W. Earle’s map of 1845 showing the Great Asiatic Bank. Reproduced with the kind permission of the Royal Geographical Society of London.
‘The Thorpe factor!’ said my friend Julie as I described H. floresiensis’ long feet. ‘Maybe it swam to Flores?’
Julie was of course referring to the famously flipper-sized feet of Australian Olympic gold medal swimmer Ian Thorpe. At size 17, Thorpe’s huge and super-flexible feet (apparently he could touch his shin with his toes) along with his six-beat kick gave him a huge boost over his rivals in the pool in the early 2000s. But what really got my attention was that Julie, with this comment she made in early 2016, had inadvertently hit upon a question that had long puzzled me: could hominins somehow have made sea crossings? I wanted to explore this idea in more depth, drawing on various archaeological discoveries in Indonesia over the years, as well as natural occurrences like tsunamis, cyclones, other weather patterns and ocean currents.
Archaeologists and human evolution researchers generally take it for granted that, of the hominins, only modern humans had the brainpower and technological ability to make sea crossings. The earliest evidence for this accomplishment was the arrival of modern humans in Australia between 40 000 and 60 000 years ago. In my experience, to human evolutionists, the idea that early hominins possessed the ability to cross water, through the use of seagoing craft or, especially, a natural talent for swimming, is like a red rag to a bull. Yet archaeological work on Flores contradicts this strongly held view. As Gert van den Bergh says: ‘The unexpected conclusion looms up: hominins were able to make sea crossings. If this conclusion will stand critical evaluation … it will alter the image of our direct ancestors.’1
When H. floresiensis burst onto the scene in 2004, researchers started asking just how this hominin could have made it to Flores. Mike Morwood and colleagues floated two ideas in the 2009 edition of the Journal of Human Evolution that was dedicated to studies on H. floresiensis. One of their ideas was that H. floresiensis could have arrived from Sumbawa, which lies to the west of Flores. The other was that Sulawesi, a large island to the north of Flores, could have been the place of origin. Morwood and colleagues favoured the Sulawesi proposal because of the Makassar Strait Throughflow,2 a strong current that flows south from the North Pacific Ocean through the Makassar Strait between Borneo and Sulawesi, and into the Indian Ocean—with some branching along the way. The researchers argued that, as a result of an extremely rare event such as a tsunami, a small colonising group could have accidentally arrived on Flores on this current while clinging to a natural raft of vegetation or an uprooted tree that had been washed out to sea.3
Robin Dennell and colleagues also proposed that H. floresiensis or its predecessor was likely to have arrived on Flores by floating on natural rafts of vegetation dislodged during cyclones or tsunamis.4 They, like Morwood and colleagues, envisaged two possible scenarios. One was that hominins had been swept off western Sulawesi following a tsunami and got caught up in the throughflow, eventually arriving on Flores.5 Their alternative suggestion was that H. floresiensis dispersed via Java, Bali, Lombok and Sumbawa.
The idea of hominins being accidentally transported to Flores via a powerful current continued to gain momentum, and by the time I became interested in exploring the idea, it had become the general consensus. But in February 2016, when I explained to oceanographer Dr Stewart Fallon on the ANU campus that I was keen to find out how long it might take for hominins to float down the Makassar Strait Throughflow to Flores, he looked somewhat shocked. I was momentarily perplexed: had I simply asked too much of this busy academic? Then Stewart’s face relaxed into an amused smile and he said, ‘The Indonesian throughflow flows 200 metres under water.’ He brought me out of my stunned silence by saying that it was the surface currents I’d be interested in. ‘The surface currents are driven by the monsoon winds. Look …’ he said, uploading maps to his computer screen. I could see that during one of the two monsoon seasons, the currents would flow south, and during the other monsoon season they would flow north. I left Stewart’s office clutching a list of articles to track down.
My reading confirmed that while the throughflow does indeed enter the Makassar Strait as a surface current, it then plunges to become a strong, jet-like current at a depth of between 70 and 240 metres, eventually splitting when it reaches the Java Sea.6 One branch heads into the Flores Sea. As it passes further south through the Lombok Strait there is a surface component to the current, but the main core of the flow is subsurface. The other exit passages are in the Ombai Strait and Timor Passage,7 where the water moves independently of surface flows.8 The Indonesian throughflow current simply couldn’t have taken hominins anywhere from Sulawesi.
