For the clerk who typed it 39 years ago, Technical Sheet No 4 was no doubt just another piece in the mountain of secret paperwork that piled up in early 1966 as France raced to detonate its first nuclear bomb in the South Pacific.

The document gives no indication that, within months, the decision it so dryly recorded would have a resounding impact on the lives of hundreds of people on the other side of the world, in French Polynesia.

Just hours after France detonated its first device at Mururoa, homes, food, water and soil on the Gambier Islands, 500km southeast of the test site, would be drenched with radioactive rain.

Yet, in line with the safety decisions taken in Paris and faithfully narrated on Technical Sheet No 4 and other documents, no action would be taken to evacuate or shelter local people or Europeans or even advise them of any danger.

Buried in defence ministry files for nigh on four decades, these documents have been anonymously mailed to an activists' group demanding transparency over France's nuclear programme.

Copies have been passed to the Herald. They have been confirmed as genuine by a French Defence Ministry spokesman, who stands by France's position that the test programme was safe.

The tale these papers tell is of fumbling, indifference and kneejerk secrecy as France hastened to build its nuclear strike force.

For campaigners, the archive trail confirms what islanders and military veterans have claimed for years: the tests caused death and sickness yet France smothered the evidence.

The documents date from 1966 and 1967, when France prepared for and then carried out its first tests in French Polynesia, a move prompted by the loss of a test site in the Sahara after the independence of its North African colony, Algeria. With President Charles de Gaulle clamouring for results, French scientists took the quick, easy and cheap option - which three years earlier Britain, the Soviet Union and the United States had already decided to ban.

And the first batch of nine nuclear explosions would use the quickest, easiest - and dirtiest - technique of all.

The bombs would be detonated on barges and on towers, creating a blast that would vaporise thousands of tonnes of seawater. The mushroom cloud of radioactive droplets and debris would be blown wherever the winds saw fit.

As the months ticked away towards the first test, French scientists and ministry officials - their experience limited to the dust and weather patterns of the Sahara - struggled to devise a net of safety procedures to cope with this fallout risk.

A document dated January 13, 1966, more than six months before the first test, reporting on a meeting of a panel called the Consultative Safety Commission, shows deep concern for the Gambiers.

People on this small archipelago, especially the 570 on the main island of Mangareva, had "particular characteristics", the commission warned.

Many were very old or very young, there was a high percentage of women who were pregnant or of child-bearing age. It suggested the maximum safe dose should be 0.5 rem a year, equivalent to 5.0 millisieverts annually.

"The slightest fallout will cause the population to absorb a dose above the given threshold," the commission warned.

In the event of fallout, the population could be sheltered in churches for a few hours until it was over, another document suggests. If the soil, vegetation and homes became too contaminated, people could be taken to French warships moored offshore or evacuated to Hao island.

But, as Technical Sheet No 4 makes clear, the authorities decided against evacuating the Gambiers.

"A preventive evacuation of the Gambiers before a test explosion is ruled out for political and psychological motives," the document says. "On the other hand, it is perhaps possible to envisage a preventive evacuation for other atolls." At 5.35am on July 2, 1966, the first French Pacific bomb, a Hiroshima-sized device codenamed Aldebaran, was detonated at Mururoa, transforming the dawn sky into a dazzle of white, orange and indigo.

Borne by treacherous, shifting winds from the northwest, fallout rained on the Gambiers.

At 6.38pm, the command post received a telex from monitors on Mangareva, the chief island of the Gambiers, warning them that radioactivity levels had reached 0.1 millisieverts an hour.

At 9.30pm, another telex warned that radioactivity measurements now varied from 0.04 to 0.58 millisieverts an hour.

The tone was panicky: "Minister informed radioactivity not negligible stop slow decrease stop soil contamination stop request instructions for decontamination and controlling food and fishing stop ends."

The minister referred to was the then minister for overseas departments, General Pierre Billotte, who was on Mangareva for ceremonies to herald the first test.

A French state radio journalist who was on Mangareva at the time, said that Billotte was quietly approached by a military officer just before he was to attend a welcoming party in the village of Rikitea.

