Arachige, Darshi, Do the Estimated Admixture Times Confirm the Proposed Holocene Gene Flow from India to Australia? (March 10, 2013). Available at SSRN: http://ssrn.com/abstract=2231116
This paper argues that the current estimates for the time of influx of Indian genes into some sections of Australian Aboriginal population during Holocene bear large uncertainties which make elimination of the probability of a more recent gene flow less likely. It also highlights that indications for the plausibility of a later gene flow exist and can also be placed in a likely archaeological perspective.
In the recent proceedings of the national academy of sciences, a team from Max Plank Institute published a research paper to confirm the recent genetic connection between Indians and Aboriginal Australians (Pugach et al, 2013). The team based their conclusion on a study involving single-nucleotide polymorphism. This research paper generated much interest and was widely written about (The Economist, Jan 2013; Nature, Jan 2013). The authours concluded that there was a gene flow from Indian Subcontinent as late as 4230 years ago. This confirmation of the previous research by Redd et al (2002) and Redd and Stoneking (1999) can become crucial in establishing this gene flow hypothesis as fact. In this investigation, it will be attempted to have an unbiased look at their analyses and conclusions by comparing them with other relevant studies. It would also be argued that an Indian genomic influence more recent than the proposed one is not improbable if the available analyses and arguments are further looked into. One of supportive arguments, i.e. the time of admixture, which can be used to exclude this possibility, can be shown as inadequate due to the reasons pointed out in the first few sections of the present paper. Two key points that will be looked at are the sample selection and the results from the Principal Component Analysis as they might have had an adverse impact on the estimation of the time of gene flow, demanding some error calculation around the estimate. The possibility of a similar gene flow taking place in a more recent time will also be explored. This later investigation will use the errors associated with various estimated times for the proposed genomic contact to argue for a more recent encounter between Indian visitors and Australian Aboriginal people.
Sample of the Pugach Study
Pugach et al (2013), following similar studies prior to theirs, assembled data from published sources including the HapMap Project. The samples used include twenty individuals from each of the following groups: Yorubans from Ibadan, Nigeria (YRI); the individuals of northern and western European ancestry living in Utah (CEU); Han Chinese individuals from Beijing, China (CHB); and Gujarati Indians from Houston, TX (GIH). The study used twelve Aboriginal Australians from Northern Territory (AUA) and twenty five individuals from the highlands of Papua New Guinea (NGH). Looking at the samples from various populations as presented in the Fig S1, one cannot stop wondering about the unusually homogeneous samples representing Europeans (CEU), Chinese (CHB) and Africans (YRI). To see such tightly packed clusters in Principal Component plots, the original data points for individual contributors, if representative of the wider population, should have very slight variability which is not explained in the paper or the coefficients from a principal component are having variable impacts on different elements of the data vectors to give similar overall scores for the small number of individuals in the group. This begs caution when interpreting the PCA plots.
Furthermore, it is good to be conscious of what the International HapMap 3 Consortium (2010) has to say: “None of the sample sets can be considered completely representative of a larger population, nor certainly of an entire continent. Thus, for example, references to the “African,” “Asian,” or “European” “populations” should be avoided when referring to these samples.” Thus, it requires caution before concluding the link between the broader Aboriginal Australia and Indian subcontinent. The people from prehistoric archaeological sites where the early inhabitants of Australia (McEvoy, 2010) were discovered had not been included in the investigation. Thus, the conclusions from Pugach et al (2013) should not be applicable to all of Australia.
It is also important for one to look at the analyses which helped the authours to arrive at the link between Indian genomes and the Australian ones. This would help us interpret the robustness of the estimate, given in the paper, of the time to admixture in the proper context.
Thus, as the authours relied on the first principal axes from a series of Principal Component Analyses (PCAs) for dimensionality reduction in the time estimation phase via stepPCO (Pugach et al, 2011), the estimation of admixture times might have been impacted by the less clear signals coming from these repeated PCAs. The robustness of these repeated PCA results can be distantly judged from the related analyses where they employed PCA. The following review, emphasising each PC, is an attempt to see the rigour of the results from such related analyses.
Based on one of the initial PCAs, interpreting Fig S1, Pugach et al (2013) commented:
“AUA are close to NGH but extend toward the European/Indian/Asian grouping, suggesting a common origin with the former and admixture with the latter.”
