Underworld: The Mysterious Origins of Civilization

by Graham Hancock

  Moreover, now that I had the maps at virtually millennium intervals, it was possible to pinpoint periods when the extent of the ongoing loss of land to the sea had been particularly rapid and to note any correlation between these and (1) John Shaw’s cataclysmic chronology for the post-glacial floods; (2) the relevant mythology; and (3) the accepted dates for the so-called ‘Neolithic revolution’ in India (i.e., the beginnings of food production at Mehrgarh and other sites).

  The north-west

  In the north-west, around Gujerat, the maps show that a huge land area was inundated between approximately 17,000 and 7000 years ago – an area contiguous to the domain in which archaeologists believe that the first recognizable roots of the Indus-Sarasvati civilization were planted during the last three millennia of the same period. As we saw in chapter 7, the submerged lands are at their most extensive around the modern Gulf of Cambay – south of which the map for 16,400 years ago shows an extensive depression, very likely to have been filled with a large freshwater lake, bounded by a further tract of land at least 100 kilometres wide and beyond that the Arabian Sea.

  The next map in the sequence – 13,500 years ago – reveals that major changes occurred during the intervening 2900 years. The landmass around the Gulf of Cambay was much reduced in area and a large island, almost 500 kilometres long and 100 kilometres wide at its midpoint, was marooned off-shore in the Arabian Sea. Between the island and the mainland a marine strait, also 100 kilometres wide in some places, opened up through the basin of the former freshwater lake.

  These rather dramatic land-losses between 16,400 and 13,500 years ago correlate well with the first of John Shaw’s proposed episodes of global superfloods, which falls midway through the period at around 15,000 years ago.

  Over the next 6000 years – between 13,500 years ago and 7700 years ago – the maps show that the large off-shore island and the coastal strip masking the outline of the Gujerat peninsula were continually nibbled away at by the rising seas, but that these events were gradual, extended over many lifetimes, and would have been unlikely to have been perceived as cataclysmic. As late as 7700 years ago the Gulf of Cambay was still the ‘pleasant valley’ that it had been, uninterrupted, since at least the Last Glacial Maximum and the island lying off-shore, though reduced, was still of formidable size – perhaps 300 kilometres in length and close to 80 kilometres wide.

  This pattern for the Gujerat area, therefore, does not correlate well with the second of John Shaw’s proposed episodes of global superfloods around 11,000 years ago. Nor does it suggest a motive for any memorable panic-migration of flood refugees out of this area at any point during this period – which straddles the supposed date of around 9000 years ago for the first settlement of Mehrgarh.

  What happens next, however, provides a close match to Shaw’s chronology of around 8000 years ago for the third flood. The maps for 7700 years ago and 6900 years ago show that in this relatively short period of 800 years the large remnant island below the Gulf of Cambay was completely wiped off the map and the Gulf itself was fully and permanently inundated to its modern extent. For any hypothetical coastal culture that had been forced to retreat and compact into the Gulf’s pleasant valley over the previous 6000 years, or that had lived on the island, it goes without saying that these events would have been more than cataclysmic.

  They would have looked like the end of the world.

  The south

  As we would expect, the inundation maps for 21,300 years ago and 16,400 years ago show that few significant coastline changes took place in the south during the five millennia or so of the Last Glacial Maximum. At that time Sri Lanka was joined to the mainland, as we have seen, and ‘a substantial integrated area – an entire sub-region of India’ that is today submerged59 – was above water in the south and the south-east (and indeed all along the Malabar coast in the west also). This lost antediluvian realm accords extremely well in a general sense with the central claim of the Kumari Kandam tradition that a large landmass did exist around the south of India in ancient times and that it was swallowed up by the sea in a series of floods.

