In my post yesterday, I dismissed the possibility that the presence of Denisovan DNA in populations to the east of Wallace line and its absence to the west of the Wallace line was not easily explained by dilution of the Denisovan admixed population because East Indonesia has a substantial Paleolithic substrate and because the level of dilution required (on the order of a factor of sixty for South China relative to Papuans) would be so great. Have I dismissed this possibility too easily?
The evidence can be marshalled to give some support to a dilution narrative.
A Dilution By Paleolithic Y-DNA Haplogroup O Men Scenario
There is a stark difference in Y-DNA haplogroup frequencies on either side of the Wallace line. To the east are high frequencies of populations typical of Sahul populations (i.e. Australia and Papua New Guinea), with a modest amount of Austronesian (i.e. Neolithic seafarer) contribution that there is very strong evidence to suggest dates to the last five thousand years or so. On the other sided of the line, for example, on the island of Bali, the percentage of Y-DNA that is something other than Y-DNA haplogroup O is quite small, about 4%, with some of the Y-DNA haplogroup O attributable to Austronesian sources and some attributable to some other migration into Western Indonesia. While the ratio isn't quite sixty to one the cline is certainly steep and could explain much of the variance in Denisovan admixture.
One plausible time for the non-Austronesian lineages of Y-DNA haplogroup O to arrive in Western Indonesia would have been when Sundaland was attached to mainland Asia. The region where non-Austronesian lineages of Y-DNA haplogroup O are common correspond to the territory of Sundaland which was a continuous land mass attached by land to mainland Asia during the Last Glacial Maximum around 20,000 years ago. Even for people with some seafaring abilities more primative than that of the Austronesians, mass migration across land is easier than mass migration across water and the distribution of these lineages closely matches a terrestrial passage at that point in time.
This isn't the whole story, however. We know from a wealth of historical examples that when farmers and herders encroach onto the territory of a hunting and gathering population, that the former's advantages in terms of numbers and technology almost always overwhelms the hunters and gatherers resulting in major demographic shifts towards to food producing migrants. But, it is not at all obvious, in general, that one hunter-gatherer population ought to be able to overwhelm another hunter-gatherer population without a good reason. And the farming and herding were not developed until long after the last glacial maximum that lowered sea levels of join Sundaland to mainland Asia (nor, for that matter, was pottery).
One can easily imagine good reasons in a first contract between modern human hunter-gatherers and Denisovan hunter-gatherers. But, how can we explain how first wave migrants kindred to the Sahul population wave, could be totally dominated by the people who were bears of non-Austronesian lineages of Y-DNA haplogroup O?
After all, the kin of the Sahul population were notable for their superior maritime accomplishments relative to all hominins who came before them and to all who came afterwards until the Austronesians. The people of Sahul were able in a historical eye blink to bring about the extinction of megafauna every bit as frightening and huge as the monsters in fairy tales, without even having metal weapons or armor. Arguably the people of Sahul were exceptionally advanced and hence an unlikely candidate to fall en masse to the migration of another Paleolithic hunter-gatherer population into their territory. They managed to island hop across Western Indonesia when it was not connected by land into a single Sundaland penninsula, to cross the Wallace line, and to cross the sea from Wallacia to reach Sahul. In contrast, the non-Austronesian lineages of Y-DNA haplogroup O apparently needed a land bridge and never managed to cross the Wallace line in large numbers and didn't cause the extinction of ferocious giant predators.
I put aside, for the moment, the argument that these were non-Austronesian Neolithic migrants at the time that they came to Western Indonesia, because it stretched estimates of the age of these lineages more than one would like, is not associated with any known Neolithic population, and may not be necessary for the reasons set forth below.
The Dog Domestication Advantage.
So, what could the Paleolithic Y-DNA haplogroup O people have had twenty thousand years ago that gave them a decisive advantage over the awesome hunters of the first wave migrants who were kin to the Sahul people in Western Indonesia that left the first wavers totally marginalized even before the Austronesians arrived?
The most plausible explanation I can muster is a simple one. The Paleolithic Y-DNA haplogroup O people had domesticated dogs; the people who were vastly diluted by them did not.
The Sahul people did not have dogs until an archaeologically well documented moment when a handful of dingos migrated to Australia around 8,000 years ago from a stock found in Southeast Asian wild dogs. The arrival of the dingo in Australia caused a continent wide secondary mass extinction that eliminated many species that had survived the megafauna extinction that accompanied the appearance of modern humans in Australia. This mass extinction is a testiment to the impact that domesticated dogs could have on hunting and gathering economics.
Sahul populations certainly didn't have dogs twenty-thousand years ago, long before the dingo was introduced. So it stands to reason that their kin in Island Southeast Asia didn't either.
The domestication of dogs took place long before any of the other plant or animal domestications associated with the any of the several independent Neolithic revolutions. But, dog domestication had to have happened by the time that the founding population of the Americas arrived in Beringia right around the time of the Last Glacial Maximum and the existence of the Sundaland penninsula as a terrestrially connected land mass connected to the Southeast Asian mainland. In other words, dog domestication probably happened before peoples with men predominantly from Y-DNA haplogroup O swept across Western Indonesia.