This comment seems to set the tone for the arguments, the paper is addressing. More obvious deduction from Fig S1 is that given the size of Eigen values, the first two principal components are the important ones. To use a broader generalization, PC1 is clearly representing Australian and New Guinean cluster versus the others while PC2 is polarised as African/Australian/ New Guinean and others. When the two components are taken together, we have three distinct clusters, namely, African as a stand-alone cluster, Australian/New Guinean cluster and Chinese/Dravidian/Gujarati/European cluster. Any further interpretation may amount to treading on unstable ground. As interpreting diagrammatic representations can be associated with a degree of subjectivity, the thesis about admixture and common origin based on these PCs seems a tenuous deduction.
In an immediate sentence, the authours made another comment;
“AUA and NGH are separated along PC4, after the separation of CEU and CHB along PC3 (Fig. S1B). The prior separation of CEU and CHB could suggest that AUA and NGH diverged after European and Asian populations,..” Given the Eigen value distribution, it might be reasonable for the authours to look at PC3. However, in the light of large dimensionality of the dataset which makes identification of contributing dimensions hard, it is very difficult to understand the inclusion of PC4, which is almost similar to PC5 in magnitude. Furthermore, it may not be justifiable to interpret PC3 and PC4 in a temporal sense i.e. AUA and NGH diverged after European and Asian populations. In Fig S1C, PC1 shows the separation between AUA/NGH groups from other groups. If we wish to include PC2 as well, the conclusion would be the separation of AUA and Pilipino Negrito (MWA) groups from the South East Asia and New Guinea groups. If both PC1 and PC2 are considered together, then, there are five clusters to be seen such that AUA, NGH and MWA clusters individually stand out from the rest. Looking at Fig 3A, obvious conclusion based on PC1 which is associated with the most substantial Eigen value, is the differentiation between AUA/NGH and other groups. PC2 which is substantially less important, place Indians, Australians and New Guineans together. All the PC graphs indicate the closeness of AUA and NGH groups more than anything else.
Furthermore, if we carefully look at Fig S8 A, it is obvious that PC1 separates Indian/European cluster from the Chinese/South East Asian/New Guinean/Australian cluster obscured by the individuals from Nysha. PC2 does not seem to be interesting unless the Eigen values, which are not given, show otherwise. If we carry through the arguments put forward by the authours in the previous sections of the article, the scores spread along PC1 shows more affinity between South East Asian and Australian samples which are closer together.
When using similar analyses, it is always good to remember the advice by the statistician F.M.C. Marriott.
“It must be emphasised that no mathematical method is, or could be, designed to give physically meaningful results. If a mathematical expression of this sort has an obvious physical meaning, it must be attributed to a lucky chance, or to the fact that the data have a strongly marked structure that shows up in the analysis. Even in the latter case, quite small sampling fluctuations can upset the interpretation..” (as quoted in p.53, Everitt and Dunn, 1991).
Main thesis Pugach et al (2013) are trying to establish is a Holocene gene flow from India to Australia which did not pass through South East Asian region or Papua New Guinea. However, PCA graphs alone do not seem to carry the thrust of the argument in the paper as the closeness of NGH and AUA is very pronounced. Thus, the use of PCA for dimensionality reduction in the context of further deductions i.e. admixture time should be treated with some caution due to possible differences in expected and observed information contents.
Time of the Indian Gene Flow
Furthermore, the Australian sample used is from an area very vulnerable to external influences. Population in the area had been in contact, for an example, with Indonesians for a long time so that their periodic arrival in boats were woven into the Aboriginal culture in Northern Australia (p412-420: Mulvaney and Kamminga, 1999). Thus, conclusions drawn from the study should be confined to the parts of Australia contributing the study sample, i.e. Northern Territory and should be viewed in perspective to sampling limitations. As the authours assert, Fig 3B, the ADMIXTURE chart for four group scenario, showing colour signals from Indian samples spreading into AUA sample can be interpreted as indicative of the admixture between Indian visitors and Australian people living in the area. Thus, the next step of determining the possible time during which such admixture could have taken place naturally follows the above.
