Australian Ancestry

   In Age and Origin of the Human Species I pointed out that fossil evidence shows the earliest inhabitants (Mungo Man) had a more modern morphology than later Australians. I have suggested that the first humans to settle Australia were derived from Eurasian sapiens, albeit they became somewhat hybridized with erectus during their radiation through the tropics. The much more recent Australian population typified by the Kow Swamp specimens was so robust, and the skull features so primitive, that some have argued they were H. erectus. While the weight of opinion seems to be that they are not erectus, it can’t be denied that they manifest many diagnostic features of that taxon. In fact they have strong affinities to the Ngandong skulls, which are erectus, but show so much sapiens influence that Rightmire excludes them, as atypical, when considering the spectrum of conserved features that characterize the erectus taxon: The Evolution of Homo Erectus; Cambridge, 1990; see, for instance, Figure 40, page 196. I have suggested that the Ngandong specimens were an erectus-sapiens hybrid type, and representative of the population that migrated into Australia and left the Kow Swamp fossils. Having lived in close association with tropical S E Asian erectus for over 20,000 years, the Kow Swamp type was much more affected by hybridization than the population that originally settled Australia.

     Lightly shaded areas, on the map below, were above sea-level during the last glacial epoch, when so much water was locked up in ice. From after 30 kya until 12 kya, S E Asia was almost connected to Australia, which was joined to New Guinea until 8 kya. Of course, sea levels were lowest during and shortly after glacial maximum, and the channels separating Australia from Asia were then at their minimum width. It was during this era that the Kow Swamp type, erectus-hybrid population was able to migrate into Australia.

Figure 1

[Map from Ingman and Gyllensten; Mitochondrial Genome Variation and Evolutionary History of Australian and New Guinean Aborigines; Genome Research 13:1600-6, 2003]

     The original colonists of Australia arrived much earlier. Dates as early as 62 kya have been attributed to the oldest human fossil in Australia (Mungo Man; LM3) but in Nature (Feb 20, 2003) a multidisciplinary team reported 42 kya. It is evident, just from the location of the find, that Mungo Man was not among the first to arrive; nor is it likely that era's only human bodies to be fossilized and discovered would have been the very earliest people to step ashore. There was a glacial epoch between 73,000 and 55,000 years ago (Oxygen Isotope Stage 4) and an ice sheet covered much of the central Asian plain during that era. About 74,000 years ago Mt. Toba erupted, creating a broad extinction zone in India and a 'nuclear winter' that lasted for several years. The effects of Toba and the glacial era seem to have driven the inhabitants of central Asia south. Hoffecker states, in Desolate Landscapes: Ice Age Settlement in Eastern Europe, on page 19, that, "The scarcity of artifacts in the loess bed that overlies the [central Asian plain] suggests that much of the plain was abandoned between 73,000-55,000 years ago." I have suggested that there was a massive migration of advanced Eurasian archaic sapiens, out of central Asia, into S W Asia, N Africa, and India. I believe that elements of this radiating population reached and colonized Australia during the OIS 4 era, which is plausible, as sea-levels were relatively low then. Conversely, suggested later dates for first settlement are less credible, due to the difficulty of crossing wider ocean expanses as the ice melted and sea-levels rose. 

    There is genetic evidence for Australian colonization during OIS 4 in Mitochondrial Genome Variation and Evolutionary History of Australian and New Guinean Aborigines (Ingman & Gyllensten,2003). They state that the "estimate of a genetic coalescent for Australian Aborigines with individuals from outside Australia based on mitochondrial HVS1 sequences ranges from 60,000 to 119,000 years, depending on which substitution rate is used". They note that other research, on mtDNA D-loop sequences yields a range of 51,000 to 85,000 years ago, for an "expansion date". The expansion date is when a small founding population began to multiply significantly. They give an mtDNA coalescence date for the N macrohaplogroup, which includes all but one Australian lineage, as 71,000 years ago, with a confidence interval of plus or minus 12,000 years. The M haplogroup's was 78,000 years ago, with the same confidence interval. So the evidence indicates that Australia was settled during OIS 4.