  The maps of 21,300 and 16,400 years ago reveal the full extent of the continental shelf that was exposed during the Ice Age, but a specific feature of great interest is the snout-shaped peninsula shown to have extended approximately 150 kilometres southwards into the Indian Ocean below modern Kaniya Kumari. As the reader will recall, such a peninsula in exactly this location is spoken of in the Kumari Kandam tradition:

  In former days the land … extended further south and … a mountain called Kumarikoddu, and a large tract of country watered by the river Prahuli had existed south of Cape Kumari. During a violent irruption of the sea the mountain Kumarikoddu and the whole of the country through which flowed the Prahuli… disappeared.60

  The peninsula that Glenn Milne’s calculations place on the inundation maps is not as large as the one described in the tradition (which was said to have been ‘ 700 Kavathams’, about 1500 kilometres, in length). Still it is there – precisely where the Kumari Kandam tradition says it should be, and in the correct time-frame. Moreover, the maps show another antediluvian landmass that has also for the most part disappeared beneath the waves standing in the open ocean to the south-west – the greatly enlarged Maldive islands as they looked at the Last Glacial Maximum.

  What if the civilization of Kumari Kandam had been partially based along the coastal margins of southern India and Sri Lanka and partially on the antediluvian Maldives archipelago? If so, then the idea that Kumari Kandam once extended 1500 kilometres to the south of Kanya Kumari does not seem so far-fetched. Nor does the notion that a civilization that had once existed in this area could have been destroyed by recurrent cycles of catastrophic floods.

  The tradition says that the last of these floods occurred 3500 years ago (supposedly the flood that destroyed the Second Sangam at Kavatapuram), and the one preceding it 7200 years ago (supposedly the flood that destroyed the First Sangam at Tenmadurai). In addition N. Mahalingam has cited further Tamil sources that speak of earlier floods: one around the date of foundation of the First Sangam, approximately 9600 years ago, one just over 16,000 years ago and the earliest 18,000 years ago.61

  Once again there is a good general correlation between what scientists now know about the meltdown of the Ice Age (particularly the episodic and recurrent nature of the post-glacial floods) and what the Kumari Kandam tradition claims was happening in the world in precisely the same period (episodic and recurrent floods). There is by no means one-to-one agreement on the dates at which particularly severe inundations occurred – as is to be expected given the margins of inaccuracy that surround the estimating processes used by both Shaw and Milne, not to mention the scope for error and exaggeration in the tradition itself. Still, there is more than enough agreement on the general course of events to give us pause for thought. After all, how many times can we reasonably cry ‘coincidence’ when the medieval Tamil ‘fabulists’ keep on getting their palaeogeography right? Or did they in fact – as Shivaraja Pillai asks sarcastically – ‘come upon some secret archive which had escaped the deluge’?62

  Glenn Milne’s inundation map for 13,500 years ago shows a dramatic change in the south Indian landscape since the previous map of 16,400 years ago: the coastal margins have been greatly reduced and the peninsula below Kaniya Kumari has been severed by the sea, leaving an island off-shore. In the Indian Ocean to the south-west the land area of the antediluvian Maldives archipelago has been reduced almost by half.

  The map for 12,400 years ago shows little significant change, but in the map for 10,600 years ago the island to the south of Kaniya Kumari has been reduced to a dot, the Maldives have been further ravaged, and, for the first time, a neck of sea is shown separating Tuticorin on the mainland and Mannar in what is now Sri Lanka. This incursion seems very close to what is described in the Sri Lankan myth of the flooding of Ravana’s kingdom (said to have extended between Tuticorin and Mannar ‘in a former age�
�).63 Moreover, the timing – between 12,400 and 10,600 years ago – coincides with Glenn Milne’s date for the submersion of the U-shaped structure at Poompuhur and accords well with the second of John Shaw’s episodes of post-glacial flooding around 11,000 years ago.

  The map of 8900 years ago shows further minor erosion all around the south Indian coastal strip and a deepening of the marine incursion beyond Tuticorin and Mannar into what is now a bay beneath the war-torn Jaffna peninsula. However, the Palk Strait was still dry land 8900 years ago and, though much diminished in size, the land-bridge connecting Jaffna to the mainland was still in place at that date (and indeed was to remain there for another thousand years).