Turning to the issue of time of admixture between Indian genomes and Australian genomes, Pugach et al (2013) claimed that it took place 141 generations or 4230 years ago. The authours used 30 years as the length of a generation. Another study addressing the same issue about Indian – Australian gene flow (Redd et al, 2002) used a generation length of 25 years. If we use this in Pugach study, we end up with an admixture time of 3525 years, which is 700 years later. The estimate done by Pugach team is smaller than 4875 years which was estimated using Y chromosomal data and a Bayesian method by Redd et al (2002). As usual with Bayesian methods, this Time to Most Recent Common Ancestor estimate is based on prior information, a value for effective population size, an unknown. The 95% CIs for two methods Redd et al (2002) employed were: (300–2,775 years) and (1,255–13,300 years). The first estimate was based on the assumption of a linear trend and the second was based on the Bayesian method. According to Redd et al (2002), the widely used linear trend method is considered as a robust method to estimate times of divergence. It is also noteworthy that the authours hadn’t used any sample from the Aboriginal populations who were less exposed to northern neighbours. One of the pioneering papers to promote the idea of gene flow from the Indian subcontinent was written by Redd and Stoneking (1999). They used the Multidimensional Scaling, which under some circumstances can be equivalent to PCA, and tree diagrams to investigate the proximity of Australian and Indian groups. The estimates for the time of separation for Australian and Indian populations were (1,686–5,093 years) based on net genetic divergence estimate of 0.03% (+/-0.03%). Note the large standard error associated with the genetic divergence estimate. Now we can clearly see, apart from Pugach et al (2013), the other two teams provide confidence intervals for their estimates which are substantially wide. Noting that the wavelet method uses first principal component, which only captures a part of the variability from the available SNP information, it is not unreasonable to expect a range of values rather than the absolute estimate of 4230 years.
Form a statistical point of view if we choose a time point between 1700 and 2000 years prior to present generation, that time point would sit snugly within any of those confidence intervals. This is especially so when Y chromosomal data and the linear time trends are used (i.e. 95% CI: 300–2,775 years) or a slightly faster mutation rate is used with the divergence values from Redd and Stoneking (1999). For the following discussion let us focus on 1700 years prior to the study generation. Now we have to see whether there is other evidence provided to support the proposed Holocene gene flow hypothesis.
Holocene contacts between India and Australia
All three research teams, namely, Pugach et al (2013), Redd et al (2002) and Redd and Stoneking (1999) use the same set of archaeological evidence to support their theory.
- change in stone implements technology during Holocene
- new ways of food processing
- appearance of dingo in Australia
- the expansion of the Pama-Nyungan languages
However, none of the authours could directly link these changes to Indian influence. There would be many questions one would like to ask. For an example, if the Indians could come to Australia by-passing South East Asia during the proposed times in Holocene, they had to have very good navigational skills perhaps ahead of their times. What did cause such a navigational adventure? According to the proponents of the theory, Dingo had to come on board a vessel (Redd et al, 2002) navigated by those early Indians. Why did such an advanced group of people only bring small tool technology when they themselves had emerged out of Stone Age in their homeland? Did Indians detoxify and use cycads as food, the way the Aboriginal Australians did?
According to Morwood (2002), some 5000 years ago dingo was introduced to Australia presumably from Timor. Dingo like skeletal remains of this age had been found in Timor (p22). Furthermore,
“The dingo originated from a population of East Asian dogs. Type A29 was one of several domestic dog mtDNA types brought into Island Southeast Asia, but only A29 reached Australia. The dingo population was probably founded from a small number of animals, as the last trickle of domestic dogs through a series of bottlenecks, or even by a single chance event and has since remained effectively isolated from other dog populations. The dingoes may have arrived in connection with the expansion, starting ?6,000 yr ago, from south China into Island Southeast Asia of the Austronesian culture (Savolainen et al, 2004).” Their results were further confirmed by Ardalan et al (2012) stating that on genetic evidence, dingo had to come via Papua New Guinea. It would be going against the evidence based on genetic studies to conclude that dingo originated from a dog with an Indian ancestry and arrived in Australia on a boat. Morwood further assures that the most aspects of cultural change over the past 5000 years can be explained in terms of indigenous Australian developments (p.24: Morwood,2002). Let alone the Indian influence, Mulvaney and Kamminga (1999) claim that new implement types, arrival of dingo and detoxyfication of cycads lack evidence as “a single cultural package from overseas (p258).” Thus, it is very unlikely that there exists an irrefutable connection between the aforementioned changes and purported Indian influence during the proposed time in Holocene. Thus, the aforesaid cultural changes alone do not substantiate a Holocene gene flow of the said nature.
