The Foundation has existed for 26 of these 75 years. The rest belongs to people who were paid quietly, published anyway, and in several cases never learned what the money was for.
1951–1968
Before the Foundation
The Foundation did not exist, and nothing in this era is its achievement. What existed was a series of private endowments, made quietly, to people who were doing the right work in places where the money had run out. Our archive catalogued them in 2001 as the prior record, and the Foundation claims no credit for any of it.
They are listed here for one reason: every specification the Foundation writes today cites at least one of them, and the work would otherwise be lost. Sums are restated in credits at constant 2000 value.
Three winters of photographic spectra of Mars at opposition, funded through an observatory visitors’ fund after the station’s own grant lapsed. Forty-seven usable plates.
The result was the first internally consistent seasonal carbon-dioxide exchange estimate taken from a single site, and it is still the earliest entry in our atmosphere series.
Nine motors built, six destroyed on the stand. The three that survived are the reason anybody kept reading the reports.
Seibel published the failures in the same volume as the successes, at his own insistence and before it was the fashion.
The endowment paid a lease, a power bill and one wage. It bought no equipment and produced no paper.
Every primary mirror in the prior record was ground on Ruth Hollis’s bench, including the Mount Sabel replacement of 1957 and the two Vesterhavn plates cameras.
Farkas established that a mechanism which works in air will seize in vacuum for reasons that have nothing to do with lubrication.
Her 1958 note is cited verbatim in the Foundation’s current bearing specification, MF/MAT/1958-A, which is numbered for the year of the note rather than the year of the specification.
Minimum-energy transfer opportunities for eighty years, computed by four people using mechanical calculators over nineteen months.
The tables were checked against numerical integration in 1991. The 2027 line is off by eleven hours.
Eighty years of windows, and the one we are using is wrong by eleven hours.
A coupon library of white and second-surface coatings with absorptance and emittance measured to a discipline nobody was requiring at the time.
Twenty of the coupons were retained. Nineteen of them flew, twenty years later, under LODESTONE tranche III.
Stereo photogrammetry applied to opposition plates taken eleven days apart, producing the first quantitative relief estimates for the classical albedo features.
The method, considerably rewritten, is still the front end of the Foundation’s site-survey pipeline.
Banich concluded that the radiation problem is a water problem: that the only shield mass anybody would agree to carry is mass they were going to carry anyway.
Nobody argued with him. Nobody solved it either. The figure he wanted was produced numerically in 1996 and confirmed on the surface in 2013.
A 31 m³ enclosure, three occupants, nine days. It failed on carbon-dioxide scrubbing on the ninth morning and the run was ended by the occupants rather than the clock.
Thorn documented the failure at greater length than he would have documented a success, which is the reason the record is useful and the reason the laboratory now carries his name.
A triboelectric charging rig for basalt simulant, built out of a converted mine hoist because nothing suitable was for sale.
Rhee’s adhesion curves were reproduced on the lunar surface in 1979 and on Mars in 2011, in both cases within the scatter of her original figures.
An archive item rather than a grant: 1,140 letters, held in Ruth Hollis’s shop ledger, between nine of the people named above — most of whom had never met and several of whom worked on different continents in fields that did not then speak to each other.
The letters are how Banich’s water conclusion reached Thorn’s enclosure, and how Farkas’s galling note reached Seibel’s motors. Somebody was paying the postage.
The work was not scattered because somebody had spent sixteen years making sure it was not.
Thirty-four endowments, nineteen institutions, seventeen years, and no announcement of any kind. The last cheque was written in November 1968.
1969–1983
Programme LODESTONE
For fifteen years the endowments went to the Moon — three separate times, for three unrelated reasons. Ranging had to be proved before anybody could navigate; the interior had to be heard before anybody could trust a foundation; and the dust had to be understood before anybody could plan to live in it.
None of this bought a ride. The instruments were built and handed to payload integrators operating vehicles that were not ours, on the single condition that has applied to every award since 1951: the data is published, in full, without embargo. It always was, and we never once had to ask for it.
1984–1987: the waiting years. The record contains no new commitments for four years. No programmes were opened, no grants awarded, no instruments built. The archive’s standing explanation is the one given at the time: the machines were not good enough yet, and there was no honest way to spend money on the problem until they were.
Design freeze on a passive corner-cube retroreflector array: 104 fused-silica cubes in an aluminium frame, no power, no moving parts, nothing to fail.
Ilic’s argument for it was that a programme which cannot measure a distance cannot claim a schedule.
Delivered to a payload integrator, crated, with a calibration certificate and a two-page instrument manual.
