This is the first part of a three post blog discussion.
Part I provided the Preface and Disclaimers
that apply to this post and the next one.
This Part II provides a work in progress narrative of the Out of Africa story of hominins from the first hominins to leave Africa until the extinctions of archaic hominins outside of Africa (as noted in the first part, the discussion of archaic hominin populations in Africa is mostly beyond the scope of this post.
Part III will provide a work in progress narrative of the Out of Africa story following the extinction of archaic hominins. A link to that post will be provided here when the third part is posted.
From Hominins To Humans
Hominins Before The Genus Homo
Starting around 7 million years ago, a new kind of primate, for which the chimpanzee and bonobo are the most closely related species that survive today, started to evolve as a distinct species. From about 7 million years ago until about 2.5 million years ago, multiple species of these primates arose. Modern humans are directly descended from some of these species, while others species in this clade of the primate evolutionary tree evolved in parallel to our ancestor species, possibly with some admixture with the species from which we are descended, before going extinct. There are perhaps a dozen named hominin species in this time period, but this partially flows from the tendency to identify each net set of fragmentary hominin remains at a new location or time frame as a new species without rigorous efforts to confirm or deny the possibility that some may belong to the same species.
Modern humans descend from some of the more gracile, rather than from some of the more robust of these species. These transitional species apparently did not manufacture stone tools or any other kinds of tools that left traces in the archaeological record.
Some of the most famous representatives of these "missing link" species between Great Apes and members of the primate genus Homo, are "Ardi" a.k.a. Ardipithecus
ca. 5 million years ago in Kenya, and "Lucy" a.ka. Australopithecus afarensis,
ca. 3.2 million years ago in Ethiopia.
Homo Erectus and Homo Ergaster
Around 2.5 million years ago in Africa, most characteristically in savanna and water's edge environments, the first hominins that we classify in our own genus, Homo
called Homo Habilis
, which are the first hominins know to have made tools became distinct. By around 1.9 million years ago in Africa, a species called Homo Ergaster
for which begins to appear in the fossil record outside of Africa around the same time as the physically similar the archaic hominin species Homo Erectus
appears in Asia, arose. I will describe both H. Erectus and H. Ergaster as "Homo Erectus" in this narrative, eschewing the distinction between H. Erectus and H. Ergaster, as it seems most likely to me that they were probably really only subspecies of a single species of hominin.
By around 1.9 million years ago in Africa, Homo Erectus lived a lifestyle somewhat similar to contemporary modern human hunters and gatherers (although not as sophisticated), using primitive stone tools to secure and process food by hunting small game and larger herbivores, scavenging meat, and gathering plants and sessile animals (like shellfish). They walked on two legs, lacked the tails and heavy fur covering of the Great Apes, has some sort of language skills although those language skills may have been primitive, were recognizably humanoid, were smaller than modern humans although not tiny, were smarter than all or most of the hominin species that arose before them, and ultimately some of them developed more advanced stone tools (Achuelean industry ca. 1,500,000-1,600,000 years before present) and used fire in a controlled manner to keep themselves warm and cook food (ca. 500,000 years ago in Europe, 750,000 years ago in the Levant, and perhaps earlier in Africa). They lived in semi-nomadic groups of about twelve to twenty individuals with very few of them living long enough to be alive at the same time as their grandchildren for any meaningful length of time. They had a range far greater than that of any previous hominin species both within Africa and outside it. They had little that we would call "art" and don't appear to have initially crafted tools out of bone or wood (which is not to stay that they might not have used an entire bone or tusk as a club or spear or fire prod, for example). Many popular stereotypes of the life of "cave men" are a better fit to Homo Erectus than to the Neanderthals or earliest modern humans that evolved later.
Homo Erectus (including Homo Ergaster) was the first hominin species to leave Africa and did so no later than 1.8 million years ago, by which time Homo Erectus remains prove they were present in Java, Indonesia. In Southeast Asia and East Asia, Homo Erectus managed with the more primitive Oldowan stone tools invented by Homo Habilis in Africa and continued to use this tool set until at least about 200,000-300,000 years ago when they disappear from the fossil record.
In South Asia, Homo Erectus had begun to use more sophisticated Achuelean stone tools by somewhere in the time frame of 1 million to 1.5 million years ago, and a perhaps as early as 1.6 million years ago in Africa. Archaic hominins called Homo Antecessor, sometimes called a new species, are found in Spain 870,000 years ago, and by 400,000 years ago or so, Homo Erectus remains and Achuelean industry are present and common in Europe.