      Mishmar (2003) gives a coalescent for both M and N as 65 kya, which is within the confidence interval op cit. I interpret that data to mean that most of the ancient Eurasian mtDNA lineages were lost, either during the era of Mt. Toba’s eruption, or by subsequent lineage sorting. As an example of the latter, the LM3 mtDNA lineage, which is the most ancient human mtDNA lineage known, survived at least until 42 kya among migrants to Australia, but was lost sometime before the present.

      Afrocentrists have argued that the original Australian population came 'Out of Africa', while, on my view they came from Eurasia. It is indisputable that the Australian mtDNA derives from macrohaplogroups M and N (Ingman & Gyllensten, 2003), which are Eurasian lineages. It is true that a few African populations' lineages have affinities to M and N, and afrocentrists interpret that to mean M and N diverged from the African genome. I argue that the evidence and time constraints indicate M and N lineages were carried into Africa by Eurasian radiation: see Plural Lineages in the Human mtDNA Genome (on site). Note that all but one mtDNA lineage in Australia is derived from the N macrohaplogroup, which contains the European lineages. In other words, to argue that the first Australians were Africans, requires one to assume that M and N are derived from the African genome, rather than being indigenous to Eurasia: I contend Plural Lineages falsifies the assumption of a single African source for all modern mtDNA.

     There are a few African populations that have some mtDNA with affinities to M and N, and you can see examples of them on the branches under box 2, such as the Yoruba. The afrocentrist who drew the 'tree' below has assumed that the African mtDNA is ancestral to the Eurasian. However, LM3, the oldest known mtDNA, is Eurasian! 

       Macrohaplogroups M and N are in the dotted-line boxes of the graphic below: N is 1; M is 2. You’ll note that ALL the lineages within the boxes are Eurasian. That sole M lineage is most plausibly associated with a more recent radiation that occupied New Guinea. It is remarkable that the migrants who settled Australia were derived specifically from the population that later became Europeans, suggesting that the first settlers of Australia began their migration AFTER the M and N haplogroups diverged. Of course, it is possible the M haplogroup could have been present in the proto-Australian population and was lost by lineage sorting. After all, Mungo’s mtDNA is unlike ANY modern lineage! In other words, the population that first came to Australia included at least one type of mtDNA that has since been lost, and does NOT correspond to any of the extant, and supposedly ancestral, African lineages. One sequence from Mungo’s mitochondria found its way into the nuclear DNA of ancient humans and is carried, on chromosome 11, by many contemporary people. The extraction of mtDNA from LM3 was reported by Adcock, et al. and published as Mitochondrial DNA sequences in ancient Australians: implications for modern human origins. It was printed in Proceedings of the National Academy of Science, USA 98 (2001) page 540. Figure 2, below from Ingman and Gyllensten, 2003, loc. sit; reproduced by them, from Saitou and Nei, 1987.

   Fig. 2

      The ''implications for human origins'' are that modern humans are NOT 'all Africans'! African replacement is contraindicated by the population distribution of the relic mtDNA sequence inserted in chromosome 11. In A Nuclear ‘Fossil’ of the Mitochondrial D-loop and the origin of Modern Humans, by Hans Zischler, et al, published in Nature 378:489-492, page 491, we read,“Overall, 39% of chromosomes tested carried the insertion. In four African populations, the frequency of chromosomes carrying the  insertion ranges between 10 and 25%, whereas it varies between 38%  and 78% in populations tested in Europe, Asia, Oceania, and South America.” The highest frequency of insertions were among the Japanese @ 65%; an Amerind tribe (Surui) @ 78% and the Melanesians @ 68%: Europeans were @ 54%. Clearly, this insertion of ancient, indigenous Eurasian mtDNA, into the nuclear DNA, occurred outside of Africa, probably in N E Asia. Since the proto-Australians carried the mtDNA lineage that was inserted, it is more likely that they were an Eurasian population. Nuclear DNA, as discussed below, practically excludes the possibility that Australians are derived from Africans. Therefore, since Australian mtDNA is derived from the N lineage, and LM3 from the insert, there is no justification for the assumption that M and N are of African origin. This evidence is more consistent with the view that those lineages were carried INTO Africa by the radiation of Eurasian humans (H. sapiens sapiens). Only the unjustified assumption that ALL modern mtDNA comes from a single, African source allows the presumption that M and N are African-derived.

