The Astonishing Micropygmies

See allHide authors and affiliations

Science  17 Dec 2004:
Vol. 306, Issue 5704, pp. 2047-2048
DOI: 10.1126/science.1107565

By now, every Science reader will have read about the discovery of skeletons representing a primitive human micropygmy population that survived until about 18,000 years ago on the Indonesian island of Flores (1, 2). These creatures were barely 3 feet tall, and had an estimated body weight of 20 kg and a brain size of 380 cm3 (smaller than that of a chimpanzee). They seem to be more similar to Homo erectus than to Homo sapiens, and are thought to have descended from H. erectus independently of sapiens' descent from erectus. When I first learned of this discovery, I thought it the most astonishing in any field of science within the last decade (see page 2013 of this issue). On reflection, paraphrasing Elizabeth Barrett Browning, let me count the ways in which it is (and is not) astonishing.

In situations like this one, I've found it useful to get the perspective of a green extraterrestrial friend visiting Earth from the Andromeda Nebula. My friend remarked, “Once again, you humans are prisoners of your ingrained species-centric biases. You already know that large mammals colonizing remote small islands tend to evolve into isolated populations of dwarfs. You have examples of insular pygmy hippos, buffalo, ground sloths, true elephants, stegodont elephants, mammoths, “Irish” elk, red deer, and even dinosaurs. So, now you have 10 examples instead of 9. What's so astonishing? Since when aren't humans subject to natural selection?”

E.T.'s response forced me to reflect. One surprise, I realized, is that we're uncertain exactly which selective pressures do select for insular dwarfs. A favorite theory is ecological release from competition, when a big species reaches an island lacking the mainland suite of smaller related species. According to this argument, the Flores micropygmy would have evolved to occupy a niche of abundant food left vacant by the lack of native apes, monkeys, and other small flightless mammals (except for rodents and a dwarfed elephant) on this island. Another favorite theory is the supposed resource poverty of islands, such that small-bodied animals will be less likely to starve than large-bodied animals. At the level of individual selection, that argument won't work: Flores and other islands with dwarfed mammals have productivities per hectare at least as high as those of continents. But the argument could work at the level of group selection and could explain the regularly increasing relation between body mass of an island's or continent's top carnivore (or herbivore) and the area of the land mass (3). What counts is the island's total productivity rather than its productivity per hectare: An isolated population of 100 full-sized human hunter-gatherers on Flores would have been at a much higher risk of extinction than an isolated population of 700 micropygmies.

E.T.'s blasé reaction then made me think further: Flores is just one of hundreds of islands in its size range, so why weren't there micropygmies on many other islands? The catch is that, for dwarfing to evolve on an island, you need humans just barely capable of reaching the island: If they could reach it too easily, the continuing arrival of full-sized colonists would prevent evolutionary divergence. Once modern H. sapiens developed the technology to reach islands, the resulting insular populations were constantly faced with new arrivals and were no longer isolated. Hence the only examples of effectively isolated insular sapiens populations known to me are from so-called land-bridge islands (like Britain and Japan) formerly connected to adjacent continents at Pleistocene times of low sea level, and isolated around 10,000 years ago when world ice sheets melted and sea levels rose. Some recent sapiens populations on those land-bridge islands were descended from ancestors who walked to the island during land-bridge times, lacked watercraft, and thus became completely isolated when the land bridge was severed.

For instance, the Australian land-bridge island of Tasmania is known to have supported a human population that survived in isolation for 10,000 years after Tasmania became cut off from Australia (4). Tasmania was large enough that full-sized humans are predicted from regression equations (3) to have lived there—and modern Aboriginal Tasmanians were indeed full-sized. However, the much smaller Australian land-bridge island of Flinders supported a human population that succumbed to isolation only after about 4000 years (5): I am unaware of skeletal remains that indicate whether these humans became reduced in size. Promising locations to search for erectus micropygmies are other Indonesian islands besides Flores: surely Lombok and Sumbawa, through which erectus colonists from the Asian mainland must have passed to reach Flores; and perhaps Sumba, Timor, Celebes, and others (see the figure). My first bet is on Celebes.

Island hopping in the Late Pleistocene.