Could surface currents driven by monsoon winds have propelled hominin tsunami victims from Sulawesi to Flores? More research revealed that, from January to March, winds push currents in the Makassar Strait southwards into the Java Sea, while from July to September, the currents flow eastwards from the Banda Sea, northwards into the Makassar Strait and westwards into the Java Sea.9 It is only the southward-flowing currents that could potentially have dispersed hominins from Sulawesi in the direction of Flores.
The question now was whether monsoonal currents flowed in these same directions a million years ago, when early hominins were getting to Flores. At first this seems like an impossible question to answer: no modern humans were around at that time to bear witness to what was happening. There are records, though, which have resulted from the scientific examination of microscopic plant pollen remains and other soil components from archaeological excavations. From these, it has been possible to figure out ancient climate conditions. For the Sangiran region of Java, for example, where the 1–1.5-million-year-old H. erectus fossils were discovered, Brasseur and colleagues found that the climate back then was just as it is today: mainly influenced by South-East Asian monsoon cycles.10 Studies in China by T Liu and Z Ding found the same thing, except that their efforts revealed that consistent monsoon weather patterns go back at least 2.5 million years, although they varied in intensity from time to time.11
We can assume, then, that monsoons, their attendant winds, and the currents those winds generated, were occurring when the ancestors of H. floresiensis reached Flores. So let’s look more closely at how floating to Flores could have panned out.
As mentioned above, Morwood and Dennell and their respective colleagues suggested that hominins could have floated to Flores by clinging to a raft comprising vegetation that was dislodged by a tsunami-like event. Dean Falk suggested an alternative scenario: perhaps, based on a foot morphology that indicates partial arboreal habits, H. floresiensis slept on tree platforms like orangutans. If so, they may have nested in groups each evening. Imagine them gathering in trees on the outside curve of a Sulawesi river swollen in the wet season. Big flows might have eroded the banks, with occasional large chunks falling into the water. If such a chunk had multiple trees with nesting individuals, this could have been the large, tangled natural flotsam on which H. floresiensis was swept away.12
The three primary elements in this scenario are a tsunami (or storm) event, the formation of a natural raft of vegetation, and surface flows. We also need to consider the probability of early hominins surviving the voyage to land on Flores and going on to establish a viable population.
Tsunamis are caused by volcanic activity that affects the ocean floor and generates a powerful surge of water. As a tsunami approaches land, it appears as a violent onrushing tide. The water movement can be very complex, with successive waves becoming larger and larger and often colliding with each other. These waves are faster, higher and contain much more energy than cyclonic storm surges. They may inundate hundreds or even thousands of metres of inland terrain, and the onslaught of successive waves can last for many hours. The toll in injury and loss of life, both onshore and as people are washed out to sea, can be catastrophic.
We have witnessed this level of destruction in recent times. On Boxing Day 2004, a 30-metre mountain of water hit Aceh, a densely populated coastal area of Sumatra characterised by many thousands of settled communities of fisherfolk and farmers. The tsunami killed 100 000 men, women and children. Buildings folded like collapsing stacks of cards, while trees and cars were swept up in the oil-black rapids—very few of those people caught up in the deluge survived.13
In 2006, another tsunami hit a 300-kilometre coastal stretch of southern Java, with most of the damage centred on the town of Pangandaran. According to witness accounts, the sea level fell before a 3-metre wave struck the shoreline; the next wave was 5 metres high. The water surged 500 metres inland.14 Some 525 people died and 273 were unaccounted for; more than 50 000 people were evacuated.
