The genetics of American settlers vs. immigrants

Cite: Bryc, Katarzyna, et al. "The Genetic Ancestry of African Americans, Latinos, and European Americans across the United States." The American Journal of Human Genetics (2014).
Cite: Bryc, Katarzyna, et al. “The Genetic Ancestry of African Americans, Latinos, and European Americans across the United States.” The American Journal of Human Genetics (2014).

The recent paper, The Genetic Ancestry of African Americans, Latinos, and European Americans across the United States, brings together a lot of results which 23andMe has been letting slip in bits and pieces over the years. Most of the press coverage has focused on racial dynamics at the level we’re used to talking about today in the United States. White, black, and Latino (of whatever race). But as I told the first author at BAPG a few weeks ago the dynamics among white Americans is probably where their massive data set can shine. You see it in the figure above, which confirms what many have suspected: the states of the inland South have retained a predominant Anglo-American settler population down to the present. This is clear in their very high fraction of people of “British-Irish descent” in 23andMe Ancestry Composition nomenclature. Including the black American population the overwhelming majority of the population likely descends from people were already resident in the future continental United States in 1776 in this region. Additionally, you can tell that these results are not crazy because in the north Indiana has higher fractions than either Ohio or Illinois, which is exactly what you’d expect if you knew something about the demographic histories of these states. Indiana experienced less migration from European populations who were not of settler stock than Illinois (Chicago) and Ohio (Cleveland and Cincinnati). Similarly, Maine’s elevated fraction makes sense since rural Yankees are demographic more dominant in northern New England than they are in the southern states. Finally, the states of the old Yankee Empire of the northern Old Northwest have been totally demographically transformed by the massive waves of migration from Germany and Scandinavia.

The distinction between settler and immigrant Europeans is clear in relation to detectable non-European ancestry:

We find very low levels of African and Native American ancestry in Europeans with four grandparents born in Europe. We estimate that only 0.98% of Europeans carry African ancestry and 0.26% of Europeans carry Native American ancestry. These levels are substantially lower than the 3.5% and 2.7% of European Americans who carry African and Native American ancestry, respectively…Excluding countries that had major and minor ports in the Atlantic with strong connections to the slave trade (namely Portugal, Spain, France, and United Kingdom) and Malta, which has been the site of migrations from Africa and the Middle East, we obtain a data set of 9,701 Europeans, where we find African and Native American ancestry is virtually absent, with only 0.04% of individuals carrying 1% or more African ancestry and 0.01% carrying 1% or more Native American ancestry, within the margins of survey error estimates.

32081The African admixture in places like the American South is almost all a function of admixture in the first 150 years or so of settlement (the exception might be Louisiana, where Spanish Creoles may have contributed some African ancestry). The historical and genetic data seem to align there. Though the racial caste system in the American South had an early origin, it became progressively more calcified and stationary as the decades progressed. But by the late 19th century when Jim Crow laws were enacted the African admixture in many Southerners was a very distant memory and thoroughly diluted. To gauge the social and demographic import, recall that detection of genetic fragments does not reflect the total genealogy. Many people with ancestors who were black American slaves do not carry any segments from those individuals, while the ones being detected in this study exhibit a relative enrichment.

In the future I would be very curious about exploring the patterns of relationship of the Anglo-American folkways, as outlined in works such as Albion’s Seed and the The Cousins’ Wars. A major problem though is that these are genetically very close to begin with. The first author of the above work suggested to me that they would need the People of the British Isles data set to get good reference populations. Perhaps in the near future that will be feasible.