It is reasonable to conclude that the Homo Erectus of Southeast Asia and East Asia were a population that was separate from the Homo Erectus of Western Eurasia, South Asia and Africa from their arrival in Southeast Asia from Africa around 1.9 million years ago, are at least starting some time prior to the arrival of Achuelean industry in South Asia from Africa around 1,000,000 to 1,500,000 years ago, until their extinction as a result of the arrival of modern humans in region, possibly exacerbated by other "natural" factors (e.g. climatic effects related to the Toba volcanic eruption ca. 75,000 years ago).
The island of Flores, just over the non-Eurasian side of the Wallace line in Indonesia, was inhabited by Homo Floresiensis
from ca. 95,000-18,000 years ago as evidence by 12-15 sets of remains of individuals with adult heights of 3.5 feet (hobbit sized) and their material culture (including stone tools, dwarf elephant remains from their hunts and fire use) from Liang Bua cave, which they inhabited on that island. They are probably pygmy version of Homo Erectus who followed the commonly known pattern of island dwarfism. Their ancestors could have arrived in Flores as early as 800,000 years ago.
Neanderthals appear in Europe around 200,000 and ultimately have a range that skirts that includes most of Europe, Southwest Asia (i.e. Arabia and the Levant), the southern fringe of the Central Asian steppe, West Asia, and South Asia before their extinction sometime between 40,000 and 30,000 years ago. Their extinction coincides with a string of major volcanic eruptions that also led to major climate changes in their range and the arrival of modern humans in Europe.
Neanderthals were behaviorally more advanced than Homo Erectus, more robust than either Homo Erectus or modern humans, show signs of adaptation of ice age European climates, and had relatively short legs well suited to steep terrain. They are associated technologically with Mousterian industry stone tools
, which are more sophisticated than Achuelean industry stone tools associated with later Homo Erectus populations in Europe, Africa and South Asia.
Mousterian tools that have been found in Europe were made by Neanderthals and date from between 300,000 BP and 30,000 BP. . . In Northern Africa and the Near East they were also produced by anatomically modern humans. In the Levant for example, assemblages produced by Neanderthals are indistinguishable from those produced by Qafzeh type modern humans. It may be an example of acculturation of modern humans by Neanderthals because the culture after 130,000 years reached the Levant from Europe (the first Mousterian industry appears there 200,000 BP) and the modern Qafzeh type humans appear in the Levant another 100,000 years later. It was superseded by the [Neanderthal] Châtelperronian industry around 35,000-29,000 BP.
They don't appear to have engaged in harpoon fishing, although they may have used unaltered bones as tools and may have simply grabbed fish out of streams to eat. There were primarily big game hunters although there is evidence that they collected shellfish on the English coast ca. 125,000 years ago. Male Neanderthals were larger than female Neanderthals to a greater extent that male and female modern humans are different in size.
Ancient Neanderthal DNA indicates that there were probably at least three regional subpopulations of Neanderthals in Europe, that their effective population size was quite small at the cusp of their extinction (perhaps 1,500-4,500 reproductive aged individuals in all of Europe).
It isn't entirely clear if they evolved from prior European Homo species, or if they evolved from prior Homo species elsewhere (e.g. Africa) and then migrated to Europe. They are genetically closer to modern humans and more distant from the Great Apes than the only other archaic human population for which we have ancient DNA, the Denisovians.
Neanderthals are not a direct ancestor of modern humans. Neanderthals diverge from the direct ancestor of modern humans around 400,000 to 500,000 years ago. In my view, the Homo Erectus who brought the Achuelean industry to Europe around 400,000 years ago were probably the population from which Neanderthals descended.
Modern humans evolved in Africa around 200,000 to 160,000 years ago at a time when Neanderthals were found in Europe, West Asia, Southwest Asia and South Asia, and Homo Erectus was present in the lower latitudes of Asia.
We are familiar with modern humans, although these early Cro-Magnons would have initially been monoracial, monoethnic, hunter-gatherers, probably, although not certainly, in East Africa.
Proto-Eurasians, however, surely must have been East Africans, because modern humans did not originally leave Africa via the Strait of Gibralter into Iberia, and the only other ways out of Africa on either side of the Red Sea can only be reached via East Africa, a conclusion buttressed by the fact that the closest phylogenetic relatives of Eurasian specific Y-DNA (B* and BF) and mtDNA (L3*) haplogroups are found in East Africa today, have maximum diversity there, and show no obvious signs of relocation to there from somewhere else.