The 2018 Sunda Strait tsunami claimed the lives of at least 426 people and injured over 14 000 others. The Indonesian Navy stated that dozens of bodies were recovered from the sea.15
Hamzah and colleagues compiled data from several sources and found that 105 tsunamis occurred in the Indonesia region from 1600 to 1999.16 Nine of these originated in the Makassar Strait, sixteen occurred off Sumatra, and nine were in the eastern Sunda Arc, an area that includes Java, Bali, Lombok and Sumbawa. There is no apparent pattern of tsunami occurrence. For example, the earliest tsunami listed in the researchers’ compilation for the eastern Sunda Arc, a region we consider below, occurred in the decade 1810–19. Another tsunami occurred in this region in the 1850s, followed by three in the next decade. Then there was a forty-year gap until the next one (in the period 1910–19); fifty years later, in the 1990s, came the last tsunami of the twentieth century. We know from this that there can be long intervals between tsunamis, but they are just as likely to occur in quick succession. Of course, we cannot know the frequency of tsunamis before written records or oral histories were kept—we have no indication of the frequency of tsunamis that took place a million years ago.
The next important element of our floating-to-Flores scenario is the mode of transportation. Natural mats of vegetation comprise either small sections of land that have detached from the coast or a riverbank, or tangles of plants, trees and roots that have been dislodged and entered the water.17 Perhaps surprisingly, little is known about the number or frequency of such dispersals of animals through time—we don’t even know much about modern-era events, with the available evidence to a large extent being circumstantial.18 What we do know is that, typically, it is insects and reptiles that are caught in this way. Theil and colleagues noted that episodic events occur at intervals of many years, often decades or even centuries apart.19 The few accounts of animals found on rafts mainly involve single individuals, although Censky and colleagues described the arrival of at least fifteen iguanas on the eastern beaches of Anguilla in the Caribbean a month after two successive category 4 hurricanes moved through the region—they arrived on a mat of logs and uprooted trees 9 metres long.20
It would seem like a relatively straightforward notion that surface currents in the Makassar Strait might have dispersed hominins to Flores, but it is not that simple. The currents are complex systems that include eddies, which are circular or elliptical ocean currents that form at the southern entrance of the Makassar Strait, and which can vary from tens of kilometres to hundreds of kilometres wide. Although we do not know if eddies occurred in the strait a million years ago, I think it is worth keeping this possibility in mind.
Nuzula and colleagues recorded twenty-nine eddy events in the Makassar Strait between 2008 and 2012; these were between 124 and 255 kilometres wide. Eddies may swirl adjacent to coastal areas or further out to sea at any time of year, with their frequency varying from weekly to monthly. For example, in January 2009, a 132-kilometre-wide eddy developed adjacent to the south-western shoreline of Sulawesi. It circulated in a clockwise direction and possibly struck the coast.21 The following month, a 147-kilometre-wide anticlockwise eddy formed that abutted the western coast of Sulawesi.22 The horror scenario for anyone swept off Sulawesi by a tsunami is to be trapped in an eddy, where there are no means of escape.
Stories of the occasional human survivor of an Indonesian tsunami are sometimes used to illustrate, or back up, the possibility that island colonisation can happen as a result of tidal wave action.23 One of these concerns a pregnant woman who was rescued from a floating sago tree five days after the 2004 Aceh tsunami.24 The woman’s survival may sound like compelling evidence for possible island colonisation, but what was not reported in the associated research was that she could not swim and had been thrashing about, trying to keep her head clear of the water, when she chanced upon a tree trunk. She survived by eating the fruit and bark of the sago palm. The fact that she twice slipped from the tree but managed to hold on, and saw sharks around her in the water, draws further attention to the hazards of survival at sea following a natural disaster. In addition, her husband, who had been swept out to sea with her, was never seen again.
JMB Smith reported that a woman clinging to a piece of driftwood was rescued 80 kilometres out to sea six days after Hurricane Mitch struck Honduras in 1998.25 This too, sounds compelling, but this person was also taken out to sea with her husband, and with her three children, and only the woman survived.
It is worth noting that each of these rescues was of one individual only, and in each case the survivors were rescued at sea between five and eight days after the event. No-one made landfall on an island or other landmass. Such anecdotal stories, used by researchers to bolster their hypotheses of the potential for accidental colonisation of islands by hominins, in fact indicate the opposite: how unlikely it is that individuals, much less groups of individuals, who have been carried out to sea will make landfall.