      Previous sequence analyses of the chromosome 11 insert had revealed NO allelic variation in any sample from any part of the world. While exhaustive analysis indicates that the LM3 sequence derives from the same mtDNA lineage that engendered the insert, there are 14 nucleotide differences. The range of differences between the LM3 sequence and contemporary sequences is at the upper end of the range of differences between contemporary sequences. Since the number of  differences is a proxy of age, we can infer that LM3 was as old as the older African lineages by 42,000 years ago! The range of differences between the insert sequence and contemporary sequences extends well beyond the range of differences between contemporary sequences. This indicates that the lineage leading to the insert sequence diverged well before the most recent common ancestor of living human mtDNA sequences. In other words, the LM3 lineage is OLDER than the 'African Eve', and the insert is far older than LM3 ... perhaps as much as 14 mutations older!

      Because mtDNA accrues variation much more rapidly than nuclear sites, and because there was no observed allelic variation in the insert, it is plausible that those 14 differences accrued to the LM3 mtDNA lineage since the era of the insertion event. Thus, it is evident that  insertion occurred long before the coalescence of the most ancient African mtDNA lineages. Given the geographical distribution of that chromosome 11 insertion, It is hard to imagine stronger evidence for the continuity of archaic Eurasian populations. Even afrocentrists concede that by implication, as in this tree of mtDNA lineages.

Fig. 3; Klein and Takahata, WHERE DO WE COME FROM? The Molecular Evidence for Human Descent, page 281.

     The length of the line, from left to right, corresponds to the estimated time-depth of that lineage. Note that the insertion 'branch' extends far to the right of the most ancient African mtDNA lineages, as well as the Neanderthal specimen. Even so, I think the scale understates the insertion’s antiquity, as 14 mutations, plus the 40 thousand year age of the Mungo specimen, represents an immense time-span! In an equivalent sequence of that Feldhofer neanderthal specimen, Krings et al. found ''11 transitional differences from the [modern human] reference sequence''(as reported in DNA sequence of the mitochondrial hypervariable region II from the Neandertal type specimen, PNAS Vol. 96, 5581-85, May 11, 99). Approximately half those differences may have accrued through changes in the human lineage. By contrast, it is hardly possible that as many differences accumulated in the chromosome 11 insert as occurred in the LM3 lineage, because nuclear DNA changes much more slowly than mtDNA, let alone the hyper-variable region. Accordingly, we may infer that the coalescent between the LM3 lineage and the chromosome 11 insert is significantly older than the most recent common ancestor of humans and Neanderthals. Krings estimates the mtDNA of humans and Neanderthals began to diverge 465,000 years ago; add in the 40,000 year age of LM3 and it is looks as if the insertion on chromosome 11 occurred at least half a million years ago, or even a million, as Kring’s upper age estimate is 741,000 years ago!

     Here is the lineage diagram Adcock presents. Note the tree is rooted in the common ancestor of Homo and Pan; the Neanderthal lineage is separate from Hss; and one of the Australian fossils robust types is also shown as distinct from all other modern human lineages. I contend such specimens resulted from hybridization with Asian erectus.