Moving past folk biology

Noahs_ArkMost biological concepts exhibit striking clarity and intuitive accessibility at the highest levels, but engender confusion when you drill down to the details. You can see this in an understanding of evolution. Most people can grasp the idea of common descent with modification relatively easily. But when it comes to getting an good intuitive grasp of evolutionary process, and what that might entail, people are often left grasping at straws. Consider the reality that many people believe that evolution works by benefiting the species, even though the mainstream position within the discipline is that operates through variation in individual fitness. What is likely happening is that our cognitive intuitions are slipping into our understanding of science. We imagine evolutionary process as changing an entire species, rather than the mass action of individuals within the species.
519gldjJoALThis is clear when you look at the research on “folk taxonomy.” Humans have an idea of what a species is, and it often corresponds relatively closely to the biological species concept. Though higher and lower taxonomic scales such as genus and subspecies reflect genuine information about the structure of reality, species is considered special by many in that it is a clear and distinct level of organization where groups are of organisms are neatly encapsulated from other groups of organisms. It’s “real.” Here intuition and folk taxonomy align with our understanding of biology. The problem is that species is neither so general, nor so neat and airtight even in cases where one might think it applies. First, the biological species concept obviously makes sense only in the context of sexual organisms. Asexual organisms are living things too., and important ones at that. There are after all an order of magnitude more bacterial cells in our body than somatic cells. Second, even many sexual organisms (e.g., plants) engage in hybridization quite regularly. Species are useful semantic sugar, but they have no atomic Platonic reality. They are not one of the fundamental units of organization of the universe around us, but rather a term which maps upon a dynamic which is of great interest to humans (i.e., the fission of complex eukaryotic sexual organisms into distinct populations over time).

51D2RoDDkXL._SY344_BO1,204,203,200_The reality that for most biologists species are viewed in instrumental terms is not always clear to the broader public. The constant war of words with Creationists which revolves around speciation and macroevolution is actually somewhat beside the point, because the very idea of species is not coherent or consistent, despite its genuine utility. And isn’t just Creationists who have drafted our cognitive Platonism toward their ends. Disputes over the Endangered Species Act get bogged down in minutiae as to what a species is, and what sort of diversity is worth preserving (or not). That’s what happens when you take a scientific concept which is not airtight on the margins and introduce activists, lawyers, and various assorted interests with no compunction about using casuistical arguments in the service of their preferred ends.

The same general confusion seems to crop up when you move below the level of species. In my post below, Fear of Race Mixing in Biodiversity, I make a pass at the fallacy of the blending theory of inheritance, which was overturned by a little known field which goes by the name of Mendelian genetics. Nevertheless in the comments the blending theory of inheritance pops right back up, even though I dismissed it in the post itself:

What about the modern Auroch? It has been reconstructed through selective breeding though no geneticist would consider it the same as an ancient Auroch. It may look similiar but it is no more Auroch from a genetic standpoint than a spanish fighting bull. Once a breeding population is lost in its isolation and genetic uniqueness. It can not be reassembled any more than two cans of paint can be unmixed. Even if the original genes still exist in the mixed population. Your viewpoint also assumes that all phenotypes have the same reproductive and behavioral tendencies. Which does not seem to be the case for nothern euros in comparison with more tropicaly evolved peoples. So unless blondeness or ruffocity convey some selective advantage over other phenotypes they will dissapear in time, especially the platinum or ash blonde varieties.

download (1)First, let me say that I really don’t understand about half of this comment, and told my interlocutor exactly that. But the portion relating to the analogy of mixing with paint is obviously leveraging the intuition about the blending theory of inheritance, and it turns out to just be false (also, a rule of thumb might be to not engage in analogies with a geneticist about genetics; just get to the point in plain language). In the early 20th century Mendelian genetics, which traces patterns of inheritance of traits across generations and sieves it through a particular model, turned out be exactly what was needed to allow for the persistence of the variation which is the raw material of evolution. The necessity and maintenance of this variation was a paradox which confronted Charles Darwin in the 19th century, and he never quite resolved it. Basically, if offspring are the blended mix of their parents, then each generation should be progressively more blended and uniform. That uniformity removes the variation which is necessary for adaptation through natural selection. Attempts to maintain variation through processes such as high mutational rates were simply not plausible. The genius of Mendelian genetics was that inheritance was transformed into a process mediated by discrete particulate units. Genes. In sexual organisms the patterns of inheritance are governed by the law of segregation and the law of independent assortment.

