basins. In addition, some localized strong magnetic anomalies are associated with patterns of swirls – high albedo features that impart no observable topography. These swirls have likewise been suggested to form by some focusing effect of the seismic waves that resulted from the large basin-forming impacts. More work is clearly needed to substantiate this hypothesis and to understand the association of swirls and magnetic fields. The internal structure of the Moon can also be inferred by measuring the lunar induced magnetic dipole moment. This is the residual response of the lunar interior to the sudden exposure of the Moon to a uniform magnetic field in a near-vacuum environment, which happens every month as the Moon passes through the Earth’s geomagnetic tail. The external field is perturbed by an induced magnetic field caused by currents at the surface of a highly electrically conducting (iron) core, and these perturbations can be measured by orbiting spacecraft. Data gathered by the Lunar Prospector magnetometer were analyzed to conclude that the Moon likely does possess an iron- rich core, with a preferred radius of 340 ± 90km. 3.5 Heat flow The rate at which a planetary body loses heat to space is an important indicator of the level of tectonism and volcanic activity on said planet. Two measurements of the lunar heat flow are available, as measured by the ALSEP’s Heat Flow Experiment during the Apollo 15 and Apollo 17 mission’s surface operations. The Heat Flow Experiment involved drilling a hole into the lunar regolith and inserting a probe that measured temperature at several depths within the hole. The rate at which temperature increases with depth provides a measure of the total heat flowing from the Moon’s interior: 2.1 μW/cm2 at the Apollo 15 site and 1.6 μW/cm2 at the Apollo 17 site, respectively. These surface heat flow measurements are close to Earth-based estimates from microwave observations. Unlike the Earth, which dissipates most of its heat by convective volcanism at the mid-ocean ridges, the Moon transports its heat to the surface by conduction. A lack of observed present-day active volcanism or tectonism on the Moon indicates that most of its original internal heat has been lost, so any observed heat flow must be instead predominantly due to the radioactive decay of heat-producing elements, with a small percent of the total heat flow consisting of the loss of residual heat from lunar formation. If the Apollo heat flow measurements are considered to represent the average heat loss characteristic of the entire Moon, they can be used to provide constraints on the bulk lunar abundances of elements that release heat through radioactive decay. The heat-producing elements K, U, and Th were concentrated near the surface by differentiation during lunar formation. However the constraints on these abundances are only mild, as the distribution of heat-producing elements is not symmetric across the lunar surface. The heat flow measurements made by Apollo could indicate bulk lunar uranium values as high as 45 ppb, over twice the terrestrial abundances. A more likely scenario is that uranium and other heat-producing elements are concentrated in the lunar crust. This is a consequence of magma crystallization. Potassium (K), rare earth elements (REE), and phosphorus (P) (KREEP), along with thorium and uranium, are among the last trace elements to crystallize from a melt. As the early molten Moon cooled, various minerals crystallized from the melt. Heavy olivines sank to the bottom, while lighter anorthosites floated to the top. The remaining incompatible trace elements probably remained molten for a much longer period of time and were eventually exposed to the surface through impact processes. The near-side concentration of KREEP may also help explain the asymmetric distribution of lunar mare. 3.6 Compositional studies The Moon is dry, with no indigenous water having been detected at ppb levels, and lacks ferric iron (as determined by both orbital measurements and sample analyses). It is strongly depleted of volatile elements (e.g., K, Pb, Bi) by a factor of about 50 compared to the Earth, or 200 relative to primordial solar nebula abundances. Compared to the Earth, the most striking difference is in the abundance of iron that is reflected in the low lunar density. The Earth contains about 25% metallic Fe; the Moon, less than about 2–3%. However, the bulk Moon contains between 12 – 13% FeO, or 50% more than current estimates of 8% FeO in the terrestrial mantle. Along with its depletion in iron, the Moon also has a low abundance of siderophile or “metal-seeking” elements. These elements are extracted into metallic phases according to their metal/silicate partition coefficients during accretion. The lunar depletion of these elements has been used to argue that they have likely been segregated into a metallic core. The other major element abundances are mostly model-dependent. Si/Mg ratios are commonly assumed to be chondritic (CI), although the Earth and many meteorite classes differ from this value. The lunar Mg value is generally estimated to be about 0.80, lower than that of the terrestrial mantle value of 0.89. The Moon is probably enriched in refractory elements such as Ti, U, Al, and Ca, a conclusion consistent with geophysical studies of the lunar interior. This conclusion is reinforced by the data from the Galileo, Clementine, and Lunar Prospector