The Indonesian and Honduras women were very lucky to have been rescued, but what are the chances of surviving at sea if rescue doesn’t occur? Firstly, to survive, an individual must have access to fresh drinking water. Humans require fresh water for survival. We can last for weeks without food because the body has many weeks of nutritional reserves, but with no significant ability to store water we can survive only a matter of a few days.26
H. floresiensis had an adult stature of only 106 centimetres,27 equivalent to a modern four-to-six-year-old child. A child of that age needs at least 1.7 litres of water a day to survive.28 And presumably a greater amount of water would be needed in the high heat and humidity that H. floresiensis individuals would have encountered in their tropical regions, unless they had a very different metabolism to us. Now, rainfall would have been plentiful in the monsoon season, occurring daily. Had hominins been swept off land on a natural raft of vegetation, the availability of fresh water is unlikely to have been an issue. That said, the rainwater still had to be collected somehow, and as far as we know, H. floresiensis did not make any containers or other vessels. Even if they did, how likely would it have been that these were on the vegetative mat when it was dislodged by a tsunami? So for H. floresiensis to get fresh water, we would have to assume that rainwater pooled on the vegetative mat and could be scooped up by hand, or perhaps collected in a large leaf or something similar. If the hominin was floating on a tree trunk, however, obtaining fresh water from downpours would have been much more challenging.
Beyond the crucial act of staying alive in such conditions, normal functioning is also important. Mild-to-moderate dehydration can affect our level of alertness and consciousness, causing sleepiness, muscle weakness, headaches and dizziness. Severe dehydration causes arm, leg and visual weaknesses, and the body finds it hard to regulate its core temperature.29 This results in low blood pressure, confusion, rapid heartbeat, fever, delirium and eventually unconsciousness.30 For H. floresiensis to achieve landfall on Flores, it would have been essential to have enough fresh water at hand to both survive and to ensure individuals functioned well enough to retain consciousness, at the very least.
Also critical to survival is not drinking sea water, which is dangerous and can even lead to death. Sea water contains 3.5 per cent salt, but urine cannot contain more than 2 per cent salt. If we drink sea water, our kidneys are unable to flush out the salt they have filtered from the blood, unless a large amount of fresh water is subsequently consumed. In an untreated person, kidney failure ensues, followed by death. We cannot know if early hominins were aware of this ‘hidden’ danger, so we must assume that to reach Flores, H. floresiensis or its predecessors knew not to drink sea water even if fresh water was unavailable.
Survival is then dependent on making landfall. To arrive on Flores, the vegetative mat on which the species was being transported had to stay intact for the duration of the voyage. It had to be dense and strong enough to withstand buffeting by turbulence—we know seas can get rough during monsoonal storms. The sad fate of individuals should the mat have fallen apart was likely to be drowning or shark attack.
Let us assume for the moment that these ominous events did not occur and the individuals reached an island. When I think of castaways being washed up on a beach, I immediately think of Robinson Crusoe. If memory serves me correctly, he was shipwrecked within wading distance of an island; it all seemed relatively easy. But we cannot assume that such an idyllic landing awaited any hominins arriving on the shore of Flores. For one thing, what kind of coastline would they have encountered? We do not know the answer to this question, but today, some of that shoreline is made up of steep rocky cliffs that plunge deep into the ocean. In those places, unless hominins could somehow self-propel themselves into another bay, the effort to attain dry land would have been hopeless; that is, although hominins might have floated on a vegetative raft to within sight of Flores, making landfall might not have been guaranteed. On the other hand, if sandy, muddy or marshy bays were encountered, getting ashore would have been easier.
Because H. floresiensis lived for at least 40 000 years on Flores, and we have evidence establishing that hominins were at Mata Menge from at least a million years ago, we know these populations were viable. By a ‘viable population’, I mean one that has enough genetic diversity among breeding pairs to ensure a high probability of survival over a relatively long period. To achieve this on an island, the initial colonisation would have to depend on a series of interrelated events. The first is that the incident in question, such as a tsunami, dislodges male and female individuals of reproductive age. Alternatively, if a different mix of individuals, such as all males or all females, was swept away, then establishing a viable population would depend on another tsunami occurring pretty soon afterwards and carrying off the right mix of sexes in order to potentially produce offspring with the earlier group. In other words, the arrival of individuals need not be at exactly the same time, but fertile individuals would need to arrive in relatively close succession. And according to this scenario, the individuals concerned would have to actually come in contact with each other, no matter which part of the 380-kilometre northern Flores coastline they were deposited on.