The modern human population (predominantly made up of men with Y-DNA haplogroup F* and mtDNA haplogroup L3*) undergo a fission ca. 80,000-75,000 years ago when they vanish from the Levant. One of the two branches ends up in a Persian Gulf Oasis or Iran or Pakistan and produces a community with mtDNA haplogroup N that ultimately spread both West to SW Asia and Europe, and East to South Asia and beyond to all points East by about 50,000 years ago. The other of the two branches ends up in a South Asian refugium and produces a community with mtDNA haplogroup M that becomes thoroughly admixed with members of the Western branch as it expands eastward.
A fission of the proto-Eurasian population before or at around the time that the mutations that distinguish mtDNA haplogroups M and N from L3 is necessary because only mtDNA descendants of haplogroup N are found with any frequency in West Eurasian populations (and the exceptions show the clear signatures of backmigrations), while Asia, Australia and New Zealand have both.
Once two distinct lineages of mtDNA haplogroups are present in a population that has had even just a few generations to admix (perhaps four or five generations over a single century), it is extremely difficult to extricate one lineage from the other in a population that branches off from the mixed group unless the founding population of the branch is extremely small, or is a little bit larger and organized matrilocally. If one takes a scenario where the fully admixed proto-Eurasian source population has an effective population size of 3,000 (see, e.g., this 2007 paper
estimating a proto-Eurasian effective population size of about 3,100, and a Yoruba ancestral effective population size of about 7,500) and the splinter group of proto West-Eurasians has an effective population size of 150 (which is quite a bit smaller than conventional estimates), for example, and the relative frequencies of mtDNA haplogroups M (with all descendants) and N (with all descendants) are similar to those of modern South India (which is probably the closest modern comparable without an extremely small founding population to the proto-Eurasian population in the narrative I've presented here), it is extremely difficult to get all of the women in the proto-West Eurasian population to all have mtDNA haplogroup N types by random chance. This statistical relationship is just weakly related to the size of the source population (although it is quite sensitive to structure within the source population's mtDNA such as matrilocal family patterns, or regional clines that arose due to founder effects when M and N arose), but it very strongly related to the effective population size of the founding population. It is extremely unlikely for an unstructured source population in the thousands and a splinter population in the high hundreds to randomly remove all mtDNA M individuals by random chance via founder effects in a stable or expanding population.
The effective size of the founding population of the Americas ca. 14,000 years ago,
is estimated to be on the order of 70 people.
A founding population estimate of 1,000 for Papua New Guinea and Australia, respectively, would probably be significant overestimate,
although firm predictions are hard to find in the literature. There are just five root mtDNA founder lineages and not more than four root Y-DNA founder lineages (perhaps as few as one)
among indigeneous Australians (with a founder lineage defined as the most basal branch of lineage private to Australia and Papua New Guinea and not due to Holocene era admixture); there are four root mtDNA founder lineages and not more than three Y-DNA founder lineages (perhaps as few as one) among indigenous Melanesians. Two of the founder mtDNA lineages of Australia and Melanesia, and two of the Y-DNA founder lineages are shared, so there are no more than seven mtDNA founder lineages and five Y-DNA founder lineages in Australia and Melanesia combined (perhaps as few as two or three). This is similar to the number of founding lineages in the New World. The Australian and Melanesian genomes have deep connections to each other not shared by other populations, matched by their shared similar proportions of Denisovian ancestry and similar in time founding dates, and their founder lineages are both very close to the Eurasian basal lineages (probably within 5,200 years). Thus, it is not unreasonable to infer that they derive from a single wave founding population and to estimate that the founding effective population size for these populations that is on the same order of magnitude as that of the Native Americans, conservatively, in the hundreds for the combined Australian and Melanesian proto-population, which fissioned early on (probably in the vicinity of Flores, Indonesia) in not quite complete break isolating the two populations after that point, with Melanesia's founding population being slightly less diverse. Melanesia's founding population may have been less than one hundred, Australia's founding population may have been in the low hundreds.
In an expanding population, you generally don't lose significant percentage haplogroups of any kind, although a bottleneck, for example, post-Toba, could provide an alternate solution for this problem.
Indeed, it is easier from a modeling perspective, for some of the same reasons, unless mtDNA haplogroups M and N are not selectively neutral (and there is no reason to think that they have selective effects of their own), to imagine a scenario in which mtDNA haplogroups M and N and Y-DNA haplogroup F are simply local offshoots within East Africa of mtDNA L3* and Y-DNA BF that would have looked simply like minor variants of L3 and B at the time, that both migrate to the Levant, perhaps even separately, but close in time or fissioning almost immediately. Otherwise one has to figure out how to purge almost all of mtDNA L3 (xM,N) and Y-DNA A and B and BF (xF) from the source proto-Eurasian population(s).
