Adherents of the shallow-crater argument point to a simple ratio. Lunar craters can be hundreds of kilometres wide, yet none seems to be more than a few kilometres deep. An asteroid big enough to dig Gagarin on the far side, they reason, should have punched tens of kilometres into rock, so something under the surface must have stopped it: an armoured inner layer, in the strongest form of the claim the hull of a hollow, artificial Moon. The measurements of crater depth and crustal thickness are public. The question is whether shallow, wide craters are a puzzle that needs a hull, or the expected result of how large impacts collapse.
The argument goes back to a 1970s article by the Soviet authors Mikhail Vasin and Alexander Shcherbakov, reproduced in English by Anomalien (https://anomalien.com/russian-scientists-in-the-70s-is-the-moon-the-creation-of-alien-intelligence/). They cite the physicist Kirill Stanyukovich: a "missile" about six miles across "must, on collision with the Moon, penetrate to a depth equal to 4 or 5 times its own diameter (24-30 miles)." Yet the craters, they write, are "all about the same depth, 1.2-2 miles," and a 148-mile crater has "three miles at the most between top and bottom levels," with a convex floor. Their answer is a shell: "a dense armoring of about 20 miles in thickness" under a looser covering, with the craters stopping at the armour. The same text is used in discussions of the wider spaceship-Moon idea; here it matters for its crater figures.
The crater most often named has an official record. The IAU's Gazetteer of Planetary Nomenclature, maintained by the U.S. Geological Survey (https://planetarynames.wr.usgs.gov/Feature/2054), lists Gagarin at a diameter of 261.83 kilometres, centred at 19.66° south and 149.35° east, a name adopted in 1970. That is about 163 miles, a little smaller than the 170 miles often quoted. A depth of a few kilometres for a crater that wide is not disputed by either side of the argument. What is disputed is whether that depth is strange, and the gazetteer itself only records the size and position.
The International Astronomical Union also published a picture of Gagarin with a short caption (https://iauarchive.eso.org/public/images/detail/ann19056a/). It gives the diameter as 265 kilometres and notes "its characteristic nested smaller craters." The photograph was taken by the Apollo 17 crew from an altitude of about 100 kilometres, before their landing in December 1972, and released by the IAU in September 2019. The caption describes the naming negotiations between the United States and the Soviet Union over far-side features. The nested craters sit on the floor of the larger one, so the picture records later impacts on top of the original crater as well as the crater itself. Any depth measured today is the depth after that later history, not on the day Gagarin formed.
Measured depths now come from laser altimetry. J. D. Kalynn, C. L. Johnson, O. S. Barnouin and G. R. Osinski used Lunar Orbiter Laser Altimeter data to revisit depth and diameter for fresh complex craters, presenting the work at the 2013 Lunar and Planetary Science Conference (https://www.lpi.usra.edu/meetings/lpsc2013/pdf/1309.pdf). Their 111 craters span 15 to 167 kilometres. They found that "crater depth increases with increasing diameter" and that highland craters are deeper than mare craters, with fits of d = 1.558D^0.254 in the highlands and d = 0.870D^0.352 in the mare. On the highland fit, a 100-kilometre crater comes out about five kilometres deep. The authors suggest the "more fragmented and porous megaregolith" of the highlands may allow deeper craters, which is the opposite of what a hard barrier would produce.
Why big craters are wide rather than deep is the subject of impact modelling. Elizabeth Silber and colleagues, in a 2017 paper accepted by the Journal of Geophysical Research: Planets (https://arxiv.org/abs/1704.04247), simulated lunar craters 10 to 26 kilometres across, straddling "the simple-to-complex crater transition." In that size range, bowl-shaped craters give way to complex ones whose walls slump and whose floors rise. The paper studies how impactor size, speed and "acoustic fluidization" of the rock control that change. Its point for this question is that the final depth of a large crater is set by collapse after the impact, not by how far the impactor travelled before it stopped.
The crust beneath the craters has been weighed from orbit. NASA's Jet Propulsion Laboratory reported in December 2012 that the GRAIL mission had produced the highest-resolution gravity map of any celestial body (https://www.jpl.nasa.gov/news/nasas-grail-creates-most-accurate-moon-gravity-map/). Co-investigator Mark Wieczorek said the average crust is "between 21 and 27 miles (34 and 43 kilometers)" thick, "about 6 to 12 miles (10 to 20 kilometers) thinner than previously thought." Maria Zuber described "fracturing of the interior extending to the deep crust and possibly the mantle," and the highland crust proved lower in density than assumed. Fractured, low-density rock tens of kilometres deep does not match a twenty-mile armour plate.
The crater record answers the shallow-crater claim on its own terms. Gagarin is about 262 kilometres wide and a few kilometres deep, and so are other large craters, because large impacts collapse into broad, flat-floored forms. Altimetry shows depth rising slowly with size and highland craters running deeper, and gravity mapping shows a fractured crust tens of kilometres thick. Nothing in these measurements requires a hull. The twenty-mile armour the 1970s article proposes would have to sit inside a crust that GRAIL measured at 21 to 27 miles of fractured, low-density rock.
What Stays Contested
Crater mechanics at the largest scales is still being modelled, and the exact depths of old, degraded basins depend on how much fill and later cratering they carry. The armoured-Moon reading treats a uniform shallowness as unexplained. The altimetry data show depths that vary with size and terrain, and the gravity data show fractures deep into the crust. A barrier that stops impactors at one depth would leave a different signature from the one measured, and adherents have not yet pointed to data showing it.