So the establishment of a long-lasting, viable population on Flores, after voyaging there from Sulawesi by floating on sea currents, would depend on the right mix of individuals being on a section of land that has the potential to be dislodged during a tsunami or storm action, or on several victims clambering onto a natural raft; and on this floating structure being relatively stable and staying afloat rather than disintegrating under the erosional action of water. It would depend on the floating mat not becoming trapped in an eddy. It would depend on survivors obtaining fresh water and not consuming sea water while travelling to Flores, followed by the safe landing of enough individuals to multiply and thrive. While not impossible, this would all seem to defy the odds.
An alternative scenario we can consider is that H. floresiensis arrived on Flores via the sweep of islands that arc there from Sumatra. The island of Sumatra emerged between sixteen and eleven million years ago.31 The blobs of land to the east of Sumatra appeared around ten million years ago, and from five million years ago these started to increase in size—they now form the islands we know as Java, Bali, Lombok, Sumbawa and Flores.32 Some historical background on the distribution of animal species among these islands will provide some context here.
In 1845, W Earle published an article in the Journal of the Royal Geographic Society of London in which he identified two geographical areas in the South-East Asian–Australian region. One he called the Great Australian Bank, which is the shallow sea that connects New Guinea with Australia. The other he named the Great Asiatic Bank, also a shallow oceanic area, averaging about 30 fathoms (55 metres) in depth, that connects the islands of Sumatra, Java, Bali and Borneo with the South-East Asian mainland, and stretches from the eastern side of Borneo to within 80 kilometres of Sulawesi. It includes the Malay Peninsula and the northern coastlines of Sumatra and Java.33 Today, we call this underwater bank the Sunda Shelf—the greater landmass that would have been exposed during periods of low sea level in earlier times is called Sundaland.
In 1853, while studying the natural history collections at the British Museum in London, the naturalist and geographer Alfred Russel Wallace noticed that the flora and fauna of some regions of the world were not well represented. One of these regions was the islands of South-East Asia, and Wallace applied for funding to mount an expedition to investigate the area. While he specifically wanted to collect items for the museum, he would also support himself by selling specimens to private collectors and naturalists,34 a normal and respectable livelihood in those days. And so for eight years—from April 1854 to April 1862—Wallace travelled throughout the islands region of South-East Asia and the Malay Peninsula.35
W. Earle’s map of 1845 showing the Great Asiatic Bank. Reproduced with the kind permission of the Royal Geographical Society of London.
Wallace was not merely a collector of rare species, or species previously unknown to the scientific establishment. He was also a keen observer, a scientist of nature. He is famous for developing the theory of evolution, or, as he called it, ‘a Law of Natural Descent’,36 at the same time as Charles Darwin—Darwin and Wallace’s ideas were presented concurrently to the Linnean Society of London on 1 July 1858.37 But Wallace was just as absorbed with resolving the question of the odd distribution of species across South-East Asian islands as he was with evolution.
Wallace went on to identify a virtual north–south boundary that separates the faunal types of the Philippines and western Indonesia from the eastern Indonesian islands. He explained his findings in a letter to a friend, Walter Bates, on 4 January1858: ‘The boundary line often passes between islands closer than others in the same group. I believe the western part to be a separated portion of continental Asia, the Eastern, the fragmentary prolongations of a former Pacific continent.’38
Today, we call this species boundary the Wallace Line (see Figure 1 at the beginning of chapter 1). It closely follows the eastern edge of the shelf that Earle identified in 1845. But Wallace was initially unaware of Earle’s work, only hearing about it in 1859 when he received a letter from Charles Darwin in which Darwin mentions Earle’s publication.39 Later, Wallace incorporated Earle’s underwater bank into his theory and proposed that changes in sea levels must have occurred on a grand scale in the past. When sea levels fell, the shelf identified by Earle would be exposed, greatly expanding the Asian landmass,40 and thus enabling the Asian species to freely move around. Those remaining on Sumatra, Java and Bali are therefore those that arrived when the islands were connected by land.41
So now we can see how H. erectus probably got to Java: by wandering overland during a period of low sea levels. But how can we explain the presence of hominins on Flores if we believe they came via the islands of Java, Bali and Lombok? As Wallace deduced, deep channels separate the islands east of Bali from Lombok. Even when sea levels were low in the past, these islands were never joined—they never formed a single landmass. Animal species from Java did not get further east; they did not cross the Wallace Line and they are not on Flores. It is highly unlikely, then, that hominins walked to Flores.