missions, which indicate that the highland crust is dominated by anorthositic rocks. This requires that the bulk lunar composition contains about 5–6% Al2O3, compared with a value of about 3.6% for the terrestrial mantle and so is probably enriched in refractory elements (e.g., Ca, Al, Ti, U) by a factor of about 1.5 compared to the Earth. Both the Cr and O isotopic compositions are identical in the Earth and Moon, probably indicating an origin in the same part of the nebula, consistent with the single impact hypothesis that derives most of the Moon from the silicate mantle of the impactor. The Moon has a composition that is unlikely to have been made by any single-stage process from the material of the primordial solar nebula. The compositional differences from that of the primitive solar nebula, from the Earth, from Phobos and Deimos (almost certainly of carbonaceous chondritic composition), and from the satellites of the outer planets (rock/ice mixtures, with the exception of Io) thus call for a distinctive mode of origin (see Section 5). 3.7 Seismology The Apollo astronauts deployed four seismometers on the lunar surface between 1969 and 1972 (Figure 7). These instruments gathered data continuously until 1977, making the Moon the only extra-terrestrial body for which extensive seismic data has been gathered. The Moon is much less seismically active than the Earth, due to its lack of oceans and plate tectonics. Still, the Apollo network recorded several types of both naturally occurring and artificial seismic events, resulting in a total number of approximately 13,000 catalogued events over the 8-year span of the experiment. Because the Moon has no atmosphere to burn off potential impactors, there were a significant number of meteoroid impacts on the surface. The booster rockets and lunar modules from the Apollo spacecraft were also purposely impacted onto the surface after the departure of the astronauts, in part to test and calibrate the seismic array. Observed naturally occurring moonquakes include the relatively large but rare shallow moonquakes of unknown origin (similar to intra-plate earthquakes), and the relatively small but frequent deep moonquakes, (triggered with monthly periodicity by the lunar tides). Observed deep moonquakes generally had body wave equivalent magnitudes less than three, with most less than magnitude one; shallow moonquakes were larger, with the largest recorded events having magnitudes between five and 5.7. In addition, the network detected many noise-like thermal events that were associated with the large temperature fluctuations between lunar day and night. Deep moonquakes are the most numerous type of seismic event, comprising approximately half of the event catalog. They are known to originate from distinct source regions located in a wide swath across the near side, at depths between approximately 700 and 1200 km (Figure 8). Events from a single source are periodic at monthly (tidal) periods, and exhibit high degrees of waveform similarity, likely representing repeated failure on existing fault structures at depth. Compared to terrestrial seismograms, lunar seismic signals exhibit characteristics typical of a large degree of wave scattering and very low attenuation, due in part to the very fractured nature of the upper few hundred meters of lunar regolith. During seismic events, the Moon tends to “ring,” resulting in recorded signals of extremely long duration, sometimes an hour or more, with P- and S-wave codas that mask secondary arrivals. The small number of seismic stations and lack of high-quality seismic events limited the types of analyses that could be performed. Despite these limitations, data from the Apollo passive seismic network have been extensively analyzed to reveal details on the Moon’s internal structure, confirming the presence of separate crust, mantle, and core layers (Figure 9). The detailed structure of the upper kilometer of the lunar crust was determined by two additional seismic experiments: the Active Seismic Experiment on Apollo 14 and 16, and the Lunar Seismic Profiling Experiment on Apollo 17. In both experiments, the astronauts detonated a series of small explosives on the lunar surface (Figure 10). A network of geophones then recorded the ground motions generated by these explosions. On Apollo 14 and 16, up to 19 explosions were detonated by an astronaut using a "thumper" device along a 90-meter-long geophone line. Additionally, on Apollo 16, three mortar shells were used to launch explosive charges to distances of up to 900 meters from the ALSEP. On Apollo 17, the astronauts were able to position eight explosive charges at distances of up to 3.5 kilometers from the Lunar Module, with the assistance of the Lunar Roving Vehicle. Both the Apollo 16 mortar shells and the Apollo 17 explosives were detonated by radio control after the astronauts left the lunar surface. These experiments showed that the seismic P-wave velocity is between 0.1 and 0.3 km/s in the upper few hundred meters of the crust at all three landing sites. These velocities are much lower than observed for intact rock on Earth, but are consistent with a highly fractured material produced by the prolonged meteoritic bombardment of the Moon. At the Apollo 17 landing site, the surface basalt layer was determined to have a thickness of 1.4 km. The lower crust and mantle seismic velocities have been estimated using the classical