More Recent Contacts
It is also noteworthy that Kumar et al (2009) finding a link between the M42 haplogroup specific to Australian Aboriginal people and seven mitochondrial genomes identified from 966 individuals belonging to twenty six relic tribes of India concluded:
“The divergence of the Indian and Australian M42 coding region sequences suggests an early colonization of Australia,~60 to 50 kyBP, quite in agreement with archaeological evidences.“.
In a later study, McEvoy et al (2010) used a sample of Aboriginal people from Riverine Region of Australia which includes Lake Mungo to carry out a genome-wide study of SNP diversity. This study found that Australian Aboriginal people from the Riverine Region had a deep link to Papua New Guineans and Melanesians. They couldn’t detect genetic evidence for any substantial migration prior to European colonisation of Australia. This result has been further confirmed by Rasmussen et al (2011) using a genomic sequence which was obtained from a 100-year-old lock of hair donated by a Western Australian Aboriginal person in the early 20th century. These results prove that the validity of inferring a Holocene Indian gene flow covering the continent of Australia is, in the least, questionable. However, this does not exclude the possibility that there was an admixture between Indian and Australian genomes in a localised fashion in more recent times as Pugach et al (2013), Redd et al (2002) and, Redd and Stoneking (1999) deduced through their research.
The analyses by the authours other than Pugach et al (2013) demonstrate that the date of such an admixture can be more recent as the confidence intervals around the divergence times are relatively wide. Therefore, we can go back about 17 centuries still agreeing with the estimated confidence limits calculations. Do we have any evidence, though vague, for Indian presence in the region closer to the above time frame?
Based on the Song Cycles of Aboriginal people in the Arnhem Land, the anthropologists, Berndt and Berndt (1954) described the encounters between the natives and a group of foreign visitors identified as Baijini who came to their shores before Macassan traders. According to the description by Berndt and Berndt (1954), Baijini visited the Arnhem Land in ships, brought their families (with kids), cultivated rice, built stone dwellings and stayed for long periods. The women wore colourful sarongs and cultivated rice. Berndt and Berndt (p36: 1954) mentioned about references to a shipwreck as the reason for the first landing of the Baijini. They are considered as an Indonesian contact which happened in the early sixteenth century (p15: Berndt and Berndt, 1954). Aboriginal people particularly remember the Baijini apart from Macassans due to their golden copper-coloured skin and presence of women. Women manufactured cloths for their own use. Some historians say that these are about the experiences of Aboriginal people who went to the lands of Macassans on their return trips back home (p421: Mulvaney and Kamminga, 1999). No one knows for sure who these people were or whether they really existed. A detailed discussion on various viewpoints about Baijini can be found in Turner (2006).
Do we ignore the song cycle about Baijini as yarns about imaginary beings or mixed-up cultural memories? Or do we take their historical presence serious as Berndt and Berndt (1954) did? The answer seems to depend on what we wish to do. However trivial or unlikely these cultural memories may sound, is it fitting for us to discredit them as ‘time-wasters’ out of hand? Believing that the Baijini had been real can lead one to more hypotheses about the past that can be further explored. Few important characteristics of the Baijini worth repeating here are their entry into Arnhem Land’s oral history before Macassans, travelling with their families, their golden copper skin colour, rice cultivation and manufacture of cloth.