The endowment paid for the cubes, the frame and the certificate. It did not pay for the ride, and did not ask for a plaque.
Two arrays returning; residuals at 0.3 m and improving as the ground stations improved. The series has never been interrupted since.
The award’s only condition was that the range data be published without embargo. It was, within the year, and by 1974 four groups nobody had funded were using it.
Publication is not generosity. It is the cheapest possible way to have your instrument checked.
Short-period seismometers, 8.4 kg, with a thermal design that assumed nothing about the regolith it would sit on because nothing was known.
Kew’s specification required 24 months of operation. Two instruments returned data for 61.
A two-station network is not a network, and the reports say so. What it gave was a ground-truth noise floor: a number for how quiet an airless body actually is, which is what a foundation design needs and a seismology paper does not.
Dust accumulation and adhesion monitors, with witness plates and a coating library drawn from Okonjo’s retained 1959 coupons.
Solheim’s proposal is two and a half pages long and is still issued to new staff as an example of how to write one.
Nineteen of the 1959 coupons, flown and read back. Absorptance degraded faster than any ground test had predicted and the mechanism was mechanical, not chemical: dust, held electrostatically, on surfaces that were never touched.
Rhee’s 1966 adhesion curves fitted the flight data within their original scatter.
Amagai’s conclusion, in the first paragraph of the report rather than the last: a lunar programme and a Mars programme do not share a dust problem. They share a seal problem, and it is the harder of the two.
Every rotary seal on the 2027 surface plant is qualified against a test he designed in 1981 and never patented.
Fifteen years, three tranches, eleven instruments, nine of which operated beyond their specified life. The closing report is forty pages and recommends nothing.
1988–1997
Programmes MERIDIAN and LATTICE
Two programmes ran in parallel for nine years and neither of them built anything that left the ground. MERIDIAN was a model of the Martian atmosphere. LATTICE was the memory to run it in, because every machine that could hold the model was losing to single-bit errors faster than the model could converge.
It was, on paper, the least ambitious decade in the record. It is the one that made the rest possible: by 1997 there was a defensible answer to the only question a lander programme has to answer first, which is where a vehicle can be put down without guessing.
1998–1999: two years of filings. The record contains no science whatsoever for twenty-four months — only legal instruments: articles drafted and redrafted eleven times, a purpose clause, a publication covenant, an endowment transfer, and a registration application lodged in January 2000.
A general circulation model, first run: 42 days of wall time to simulate 11 days of weather.
It was wrong. Nold’s opening report makes the important point about it — that it was wrong in a way that could be measured, which is the only useful kind of wrong.
Error-correcting memory at density: a correction scheme that survived the soft-error rate the modelling machines were actually seeing rather than the rate their datasheets claimed.
Memory was the scarce resource. It is still the scarce resource, in every machine anybody has ever built, and every honest specification sheet says so.
Ramanathan rewrote the solver rather than waiting for the hardware. The same machine finished the same run thirty-one times faster, which converted a nine-month experiment into a nine-day one.
She joined the staff thirteen years later and has been president since 2016.
With the arithmetic fixed, the model spent 71 per cent of its wall time waiting on disks. Anozie fixed the disks.
The report contains the line that is quoted in the Foundation’s grant guidance to this day: find out what your programme is actually waiting for before you buy anything.
The model’s dust-lifting threshold predicted the onset of three observed regional storms to within six days, from initial conditions taken eleven weeks earlier.
Nold’s note on it is one sentence: we are no longer describing the weather, we are anticipating it.
We are no longer describing the weather. We are anticipating it.
Marchenko’s scheme was adopted by three memory manufacturers within eighteen months. The endowment took no royalty.
This was not generosity; the grant instrument simply did not provide for one, and has never been amended to.
1.9 million cell-hours, validated against thirty-one years of published opposition observations and the whole of the prior record’s atmosphere series.
It was, in Nold’s phrase, the first model anybody would be willing to land on.
Three hundred and twelve candidate sites, screened on winds, pressure, opacity and storm statistics and nothing else. Forty survived.
Those forty are the sites PLUMBLINE inherited nine years later, unchanged.
Vasilenko’s transport code reproduced Banich’s 1962 conclusion, thirty-four years later, and put a number on it: shielding mass is water mass, and the crossing is affordable only if the water is doing two jobs.
The number was confirmed on the surface in 2013 and has not needed revising.
Nine years, nineteen million credits, two closing reports, and nothing anybody could sell except a memory part that had already been given away.
The closing line of the joint report: the machines are now good enough. What is missing is the ground.
