In the summer of 1998, I received a letter from a friend named Rand Flem-Ath, suggesting that he and I should collaborate on a book about Atlantis. His proposal sounded so fascinating that I agreed without hesitation.

Eleven years earlier, Rand had written a paper suggesting that the site of Atlantis was the South Pole, or rather, the continent of Antarctica. This sounds, of course, absurd, since Antarctica is thousands of miles south of the mid-Atlantic, which Plato claimed was the location of Atlantis. But Rand's theory was based on a suggestion made by Charles Hapgood, a New England professor of history, who declared that the earth's crust is subject to periodic shifts, which can cause the continents to change their position. This apparently unlikely theory gained the support of Albert Einstein, who wrote an introduction to Hapgood's book Earth's Shifting Crust (1958).

Rand sent his paper to Hapgood, and was delighted to receive a warm reply that called it 'the first truly scientific exploration of my work that has ever been done.' Oddly enough, Hapgood did not comment on Rand's notion that Atlantis might be in Antarctica. But Rand was nevertheless encouraged to continue with his investigations. He and his wife Rose went to London, where Rand pressed on with his studies in the British Museum Reading Room.

In October 1982, Rand wrote to Hapgood to update him on his progress with the 'shifting crust' theory, and in reply, received a letter that left him staggered. Hapgood stated that he had recently made some exciting discoveries that showed that civilization, with a high level of science, was at least a hundred thousand years old. He said that he intended to publish these findings in a new edition of his book, Earth's Shifting Crust.

Understandably, Rand was wildly excited, and wrote to Hapgood saying, in effect, 'Quick, tell me, tell me!'

His letter was returned a month later stamped 'Deceased.' Hapgood had walked in front of a car and been killed.

This is the story that Rand told me, and it explains why I also became so excited. As everyone knows, civilization is supposed to have started in the Middle East about ten thousand years ago - say 8,000 BC - not a hundred thousand years. But Hapgood had written to Rand of 'a whole new cycle of civilization' - in other words, a civilization that existed before our present one. It sounded like science fiction. But Hapgood had always been a sober, hard working professor of history, without even a touch of the crank in his composition.

Rand had tried hard to solve the mystery. He had visited the Hapgood Archives at Yale, but left empty-handed. He had written to all Hapgood's relatives, and many of his friends, asking if Hapgood had mentioned his 'discovery' to any of them. No one could help him. And so finally, Rand had given up.

The question now was: Could I do any better?

I decided to go back to Rand's starting point: Hapgood's 1958 book Earth's Shifting Crust.

And here, in fact, Rand's results had been more encouraging. He had started with the question: Is there any proof that the earth's crust has 'slipped'?

Hapgood's answer was: Yes. The earth's crust is full of iron ore. And since the earth is a magnet, all the molecules of iron oxide point in the same direction, like tiny arrows. And since they all point towards the north pole, it is quite easy to see if the position of the pole has altered since the ore solidified, for in that case, the 'arrows' will be pointing in the wrong direction. And that is precisely what has happened. The 'arrows' in a great deal of the earth's crust are pointing too far south, which shows that the pole must have since moved north.

Or rather, to be more accurate, the pole has stayed where it is, but the earth's crust above it has slipped south, a little like pulling a schoolboy's cap down over his eyes. It has slipped south by about two thousand miles. Hudson Bay used to be right above the north pole. Then the crust slipped, and Hudson Bay - together with Canada and North America - moved two thousand miles south.

Do we know when this happened? Yes, we do. In Earth's Shifting Crust, Hapgood explains the various indicators that reveal that the great 'slippage' occurred about twelve thousand years ago, at the end of the last Ice Age.

And Hapgood also took note of the fact that Plato, in the Timaeus, had pointed to that date as the time when Atlantis was destroyed 'in a day and a night.' Plato mentions that his ancestor, the statesman Solon, (who lived around 600 BC) said that Egyptian priests had told him that Atlantis had disappeared about nine thousand years earlier. That is close enough to 10,000 BC.

Now although Hapgood never dared to admit this in his books, he had experienced a lifelong fascination with Atlantis. That was why he began to study the question: what kind of catastrophe could destroy a continent in 'a day and a night'?

A possible answer came to him when the washing machine in his kitchen went wrong. The wet laundry had bunched into a tight bundle, so that as the drier spun round, the drum began to shudder and bump until it ripped the bolts out of the floor. This led Hapgood to wonder: would the earth also shudder and bump if one of the continents (say Antartica) became lop-sided?

A quick look at the map seemed to support the idea: Antartica is lop-sided (i.e., off-center). But then a colleague who was a mathematician did the calculation, and pricked the bubble. The answer was that even if Antarctica was made of solid lead, it would not give the earth a lop-sided spin.