Indian Gene Flow
The possibility of a Holocene gene flow in the manner proposed in the above papers is generally doubted as the external evidence are lacking; arrival of dingoes, detoxification of cycads and small stone tool technology cannot be directly linked with such an admixture. Research conducted by Pugach et al (2013), Redd et al (2002) and Redd and Stoneking (1999) show some evidence as to a localised genomic exchange between Dravidians from Indian subcontinent and Australian Aboriginal people from some parts of Australia. Had Dravidians been near the northern shores of Australia in historical times? In fact, around 17 centuries ago, they were in South East Asia as archaeological records confirm. Javanese chroniclers mention Indians as the colonisers of their land. According to a chronicle in the possession of a Javanese ruler, the king of India sent twenty thousand families to Java who became first successful inhabitants of the land (p78: Raffles, 1830). On the other hand there is evidence for a flourishing Hinduised society in West Java in the 5th Century AD (p83: Sastri, 1949). A five letter stone inscription in a script similar to one used by the Pallavas, found in East Java has been dated to the 5th Century AD (Sarkar, 1969; p83-84: Sastri, 1949). Sastri (1949) stated that the relics of the Mahakam valley from Borneo had been among the earliest known and dating from about 400 AD (p.106). The above dating due to Jean Philippe Vogel is not challenged by more recent researchers (Supomo, 1995);
“The oldest known inscriptions of the Indonesian archipelago are those on seven stone pillars, or y?pa (“sacrificial posts”), found in the area of Kutai, East Kalimantan, some twenty miles from the Makassar Straits. Written in the early Pallava script, these Sanskrit inscriptions were erected to commemorate sacrifices held by a King M?lawarman, and are datable on palaeographical grounds to the second half of the fourth century AD …… but they are the most important evidence that we have that testifies to the emergence of an Indianized state in the Indonesian archipelago prior to AD 400 (p.310)”. Thus, we have very strong evidence to argue that there had been a substantial Dravidian presence a little further to Australia’s northern coast about 17 centuries ago.
As now we know that Dravidians were in the vicinity of Makassar, it would have been very easy for them to visit Arnhem Land or Kimberley region with the help of the northwest monsoonal winds (p.410-411: Mulvaney and Kamminga , 1999). Thus, if Haviks and Mukris (Redd and Stoneking ,1999) who are usual inhabitants of the western side of Indian subcontinent could genetically contribute to the Australian Aboriginal genome, as the researchers suggested, the Dravidians who might have started from Coromandel Coast (p62: Banerjee, 1921) also stood a similar chance of such a contribution in rather recent historical times. Thus, given the archaeological evidence from the region and the genetic connection discussed above, the probability of a Dravidian contact around the 4th Century AD is very high.
The other important part of the equation to see whether Baijini gave away anything to show some semblance to Dravidian culture is to look for clues among the Aboriginal people who talked about them. At least from the perspective of oral traditions, encounters that took place long time ago can only be a distant memory tangled with manufactured events, places and people. As no archaeological evidence has so far been uncovered, the following discussion is about a few possibilities which are mainly speculative.
Similar to Javanese people who thought about Indians as the first settlers in their land, the Aboriginal people placed the Baijini in the distant past. Baijini travelled with their families in a similar fashion to the Indians. It may be a ship full of such migrant families to South East Asia from India ended up in the northern shores of Australia with their weapons, cooking utensils, unhusked rice, cotton wool (?), their spinning wheels, looms, pots for dyeing and various tools for cultivating and to work with stones etc during one north-west monsoonal period. It is also noteworthy that Basedow reported having seen a stone phallus (page x: Basedow, 1925) on one of his trips to the far north western corner of Australia in 1916. If we assume that the Aboriginal people inherited the veneration of phallus from the Baijini, then the mythical Baijini had a very strong connection to the Dravidians who visited from India. Note what Sastri (1949) says about lingam worship in Kalimantan:
“There is also a mukhalinga of the sarvasama type in which the square Brahmabh?ga (below), the octagonal Visnubh?ga (middle), and the cylindrical Sivabh?gha (above) are of equal length; the linga comes from Sepaoek in the Sintang division of West Borneo.” p106
According to Sastri (1949), placing lingams in shrines in Java transitioned to shrines with figures only in the 9th and 10th centuries (p.64). If the migrant families landed in Australia, by chance or not, it would not be surprising to see them, bringing along their objects of worship to Australia. And they, perhaps, came before the said transitionary period. It is not difficult to imagine that the brown skinned people from the subcontinent were seen as copper-coloured. Traditionally, planting and harvesting rice had been women’s work in South Asia. Dravidian women could also have manufactured cloths.
It should be noted that it was not attempted in this paper to prove the Baijini legend to be true or otherwise. What was intended was to show that a recent gene flow between India and Australia is more than or, at least, equally plausible as, the Holocene one proposed in the Literature and the Baijini legend may have some meaning in a historical context.