  On John Shaw’s estimates, the third of the three great episodes of post-glacial flooding was unleashed on the world’s oceans around 8000 years ago – and we have seen how this correlates well with what happened at around that time when the Gulf of Cambay and neighbouring areas of the north-west of India were rapidly inundated. In the south-east the inundation maps show that in the same period between 7700 and 6900 years ago there was also significant further inundation of the Maldives, while the land-bridge between Sri Lanka and Tamil Nadu, which had clung on for so long, was at last swallowed up by the sea -leaving India looking very much as it does today.

  Occam’s razor

  What are we to conclude about the Kumari Kandam myth?

  In some respects there is no doubt that it has proved eerily, stunningly accurate. On the other hand much of it sounds wildly improbable and in places obviously ‘manufactured’. For example, when one studies the way numbers are used in the myth (something that I have not sought to tax the reader with here) certain obvious patterns emerge that are more suggestive of a mathematical game, or code, than of true reports of the number of members of, or the number of royal patrons of, or the duration of this or that Sangam.

  It will be recalled that the durations of the three Sangams were said to be 4440 years for the First Sangam, 3700 years for the Second Sangam and 1850 years for the Third Sangam.64 It is obviously not an accident that each of these numbers is a multiple of 37 (120x37 = 4440; 100x37 = 3700; 50x37 = 1850).65 What the significance or purpose of this pattern is I cannot begin to guess, but it means that the chronology of the myth is suspect and cannot be treated as a reliable historical record.

  Still, it does not follow from this and other criticisms that the whole myth must be tossed in the dustbin of history and forgotten – as it has been by most scholars. Although wildly out of line on some of the details and dates, the myth is right in the broad sweep. It is right that India’s Dravidian peninsula was formerly much bigger than it is today. It is right that a series of huge deluges occurred over a period of several thousand years and that these swallowed up the antediluvian lands in stages. And the myth selects the correct epoch – smack in the middle of the post-glacial floods around 11,600 years ago – in which to set its flood story.

  Besides, whatever one thinks of myths (and most historians and archaeologists regard them as useless to scientific inquiry)66 there is the awkward and inescapable archaeological fact of the U-shaped structure 23 metres underwater and 5 kilometres off-shore of Poompuhur – a structure that is ‘11,000 years old or older’.67 Isn’t the most parsimonious way to explain its presence there the very one that the myth itself provides – namely, that a civilization of former times once flourished in this region but was swallowed up by the sea?

  I could only learn more by diving.

  12 / The Hidden Years

  The period dreadful for the universe has come. Make for thyself a strong ship, with a cable attached; embark in it with the Seven Sages and stow in it, carefully preserved and assorted, all the seeds which have been described of old …

  Satpatha Brahmana

  An epoch of spectacular geological turmoil occurred at the end of the last Ice Age, with the most dramatic effects registered in a series of cataclysmic floods that took place at intervals between roughly 15,000 and 7000 years ago. Is it an accident that this same 8000-year period has been pinpointed by archaeologists as the very one in which our supposedly primitive forefathers made the transition (in different places at somewhat different times) from their age-old hunter-gatherer lifestyle to settled agriculture? Or could there be more to ‘the food-producing revolution’ than meets the eye? After all, most scientists already recognize a causative connection between the end of the Ice Age and the supposed beginning of farming – indeed an unproven hypothesis that rapid climate changes forced hunter-gatherers to invent agriculture presently serves as pretty much the sum of conventional wisdom on this subject.1

  But there is another possibility. Nobody seems to have noticed that in the general vicinity of each of the places in the world where the food-producing revolution is supposed to have begun between 15,000 and 7000 years ago there is also a large area of land that was submerged by the post-glacial floods between 15,000 and 7000 years ago:

  We have seen that this is true for India, one of the world’s ancient agricultural ‘hearths’,2 which lost more than a million square kilometres in the south and the west and, most conspicuously in the north-west, at the end of the Ice Age.