A possible alternative explanation for how hominins got to Flores is that they found themselves making accidental sea crossings. In that case, at least three crossings would be involved: from Bali to Lombok, Lombok to Sumbawa, and Sumbawa to Flores. More crossings, of course, would be required at periods of high sea levels when Bali was separate from Java and Java was separate from Sumatra.
The straits between these islands are relatively narrow and it is no big deal these days to travel between them in motorised boats. But hominins left to the mercy of currents would be another matter entirely. If Brasseur and colleagues, and also Liu and Ding, were correct when they concluded that monsoons were occurring as early as 2.5 million years ago, then hominins would have faced monsoon-driven currents in the straits. We cannot know the exact form or intensity these would have taken, but I can alert you to the problems that would face anyone being swept off these islands today.
Situated between Bali and Lombok is the 20–40-kilometre wide Lombok Strait. Between July and September, the south-east monsoon causes a south-flowing current running at between 3.5 and 6 knots (1.5–3 metres per second). In all likelihood, anyone carried off Bali into the Lombok Strait would find themselves rushed northwards into the Java Sea or southwards into the Indian Ocean, depending on the monsoon season in which the unhappy event occurred. An extra hazard in the strait is the strong eddies that may be encountered at its northern and southern entrances, and there are turbulent sea conditions at its narrowest part.42 In each case, without rescue, victims washed off Bali would face poor prospects for survival.
I have never experienced boating in the Lombok Strait, but Wallace did. Here is his description of a voyage that was clearly not for the faint-hearted:
The beach [at Ampanam, Lombok] of black volcanic sand is very steep, and there is at all times a heavy surf upon it, which during spring tides increases to such an extent that it is sometimes impossible for boats to land, and many serious accidents have occurred. Where we lay anchored, about a quarter of a mile from shore, not the slightest swell was perceptible, but, on approaching nearer, undulations began, and rapidly increased, so as to form rollers which toppled over on to the beach at regular intervals with a noise like thunder. Sometimes the surf increases suddenly during perfect calms, to as great a force and fury as when a gale of wind is blowing, beating to pieces all boats that may not have been hauled sufficiently high upon the beach, and carrying away incautious natives. This violent surf is probably in some way dependent on the swell of the great southern ocean, and the violent currents that flow through the Straits of Lombok. These are so uncertain that vessels preparing to anchor in the bay are sometimes suddenly swept away into the straits, and are not able to get back within a fortnight. What seamen call the ‘ripples’ are also very violent in the straits, the sea appearing to boil and foam and dance like rapids below a cataract; vessels are swept about helpless, and small ones are occasionally swamped in the finest weather and the clearest skies. I felt considerably relieved when my boxes and myself had passed in safely through the devouring surf.43
To the east of Lombok lies Sumbawa, with the two islands separated by the 117-kilometre-wide Alas Strait. The currents in this strait run in a northerly or southerly direction in accordance with the monsoons. Anyone washed off Lombok into the Alas Strait would most likely be transported into the Java Sea or the Indian Ocean, with little possibility of making landfall; similar conditions are encountered in the Sape Strait, which separates Flores from Sumbawa,44 suggesting that the same potential outcomes face individuals accidentally swept off that island.