But now that he was thinking along these lines, Hapgood pressed on. And when he discovered the iron ore evidence, proving the earth's crust has slipped, he felt he had found the basic solution. The Atlantis catastrophe had occurred because the earth's crust had moved, creating roughly the same effect as a movement of the San Andreas fault, but thousands of times greater. Hapgood had no idea why it had slipped - it might even be because the earth had been struck by a giant meteor; all he knew was that it had.

What Rand was now looking for was some evidence that would convince non-geologists, and he stumbled on it by accident when his wife brought home from the library a book called Archaeoastronomy in Pre-Columbian America by Dr Anthony Aveni. There he found a vital clue. It consisted in the statement that there are no less than fifty religious sites in Mexico that seem to be misaligned from true north.

For example, Mexico's most famous religious site is Teotihuacan, which was once as large as ancient Rome. This has a long avenue running from north to south, now known as the Way of the Dead. But for some odd reason, it does not run precisely from north to south. It is 15.5 degrees off true north. And when Rand learned that there are 49 more sacred sites that are similarly misaligned, he began to feel he was onto something.

He took a tape measure and tried extending the Way of the Dead in a straight line. The line passed through Hudson Bay.

What this appeared to mean was that when Teotihuacan was first built, the Way of the Dead pointed due north. Then the earth's crust slipped, taking Teotihuacan and the Way of the Dead with it, but twisting sideways slightly as it did so, so it now pointed 15.5 degrees out of true north.

And the same thing had apparently happened to forty-nine other religious sites. That seems to rule out the possibility that the architect of Teotihuacan was drunk when he laid out the plans.

The implications of this insight are hair-raising. It means that Teotihuacan must have been built before the great crustal shift of 10,000 BC. And according to historians, there was no civilization as early as that.

Besides, most archaeologists agree that Teotihuacan was built much more recently. Some suggest 4000 BC, while a few even think it as recent as 150 BC.

However, that would not affect the argument as much as one might suppose. For we know that most great religious sites are built on older religious sites. Many great temples are built on the ruins of half a dozen earlier temples.

But if this is correct, it means there must have been a civilization before 10,000 BC.

It is true that Plato said that Atlantis existed in those days. But that is supposed to be a legend. What Rand seems to have proved - accidentally - is that it must have been a fact.

Now Hapgood himself seems to have believed that Atlantis was a reality. But he came to that conclusion by a completely different route.

His speculations about 'earth's shifting crust' had suggested a mechanism to explain how a continent might vanish overnight, but had done nothing towards demonstrating that it had actually happened. Then, on August 26, 1956, a broadcast from Georgetown University in Washington threw a new light on the problem.

A panel was arguing about the so-called Piri Reis map, found in the Topkapi Palace in Istanbul in 1929. A copy had been presented to the Library of Congress many years ago. Piri Reis was a Turkish admiral and one-time pirate, who in 1513 had set out to construct a map of the Atlantic Ocean, showing Europe and the 'New World.' This was difficult because America had been discovered so recently, and few maps existed. Piri Reis said he had used many old maps, including one used by Columbus.

What was amazing about his map was that it not only showed the whole coast of South America - in impressive detail - but that it also showed a fragment of Antarctica in the extreme south. It also showed certain bays on its coast - bays that could no longer be seen because they were buried under about a mile of ice.

But in 1949 an expedition had used radar to penetrate the ice and show the land underneath. And what was so startling about Piri Reis's map of 1513 was that it also showed these bays. Yet the coast of Antarctica had undoubtedly been covered in ice in 1513 - and for thousands of years before that. Core samples taken by the 1949 expedition were six thousand years old - i.e., from 4000 BC.

But although civilization existed in 4000 BC - the Middle East was dominated by the Sumerians - no one had yet invented writing. And of what use is a map without writing on it?

It looked as if there must have been an even more ancient civilization than any known to history. Plato's Atlantis began to seem less and less like a myth.

Hapgood wrote to the Library of Congress to ask if they had any more of these ancient maps. They replied they had hundreds. Such maps were called portolans, which means 'from port to port,' and were used by mariners to sail by the shortest route.

The Library invited Hapgood to come and look at their collection.

What he saw took his breath away. Among the dozens of maps laid out on trestle tables, some showed the whole of Antarctica. Yet Antarctica was not officially discovered until 1818. These maps meant that ancient mariners must have sailed all around it. But they must also have explored it from end to end, for other maps showed Antarctica without the ice, including rivers and mountains in the interior. In fact, the maps were so detailed that it seemed fairly clear they must have been made by the inhabitants of the continent, since sailors, no matter how hard-working and conscientious, could never have done it.