In the preceding discussion, it was shown that the possibility of a Holocene gene flow between Indian people and Australian Aboriginal people is real. However, the external evidence quoted to support the thesis of such genomic fusion around four thousand years ago is inadequate and does not enjoy the support of many experts in the field. Given the errors associated with the estimated times of a localised admixture between these populations, it is not impossible to find a more recent time for an encounter between South Indian migrants to South East Asia and Aboriginal people from northern parts of Australia. Such an encounter is far more plausible from the archaeological evidence available in the neighbouring islands. Even though it is not possible to link the Baijini gypsies with the Dravidians due to flimsiness of the available information about the former, it is a possibility worth pursuing.
Ardalan, A., Oskarsson, M., Natanaelsson, C., Wilton, A.N., Ahmadian,A. and Savolainen,P (2012) Narrow genetic basis for the Australian dingo confirmed through analysis of paternal ancestry, Genetica, 140:65–73 doi 10.1007/s10709-012-9658-5
Banerjee, Gauranganath (1921) India as known to the ancient world, Oxford University Press, Culcutta: available at http://books.google.com/
Basedow, Herbert (1929) The Australian Aboriginal, F.W. Preece & Sons, Adelaide
Berndt, Ronald and Berndt, Catherine (1954) Arnhem Land: Its History and Its People, Cheshire, Melbourne
Bonatto, S.L. and Salzano, F.M (1997) Diversity and age of the four major mtDNA haplogroups, and their implications for the peopling of the New World. Am J Hum Genet 61: 1413–1423
Everitt, Brian S. and Dunn,G. (1991) Applied Multivariate Data Analysis, Edward Arnold, London
Geiger, Wilhelm (1912) Mahavamsa, The Pali Text Society, London: available at http://mahavamsa.org/
Genomes link aboriginal Australians to Indians: Nature News & Comment (2013) Nature, doi:10.1038/nature.2013.12219
Issacs, Jennifer (1980) Australian dreaming: 40,000 years of Aboriginal History, Lansdowne Press, Sydney
Kumar, S., Ravuri, R.R., Koneru, P., Urade, B.P., Sarkar, B.N., Chandrasekar, A and Rao, V.R
(2009) Reconstructing Indian-Australian phylogenetic link, BMC Evolutionary Biology 2009, 9:173 doi: 10.1186/1471-2148-9-173
Malvaney, John and Kamminga, Johan (1999) Prehistory of Australia, Allen and Unwin
McEvoy, B.P., Lind, J. M., Wang, E.T., Moyzis, R.K.,Visscher, P.M.,van Holst Pellekaan, S.M and Wilton, A.N (2010) Whole-Genome Genetic Diversity in a Sample of Australians with Deep Aboriginal Ancestry, Am. J. Hum. Genet. 87, 297–305
Morwood, M. J. (2002) Visions from the past: The Archaeology of Australian Aboriginal Art, Smithsonian Institution Press, Washington
Pugach, I., Delfin, F., Gunnarsdóttir, E., Kayser, M. and Stoneking, M (2013) Genome-wide data substantiate Holocene gene flow from India to Australia Proc. Natl Acad. Sci. USA http://dx.doi.org/10.1073/pnas.1211927110
Pugach , I., Matveyev, R., Wollstein, A., Kayser, M. and Stoneking, M (2011) Dating the age of admixture via wavelet transform analysis of genome-wide data, Genome Biology, 12:R19: available at http://genomebiology.com/content/12/2/R19
Prehistoric migration: An Antipodean Raj (2013), The Economist, 17 Jan
Raffles, Thomas S. (1830) The History of Java, Vol II, John Murray, London: available at http://books.google.com/
Rasmussen, M. et al. (2011) An Aboriginal Australian Genome Reveals Separate Human Dispersals into Asia Science 334, 94–98
Redd, A.J., Roberts-Thomson, J., Karafet, T., Bamshad, M., Jorde, L.B., Naidu, J.M., Walsh, B and Hammer, M.F (2002) Gene Flow from the Indian Subcontinent to Australia: Evidence from the Y Chromosome, Curr. Biol. 12, 673–677
Redd, A. J. and Stoneking, M (1999) Peopling of Sahul: mtDNA Variation in Aboriginal Australian and Papua New Guinean Populations M. Am. J. Hum. Genet. 65, 808–828
Savolainen, P., Leitner, T., Wilton, A.N., Matisoo-Smith, E. and Lundeberg, J (2004) A detailed picture of the origin of the Australian dingo, obtained from the study of mitochondrial DNA, Proc. Natl Acad. Sci. USA, doi: 10.1073/pnas.0401814101
Supomo, S (1995) Indic Transformation: TheSanskritization of Jawa and the Javanization of the Bharata in The Austronesians Historical and Comparative Perspectives edited by P. Bellwood, J. J. Fox
and D. Tryon, ANU Press, Canberra: available at http://epress.anu.edu.au
Swain, Tony (1993) A place for strangers: Towards a History of Australian Aboriginal Being, Cambridge University Press
Tamura, K and Nei, M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol 10: 512–526
The International HapMap 3 Consortium (2010) Integrating common and rare genetic variation in diverse human populations, Nature 467(7311):52–58
Turner, Phyllis (2006) The colonisation of Australia prior to European settlement, the Degree of Master of Science Thesis, University of Adelaide
 From the discussion in the following sections, the question would arise why the samples from SE Asian region didn’t show clear Indian admixture while the historical evidence show presence of Indians in this part of the world for several centuries.
