The Worldmaking Project · An atlas of rotating worlds, 1969–1977
Space Settlements
In the autumn of 1969 the physicist Gerard O’Neill put a question to a small seminar of Princeton freshmen: is a planetary surface the right place for an expanding technological civilization? Over the next eight years the question grew into a body of engineering, a set of settlements built in free space from lunar material, rotating for gravity, lit by mirrors, and large enough for towns, farms and a sky. This atlas opens five of those designs so you can examine how they work, and follows the economy, energy and industry meant to pay for them.
Choose a habitat to examine
01
Why build in free space
O’Neill’s seminar started from three observations. In space, solar energy is available all the time, with no night and no weather. Planets add surprisingly little room: settling the Moon and Mars would hardly more than double humanity’s land, and at the growth rates of the time that increase would be used up in about thirty-five years. And building material is far cheaper to lift from the Moon, whose gravity well is 22 times shallower than Earth’s.
So the land itself would be built. A rotating vessel presses its contents against its outer wall, which becomes a floor with gravity. The group considered three shapes for such a vessel, the sphere, the cylinder and the wheel, and their first calculation surprised them: a steel shell spinning for Earth gravity and loaded with soil and air could be several miles across.
In the summer of 1975, NASA’s Ames Research Center and Stanford University ran a ten-week study that worked the idea out as a complete system for 10,000 people, published in 1977 as Space Settlements: A Design Study. Its authors concluded that the obstacles were “principally philosophical, political, and social rather than technological.” O’Neill’s book The High Frontier (1976) carried the idea to a wide audience through a sequence of designs he called Island One, Two and Three.
Both works tie the settlements to an economy. The colonists refine aluminum, titanium and silicon from lunar ore and build solar power satellites, an idea Peter Glaser introduced in 1968, which send electricity to Earth. Sales of that power were meant to pay for the first colony and for the ones that followed.
1869Edward Everett Hale, The Brick Moon: a brick sphere launched by accident, its workers settling in for good
1903Konstantin Tsiolkovsky adds rotation for gravity, solar energy and a space greenhouse
1929Herman Potočnik (Noordung) proposes the 30 m Wohnrad, the living wheel
1929J. D. Bernal imagines self-contained spherical worlds in The World, the Flesh and the Devil
1952Wernher von Braun enlarges the wheel to 76 m for an orbit 1,730 km up
1969O’Neill’s Physics 103 seminar at Princeton asks where civilization should expand
1974O’Neill first publishes the space colony studies
1975NASA Ames and Stanford summer study designs the Stanford Torus
1976O’Neill’s The High Frontier presents Island One and Island Three
02
The geography of the Earth–Moon system
Gravity shapes empty space into terrain. Earth and the Moon sit at the bottom of deep wells, and between them lie five shallow valleys where their pulls balance the rotation of the whole system, the Lagrangian libration points. Three of them are saddles that need constant station keeping. The other two, L4 and L5, lie on the Moon’s own orbit, each as far from Earth as from the Moon, and an object nudged away in any direction drifts back. The 1975 study put its colony at L5 and called the choice somewhat arbitrary, since L4 is almost identical.
Drawn in a frame that turns with the Moon, so the Moon and the five points hold still. Distances to scale; Earth and Moon enlarged.
03
What every settlement has to solve
Gravity by rotation
Spin presses everything against the outer wall. A larger radius allows a slower, more comfortable spin, which is why the designs are so big.
Shielding by mass
Cosmic rays are stopped by several tonnes of material over every square meter, usually lunar soil, so shielding becomes most of a settlement’s mass.
Sunlight by mirror
Mirrors carry daylight in while the structure blocks straight-line radiation, and their angles set the length of the day.
A closed loop, and heat out
Air, water and food cycle between people, crops and animals. Everything that comes in as energy leaves as heat from radiators.
Each habitat sheet in the atlas has live diagrams for these systems, scaled to that design.
04
The system around the habitats
A habitat was only one part of the 1975 plan. The rest was an industrial system reaching from a mine on the Moon to power stations over Earth, and an economy meant to pay for it all.
While the studies were under way, NASA Ames commissioned paintings to show what the settlements would look like from outside and from within. Don Davis worked in oils with a naturalist’s light; Rick Guidice used a brighter, graphic style drawn from illustration. The paintings are in the public domain, and each habitat sheet in the atlas shows the ones that match it. Select a painting to see it larger.
The 1975 study justified the colony with one product, solar power satellites built from lunar material and sold to Earth as electricity. Its authors admitted the risk: “it is a definite weakness of the design to depend entirely on this one particular enterprise.” Their program runs year by year from research on Earth in 1976 to a completed colony in year 22, with every figure in 1975 dollars.