2000
The hinge
For forty-nine years the money had moved quietly and belonged to no institution. There was no board, no audit, no letterhead and no name — only endowments, a ledger, and a great deal of postage.
On 14 March 2000 all of it was put inside a filing. Nothing about the direction changed. What changed was that the work now had a registration number, a board that could be voted against, accounts that could be audited, and a covenant that could be enforced against the Foundation itself.
What actually changed in 2000. Before: one ledger, no name, no obligations that anybody could hold anyone to. After: a registration number, eight divisions, a board with a quorum, audited accounts published in full, and a covenant that compels the Foundation to publish results it would rather not. The direction had been fixed since 1951. The paperwork was the new part.
Filed as a non-profit by a single founder, E. Starlight, who endowed it with 1.2 billion credits and took no salary, no royalty and no reserved power beyond the chair of the board.
The registration is four pages. The publication covenant attached to it is nine.
Adopted unamended and never since amended. It is the only sentence in the Foundation’s constitution that the board cannot alter by ordinary resolution.
Five members, including Teodora Ilic, who had held a lunar-instrument endowment in 1969 and had never met the person paying for it.
The founder attended by telephone. She has attended every meeting since by telephone.
A combined Raman and laser-breakdown spectrometer bench, 2.4 million credits, to Sofia Errazuriz at the Aldern Institute of Technology.
The instrument that grew out of it flew seven years later. Errazuriz has been the Foundation’s Director of Research since 2011.
Four buildings on a disused ordnance site at Wraith Basin: a drawing office, a clean bay, a machine shop and an archive. Leased rather than bought, on the principle that the Foundation should own as little as it can.
Eleven months of archive work catalogued thirty-four endowments made between 1951 and 1968, with every surviving report scanned and published in full.
The catalogue’s preface is explicit: the Foundation did not make this work and asserts no claim to it. It is listed so that it is not lost.
2004–2021
Programme PLUMBLINE and family
The Foundation has never operated a vehicle on Mars. It has paid for instruments carried on vehicles operated by others, which is roughly two orders of magnitude cheaper and returns the same numbers. The condition of award is the one from 1951: publish.
Because the operators publish, the Foundation has never once had to request data from a surface vehicle. It reads what everybody else reads, on the day everybody else reads it, and spends its money on the instrument instead of the spacecraft. That was the whole arrangement.
The purpose of the survey was not discovery. It was elimination: three hundred and twelve candidate sites down to a number small enough that a crew could be sent to one of them on purpose.
2022–2026: qualification. LONGBOAT opened as a programme of record in 2022 and TENDER in 2023. Long-lead forgings for the first pressure hull were ordered in 2022; the reactor-electric stage passed concept freeze in 2020 and its first article test in 2024; the tender cargo stage flew its qualification flight in 2025; a full-scale wet mock-up of the orbital assembly sequence was rehearsed through 2026. Nothing was announced, because nothing had happened yet.
A survey programme with a defined end condition written into its charter, which is unusual and was deliberate: the programme closes when the shortlist reaches five sites or fewer.
Forty sites inherited from MERIDIAN’s 1995 screening, eleven criteria, seventeen years budgeted.
Errazuriz’s combined Raman and breakdown head: 1.9 kg, 1.9 m standoff, 532 nm, a 40 µm slit and a detector cooled to −40 °C. Built at the Aldern Works, crated, and given away.
It flew in 2007 and returned mineralogy for six years against a design life of two.
Pulsed-neutron and passive-gamma instrumentation, because a crew’s shielding, propellant and drinking water are the same problem and Banich had said so in 1962.
Within the first metre, at two of five surveyed bands. The Foundation’s shortlist lost eleven sites that year and the survivors gained a reason to be on the list at all.
Published in full, forty-three days after acceptance.
Bearing strength, trafficability, slope stability, anchor pull-out. Measured in five simulants at Braemore, then checked against wheel-track and scuff imagery that the vehicle operators had already published.
It is the least cited programme in the archive and the most used: it is where a landing pad’s thickness comes from.
Column opacity, deposition rate on optical and radiating surfaces, storm-onset statistics.
Solheim’s 1976 conclusion, re-measured on another planet and unchanged: the dust wins. Do not design against it; design around it, and budget for the cleaning.
The dust wins. That is not a pessimistic finding, it is a design input.
A water problem turned into a chemistry problem. Nasrallah’s wet-chemistry cell established the processing penalty for making the surface water drinkable, in kilowatt-hours per litre.
Two shortlisted sites were downgraded on the strength of it.