Figure 4, from Adcock, 2001, op cit

     As noted, the modern Australians’ mtDNA comes from macrohaplogroup N (and, to a minor extent, from M) and that they have affinities to Europeans. That is remarkable, when you consider that the original settlers must have lived in S E Asia for some time, and that later waves of migration to Australia show strong affinities to the erectus-sapien-hybrid Ngandong population. In fact, that tends to corroborate the view I expressed in Age and Origin: Eurasian sapiens radiated through the tropics, interbreeding with indigenous erectus, and later migrations displaced the most primitive of those hybrid populations.

     As the afrocentrists claim the M and N lineages diverged from the African mtDNA genome in N E Africa, it is instructive to compare the nuclear DNA of Africans and Australians. On page 75 of The History and Geography of Human Genes (Cavalli-Sforza, Menozzi & Piazza; Princeton, '94) we find that Africans and the people of Australia & New Guinea are the most UN-related populations on the face of the earth! The Bantu and N. Guineans have a genetic distance of 3372, while the Australians are 3272, exceeded only by the Thai at 3364. To put these units of genetic distance into perspective, consider that Europeans are separated by distances two or three orders of magnitude less! English and Italians are separated by only 4 units, compared to the 3,272 units that separate the Australians from typical Africans. Even the Basques and Greeks are only separated by 5 units; the Lapps and Danes by 204. There is another common method of calculating genetic distance, and it is expressed in a different numerical scale. The same authors used that method too, and it yielded equivalent results: Africans and the Australians are the most UNrelated populations on earth.

     IF Africans settled Australia and M & N macro-haplogroups are African derived, one might suppose that the Australians would be substantially more closely related to the African populations that have some mtDNA lineages with affinities to the M and N haplogroups. But, compare the people of Australia and N.G. specifically to the Africans that are descended from the population supposed to have been ancestral to the M and N macrohaplogroups: the Yoruba, for instance. The Yoruba’s distance from the most unrelated west-African group is only 398. The N E Bantu’s distance to the Bane, at 153, is less than the difference between groups of Bantu, and the Bane cluster with the Yoruba. Note how the Bantu {“BNT”} and Bane group (along the 2nd principle component axis) while all the northern African populations are separated by the first principle component. On my view, that is because N African populations are the ones most affected by gene flow from Eurasian radiations, especially since the agricultural era. By contrast, the less-mixed African populations are separated from each other by distance from the second principle component axis.  

Fig. 5, from Cavalli-Sforza, Menozzi, and Piazza, page 178, op cit.

     Bear in mind that the 'genetic distance' figure is based on sound data: the frequency and distribution of various alleles and polymorphisms. However, the 'trees' that authors construct with that data are usually based on debatable assumptions: that Africans are an ancestral population, and that modern humans are all descended only from H. sapiens. On my view, Africans are a hybrid population, created by the interbreeding of Eurasian sapiens with H. erectus, just as the Australians were. In order to accurately reflect the human lineages, such trees would have to distinguish between the ancestry of Eurasians and the tropical populations created by recent back-cross hybridization with remnant erectus. The trees would, in fact have to be networks. To give him his due, Cavalli-Sforza acknowledges that as a possibility, loc cit page 81. The authors state that, ''a tree with inter-connections would be highly desirable'', albeit he is apparently only considering the mingling of modern human populations, rather than taking erectus into account.