Annual cash flow, benefits minus costsbillions of 1975 dollars · program years 1–45 (1976–2020)
Running totalcumulative cash flow, undiscounted and discounted
Where $190.8 billion goes
Baseline estimate, research through the first colony, with 20 percent overhead. Transportation dominates.
The product
Each 10 GW power satellite earns about $1.17 billion a year sold in the US. Electricity at the busbar starts at 8.5 mils per kWh and falls to 4.8 mils by year 70, against a target of 14.1 mils set by coal and nuclear plants of 1974.
1 mil = 0.1 cent.
Learning curves
After the first, each power satellite costs about $9.73 billion, or $6.76 billion once the new shuttle flies, and each further colony about $9.24 billion. Both fall along 80 percent learning curves: every doubling of output cuts the variable cost by a fifth.
Topic 02 · Energy · SP-413 chapters 4 and 6
Power from orbit
A solar power satellite in geosynchronous orbit sees the Sun almost all the time and sends its output to the ground as a microwave beam. The idea came from Peter Glaser in 1968. The study weighed two designs, a photovoltaic one and a turbogenerator one, and judged the turbogenerator preferable because it relied on technology already in use, without making a final choice.
Capacity against demandgigawatts · program years 15–45
Capacity counts the 10 GW satellites in use from the study’s Table 6-12. US demand starts from 224 GW in 1975 and grows 5 percent a year, the study’s assumption. About a third of the power was planned for export.
Power for the settlements themselves
The torus runs on 50 MW from solar cells stretched between its spokes, 3 kW for each resident, and a 200 MW solar plant powers fabrication. The lunar base needs 200 MW, almost all of it (192 MW) for the mass driver, and two complete launchers keep shipping through the two-week lunar night. Later in the program a power satellite at L1 beams energy to the Moon on a 3 cm wavelength into a 27 km² receiving antenna.
Topic 03 · Mass driver · SP-413 chapter 5 and appendices F–I
Throwing the Moon
Ten million tonnes of lunar soil go into the first colony. The study’s mining base sits near the Apollo 16 landing site in the Cayley region, chosen for aluminum-rich ore and flat ground. Ore is compacted into small blocks, loaded into buckets with superconducting coils, and accelerated along a magnetic track at about 30 g until it leaves the Moon at 2.4 km/s, bound for a catcher at L2.
The launch trackhorizontal distances compressed; buckets run in real time
The lunar base
A pit about 2 km across and 10 m deep, worked by scooper-loaders at about 4 t a minute. A construction crew of 300 builds the base, and 150 people run it, on two-year tours in prefabricated quarters buried under 5 m of lunar soil. Its 200 MW plant is nuclear, so launches continue through the lunar night.
The catcher
A thin net of 10 m², steered by three cables across a triangle 1 km on a side, tracks each 10 kg block by radar from 2 km out and slows it from 200 to 20 m/s. Each catcher weighs about 220 t. Five were planned for 11 million tonnes in ten years, and ore carriers then take about two months to reach L5.
Rejected: the gas gun
The study also costed a lunar gas gun firing 57 t projectiles twice an hour. It needed less equipment on the Moon, but an error of ±0.5 m/s scattered payloads over a circle about 1,000 km wide, to be rounded up by 50 interceptor rockets.
Topic 04 · Construction · SP-413 chapter 6
Building the first colony
The study planned 22 years from the start of research to a finished colony: research on Earth, tests in low orbit, then a lunar base, mass driver, catcher and construction shack at L5, six years of habitat construction, and four for the shield and the arrival of colonists. Phases overlap; the years below are placed from the study’s stated durations and milestones.
Topic 05 · Studio · after SP-413 chapters 3 and 4
Design your own habitat
The study chose its torus by testing shapes against a list of human needs: 0.95 ± 0.05 g, no more than 1 rpm, 0.5 rem a year, and 67 m² of projected area per person. Set a shape and its dimensions, see which criteria it meets, and place it beside the designs the study and O’Neill compared.
Mass per person against populationstructure plus shield · logarithmic axes
Structure is a membrane estimate for aluminum at 200 MPa working stress holding 50.8 kPa of air, including the hull’s own spin weight. It reproduces the study’s sphere (3.5 Mt) and lands below its torus (150 kt), which carries additional internal loads. Shield area adds 2 m of clearance. Comparison points come from SP-413 Tables 4-1 and 4-2 and from The High Frontier.