download (2)When I teach undergraduates basic genetics and these laws I always tell them to conceptualize them physically, because we know now that genes are actually embodied in strings of base pairs. Mendelian genetics, and the abstract understanding of the genome in its essence, long predate the discovery of DNA. But through a grasp of both the abstraction of genetic inheritance, as well as its concrete manifestation in the sequence of DNA, one can infer the dynamic of microevolutionary process. Because of segregation there is variation in the ancestry inherited from the grandparents. Because of recombination the law of independent assortment holds beyond a certain genetic distance even on the same chromosome. Evolution can be thought of now not as simply phenotypic change, but rather as fluctuations in allele frequencies over generations, unto extinction and fixation. Finally, the nature of quantitative traits due to polygenetic architectures becomes much more transparent if one simply imagines the combined effect of numerous genes producing a final outcome.

The above explains why racial admixture of modern populations will not lead to uniformity and homogeneity. Consider the case of skin color, where variation in ~10 genes accounts for ~90 percent of the inter-continental variation in complexion. Populations where many of these genes are segregating, such as in Brazil or India, are not of uniform coloration, but manifest the full range of ancestral complexions. On average the complexion lies at the midpoint, but the underlying allelic variation remains. In fact, because admixed populations exhibit combinations of multi-locus genotypes not found in the ancestral populations they’re likely to be more diverse overall than the total variation in the summed ancestral groups (e.g., which group has more phenotypic diversity, Spaniards and Amerindians separately, or mestizos?). From the standpoint of anti-racists who may hope for a post-racial world where amalgamation leads to the abolition of race, that will not happen. First, the data from Latin America is clear that phenotypic race remains even after genetic admixture because individuals vary a great deal in appearance. Second, most modern races themselves are almost certainly the product of admixture events over the past ~10,000 years. Racial categorization can be useful, and reflects real history, but it is not a fundamental unit of special genetic structure. Races are neither primal nor Platonic. But given rather conventional conditions they seem to emerge out of folk taxonomies. They’re an evoked part of human culture.

product_thumbnailFollowing the comments of my interlocutor I don’t really believe the issue at heart was scientific. It’s patently false that once genotypes are scrambled they can’t be unscrambled. The law of segregation and the law of independent assortment offer up exactly the manner in which you can reassemble ancestral types from admixed populations. The root concern derives not from biology at all, but psychology, and an intersection between folk taxonomy and ideas of Platonic essences and contagion. Once the category of the Northern European phenotype is sullied in some way, it is lost forever. Rather than conceiving of Northern Europeans as a biophysical expression of discrete alleles, which are a very recent event in history due to admixture from diverse lineages and in situ evolution, the implicit cognitive assumption is that they have a particular unitary racial spirit, and that their racial category is fundamental and primordial. These might seem to be antiquated early 20th century concerns, but though the sentiment is sublimated I’m pretty sure it’s still very common, because that’s the only reason I can think that phenomena such as the “disappearing blonde gene hoax” often go viral so quickly.

When conceptualizing genetic and evolutionary processes one can frame it as a spectrum, from gradual change on polygenic characters to the emergence of single gene traits. The human default frame seems to veer erratically between the two. On the one hand blending inheritance theory implicitly underpins many intuitions about the nature of how characters change over time. Yet there is also a craving for a specific and singular concrete gene for a given condition. The latter likely emerges from the reality that DNA and modern methods, starting with linkage analysis, zero in on very specific positions in the genome and are localized to a particular gene for many Mendelian diseases. What gets lost is the Mendelian framework, which is a conceptual model which can integrate both the quantitative genetic characters and the monocausal traits and diseases conditioned upon a very specific change in one region of DNA.