The island realm from Southeast Asia to Australia and New Guinea. Solid lines denote the current configuration of land. Brown shading denotes the configuration of land in the Late Pleistocene, when the sea level was about 150 m below its present stand, and when shallow seas on continental shelves now less than 150 m deep were dry land. At that time, Bali and Java were joined to each other and to the Asian mainland, Lombok was joined to Sumbawa, and Flores was joined to Lomblen. However, reaching Flores and Lomblen from Asia still required crossing three narrow water gaps, and reaching Australia from Timor or islands to the east would have required crossing even wider gaps of water. [Adapted from (6)]


How did the ancestors of the Flores micropygmies, whoever they were, reach Flores? At Pleistocene times of low sea level, the Indonesian island chain of the Greater Sunda Islands was connected to the Asian mainland as far east as Java and Bali, but water gaps of 6, 19, and 3 km, respectively, separated Bali from Penida, Penida from Lombok and Sumbawa (joined in the Pleistocene), and Lombok and Sumbawa from Flores and Lomblen (also joined in the Pleistocene) (6). Across each of those water gaps, the island on the far side would have been visible to someone standing on the island on the near side. Hence the micropygmies' ancestors could have colonized the island by sailing toward it in a watercraft (perhaps a rudimentary raft, or a mere floating log), or they could have landed on the island accidentally when their watercraft was swept to sea by ocean currents. Perhaps they even swam to the island. Stegodont elephants reached Flores and Timor and Celebes, and monkeys and buffalo and squirrels also reached Celebes, all surely without making rafts; H. erectus presumably could have as well.

Why haven't remains of erectus-like humans been found in Australia and New Guinea, at the eastern end of the Indonesian island chain? Possibly, for the same reason they weren't found on Flores until 2004; perhaps these humans did reach Australia and New Guinea, but archaeologists just haven't looked hard enough for their remains. I doubt this answer; hundreds of Pleistocene human sites are now known in Australia, with no remains of humans other than those of sapiens. Instead, the answer probably has to do with geography: A modern map plus bathymetric charts show that, even at Pleistocene times of low sea level, a water gap of at least 87 km separated the easternmost Indonesian islands from either Australia or New Guinea, which would not have been visible across that wide gap (6). Such gaps were too wide not only for pre-sapiens humans, but also for stegodonts, monkeys, buffalo, and squirrels, none of which are found in Australia and New Guinea.

The discoverers of the Flores micropygmies conclude that they survived on Flores until at least 18,000 years ago (1, 2). To me, that is the most astonishing finding, even more astonishing than the micropygmies' existence. We know that full-sized H. sapiens reached Australia and New Guinea through Indonesia by 46,000 years ago, that most of the large mammals of Australia then promptly went extinct (probably in part exterminated by H. sapiens), and that the first arrival of behaviorally modern H. sapiens on all other islands and continents in the world was accompanied by similar waves of extinction/extermination. We also know that humans have exterminated competing humans even more assiduously than they have exterminated large nonhuman mammals. How could the micropygmies have survived the onslaught of H. sapiens?

One could perhaps seek a parallel in the peaceful modern coexistence of full-sized sapiens and pygmy sapiens in the Congo and Philippines, based on complementary economies, with pygmy hunter-gatherers trading forest products to full-sized sapiens farmers. But full-sized sapiens hunter-gatherers 18,000 years ago would have been much too similar economically to micropygmy hunter-gatherers to permit coexistence based on complementary economies and trade. One could also invoke the continued coexistence of chimpanzees and humans in Africa, based on chimps being economically too different from us to compete (very doubtful for micropygmies), and on chimps being too dangerous to be worth hunting (probably true for micropygmies). Then, one could point to the reported survival of the pygmy stegodont elephants on Flores until 12,000 years ago (1, 2): If stegodonts survived so long in the presence of H. sapiens, why not micropygmies as well? Finally, one might suggest that all of the recent dates for stegodonts and micropygmies on Flores are in error [despite the evidence presented in (1) and (2)], and that both stegodonts and micropygmies became extinct 46,000 years ago within a century of H. sapiens' arrival on Flores. All of these analogies and suggestions strike me as implausible: I just can't conceive of a long temporal overlap of sapiens and erectus, and I am reluctant to believe that all of the dates in (1) and (2) are wrong. Hence I don't know what to make of the reported coexistence.

At last comes the question that all of us full-sized sapiens wanted to ask but didn't dare: Did full-sized sapiens have sex with micropygmies? The difference in body size would not have been an insuperable obstacle: Some individual modern humans have sex with children or with domestic animals no larger than the micropygmies. I suspect that the answer is the same as the answer to the question of whether we modern humans have sex with chimpanzees. We don't, because chimps are too unlike humans to appeal sexually to most of us, and because chimps are much too strong, unpredictable, and dangerous to make sex a safe proposition for any individual humans who might find them sexually attractive. Ditto for H. erectus, even when dwarfed.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
View Abstract

Stay Connected to Science

Navigate This Article