Other maps, that included Russia and China, made it clear that these mysterious explorers must have penetrated to every corner of the globe. But who were they? Hapgood enrolled the help of his students at Keene State College, in New Hampshire, to help him study the maps, and try to infer their dates from internal evidence.

The result were summarized in Hapgood's book Maps of the Ancient Sea Kings, subtitled: 'Evidence of Advanced Civilization in the Ice Age' (1966). In this book Hapgood concluded that the old maps proved beyond doubt the existence of a worldwide maritime civilization around 7000 BC. Its arguments were all so well-documented that no one could seriously doubt that Hapgood had proved the existence of a forgotten civilization.

What Hapgood took good care not to mention - or even to breathe - was the word 'Atlantis.' He wanted to give his fellow academics no chance to dismiss him as a member of the lunatic fringe.

Unfortunately, an appalling piece of luck meant that this happened anyway. Serious works of scholarship inevitably take a while to break out of the closed circle of academia and become known to a wider audience. And this failed to happen to Hapgood because the Piri Reis map was already regarded as the happy hunting ground of cranks. In 1960, a book called The Morning of the Magicians, by Louis Pauwels and Jacques Bergier, became a world best-seller, and it argued that the Piri Reis map showed such detailed knowledge of the coast of South America that it could only have been observed from the air by a space ship. So Hapgood already had a hostile academic audience prepared to dismiss his book as fantasy.

Worse was to follow. In 1967, the year after its publication, Hapgood was mentioned with approval in a book called Memories of the Future, by a Swiss hotelier named Erich von Daniken. Translated as Chariots of the Gods, it went on to sell more copies than the Bible. Daniken argued that all the earth's greatest monuments, from the pyramids to the statues of Easter Island, were built by visitors from outer space.

The book was full of absurdities and inaccuracies, such as multiplying the weight of the Great Pyramid by five, and arguing that the Nazca lines in Peru were built as runways for space ships, (when anyone can see that they are merely scratched on the surface of the desert.) And Hapgood was once again praised for asserting that the Piri Rei map must have been based on an aerial photograph taken by these space visitors.

With that, his last chance of being taken seriously evaporated.

Hapgood must have felt bemused. He had written two of the most important books of the mid-century, and they had both been dismissed. Earth's Shifting Crust had run into trouble because his fellow academics objected that Hapgood was not a geologist, and said he had no right to venture outside his field. But it was his field - the history of science. It should have made him world famous. Instead it was ignored or derided.

After a period of discouragement, during which he retired from Keene State, he took up again the problem of the whereabouts of Atlantis. On the Piri Reis map he had noted an island in the mid-Atlantic, opposite the coast of Venezuela. It has now vanished, but since it appears on two other ancient maps, it cannot have been a mistake. Nowadays all that remains there are two tiny islands balanced on the mid-Atlantic ridge. They are known as the islands of St. Peter and St. Paul, and are obviously the tips of submerged mountains. These, Hapgood was convinced, was all that remained of Atlantis. He continued to hope that someone would back his hunch and go and look. He had even approached President Kennedy in 1963, hoping he would send a naval craft to explore the spot. Kennedy agreed to see him - Hapgood had once worked for the Secret Service - but was assassinated at Dallas not long before the appointment was due. And when attempts to interest Walt Disney and Nelson Rockefeller also came to nothing, Hapgood gave up.

This explains why he never took Rand Flem-Ath's Antarctica theory seriously. He was convinced he had already pinpointed the site of Atlantis. Besides, his interest had already moved to even stranger and more fascinating matters - as we shall see.

So Hapgood died in 1989, after being hit by a car. But others had taken up the quest where he left off. One of these was a maverick Egyptologist named John West, who became convinced that the Sphinx had been eroded by rainfall rather than by wind-blowm sand. He persuaded an academic geologist, Robert Schoch, to go with him to look more closely, and was delighted when Schoch agreed that the Sphinx was water-weathered. Their result was announced at a conference of the Geological Society of America in 1993, and many geologists were convinced.

Schoch suggested that his evidence indicated that the Sphinx was built in 7000 BC, not 2500, as is usually accepted. This date, of course, is the same as that suggested by Hapgood for his worldwide maritime civilization.

By this time, I had also become fascinated by the controversy, my interest sparked by a request by film producer Dino de Laurentiis to write a film script about Atlantis. As a result I met John West in New York, and learned that two more writers were working on the same subject. One, John said, was a Canadian librarian called Rand Flem-Ath, who had written an unpublished book, arguing that Atlantis was Antarctica. The other was Graham Hancock, who was writing a book called Fingerprints of the Gods, whose thesis was that civilization is many thousands of years older than historians believe.

