I've been wary of commenting on this paper as I seem be the least qualified commentator so far. I'm not a geneticist nor do I follow early Palaeoindian settlement. As far as any interest in American archaeology goes I tend to look at periods around the 1st millennium when it's taken for granted that the colonisation of the Americas has occurred. What I do some experience of, especially with my work with people at the Interdisciplinary Science centre at Leicester, is working with people when I don't really know much about a topic. At a result I have some experience to discussing things I know little about and I'll bring the full force of my ignorance to bear in these comments.

Briefly I think ignorance of associated fields is a factor which needs to be addressed when taking interdisciplinary work and presenting it to colleagues in home disciplines. I also think this may become for practical purposes a permanent feature of interdisciplinary research and it may be necessary to embrace a post-positivist approach to tackle such problems. Finally, I think that to be able to meaningfully discuss interdisciplinary work, it must be accessible to the disciplines the work draws upon, and that probably requires Open Access. To a greater or lesser extent this paper addresses all these problems and for that reason not only is it a fine discussion of the colonisation of the Americas, it's also a useful exemplar for other interdisciplinary publications.

Selling genetic explanations to archaeologists

Acceptance of genetic explanations in archaeology is patchy. They're most warmly welcomed when they confirm what the archaeologists were thinking anyway, which makes me wonder how meaningfully archaeologists engage with the material. This isn't a blanket criticism. Archaeologists who are unconvinced by past genetic explanations are unlikely to take the time to look deeper into the subject, judging that they could spend the time studying something more useful. Equally archaeologists excited by genetic research may spend more time looking deeper at the methodology, or may accept it regardless. I think this means there's a large pool of archaeologists who lack the skills to critique the genetic analysis of this paper and I'm one of them.

Pessimistically I don't see this improving to a huge degree in the future. I currently have Mark Jobling's textbook on Human Evolutionary Genetics on my bedside table for reading. The impression I get is that getting up to speed Evolutionary Genetics is going to take more than a weekend. As research progresses so the horizon to be able to meaningfully judge the validity of research is going to progress.

The answer is *not* to simplify the paper. PlosOne is peer-reviewed, so I satisfied it's plausible and there are plenty of geneticists with an interest in population migration. I'm hoping that as this paper ages we'll see comments from geneticists upon the methodology. I don't expect to understand those comments fully either, but seeing that kind of analysis will lend confidence to those results and might even be transparent enough that if there is a problem then another comment asking "Would this new archaeological result from Alaska make a difference?" might help push discussion of the model forward.

There's nothing the authors can do about this. They've taken the step of putting the paper here and we have to wait.

Positivism vs. Post-Positivism

Another, wider, problem may be the issue of how the scientific method works. This may be a personal peculiarity, I studied in Europe where archaeology is often seen as allied to classics or history rather than anthropology, but the view of some archaeologists over here of science is a little simplistic. For some people processual archaeology is still basically hypothetico-deductive reasoning which takes the view: hypothesis to test to possible falsification. Working within Archaeology this paper could be said to be weak because much of what it proposes is archaeologically untestable,

It may be tunnel-vision to work within one discipline but it there are similar situations in genetics. For example Capelli et al.'s 'A Y Chromosome Census of the British Isles' (Current Biology, Vol. 13, 979–984, May 27, 2003, http://www.ucl.ac.uk/tcga... ) concludes that Norfolk is one of the most Viking places in the British Isles based on DNA studies. The paper doesn't mention that in the 16th century a quarter of the population of Norwich was Dutch and it doesn't have a Dutch comparison sample. It's possible this is a trivially easy effect to allow for, but the lack of historical context makes the work unappealing to historians no matter how sound the genetics. I'm happy that the work of Capelli et al is a Rolls-Royce, but it seems to be driven into a cul-de-sac. This paper in contrast tackles the problem superbly, by placing the discussion of the genetic data in an archaeological context. What is proposed is a challenge to the Clovis First model. While the archaeological sites mentioned would seem to demolish that idea there seems to be curious lack of pre-Clovis sites.