The means by which early hominins might have arrived on Flores stirred the imagination of the National Geographic documentary-makers who were filming at Liang Bua in 2004.45 They thought that Sumbawa was the most likely place from which the early hominins set off for Flores, with bamboo rafts being the most likely means of transportation. To test this theory, they set up an experiment with the Liang Bua team. They would build a raft and paddle it from Sumbawa to Komodo. Robert Bednarik, an experienced rafter and a colleague of Mike Morwood, was engaged to design and organise the construction of a 12-metre bamboo raft, which he did with the help of seventeen Sumba people after sourcing 5 tonnes of bamboo from Bali. Aboard for the trip were Thomas Sutikna, Wahyu Saptomo, Mike Morwood and Bert Roberts. Crewed by a dozen fishermen, it took the raft eleven hours to cross the 22-kilometre-wide channel.
The trip took a toll. Strong south-flowing currents swept the raft south of Komodo, resulting in an exhausting paddle to get back to the leeward side of that island. One of the passengers became violently seasick and the support team had to remove him from the raft. Two other passengers became exhausted after six hours and they, too, were rescued from the situation. Then, as the raft approached the Komodo shoreline, the crew was faced with wild conditions, with waves crashing over rocks, and there was no obvious landing place in sight. This fraught situation was further complicated when Mike tripped and broke his ankle.46
Bert Roberts later made an interesting observation about the voyage. The raft, constructed from fresh-cut bamboo, became weighed down and heavy to manage as it absorbed sea water. ‘If you are going to build a bamboo raft, it needs to cure for about a year to dry out. Holes in bamboo, the sap channels, seal up when the bamboo dries out,’ said Bert, adding that when ‘the raft we were on was being returned to Sumbawa, it sank’.47
Putting aside the assumption that early hominins could, or did, make rafts, this experiment highlighted the potential dangers. Had the twelve-person crew not been able to turn the raft around and get to the leeward side of Komodo, it is likely they would have been taken out to sea. Had they not been able to manoeuvre away from the hazardous rocky shoreline, their vessel was in danger of being smashed to pieces and its occupants thrown into the ocean. How much more dangerous would this journey have been for hominins entirely at the mercy of currents?
Yet hominins did get to Flores. So let’s imagine how this could have happened, ignoring for the moment the danger of the north–south monsoonal currents.
Say, for example, that a number of hominins are on a beach on Java, gathering shellfish. A tsunami races in, or a gale occurs such as the one Wallace described, and the hominins are washed away along with uprooted trees and debris from the coast and the island’s hinterland. As luck would have it, all the hominins manage to cling to the debris until, with an extra dash of good luck, they are deposited on a neighbouring island. Years later, the same group endures another tsunami and are lucky, again, to clamber onto some floating debris and, again, are washed up onto the next neighbouring island. Could this misadventure have happened to the same group on four successive occasions, facilitating them crossing four straits, to eventually arrive on Flores? I think not.
A more plausible scenario is that a group comprising males and females of breeding age are swept out to sea following a tsunami and, luckily, they all land on the next island. The descendants of the ever-growing group then populate the island. At some stage, a group of these descendants become the victims of another tsunami, with some arriving on the next island to the east. If both males and females of breeding age are again part of this group, it would grow in numbers and, again, the island would become populated. And so the occupation of each island in turn would happen over the generations until a descendant group of hominins found themselves on Flores. Yet even this scenario has its flaws. How likely is it that the groups washed off the islands included males and females each time? Also, if the survivors could not see land, how could they know that they needed to somehow counteract the north- or south-flowing current they found themselves in? Would they even have had the strength to do so?
Furthermore, we have seen from historical records that tsunamis can occur in close succession or they can take place as many as thirty, forty or more years apart. Again, we do not know the frequency of tsunamis at the time hominins first lived in the region, but if there were thirty-year intervals between these events, it would seem unlikely that castaways would wash up on Flores within one generation such that they could produce offspring.
I have one last thought to share on the vexing question of how H. floresiensis arrived on Flores. At the time of the first toolmakers, there were several animal species on Flores, including a giant 1.8-metre-tall bird;48 stegodonts; a giant tortoise; a small crocodile; a giant rat the size of a small dog; and a large predatory reptile, the komodo dragon, which still lives on Flores.49 Meijer and colleagues and van den Bergh and colleagues have noted that each of the early animal species on Flores was capable of making sea crossings. The ancestor of the giant stork can fly.50 Tortoises can float. Stegodonts, if they had the abilities of their modern relation the elephant, could have swum. Komodo dragons, too, can swim. In this regard, the odd species out is H. floresiensis.