I contacted both, and when Rand Flem-Ath told me he could not find a publisher for his book, When the Sky Fell, I offered to read it and write a preface. This seems to have done the trick, and the book was accepted.

Graham Hancock also sent a warm reply, followed by the typescript of the original version of his book . My first reaction was to wonder whether Graham had left me anything to write; but I soon noted important points that he had left uncovered – like ancient India - and returned to planning my own book, then called Before the Sphinx.

Graham's book appeared in the spring of 1995, and became an instant bestseller, making him a millionaire. Rand's book, which had appeared a few months earlier, was also warmly received, although it made nothing like the same impact. My own book, now called From Atlantis to the Sphinx (because the publisher insisted on Atlantis in the title) came out in May 1996, and sold out its first printing before publication.

One result was that in September 1996 I was invited to a symposium on ancient knowledge at the University of Delaware. Graham Hancock, John West and Robert Schoch were there. So was Robert Bauval, a Belgian engineer, whose remarkable book The Orion Mystery had thrown a completely new light on the Atlantis question. In the Cairo Museum, Bauval had seen an aerial photograph of the pyramids of Giza, and noted idly their oddly asymmetrical arrangement on the ground. The two main pyramids - of Cheops and Chefren - were arranged, so to speak, 'in line.' But the smallest pyramid, of Men-kau-ra - had slipped well below the line. That seemed odd for builders as obsessed with symmetry as the Egypian were.

Some time later, camping in the desert, he noticed that the three stars of the belt of Orion have precisely the same asymmetrical structure, with the third - and smallest - 'out of line.' And since he knew the ancient Egyptians regard earth as 'heaven's mirror,' he wondered if the Giza pyramids were built as a 'reflection' of Orion's Belt.

But in our own time, this reflection is not exact. Due to a phenomenon called 'precession of the equinoxes,' it is as if someone has twisted the mirror slightly askew.

He consulted his computer, and discovered that the last occasion when the stars and pyramids 'reflected' each other was about 10,500 BC. But why should the ancient Egyptians have wanted memorialize that particular date?

Then it struck him - Atlantis! He recalled the legend that ancient Egypt was settled by Atlantis survivors. That surely had to be the answer to the Orion mystery.

Rand and his wife Rose were also present at the Delaware Symposium - Rand was about to deliver his first public lecture.

I took to the man immediately. He was a bullet-headed, bearded fellow who gave an impression of strength and determination. Rose, who was writing a novel, was a charming and intelligent girl, and I could see they formed an ideal partnership.

That weekend I was approached by a TV producer named Roell Oostra, who wanted to know if I would act as the presenter of a program called The Flood, arguing that the 'great flood' recorded in so much ancient literature was the catastrophe that destroyed Atlantis. It would involve traveling to Mexico, Egypt and South America. Being a typical Cancer, I hate travel, but my wife Joy loves it, so I agreed to go, so long as I could take her.

The first leg of the journey, to the La Brea Tar Pit in Los Angeles, then to the Library of Congress in Washington, would only take a weekend, so I went alone. And, to my delight, I met Rand again - at the Library of Congress - and was confirmed in my view that he was one of the nicest people I had ever met. As I finally said goodbye to him - on my way to the airport - I felt a twinge of sadness at the thought that I might never see him again.

The program took six months to make, and was broadcast in the spring of 1998. I wondered if Rand had seen it, and emailed to ask him. It seemed Roell had forgotten to send him a copy, so I sent him one myself. And a few weeks later, I received a letter from Rand asking me if I would be willing to work with him on a book called Blueprint for Atlantis. His letter was accompanied by a magazine article he had published about his theory.

The theory, quite simply, was that before the great Atlantis catastrophe, scientists had become aware, from volcanic eruptions, that the earth's crust was beginning to split and drift. Aware that this had happened before, they constructed 'markers' at various points along the 'lines of strain,' to be able to anticipate disaster. These 'markers' were laid out symmetrically all over the earth's surface, for the Atlanteans had a worldwide civilization.

After the catastrophe (which Rand, like Plato, dated to 9600 BC), later civilizations discovered the 'markers,' and assumed they had been placed there by gods or superhuman beings. So they treated them as sacred sites, and built temples on them -- which is why sacred sites are placed so symmetrically all over the world.

This latter was the essence of Rand's theory. He concluded that sacred sites are built at certain definite latitudes and longitudes, very often in a grid pattern. They are evidence that the earth was once covered by a great pre-Atlantean civilization.