There may have been a minority of men belonging to Y-DNA haplogroup DE* in the original group that fissioned, or those men may venture, unaccompanied by women, to South Asia to a community dominated by mtDNA haplogroup M that has not yet admixed with mtDNA haplogroup N women in Eastern India.
The Proto-Eurasian populations were quite homogeneous, having only two matrilines and one or two patrilines, so they too were probably monoracial and probably comprised no more than a couple of ethnicities or languages. Conventional estimates put the effective population size of the proto-Eurasian population in the low single digit thousands by more than one partially independent method (some of those methods rely on Y-DNA and mtDNA diversity, rather than autosomal DNA, and hence wouldn't be distorted by Neanderthal admixture in the autosomal genome which would exaggerate effective population size).
Given the fact that there are human remains to prove that modern humans were in the Levant ca. 100,000 years ago (although we can't definitely show that later Eurasians were actually descended from then as opposed to a later wave), and that there is pretty much no evidence whatsoever to definitely favor a Gate of Tears and Arabian coastal route migration, an Out of Africa route for modern humans up the Nile Basin (or parallel to it on the Lake Chad side in a wet sahara era), which are down hill routes with good fresh water supplies that naturally attract game and provide a fishing resource, and across the Suez to the Levant seems more plausible than the Gate of Tears route. Indeed, even today, the lack of continuity in populations on either side of the Gate of Tears is quite striking. Since we know from digs in India that produced charcteristically modern human tools in India prior to 75,000 years ago (pre-and post-Toba ash deposits), we don't need an Arabian coastal route to West Asia theory to get modern humans to South Asia and beyond in the right time frame, if we can get them from the Levant to West Asia between 100,000 and about 76,000 years ago via the Fertile Crescent, which we can just as easily as we can via the Arabian coastal route, in a matter of a few centuries. Recent discoveries from Arabian digs really do nothing to disturb that analysis decisively.
Humans Meet Archaic Hominins Outside of Africa
Neanderthals had an overlapping presence with modern humans in the Levant from about 100,000 years ago to 75,000 years ago, were the only hominin species in the Levant from about 75,000 years ago until 50,000 years ago, and after thousand years of another overlapping presence with modern humans in the Levant around 50,000 years ago, were no longer found in Southwest Asia. The last Neanderthals went extinct around 29,000 years ago.
Neanderthals were outnumbered 10:1 by the modern human hunter-gatherers that immediately followed them in residing at the same locations in Europe where adjacent layers are present. There is no convincing evidence that there were ever communities in which both pure modern humans and pure Neanderthals lived side by side, although there are a few remains that arguably suggest the presence of hybrid individuals, something that the genetic record strongly supports the existence of historically, and one interpretation of a brief period
Neanderthals admixed with the proto-Eurasians leaving a roughly 1-4% admixture percentage (average about 2.5%) of Neanderthal genes in all non-African modern humans. Neanderthal admixture in Western Eurasia overlaps only slightly and in the most common surviving types of Neanderthal DNA in modern humans with Neanderthal admixture in Eastern Eurasia. These admixture estimates are based on direct comparisons of several ancient Neanderthal DNA samples from the period shortly before Neanderthals went extinct, and are corroborated by indications from the nature of the genome sequences found there that the Neanderthal sequences are mutationally and recombinationally distant in time depth from the rest of the modern human genome.
Neanderthals did not contribute any Y-DNA patrilines or mtDNA matrilines to modern humans.
I suspect that what happened is that there was admixture in both directions in some quite specific and parallel patterns.
Neanderthal men impregnated modern human women outside lasting pair bonds, and the modern human women gave birth to hybrid children nine months later in their own modern human tribes. Those children had their mother's mtDNA and due to Haldane's law, the fertile interspecies hybrid children that were produced were overwhelmingly female. Some of the more robust arguably modern human remains may have had significant Neanderthal admixture.
Neanderthal women likewise would have been impregnated by modern human men outside lasting pair bonds and the Neanderthal women gave birth to hybrid children nine months later in their own Neanderthal tribes, the last of whose descendants died out when Neanderthals went extinct around 29,000 years ago. They hybrid children would have had Neanderthal mtDNA and due to Haldane's law, the fertile interspecies hybrid children would have been overwhelmingly female, so they would lack modern human Y-DNA.