Rather than seeing this as speculation in the absence of evidence, I'd follow David S. Whitley's arguments for a post-positivist approach which is mentioned and referenced in a few articles, all of which I cannot access as my university does not have online subscriptions for the articles and I'm working from home. Instead I'll mention the three relevant points which he raises in his chapter 'Issues in Archaeoastronomy and Rock Art' from 'Viewing the Sky Through Past and Present Cultures: Selected Papers from the Oxford VII International Conference on Archaeoastronomy' (eds. Todd W. Bostwick and Bryan Bates, Pueblo Grande Museum Anthropological Papers 15. City of Phoenix Parks and Recreation Department. 2006:85-102 esp.87, http://worldcat.org/oclc/... ). He argues a post-positivist approach is based on Inference to the Best Hypothesis ( http://worldcat.org/oclc/... ), and this comes from Convergent Methodologies ( http://worldcat.org/oclc/... ) which are independent from each other which means that methods work as Intertwined Cables ( http://scholar.google.com... ) rather than chains which are as strong as the weakest link (Whitley's citations, not mine). This comes out clearly in Connie Mulligan's response to Greg Laden when she says: ( http://www.plosone.org/an... ) "Indeed, the point of integrating genetic, archaeological, geological, and paleoecological data was to provide a robust model that did not depend on a single dataset." It's a complex approach, but it's clearly not a method for cherry-picking results to push a conclusion.

Beringian Archaeology?

The post-positivist perspective the results are exciting because while the archaeological evidence may be slender it's reinforced by the genetic data and the palaeontological data regarding Beringia. I think Afarensis is absolutely right in highlighting the importance of this middle phase of the colonisation ( http://www.plosone.org/an... ). From elsewhere in the Upper Palaeolithic we have evidence of long range movement across landscapes and seasonal utilisation of resources. For the period of time given we are clearly talking about a population that was thoroughly at home and successful in Beringia. This suggests that questioning the colonisation of the Americas perhaps has shades of a telelogical perspective and instead we should also consider what caused a successful group to abandon Beringia and phrase the question as "The Decolonisation of Beringia'. In light of this while I agree that there was likely a rapid (for a certain definition of rapid) expansion into North America, would it not also be reasonable to assume there was a parallel expansion into Asia as well? I assume this is where the second genetic bottleneck at 15,000 kyr occurs with a division of the Beringian population. I also assume such an expansion cannot be traced genetically on the Asian side due the the time passed and population size, but it could suggest that there may be archaeological factors to look for in Asia around this period if Beringians are moving into the area.

I know even less about East Asian archaeology, but the reason I ask this question is what percentage of people in Greater Beringia became Amerindian ancestors? A small percentage would suggest that inter-regional contacts were important in this period. Alternatively a large percentage might tell us something about population pressures, and I'm sure there are plenty of other plausible explanations to test.

I'm also interested in the presence of haplogroup X. Is this the kind of residual you have an all such tests, is it simply the way things are for Amerindian genetics, or it is possibly evidence of a different phase of colonisation. I assume it's not the latter, given the conclusions, but I'm not entirely clear how this fits into the model. It's very likely I'm asking a question of trivial importance, but I can't tell.

The importance of Open Access

I think both the scholarship and the presentation in PlosOne contribute to the importance of this paper. To some extent geneticists can ask why archaeologists aren't reading their papers. The answer is library budgets are limited and if an archaeological department has a choice between subscribing to a biological journal and an archaeological journal then it's likely that the archaeological journal is going to have more relevant articles in it and get the subscription. The result is there's almost certainly very good work going unnoticed and some quite good work which might become highly useful if it was tempered by some historical questions being left untouched.

If this paper hadn't been in an open access journal I wouldn't have read it. As is probably clear from above while I'm not hostile to genetics answering historical questions, I'm sceptical of quite a bit of the work I've read. I don't question the rigour of the work, but like Martin Rundkvist I tend to feel it's not hugely relevant to the questions I'm asking ( http://www.plosone.org/an... ). In contrast archaeological and historical research often is relevant, which is why I read those kind of papers. Articles like this show me I can't write off input from genetics. If this gets picked up by geneticists as well as archaeologists then my confidence in it will grow. With it being in an open access journal there's no economic reason why geneticists shouldn't incorporate insights from this paper into their own research. It also shows how to integrate very different methodologies when tackling one problem.