While different studies reach different conclusions, some of the efforts to identify the principle ancestral source of the overwhelming majority of modern domestic dogs suggest a Southeast Asian progenitor, near the likely geographic point of origin of the non-Austronesian lineages of Y-DNA haplogroup O.
Dogs can explain how a Paleolithic hunter-gatherer population with dogs could totally marginalize a Paleolithic hunter-gatherer population without them. Since the advantage provided by dogs to hunter-gatherer populations are largely limited to terrestrial hunting and gatherering, the domestication of dogs could also shift the relative desirability of a terrestrial hunting and gathering strategy relative to a maritime fishing strategy, causing the maritime capabilities of populations that domesticated dogs to wither from disuse, particularly in areas where fishing had previously been only marginally superior to terrestrial hunting and gathering as a food collection method.
Y-DNA haplogroup O must have arisen sometime after first contact with the Denisovans. Its phylogeny makes clear that it isn't as old as the Sahul Y-DNA lineages.
If Y-DNA haplogroup O arose in an area that was temporarily purged or nearly purged of Denisovan populations by the direct effects of the Toba eruption when humans first settled there (i.e. leaving Denisovans in numbers not significant enough to have a meaningful interaction or impact on the modern humans migrating into the area anywhere but Java island and points to it east in Island Southeast Asia and in East Asia), then the ancestors of the modern humans who settled in mainland Southeast Asia, Borneo and Sumatra without going further, would not have been Denisovan admixed except from slight percentages due to back migrations from the kin of Sahul-like populations. And, one of these mainland Southeast Asian populations probably gave rise ot the Y-DNA haplogroup O lineages.
The high level of homogeneity of the Y-DNA haplogroups of East Asia similarly argue the point that extreme levels of dilution of prior Sahul-like waves of migrants though multiple waves of Paleolithic and Neolithic migration of populations who can trace their origins to Denisovan-free Southeast Asia, making any Denisovan ancestry that remains almost imperceptable, isn't all that implausibe in East Asia.
Other Possible Demographic Impacts Of Y-DNA O Expansion
Of course, the dog domestication advantage I suggest as a possible source for the expansion of men with Y-DNA haplogroup O could also have directly devistated any relict Denisovan populations, in addition to diluting the contributions of first wave modern humans admixed with Denisovans in the modern human gene pool in Asia.
An expansion of a population predominantly made up of Y-DNA haplogroup O men that really did have such a great advantage over other Paleolithic peoples could also have dramatically marginalized populations where Y-DNA haplogroup D was predominant except in places like the Andaman Islands, Japan, and LGM refugia in Tibet, that this population, which had inferior martime capabilities, could not reach.
Limits On Y-DNA O Expansion
The same jungles and mountains that provided an effective barrier to modern human migration into Denisovan territory in Asia prior to the Toba eruption, however, may also once it had grown back tens of thousands of years later, have also mutated the capacity of Y-DNA haplogroup O men to back migrate en mass and dilute into irrelevance the populations of South Asia until they received the further boost of Asian Neolithic technologies like rice cultivation that brought Austroasiatic peoples into Northeast India.
The Limits Of Early Sahul Wave Maritime Capabilities
It is also worth keeping in mind that proto-Sahulean maritime capabilities should not be unduly exaggerated. Yet, they managed to island hop across the small straights of Western Indonesia, across the Wallace line, and then from Wallacia to Papua New Guinea and Australia, respectively. Yes, there is some evidence of very early deep sea fishing in Melanesia at around that point in time. But, the developments we don't see also speaks volumes.
The proto-Sahuleans did not settle the remainder of Oceania. They also did not settle Japan or the Andaman Islands. The Azores Islands in Europe were not settled by modern humans until the Bronze Age, long after the Upper Paleolithic revolution in which the proto-Sahuleans seems to have participated. There is considerable genetic evidence that Papuans and Aboriginal Australians did not experience much population exchange at all starting shortly after their founding populations were established. It took twelve thousand years or so from the domestication of the dog to the introduction of the dingo to Australia and the introduction of the dingo to Australia shows strong genetic traces of being a single event. There is also no sign of meaningful modern human population exchange between Sahul (or its successors) until sometime after the introduction of the dingo (and perhaps no more recently than the expansion of the Austronesians).
This evidence suggests that the possibility that apart from just a handful of isolated instances of one way voyages beyond the line of sight involving only a very small Sahul founding population, perhaps numbering something on the order of a hundred people (Australian aboriginal tradition speaks of seven canoes at a time when there were no domesticated animals, even dogs), that their navigational capacabilities may have been limited to navigation within a line of sight to the mainland. It appears from the evidence that there was virtually no intentional maritime trade between Sahul and the rest of Southeast Asia from this twelve thousand year time period, and probably from as far back as 40,000 years ago until sometime quite a bit later than 8,000 years ago, perhaps as late as the point at which contact with the Austronesians began. Even if the original founding populations of Sahul and Wallacia were competent deep sea mariners, the Sahuleans themselves appear to have promptly lost this technology. When the sea ultimately separated Tasmania and the rest of Australia there was no maritime travel between the two across the relatively shallow straight. The much later development of martime capabilities in the waters between the Island of Formosa and mainland China that ultimately culminated in the Austronesian's extraordinary maritime capabilities seems to have been an independent development not traceable to the Sahulean martime tradition.