Transit and surface dose reconstructed from published detector records and cross-checked against Vasilenko’s 1996 transport code.
Banich was right in 1962, and the Foundation now knows by how much: 0.31 Sv for the crossing, and a surface stay that is survivable only under regolith or water.
Sintering simulant into structural slab at power levels a surface plant could actually supply. First 400 × 400 mm slab in 2016 at 1.2 kW, with a flexural strength inside the shielding specification.
The programme’s first run deliberately reproduced Thorn’s 1964 carbon-dioxide failure, in the same enclosure volume, in order to calibrate a modern scrubber against a known collapse.
It is the only grant in the Foundation’s history whose stated deliverable was a failure, and the Cold Lab still runs it annually as a standard.
The Foundation’s third president, and the first who had been a grantee before it existed: her sparse-solver award was made in 1990, ten years before there was a Foundation to make it.
Ranked against eleven criteria: water within a metre, slope, bearing strength, opacity, thermal amplitude, entry corridor, dispersions, sinterable feedstock, dose, communications geometry, and distance from anything worth not disturbing.
The nine: Thessaly Bench, Ardea Plain, Corvine Shelf, Little Mourne, Kestrel Gap, Marrow Flats, Vane Terrace, Quillon Rise, Harrowgate.
The covenant had been honoured since 1951 by habit and since 2000 by contract. In 2018 it got a budget, a staff of four and a median: forty-one days from acceptance to open publication.
Negative results included. Retractions included. Nil embargoed, nil paywalled, nil patents retained.
A structures trade study for a crewed transfer vessel. Not a vehicle, not a programme, not an announcement — a spreadsheet, six configurations, and a recommendation to do nothing until the survey closed.
The board accepted the recommendation to do nothing.
Cumulative disbursement since 2000 passed two billion credits. No anniversary event was held, no statement issued and no photograph taken.
The audited accounts were published on the usual date, in full, as they have been every year.
Thessaly Bench (primary), Ardea Plain, Corvine Shelf. Seventeen years, forty-six funded instruments, eleven of which flew, three hundred and twelve candidate sites reduced to three.
PLUMBLINE reached its charter end condition and was closed rather than extended. The Foundation stopped surveying on the day it had enough survey.
A finished survey should be closed, not extended.
2027 →
Programmes LONGBOAT and TENDER
The Foundation is not going to Mars in 2027. In 2027 it begins building the thing that goes, and begins putting the plant on the ground that has to be working before anybody is asked to live beside it.
The schedule below is tied to the transfer windows, which are not negotiable and are known to eleven hours — a figure the Foundation owes to four people with mechanical calculators in 1958. Cargo precedes crew by two windows in every case. Nobody is sent to a site that does not already have power, water, shelter and a spare of each.
None of the above is a promise. It is a schedule, which is a different kind of document. The Foundation has held one since 1951 and has missed dates on it before — the 1984 to 1987 gap is four years of missed dates — and it publishes the misses in the same place and the same type as the rest.
Board authorisation for the first hull: LB-1, 68.4 m overall, a 1.4 MWe reactor-electric stage, two 148 m² radiators, a pressure hull of 412 m³ with six berths, and 142 t of water forward doing two jobs at once.
Articles are built at the Aldern Works and assembled in orbit. The vessel is never intended to touch either surface.
Three tender flights to Thessaly Bench: the power plant, the sinter works that makes the landing pad and the shielding slab, and a consumables cache with a spare of everything on the critical list.
Manifest TDR-01 runs to 11,904 line items. The Foundation publishes all of them, and is aware that no inspector has yet finished reading one.
Habitat shells one and two, the water plant, and LB-1 itself flying the crossing uncrewed with the full consumable load aboard and nobody in it.
If the shakedown transit fails, the 2031 crew does not fly. That is written into the programme rather than left to judgement.
Six people, 258 days out, to a site with two habitat shells, a working power plant, a sintered pad, a water plant and four years of cargo already on the ground.
They are not settlers and the Foundation does not use the word. They are the first crew of a permanently occupied station, and they come home in 2033.
LB-2 carries the second crew out and LB-1 brings the first crew home. The station is occupied continuously from this window onward, by rotation rather than by growth.
Third rotation, expansion shells four and five, and the point at which the station’s consumables are produced locally faster than they are consumed.
That is the condition the Foundation was filed to reach. The charter says nothing about what happens afterwards.
What happens next
The survey closed in 2021 with three sites left standing and one chosen. Construction is authorised in 2027, cargo flies in the same window, and the first crew of six leaves in 2031 on a 258-day transit.