The interspecies liaisons were probably sometimes rape and some times "flings" during seasonal periods when Neanderthal and modern human tribes were near each other. But, the great gaps between the two species probably impaired any hope of successful long term relationships. (Matrilocal long term pairings would lead to the same result, but the little evidence that we have suggests that Neanderthals were patrilocal and that patrilocality was probably more common than matrilocality for modern humans.)
The Neanderthal Châtelperronian industry around 35,000-29,000 BP, about 5,000 after a dramatic transition from Neanderthals to modern humans that reached a tipping point as a result of volcano driven climate change, is sometimes described as a crude attempt by Neanderthals to copy modern human tool making methods. But, I suspect that there was more to it than simply Neanderthal efforts to copy modern human tool making. My suspicion is that this industry was also a product of modern human admixture with late Neanderthals, after hundreds of thousands of years without innovation in tool crafting technologies. Modern human admixture may have made some level of innovation possible, but without membership in a modern human tribe, the necessary training to perfect those methods would have been absent.
The universality of Neanderthal DNA at similar levels in all non-Africans, and the apparent lack of overlap, suggest the following narrative.
Neanderthal admixture happens early once modern humans leave Africa. It may happen in the Levant in the time period from 100,000 to 75,000 years ago, or more likely in my view, given the lack of overlap, the Western group and the Eastern group admix with Neanderthals separately and in parallel admix at similar rates, with Neanderthals who are initially present in both areas, over thousands of years, until the Neanderthal go extinct locally. Two separate sources of Neanderthal genes would explain the limited overlap of specific Neanderthal source genes between Western Eurasia and Eastern Eurasia, and the processes involved in each case would probably be quite similar since they involve two parts of a recently split population with similar cultures interacting with the same species in similar environments.
Members of the West Asian branch of modern humans, predominantly Cro-Magnons of certain subgroups of mtDNA haplogroup U, colonize much of Europe once the population barrier created by the presence of Neanderthals in Europe weakens due to volcano driven climate change opening the flood gates to a surge of more effective modern human hunters with better tools who outnumber them, admixing at low rates with the shrinking Neanderthal population as they advance and developing higher levels of Neanderthal admixture than other modern humans. But, after the first 10,000 years or so of this, Neanderthals are extinct, but the influx of modern humans from West Asia and Central Asia and Southwest Asia to Europe continues, their population is reduced and diluted as they retreat to refugia during the Last Global Maximum ca. 20,000 years ago, and further replacement and dilution takes place in the Younger Dryas repopulation of Europe and in subsequent Neolithic, dairy farming Neolithic and Indo-European migrations, diluting the elevated levels of hybrid Neanderthal admixture in Cro-Magnons from the time period when the two hominin species co-existed for long periods among people whose lines of descent stayed on the boundaries of the expansion to unmeasurably low levels.
Homo Erectus, Denisovians and Hobbits
In my view, the most plausible possibility is that Homo Erectus to the East of South Asia (separated from other hominins from ca. 1.9 million years ago until the encounter modern humans ca. 100,000 years ago) are the same species whose ancient DNA was extracted from a cave in South Siberian Denisovia from ca. 40,000 years ago - probably a representative of the Northern extreme of this population when forced by modern human population pressures to a refugium outside their comfort zone.
A first wave of modern humans arrives not earlier than 100,000 years ago, and probably not in large numbers until they are emboldened by weakness in the Homo Erectus population due to climate change caused by the Toba volcano ca. 75,000 and possibly facing push pressure that started their migration from a West Asian/South Asian start by the exile of modern humans from the Levant in favor of Neanderthals, and possibly not even until 50,000 years ago, perhaps emboldened by cultural developments within their West Asian and South Asian refugia.
Due to their superior competition for resources, larger numbers and possibly outright warfare, modern humans prevail over Homo Erectus sending them to extinction outside relict populations. Modern humans are smarter and more adaptable than Homo Erectus, are stronger, have better tools, are probably better organized as team players, and are superior hunters who also make more systemic use of fire. And, these Homo Erectus, who had not even acquired Achuelean industry to replace their more primate stone tools, perhaps because the abundance of their environment made innovation unnecessary to survive, were even more vulnerable to the modern human onslaught than archaic hominins elsewhere.