The latitudes and longitudes are in round figures, like 50, 40, 25, 10, and so on. Our modern measurements disguise this, for they might give the longitude of a sacred site as 127.153. But all we have to do is to measure from the Great Pyramid, which Rand believes the ancients used as their line of 0 degrees longitude, and we immediately see that other sacred sites - say Delphi or Teotihuacan - fall into neat round figures.

I was totally convinced, and agreed to collaborate. But the first task was to find a publisher, and get the book commissioned. I approached Bill Hamilton, who was Graham Hancock's agent, and he suggested that we should do a series of 'presentations.' Rand and I would go to various publishers, and give short lectures, accompanied by slides, presenting the 'Blueprint theory.' Then we would accept the offer of the highest bidder.

This would involve Rand coming to London. After a few days in my home in Cornwall, during which we rehearsed our 'act,' we went to London, and did our presentations to half a dozen publishers. The offers were larger than we had anticipated, and we were both well satisfied. Then we went to New York, this time under the guidance of my American agent Al Zuckerman, and we did it all over again. Once more, the offers were higher than we expected. So I began writing in December 1998. The delivery date was January 2000, which seemed far too soon. But we had such a vast amount of material that I felt that all I had to do was to arrange it in an orderly fashion.

There was still, of course, one major problem: Hapgood's letter to Rand asserting that he was convinced that civilization was more than a hundred thousand years old. Unless I could prove that, the arch would lack a keystone. But I didn't let this worry me too much - there was almost a year to go.

I began by contacting as many of Hapgood's friends and relatives as I could locate. These included his cousin Beth, and his friends Elwood and Daria Babbitt. But they were unable to help me track down the missing notes that might explain the claim of a 'hundred-thousand-year-old civilization.' These notes were obviously essential, for without them, the book would remain a kind of unfinished detective story, like a Poirot novel without its last chapter. But as the weeks turned into months without result, I felt my optimism begin to fade.

In spite of this, I worked on steadily. I was doing the actual writing, while Rand provided me with material. I also made interesting discoveries of my own - such as the amazing story of the 'Nineveh number,' which evidence shows to have been discovered about sixty-six thousand years ago.

And then suddenly, my luck changed. It happened on February 28, 1999.

The previous day I had received an email from a friend of Hapgood's, telling me he had mislaid the material I had sent him; so I sent it again. That morning - Sunday - he had written to thank me, and mentioned, as a postscript, that it might be worth contacting a retired academic who lived in New Hampshire, who had also known Hapgood. Accordingly, when I had finished my day's work, I gave him a ring.

The phone was answered by a man with a pleasant New England voice. I explained about Hapgood's last letter to Rand, and said that I was trying to find how Hapgood had become convinced that civilization was more than a hundred thousand years old.

The man on the other end of the line said: 'Oh, sure, that was me.'

The admission was so casual that it took a moment to sink in. 'What was you?'

'I gave Charlie Hapgood the information that persuaded him that civilization was a hundred thousand years old.'

My heart skipped a beat.

'And how did you do that?'

'Well to begin with I told him that Neanderthal man had calculated the equivalent of Thom's megalithic yard, and Cro-Magnon inherited it from him.'

'And what made you think so?'

'Lot's of things. To begin with, the La Quina disc, the Disque en Calcaire.'

I was dizzy with relief and happiness. What would Rand say when I told him I had solved our problem, and now had the ending of our book?

What Carl said made complete sense. An old friend of mine, the psychologist Stan Gooch, had written a book called Cities of Dreams (1989), arguing that Neanderthal man was a highly intelligent creature who had created his own civilization; his arguments totally convinced me. And Carl's mention of the megalithic yard also made sense. The astronomer Alexander Thom had claimed that megaliths all over the world had been built using the same basic measurement, the megalithic yard. What amazed me was Carl's assertion that Neanderthal man had used it too.

But during that conversation, it soon became clear that there was an unforeseen problem. Although our talk lasted two hours, I couldn't understand more than one sentence in ten. Like certain brilliant people, whose heads are crammed with knowledge, Carl was unable - or unwilling - to express himself clearly and to the point. It was obvious that when I asked him a question, he wanted to say thirty things at once, and it was like a crowd trying to push through a narrow doorway. So I sympathised with Carl.

It was almost impossible to stop him talking. When I said I had to go, he just said 'Oh sure,' and went on talking.

This was real luck, a proof that the gods were on my side. When I got home and consumed half a dozen oysters and several glasses of Chablis, I felt as if I was celebrating a win on the lottery.