     All but 2% of the 'bootstrap' trees generated by Cavalli-Sforza et al. showed the first separation between Africans and all other populations, but the way ''all other non-Africans'' grouped is significant. 1) in 25%, the Australians+NG+Pacific islanders & S E Asians vs. all other non-Africans; 2) in 24%, Pacific islanders+S E Asians vs. all other non-Africans; 3) in 15%, Australians+NG vs. all other non-Africans; 4) in 8%, Europeans+extra-European Caucasoids vs. all other non-Africans; 5) in 8%, Americans vs. all other non-Africans. In the other 18% of trees the sub-clusters disintegrate. The authors say, ''For instance, in some of them New Guineans and Australians part from each other, one of the two being separated from all other groups in the second fission''. What is really indicated, as the authors acknowledge, is ''a trichotomy, separating the three sub-clusters of NG+Australians; S E Asians + Pacific islanders; and all other non-Africans. The bottom line: NG/Australians and Africans are the most unrelated populations on earth, but BOTH of them are very different than all other non-Africans, with the reservation that there are affinities between S E Asians and NG/Australians, as might be expected due to relatively recent population mixing. I contend that the 'out of Africa' hypothesis cannot account for this. By contrast, the hypothesis that Australians and Africans are erectus-sapiens hybrid populations, is completely consistent with such data, even explanatory of it.

     A 'principal component map' shows the genetic difference between European and African populations in a graphic manner that facilitates appreciation of their minimal relationship, and avoids the unwarranted assumptions inevitable in the 'trees' that ignore the recent, as well as ancestral, influence of erectus. Ancient Eurasian sapiens' radiations led to erectus-sapien hybrid populations, such as represented by the Herto skulls in N Africa, and even more advanced forms in S W Asia. But the well-adapted tropical populations were very much larger than the radiating Eurasian's, and they absorbed the early migrations, preserving  a great deal of the erectus genome. Later sapien radiations, especially since the agricultural era, displaced the most primitive groups, but, to a great extent, those large, hybrid populations still exist.


Fig. 6, Cavalli-Sforza, et al., page 81, op cit.

     Since there is less genetic affinity between Africans and the populations of Australia and N. Guinea than anyone else in the world, it is obvious that Africans didn’t settle Australia. That casts doubt on the afrocentrist assertion that macrohaplogroups M and N are derived from the African mtDNA genome, and tends to corroborate the view advanced in Plural Lineages, that M and N are indigenous to Eurasia.

      The question naturally arises: why are Australians and Africans so genetically different if they were both created by Eurasians interbreeding with erectus? I contend that it is because the S E Asian and African erectus were so long separated, with so little gene flow between them. There, as in the mtDNA insert on chromosome 11, are the 'deep genetic lineages' the afrocentrists say should exist if the multiregionalist model is correct! The S E Asian and African erectus genomes were so divergent because they had been separated for over a million years, and that deep division is only partially unified by both interbreeding with Eurasians, principally with the population that became Europeans. That is also why there are some affinities between Europeans and Africans, in spite of their vast phenotypic differences, and deep genetic separation. The afrocentrists try to interpret this as meaning that Africans are ancestral to Europeans, but, for all the reasons documented in Age and Origin, Plural Lineages, and elsewhere, that is impossible. In contrast, those affinities are well explained by my hypothesis of Eurasian radiation into Africa, with interbreeding between sapiens and African erectus.

     Yet another method of graphic representation reveals the genetic relationships  between Eurasians and Africans. A 'neighbor-joining-tree' by the cited authors is particularly valid because it is NOT 'rooted'; in other words, no assumptions are made about which population is ancestral.


Fig. 7, Cavalli-Sforza, et al., page 91, op cit.

     Three significant aspects of the topology are that the Europeans have a CENTRAL location, on VERY SHORT branches (d & e), and there is MAXIMUM SEPARATION between Melansians (which includes Australians) and Africans. With respect to branch 'e', its minimal length indicates how closely related the Europeans are to the COMMON ancestor of Asians. One might interpret the very short 'd' branch as indicating how similar Europeans are to the original H. sapiens sapiens type, while their central position could support the same view. One might also view it as consistent with slight introgressive hybridization (of a common ancestral, Eurasian Hss type) with archaic sapiens in central Asia and Europe, leading to differentiation of the Euro-type. I attribute the long 'c' and 'g' branches to the influence of the archaic, and widely diverged, remnant erectus genome on both African and Australian/ New Guinean populations.

[Australian Ancestry Printer Friendly]