After a few words whispered in his ear, he left abruptly, flying off with aides on a Catalina seaplane.

What, though, did the French authorities do for the local Polynesians as the fallout alarm sounded?

The answer appears to be: nothing.

They went into crisis mode, dispatching a technical ship, the Coquille, to measure radioactive levels but doing everything they could to ensure no word of the accident leaked out.

No one was evacuated. No bans were imposed on food and water.

No one was even told about the danger. People walked about bare-chested or in their shirtsleeves. Children played barefoot and rolled on the ground.

The measurements made by the Coquille's technicians are eloquent. On July 6, an unwashed lettuce had 666 times normal radioactive levels. After it was washed, the level was reduced to 185 times. Drinking water was six times the normal levels of radioactivity. On July 8, radioactive levels in lettuce were 359 times (unwashed) and 148 times (washed) those of normal levels. The same day, soil taken from drains after 12 hours of heavy rain had more than 50 times its natural level of background radioactivity.

Tight lips and public calm were the priorities. "No restriction measure is envisaged," Dr Philippe Millon, a senior officer aboard the Coquille, reported to base. "The Tahitian population is perfectly unaware, carefree and showing no curiosity ... [the military personnel] are aware [of a fallout problem] but, obviously, most of them do not know the figures."

After reporting that the handful of Europeans on Mangareva were also ignorant of developments, Millon comes to a reassuring conclusion: "The psychopolitical situation in the Gambiers does not seem to present any problem for the immediate moment."

Bruno Barrillot, of the Lyon-based Centre for Documentation and Research into Peace and Conflicts, which received the documents, says the picture is "terrifying ... the services concerned manipulated information to minimise their impact on the environment, personnel and local population. On July 2, 1966, the soil contamination in Mangareva was 142 times higher than in the forbidden zone around Chernobyl," he calculates.

The French Defence Ministry told the Herald the documents had been scrutinised by its experts, who confirmed the papers were genuine.

"These documents were stolen from us. We intend to find out who and how this happened, but that's our problem," a spokesman said.

"The important point is that the person who took them obviously selected a few documents out of miles and miles of files, and they are documents that are sparse, whose selection is biased and which do not give the full picture of what happened."

The spokesman said the dose of radioactivity on Mangareva on July 2, 1966, was 5.5 millisieverts, very slightly above the threshold at the time of 5.0 millisieverts.

"0.5 millisieverts is not a sufficiently big risk to evacuate an entire population which would have been traumatised, in an operation which would have created a whole range of problems," he contended.

Barrillot angrily retorts that this is spin. The dose of 5.5 millisieverts, he says, was not measured at the time. It was an estimate made in 1998 by the IAEA in its report into radioactivity at the Mururoa and Fangataufa test sites and at other atolls in French Polynesia.

The IAEA calculated this dose on the basis of data from France, which told the agency the fallout had lasted for one hour 20 minutes, Barrillot says.

Yet the telex entries in the documents suggest that fallout lasted more than twice as long - two hours and 52 minutes.

Barrillot says further evidence that the dose was much higher can be found in the "frightening" levels of radioactivity measured by the Coquille technicians several days after the blast.

In any case, says Barrillot, 5.0 millisieverts was the maximum permitted dose a year - not just for a day - and the papers show further fallout occurred on September 24, 1966, with Rigel, the fifth blast in the series, which was more than 16 times the size of the Hiroshima bomb.

The documents, while still giving an incomplete picture, greatly boost the Polynesians and former military personnel campaigning for compensation for what they say was ill health caused by the tests.

Organisations representing these groups say cancer among those exposed to the test programme is twice as common as among the general French population and certain cancers of the blood systems are up to 100 times more frequent.

Sue Roff, an expert on nuclear tests at the University of Dundee Medical School in Scotland, says the documents are valuable but may not be enough to force the French Government to admit any liability.

She predicts that yet more classified documents could start turning up mysteriously at research institutions and activists' groups.

"Somebody was wise enough back then to keep these documents, they sat on them and perhaps now when they don't feel so threatened by various national security laws they say, well, I will make them public one way or another."