I could hardly wait to tell Rand that our troubles were over, but I restrained myself from telephoning - largely because the time gap between England and Vancouver was nine hours - and instead wrote him a long email.

His response was disappointing. He was clearly unconvinced by Carl, and made no bones about calling him a fraud and a con man. This staggered me. What possible motive could Carl have for pretending to be a friend of Hapgood's and telling me a string of lies? He hadn't asked me for money or anything else. He was obviously an intelligent man, possibly a man of genius. Why should he tell me anything but the truth?

But Rand pointed out that he had found no reference to Carl in Hapgood's papers at Yale. I replied that Hapgood's last letter talked about 'certain recent discoveries. Probably Hapgood had not had time to write about them yet. But Rand simply would not have it.

Rand's attitude struck me as unreasonable. He had not spoken to Carl, and from what little I had told him, had no evidence whatever to accuse him of being a fraud. Rand gave me the impression that he simply did not want a third person involved in this Hapgood mystery; he felt that three was a crowd. It was all, so to speak, a matter of intellectual 'territoriality.'

I had arranged to ring Carl back in two weeks, and to install a recording machine that would play for an hour. But this proved to be quite inadequate. Carl simply talked non-stop for an hour, and when I told him the tape had ended, just went on talking - for almost another hour.

But at least he said some fascinating things. He said, basically, that the antiquity of civilization was proven by its measures. And if these measures could be shown to date back to the La Quina disc, carved by a Neanderthal man a hundred thousand years ago, then the point was proven. I had to agree. He also talked about linguistic evidence in Greek, Semitic, Sumerian and Sanskrit, and cited the exact words. I had never come across a man of such immense erudition. His theory was incredibly difficult, involving music, planetary distances, archaeology and atomic numbers. His articles – of which he sent me several - ranged from the Great Pyramid, ice age art, and Chaco Canyon to alchemical symbolism.

During subsequent calls, I found that one of the most difficult things to endure was the sheer length at which he talked. I simply didn't have time to listen for two or three hours. I suggested that he should ring me, so that perhaps the awareness of a mounting phone bill would act as a brake, but it made no difference.

The explanation was simple. He was lonely, and wanted intellectual companionship. He felt - probably rightly - that he had one of the great minds of the age, and he wanted to share his abundance. Living in a high-rise in the city must have been very frustrating.

But I soon realized that I could not simply present him to the reader as an unrecogniszd genius, for some of his views left him wide open to the accusation of being a crank. He not only accepted the reality of the notorious 'Face on Mars,' but believed it had been created by human beings, and that one of the satellites of Mars was some kind of artifact. I could understand his point, for I was preparing to write a book on UFOs called Alien Dawn, and had ended by recognizing that total skepticism is simply a sign of that odd panic that rationalists experience in the face of mysteries that seem to contradict all their dogmas. But Carl's views seemed to go too far. And it was impossible to get him to explain himself.

Not that he was evasive - on the contrary. If I succeeded in interrupting the torrent long enough to ask him a question, the answer was so full, so comprehensive and encyclopedic, that it was impossible to digest, and a request for elucidation simply led to another equally long and mind-boggling discourse.

Just as I was beginning to wonder if Rand could be right, and Carl might be a fraud, I was confronted with evidence of his genuineness. An old friend, Andy Collins, came past our house on his way to see the eclipse in Cornwall, and when he overheard me telling someone in the pub about Carl, he said he knew him. Fascinated, I asked for details. It seemed Andy had met Carl at a London party, and that Carl had quickly monopolized the conversation, until he held the whole room enthralled. Andy agreed that Carl could be described as brilliant.

He mentioned a friend of his who lived in the Midlands, who had been on an archaeological expedition with Carl in Mexico. I rang him up, and as a result received some more interesting first-hand information about Carl. As a traveling companion he could apparently be exacting, obsessive, and infuriating. In spite of which he was - as I had deduced from those long phone conversations - erudite and brilliant.

Was he, I asked my informant, strictly truthful?

There was a long pause before he answered: 'Not exactly. He was too imaginative for that. He would say things that you realized weren't quite true. But you sensed he had got carried away, and believed them himself.'

I understood perfectly - all writers do. As Shaw said, 'We all bid for admiration with no intention of earning it.' Carl was a brilliant thinker, but he was unrecognized; no wonder he was inclined to bid for a little of the admiration he thought he deserved.

Naturally, Carl was enraged and deeply offended when he learned that Rand thought him a fraud. I didn't blame him; in this matter I thought Rand was a dogmatic fool. Carl was even more enraged when he learned that Rand had been making discreet enquiries about him, and had concluded that none of the qualifications he claimed were genuine.