This relative weakness means that the replacement of Homo Erectus by a wave of incoming modern humans happens more rapidly than it does with Neanderthals vis-a-vis whom modern humans have less of a competitive advantage, so the admixture percentage for the most part is much lower than it was with Neanderthals to the point of being almost negligible except in rare cases like HLA complex genes that have selective advantage and survive from a very small percentage of Homo Erectus/Denisovian genes. But, in this pre-LGM, high sea level era, Indonesian islands and territory across the mountains in Siberia falls much less rapidly to incoming modern humans than the rest of the mainland, because they are not so greatly outnumbered there.
Probably over a couple of thousand years or less, a wave of modern humans including the proto-Papuans and proto-Australians, who are at that point a single founding population that will also seed Filipino Negrito populations in part, advance along the island chain somewhat haphazardly, and their very small founding population admixes with relict Homo Erectus giving rise to a fixed 8% admixture percentage, probably before possibly absorbing a small second wave of mainland modern human migrants (that is large relative to their population), and before fissioning into proto-Papuans and proto-Australians.
The exception to the pattern of rapid extinction of Homo Erectus in the face of advancing waves of modern humans is in Flores, Indonesia where the Hobbits co-exist with modern humans from about 40,000 to 18,000 years ago, possibly because they required fewer resources to survive, their small size made them seem like less of a threat, and the Hobbits may have been co-opted into a servant-master relationship with the modern humans. Hobbits probably make first contact with modern humans in the form of the proto-Papuan/Australian population. Also, the Wallace line gap insured that the number of modern humans who got this far was much smaller and less intense than it had been relative to Homo Erectus in mainland populations.
The much larger Denisovian admixture percentage in this proto-population (probably about 8%), the fact that Flores is a necessary stop on the path to Papua New Guinea and Australia, and the lack of notable Denisovian admixture percentages in mainland populations, and the 22,000 years of stable co-existence between the two species on the relatively small island, suggests that perhaps the principal source of Denisovian admixture in these populations was actually with the Hobbits of Flores and not elsewhere, as relatively long term relationships and higher levels of interaction might have been possible on Flores than elsewhere. Homo Erectus did not make it to Papua New Guinea or Australia (at least prior to modern human arrival), so Flores may well have been the very most distant refugium of Homo Erectus in the world to the South of Siberia before their extinction. Perhaps hobbit servants even accompanied modern humans to Papua New Guinea and Australia and admixed within these very small founding populations more readily as the available choice of mates was scarce.
There is even some indication that Flores, linguistically, shows the simplifying elements associated with children learning a language from non-native language speakers that could be a legacy of thousands of years of Hobbit nannies not fully equipped to learn modern human languages and of interaction with Hobbits in day to day life that influenced language development there that is not seen in neighboring islands, although the age depth of an apparent Hobbit presence looks a bit great for that hypothesis to fit.
As in the case of Neanderthals, Haldane's law would insure that most fertile interspecies hybrids (and the species distance would have been even greater for Denisovians than for Neanderthals) were female (and hence lacked Homo Erectus Y-DNA). In the case of modern human mothers, the Denisovian mtDNA would have been absent even in first generation modern humans. Hobbit mothers may not have been able to give birth to hybrid children given the impact of the immense size disparity between the mothers and fathers, perhaps instead dying in childbirth in a death that also kills the baby or having stillbirths.
The strong genetic similarity of Indonesian Homo Erectus (as evidenced by their genetic legacy in Australian aborigines and indigenous Papuans) and Denisovians suggests that the Eastern Homo Erectus population had by diffusion over a couple of million years remained a fairly coherent single genetic population. Their lack of expansion beyond Southeast Asia and Southern East Asia until the very end suggest that they were unable to cope with cold climates with the level of success that Neanderthals did and were not inclined to try absent great duress.
Millions of years of exposure to Homo Erectus helped megafauna in the region, which was in any case not as marginal in a biologically rich environment, survive and co-evolve to the point where they were able to survive the onslaught of modern humans. But, in Australia and Papua New Guinea and the Americas and Northeastern Eurasia, a lack of exposure to modern humans meant that they had not co-evolved to handle these threats and megafauna were quickly hunted to extinction when moder human arrived.
Subsequent waves of migration into Southeast Asia and East from South Asia (and eventually from further East), after Homo Erectus was extinct in almost all of its range, diluted and replaced the original modern human populations that had co-existed with Homo Erectus and thus had an opportunity to admix with them. This would greatly dilute the percentage of admixture in later populations in mainland Asia and Indonesia to the West of the Wallace line (much as later waves diluted excess Neanderthal admixture percentages in Europe).