They’re not a gym

This week Planet Money had an episode about the economic angle of how gyms make money off you not using your services (transcript). The basics are easy enough to understand. If you buy a memership to a gym, and actually showed up every day, then you’d wear out the equipment. If you are a member and pay dues and don’t utilize the services then the costs are kept down. In fact it is obvious that gyms have many more members than they have capacity to service at any given time. The radio piece quotes one manager as saying that her Planet Fitness has 6,000 members, but a capacity of only 300 (and often it is rather empty). In fact the whole idea of gyms like Planet Fitness is to mix a portfolio of projecting the ideal of fitness while enabling your more slothful tendencies. People join for the cardio machines, but they stay for the pizza nights and socials.

In contrast another gym they profiled, Precision Athlete, serves an opposite clientele. While Planet Fitness charges $10 a month, Precision Athlete charges $500. Some of this has to be the fixed costs of a small operation with high level trainers, but it is also clear that their equipment gets used, a lot. This puts in stark relief the peculiarities of the normal gym, where there’s a power law distribution of utilization. About 10% of the members probably use 90% of the “gym time.”

Addendum: In some ways Planet Fitness perfectly encapsulates are our contemporary ethos among a certain set of middle class Americans. Since actually working out is “triggering,” they ban or discourage it.

Open Thread, 12/21/2014

165280Currently reading Mathematical Models of Social Evolution: A Guide for the Perplexed. Though there is a focus on the social aspect, this is really a book on evolutionary modeling more generally. And, it’s less daunting than many other works that I’ve run into. The authors also make a nice case for why analytical modeling is still important in an age of simulations.

Been very busy with stuff, trying to get as much work done as possible before Christmas. You know how it is.

Fear of race mixing in biodiversity

lean
Abolishing diversity, one child at a time!

One of the often overlooked historical oddities in the development of the environmental movement in the United States is its past close relationship to what we would today term white supremacy. Though many praise Teddy Roosevelt for his embrace of conservationism and evolutionary theory, he also adhered to the normative racial beliefs of the day, which presumed the superiority of Anglo-Saxon people, and couched that superiority in Darwinian terms. Even less well known is the activism of race theorist Madison Grant, who was as much a conservationist as the intellectual doyen of white supremacy that he is remembered as today (see Defending the Master Race: Conservation, Eugenics, and the Legacy of Madison Grant). In some ways the connection is reasonable and not surprising, in that both are fundamentally conservative preservationist instincts. To preserve the environment and the racial order of the day. The association was clear well into the 20th century, Charles Lindbergh was a prominent eugenicist, but later became an environmentalist, while Garrett Hardin, who originated the term “tragedy of the commons,” opposed high immigration levels and was skeptical of racial diversity.

cover_passingBecause of environmentalism’s place within the cultural Left in the United States these corollaries no longer apply. In fact, the Sierra Club and other such organizations tend to be careful to not oppose immigration on environmental grounds any longer because of its racial implications. But, I’ve noticed that many people with an environmental orientation still use what strikes me as quite racialist language in the context of animals. I don’t think it is a problem.  Different moral and ethical standards apply to animals. We eat them. We don’t eat humans. But I also think it is funny, as well as somewhat wrong-headed. This came to my attention again because of an article in Nautilus, A Strange New Gene Pool of Animals Is Brewing in the Arctic. There’s a lot of talk about issues like hybrid zones, and pre- and postzygotic isolation (at least implicitly). But this section is just totally confused:

In September, in an inlet some 1,800 miles north of Fargo, North Dakota, where the North American landmass dissolves into the Arctic Ocean, the whales met in the middle. They spent two weeks together, and although not much happened before they turned around, the meeting was historic. The fossil record indicates the last time Pacific and Atlantic bowhead whales came into contact was at least 10,000 years ago.

While it’s tempting to imagine a strange new Arctic teeming with “grolar bears” and “narlugas,” hybridization comes at a cost. Arctic biodiversity will be reduced through gradual consolidation, taking with it a blend of genes that have evolved by natural selection over millennia. “There’s going to be a whole bunch of organisms containing genes that we’re going to lose,” Kelly says. Which genes, exactly, is unclear….