Here, we return to my friend Julie’s speculation: could H. floresiensis have swum to Flores? If we want to get some idea of whether the hominins might have been swimmers, the best we can do is look at our closest relatives, the apes, and ourselves. We modern humans are not innate swimmers. Once we’re out of our depth in water, we will drown unless we’ve been taught the basics of swimming, or at least of staying afloat. Nor does it seem that apes are innate swimmers. But evidence of their potential swimming ability comes from observations made of a captive orangutan and chimpanzee studied by the researchers Renata Bender and Nicole Bender (University of Bern).51
The orangutan had been regularly exposed to a pool as an enrichment activity at a private zoo; specifically, it had been taught to dive and swim. At the age of seven years, it began to swim underwater from one keeper to another, and it was later able to fully submerge and swim unassisted for a short distance in the pool. The chimpanzee, meanwhile, was brought up in a human environment. He became accustomed to water play and would submerge his head, always covering his eyes and nose—the longest time he spent underwater was fifteen seconds. When fully submerged, he displayed a great variety of play behaviours. To return to the surface, he vigorously kicked with his legs.
It is interesting that the orangutan and the chimp seemed happy exploring a water environment, but, as Renata and Nicole Bender point out, this does not mean that they are natural swimmers. So it is difficult to argue that the ancestors of H. floresiensis had an innate ability to swim. In turn, it would seem unlikely that H. floresiensis individuals swam to Flores, despite those long feet.
The fact is that none of the scenarios explored here work as a model for the arrival of H. floresiensis on Flores. Whether it’s travelling between Sulawesi and Flores via the Makassar Strait, or crossing sea barriers between Bali, Lombok, Sumbawa and Flores, the proposed solutions to this mystery remain unconvincing. How, then, can we resolve this?
Perhaps we cannot because we simply do not have the full picture yet. We know that the animals on Flores in H. floresiensis’ time could have reached the island by sea. If we could work out where each species came from, we could assess if any of those routes might also have facilitated the arrival of H. floresiensis.
It was while still immersed in an early draft of this chapter in June 2019 that I attended the inaugural Asia Pacific Conference on Human Evolution, held at Griffith University in Brisbane. There I met Glenn Marshall who, along with his colleagues in The First Mariners team, tests sea current directions using a man-made structure that replicates flotsam. Inspired by the question of the arrival of H. floresiensis on Flores, Glenn suggested that we test for different possible routes by which the species may have voyaged by launching the flotsam device from the southern shore of Sulawesi and that of other islands. This project is still in the pipeline. We could supplement this approach by using advanced computer modelling of possible past ocean currents. This method has already been applied to help figure out how Indigenous peoples may have voyaged to Australia.52
I recall Mike Morwood occasionally musing about H. floresiensis arriving at Flores via an alternative route: floating from the Philippines along the currents that flow to the east of Sulawesi. This would have entailed a series of very long and probably hazardous accidental journeys. But because of the baffling issue confronting us, we need to consider anything we can that might resolve it.
That includes thinking outside the square and contemplating whether hominins may have fashioned rafts or other watercraft with the intention of making sea crossings. Here, Bert Robert’s observation may be pertinent. If hominins did make bamboo rafts, it’s possible they would have had to plan the voyage a year ahead to allow time for the harvested bamboo to dry out. Of course, the idea that early hominins could think so far ahead to realise the ambition of voyaging is a radical proposal for palaeoanthropologists, including me, to consider. It takes us back to the heart of the enigma: if hominins were making deliberate, pre-planned journeys between islands on seaworthy watercraft a million years ago, why would H. floresiensis have become isolated? How could the population have become cut off on the island for a sufficient length of time to evolve into an entirely distinct species if it was possible for them simply to sail back to where they came from, or if multiple generations of other seafaring hominins from the same species could also reach Flores?53
Thinking outside the square, I cannot get those long flipper-like feet of H. floresiensis out of my mind. Now, if only we could come up with some way of testing if the species could swim …