nonlinear inversion of compressional and shear wave arrival time readings made from the Apollo seismograms (Table 2). Shallow structure is constrained largely using surface and near-surface events (impacts and shallow moonquakes), while deep structure is constrained using mid-mantle and deeper events (deep moonquakes). Seismic velocities increase steadily down to approximately 20 km. At that depth, there is a change in velocities within the crust that probably represents the depth to which extensive fracturing, due to massive impacts, has occurred (see Figure 12 and Section 4.1). At an earlier stage of seismic data analysis, this velocity change was thought to represent the base of the mare basalts, but these are now known to be much thinner. The main section of unbroken crust from 20 to 60 km has rather uniform velocities of 6.8 km/s, corresponding to the velocities expected from the average anorthositic composition of the lunar samples. Very few seismic rays detected by Apollo traverse the region below the deep moonquake zone. Evidence for a highly attenuating region in the deep interior such as a layer of partial melt or a fluid lunar core is implied in part by a lack of observation of seismic signals originating from the far side of the Moon (Figure 11). Since deep moonquakes are generally small, their energy cannot penetrate the attenuating region to reach the nearside Apollo array. An additional interpretation of the lack of farside signals is that the farside is aseismic, which given the other global nearside/farside asymmetries (e.g. crustal thickness, mare distribution), is not outside the realm of possibility. Further seismic exploration is needed to resolve this issue. 4. Lunar internal structure Decades of research following the Apollo era has led to a model of the lunar interior that consists of a silicate crust and mantle, and a small iron core (Figure 1). Overlying the crust is a very thin layer of extremely pulverized material known as the regolith, resulting from the initial heavy and continued bombardment of meteorites on the lunar surface. As discussed previously, the crust is globally asymmetric in thickness, with the nearside on average thinner than the farside. The mantle is considered to be largely homogeneous, with increases in seismic velocity and density of only a few percent from the base of the crust down to the partial melt boundary later between the mantle and core. The core itself likely consists of a fluid outer layer and a solid inner layer. This section will provide details on the structure of the lunar interior. 4.1 Regolith The surface of the Moon is covered with a debris blanket, called the regolith, produced by the impacts of meteorites (Figure 12). It ranges in scale from fine dust to blocks several meters across. Although there is much local variation, the average regolith thickness on the maria is 4–5 m, whereas the highland regolith is about 10 m thick. Seismic velocities are only about 100 m/s at the surface, but increase to 4.7 km/s at a depth of 1.4 km at the Apollo 17 site. The density is about 1.5 g/cm3 at the surface, increasing with compaction to about 1.7 g/cm3 at a depth of 60 cm. The porosity at the surface is about 50% but is strongly compacted at depth. The individual crater ejecta blankets that comprise the regolith typically range in thickness from a few millimeters up to about 10 cm, derived from the multitude of meteorite impacts at all scales. These have little lateral continuity even on scales of a few meters. Most of the regolith is of local origin: lateral mixing occurs only on a local scale so that the mare–highland contacts are relatively sharp over a kilometer or so. The rate of growth of the regolith is very slow, averaging about 1.5 mm/million years or 15 Å/year, but it was more rapid between 3.5 and 4 billion years ago during the late heavy bombardment. Five components make up the lunar regolith: mineral fragments, crystalline rock fragments, breccia fragments, impact glasses, and agglutinates. The latter are aggregates of smaller particles welded together by glasses. They may compose 25–30% of a typical regolith sample and tend to an equilibrium size of about 60 μm. Their abundance in a sample is a measure of its maturity, or length of exposure to meteoritic bombardment. Most lunar regolith samples reached a steady state in particle size and thickness. Agglutinates contain metallic iron droplets (typically 30–100 A ̊) referred to as “nanophase” iron, produced by surface interaction with the solar wind during melting of the regolith by meteorite impact. A “megaregolith” of uncertain thickness covers the heavily cratered lunar highlands. This term refers to the debris sheets from the craters and particularly those from the large impact basins that have saturated the highland crust. The aggregate volume of ejecta from the presently observable lunar craters amounts to a layer about 2.5 km thick. Earlier bombardment may well have produced megaregolith thicknesses in excess of 10 km. Related to this question is the degree of fracturing and brecciation of the deeper crust due to the large basin collisions (Figure 12). Some estimates equate this fracturing with the leveling off in seismic compression-wave velocities (v ) to an approximately constant 7 km/s at 20–25 km depth. In contrast to the p highlands, bedrock is present at relatively shallow depths (tens of meters) in the lightly cratered maria.