All this caused Carl's attitude to me to cool perceptibly. One day, he sent me an email accusing Rand of inventing the whole story of Hapgood's interest in Atlantis. In reply, I printed up the first chapter of the book I was writing, called The Atlantis Blueprint, and sent it to Carl, with chapter and verse, with its detailed account of the material Rand had found in the Yale archive. That, I thought, would leave him in no possible doubt.

A few days later, looking through some old emails, I realized that I had sent Carl this first chapter many months before. He had simply failed to read it. Carl's problem was that he was so obsessed by his own ideas that he simply was not interested in anyone else's.

In his reply, Carl explained that the whole quest for Atlantis was a waste of time, because Atlantis was not a place, but the name of a mother goddess encoded in Gematria. And when he added that if I persisted in my nonsensical claims, I could remove his name from the book, I saw we had reached the end of the line. Since the whole book was about Atlantis, to ask me to remove all mention of the word was tantamount to asking me to tear it up.

I was unhappy at the idea of removing Carl from the book, for it was obvious to me that he had to be the person who had told Hapgood about the 'hundred-thousand-year-old civilization.' It was obvious to me that Carl had told Hapgood about the disque en calcaire and the ancient measurements, and that Hapgood had seen as clearly as I had that the answer had to lie in Neanderthal man.

There seemed to be two possible solutions. Carl's ideas about Neanderthal civilization had already been put forward by Stan Gooch in Cities of Dreams, while his arguments about ancient measurements could be found in a remarkable work called Historical Metrology (1953), by A. E. Berriman. I could use these instead of citing Carl - but this would mean dropping the whole point of the book: telling the reader where Hapgood had come upon the idea of the hundred-thousand-year-old civilization.

The alternative was to remove Carl's name from the book and use a pseudonym. Carl himself could hardly object, since he insisted that his name should be removed. And provided I made it quite clear to the reader that 'Carl' was not his real name, I could not be accused of infringing his intellectual copyright.

At this point, it suddenly looked as if the whole matter might be amicably resolved after all. Carl offered to allow us to use his name, provided we devoted a full chapter to his ideas, and several appendices on such questions as primitive art and the moons of Mars.

Rand almost had apoplexy when I suggested it. He was not, he said, going to allow his book to be hijacked by this fraud and intellectual con man. The very idea made him breathless with rage.

The absurd thing was that Rand and I were now basically in agreement about Carl. Even I could see that Carl had an innate habit of embroidery and exaggeration. This was proven by an episode that finally led me to break off contact. In an email, Carl had mentioned that he had an inscribed copy of Hapgood's Maps of the Ancient Sea Kings, and he even quoted the fulsome inscription. I was delighted, for this would obviously leave Rand no alternative than to admit he was wrong. So I asked Carl to fax me the title page with its dedication. When there was no reply, I faxed the girl who acted as his unpaid secretary, and asked her to do it.

The result was an explosion. I had tried to go over his head - Carl declared through an intermediary - and he would not communicate with me any more.

The inference seemed obvious. The inscribed copy almost certainly did not exist, otherwise Carl would have been only too delighted to do as I asked, and prove once and for all that he and Hapgood had been friends.

Yet even this did not alter my view that it was Carl who had told Hapgood about the 'hundred-thousand-year-old civilization.' There was simply no logical alternative - no one else who knew that much about Neanderthal man and ancient science. In that, at least, Carl had to be genuine.

Which was a relief. Now I need have no reservations about telling it all, and handing Poirot his moment of triumph in the last chapter.

But Rand objected to this too. He was not, he declared, going to have Carl in the book at all, pseudonym or not. He felt that introducing a mystery informant into last chapter would only divert attention from his own ideas.

It was now mid-December; and Atlantis Blueprint was due for delivery in the New Year. I was writing for eight hours a day, and felt like a racing driver, hurtling along at top speed, while his co-driver tries to wrest the steering wheel out of his hands. Rand seemed to have forgotten that he had no right to overrule an idea I considered central to the book. We were equal partners, and had a collaboration agreement. This stated that he had no right to tell me what I could or could not include. He could insist on something of his own going into the book even if I objected, but I could also put anything I wanted into the book whether he agreed or not. I told him that I would compromise to the extent of removing Carl's name. But if he still refused, then I would simply withdraw from the book and repay my part of the advance.

It was late afternoon when I wrote Rand an email to this effect. Because I was feeling totally frustrated and exhausted, I decided not to send it that day, but to sleep on it first. But when I looked for it half an hour later, it had gone. Another email had come in, and mine had been sent automatically.

But it seemed to produce the effect I had hoped for. Rand replied the next day saying that since he could not afford to repay his part of the advance, he would accept that I should include the story of Carl. So we ended in agreement after all.