  It is true for China and for south-east Asia, both important centres of palaeo-agriculture. Immediately adjacent to them, but now under as much as 100 metres of water, lies the Ice Age continent of Sundaland. Prior to its final inundation of about 8000 years ago, this consisted of more than 3 million square kilometres of prime antediluvian real estate extending from the Malaysian peninsula through what are now the Indonesian islands and the Philippines. Taiwan was incorporated with the Chinese mainland and northwards from there the coast expanded almost 1000 kilometres to the east to fill what is now the Yellow Sea and incorporate the Korean peninsula fully with the mainland.

  It is true for the so-called Fertile Crescent – the prime agricultural ‘hearth’ of the Middle East, centred around lands watered by the Tigris and Euphrates rivers, that forms a rough semi-circle through parts of modern Israel, the Lebanon, Syria, Turkey, Iraq and Iran and ends up near the Persian Gulf. For not only was the Gulf previously dry – and flooded at the end of the Ice Age, as we saw in chapter 2 – but a glance at the wider map also shows several other inundated areas near by in the Indian Ocean, the Red Sea and the eastern Mediterranean.

  And it is true for Central America, where agriculture is thought to have sprung up spontaneously, independent of developments in the Old World. Off the Gulf of Mexico, the Yucatan, Nicaragua, Florida and Grand Bahama Banks were imposing landmasses during the Ice Age that were swallowed by the post-glacial floods around 7000 years ago. Evidence from Mexico and Panama, published in July 2001, indicates that ‘agriculture in the Americas began around 7000 years ago’. It is notable that: ‘On the Gulf coast pollen evidence suggests that forest was being cleared around 5100 BC and domesticated maize plants were being grown only a century later … The San Andres site near the famous Olmec centre of La Venta showed that maize had been introduced and grown in a region of beaches and lagoons.’3

  My curiosity about coincidences like these developed as I researched Underworld - because the sudden appearance of village farming communities at the end of the Ice Age was the first step on the road to modern civilization (so the stakes in this inquiry are high), and because the Ice Age lands that went under the sea cover an area of more than 25 million square kilometres of the earth’s surface where, for obvious practical reasons, almost no archaeology has ever been done (so important evidence could very easily have gone undetected). Since many of the coastal lands that were inundated would have offered desirable refugia from inhospitable and unpredictable Ice Age conditions, the possibility surely has to be considered that the real story of the origins of food production and of civilization may yet await discovery because the evidence is underwater.

  I decided to explore this neglected possibility with all the resources at my disposal, knowing when I did so that it would commit me to an exhausting and expensive
schedule of travel and diving – much of which might prove fruitless – and that I would have to enter arcane areas of inquiry, ransack obscure libraries and rack my brains on uncompromising sciences if I was to have any hope of success.

  Long shot

  I needed a good research assistant and in August 2000 I found one – Sharif Sakr, who has proved to be the very best of the many good researchers I have worked with over the years. Right at the beginning, I asked Sharif to find me an authoritative scientist at a major university who could produce high-resolution inundation maps for us, virtually on demand, for any point on earth at any time during the meltdown of the Ice Age. This was the start of our long and productive working relationship with Glenn Milne.

  Then, as the inundation data began to pour in during the last quarter of 2000, I set Sharif another closely related task. This was to comb through collections of ancient maps from the sixteenth century or earlier – i.e. before the world had been fully explored – to see if he could find any that showed correlations with Glenn Milne’s reconstructions of Ice Age coastlines.

  This touches on a problem – and a mystery – that I have long had an interest in and to which I devoted three chapters in my 1995 book Fingerprints of the Gods. To put matters at their simplest, it has been claimed by Charles Hapgood and others that certain maps dating roughly between the fourteenth and sixteenth centuries show Antarctica and other areas of the world not as they look today, but as they may have looked during the Ice Age when sea-levels were 120 metres lower. Moreover, many of the areas in question had not even been discovered when the maps were drawn (Antarctica was not discovered until the nineteenth century).