 Computed using average divergence between Australian and Indian C* Chromosomes (0.256) with its standard error (0.105). i.e. 0.256+/- 1.96*0.105 which was translated to a time of divergence based on mutation rate of 2.08E-03.
 The authours reported a mean of 3390 years which was based on net average divergence/mutation rate (=0.0003/ 8.85E-08). The net divergence was calculated using Tamura & Nei (1993). Also note that the use of the faster mutation rate for HSV1+HSV11 regions as quoted in Bonatto et al. (1997) (i.e. 11.5E-08) with the same coefficient of variation as derivable from the corresponding CI calculation in Redd and Stoneking (1999) yields following 95% CIs (1298 – 3920 years)
 Even around 5th century BC, Indians were likely to have navigated close to the land. As narrated in Mahavamsa, the 700 men deported from Bengal landed near Bombay first and then, after embarking from there, in Lanka. (p.xxxvi & p54: Geiger, 1912). Their women and children deported separately also landed in islands near the Indian coast.
 See p171 of Swain (1993). The poem has the following:” from the young Baijini playing?”
 This assertion by Berndts is contentious as, following Macknight, Mulvaney and Kamminga (p415: 1999) place Macassans’ arrival around 1700 AD. If Baijini were before Macassans, how did they find themselves with an Aboriginal name with a Macassan root meaning women (p 170: Swain, 1993; p16: Berndt and Berndt, 1954). The way this was possible is by Baijini being contemporary with Macassans or by Song Cycle being post-Macassan or by the Aboriginal name being not of Macassan origin. If, instead of accepting and contradicting it, we accept the Aboriginal myth that the Baijini had been there in the time of the mythical ancestors, Djanggawul sisters and Laindjung (p170-171: Swain, 1999), the origin of the name ‘Baijini’ or ‘Baiini (p261: Issacs, 1980)’ is less likely to be Macassan.
 Note that he saw a stone phallus erected in the ground and “surrounded by a cleared cirque where much blood had been spilt at a recent ceremony (p x.: Basedow, 1929).” Similarly he had seen another vertically standing three feet long stone placed in the ground by replacing a natural stone, in the Gleneleg district of the north western corner of Australia(p.288: Basedow,1925). Basedow discussed seeing the phallus worship in other parts of the continent in the form of using the phallic symbols in ritualistic dancing etc (p.282-283: Basedow, 1925). However, nothing came close to a worship than the one described in the preface to his book. According to him, phallus worship slips away from new generations. But if what was practised in the past was similar to what he saw, then the above stands out as more devout way of worshiping it; probably similar in spirit to what Hindus did. However, with much blood around what he had seen could also represent a form of sacrificial post similar to yupa stambas in Kalimantan.
 In the poem quoted in Swain (1993) we find “..hiding the ladle beneath our arm. It is sacred (p.171)”. Thus, according to the legend, the Aboriginal people had used things left out by Baijini and treated them as sacred.
 In a fifth century compilation of earlier works called Mahavamsa, there is mention about a native woman of Lanka spinning when the first Sinhalese people arrived in the island (p56: Geiger, 1912). According to the Chronicles this should have taken place around the 5th century BC. Traditionally, women in the subcontinent might have been manufacturing cloth.