That made me feel conciliatory, so I decided to reduce the Carl section to an absolute minimum. And The Atlantis Blueprint was delivered on time after all.

About six weeks later, I received the edited typescript. On the whole, I was well satisfied with it. I was not entirely happy with the last chapter; it was too long, and I could see signs of the rush and strain in which it had been written. But the section on Carl struck me as balanced, truthful, and above all, brief, and I felt that Rand had every reason to be pleased with it. It answered the reader's question about the hundred-thousand-year-old civilization, but placed it all in a far wider context. And the book, I felt, had turned into one of the best I had ever written. I would be amazed if it did not get on the best seller list.

Rand and I had communicated little between that time and the day in late July when I received the proof. And when the proof finally arrived, I left it for twenty-four hours while I finished another piece of work.

But when I finally looked at it, I was baffled. This was simply not my book. Its final two chapters - my own conclusions about ancient civilization and cosmology - had been removed completely. Carl, of course, had been removed too.

My first thought was that it was a simple mistake - someone had somehow forgotten to send its last two chapters to the printer. Then I looked more closely and saw this could not be true. The remainder of the book had been rearranged like a jigsaw puzzle. Every reference to Neanderthal man had been excised, and in one paragraph, had been altered to 'people like us,' implying that I was talking about Cro-Magnon man. There could be no doubt: this butchery was deliberate.

I rang my publisher, Alan Sampson, at Little, Brown. He professed to be baffled, and suggested I ring Caroline North, the editor responsible for the final version.

I rang Caroline. She told me that she had indeed made the changes - having been assured by Rand that they were made with my full knowledge and permission.

So I sent Rand an email, asking him what had prompted him to this act of incredible treachery. His reply was a scream of rage which began: 'LIBEL!' and demanded how I dared accuse him in this outrageous manner. He had, he claimed, submitted every tiny alteration to me, even to commas, and was it his bloody fault if I didn't even do my fellow author the consideration of reading his suggestions?

This clearly implied that he had submitted suggestions for change. But when and where? Joy and I spent two days looking through all the emails Rand had sent in the past six months. We could find no such 'suggestions.'

Finally, I was forced to write to Rand: 'What was the date of the 'suggestions' email?' He replied: 'June 1.'

I turned to his email of June 1 and found a ten-page communication headed: 'Last of it.' It was a harmless-looking document saying he had tracked down the remaining footnotes, and the book could now be considered finished.

He went on to tell me that they had just had a film crew there, 'which was a really nice break,' then had lost his dogs when a neighbor let off fireworks. He had spent hours searching the woods. This story was told in considerable detail. Then there were comments on problems at his library. And now, he concluded, he was looking forward to relaxing and getting back to other work, and he was sure I felt the same.

It concluded: 'With affection, Rand.'

This was followed by 'Last Appendices, notes, etc,' page after page after page of latitudes and longitudes, which of course I had skipped.

But as I now read it, I discovered that, after five pages, and without even a single-line break in the text, he went on: 'The impact of this book could be improved with a slight restructuring.'

Then, after a page of suggestions, and again without a line break, back to more appendices and notes - page after page.

No wonder I had not seen his suggestions for a 'slight restructuring.' He had camouflaged them in a great mass of stuff he knew I would not read. Moreover, I had sent back a reply: 'Marvelous! Congratulations!' He must have chuckled with satisfaction.

I could, of course, have refused to allow this travesty to go into print, and in all probability, Rand would have been liable for the costs of reprinting the book. But this would undoubtedly cause major problems for my publisher.

It was Joy who pointed out to me that Rand's deception was described in Chapter Five of our book, where he talks about reading a children's book called The Magic of Alkazar, which explains about the basic magician's techniques of misdirection. 'The audience will always look where the magician looks. The audience will treat as important what the magician treats as important, and as unimportant what the magician treats as unimportant.' Rand's 'sting' had been a classic case of misdirection.

And so, in due course, The Atlantis Blueprint was published in its hacked and truncated form. No reviewer seemed to notice that it failed to fulfill its promise to explain Hapgood's hundred-thousand-year-old science.

But I was not surprised when it failed to cover its advance.

I was saddened that the friendship with Rand had ended so badly, for I had been genuinely fond of him. Then I was struck by a consoling thought. Since he had hacked out so much of the book, he had left me plenty of material on which to base a sequel. And in fact, in the two years since publication of Atlantis Blueprint, I had come upon important material that strengthened my argument – for example, that Neanderthal invented superglue, and played music on a bone flute, using the same musical scale we do.

BACK