The problem here is that the terms are being mixed up. “Biodiversity” is often applied at the level of species or races, with a diversity index calculated from discrete numbers of population types. If you calculate a diversity index based on Swedish, Nigerians, and Chinese, you start out with three populations and look at their proportions (the more skewed the proportions, the lower the diversity). If you take them all and mix them so they are one random mating population obviously the ecological diversity index is going to go down. But the genetic diversity is not going to down, because genes don’t “mix”. Mixing implies a blending theory of inheritance, what Mendelian genetics overthrew with its understanding of discrete and particulate units of inheritance. The same confusion crops up with the ideas of “disappearing blondes” and “disappearing redheads.” The phenotypes may change in frequency, but the understanding alleles, the genetic variants, remain. From a genetic perspective if you wanted to you could probably pull back out the original populations through selective breeding. Not only does the allelic diversity of the pooled populations not change, but the genotypic diversity increases, because of elevated heterozygosity. Finally, new potential combination genotypes arise from the mixing, so the phenotypic diversity in totality also probably increases (e.g., Brazilians exhibit a wider range of skin color variation than Africans or Europeans).

Of course this is predicated on racial/subspecies level variation and divergence. If the populations are separated long enough then there will be barriers to easy gene flow. This is evident in the modern human-Neandertal event, where the X chromosome seems to have been purified of Neandertal alleles (this is a common tendency with hybridization events). But please note above that the people in the piece are concerned about populations of whales separated for 10,000 years. There are plenty of human populations separated for 10,000, and even 100,000 years. So this isn’t really a terrifying number of generations.

Most horse lineages are in heaven

Citation: Prehistoric genomes reveal the genetic foundation and cost of horse domestication
Citation: Prehistoric genomes reveal the genetic foundation and cost of horse domestication

wuwei_horseThe present is often only a faint echo of the past. That is why ancient DNA has totally revolutionized our understanding of the evolutionary past of many branches of the trees of life. The tips of the tree that we see around us today are all that remains of diverse and chaotic brambles which have been thoroughly pruned by chance and necessity. Utilizing present genetic variation researchers have been able to make some very interesting inferences, but you can’t infer that which you lack all evidence for. When it comes to the “megafauna” (i.e., anything bigger than a rat) the past two million years have been very trying, with cold alternating with short warm spells. This has no doubt resulted in thinning every so often as populations go extinct across vast swaths of Eurasia in the face of advancing glaciers. In time the range expands, as populations are re-founded by surviving lineages. But these oscillations drive down long term genetic diversity. In addition, the rise of modern humans has resulted in a great wipe out of whole lineages due to our predatory and avaricious behavior.

Domestic animals are arguably the most extreme case of this dynamic. A recent paper on the genomics of the domestic dog highlighted just how wrongheaded previous assumptions were. The standard thinking was that modern dogs are a derived form of wolf. In other words, dogs are simply a specialized subset of wolves. A domestic wolf as it were. Using whole genome sequencing it turns out that wolves that we see around us today are a sister lineage to dogs. Using wolves as the ancestral form of the dog may not therefore be quite as obvious as we’d thought. Of course it seems likely that the ancestors of the dogs were a wolf lineage of some sort, but we can’t assume that they resemble modern Holarctic wolves.

A similar, more complicated, dynamic is being illustrated by ancient DNA for our own lineage. And now a new paper in PNAS highlights similar dynamics to dogs in horses, Prehistoric genomes reveal the genetic foundation and cost of horse domestication. The authors sequenced two horses from the Taymyr regio
n
in Siberia with medium and high coverage. The horses were from 15 and 40 thousand years ago. That means they well predate domestication, which probably occurred in the 5 to 10 thousand year interval.

These results confirm that the wild Przewalski horses are not ancestral to the domestic lineages, and that rather they are simply the single wild lineage which persisted down to the modern period. The horses from Taymyr are more distantly related to modern horses than the Przewalski are, but intriguingly tests of admixture indicate that there was gene flow from lineages more closely related to the Taymyr individuals to the modern lineages. This gene flow has to be very close to the root of the origination of modern domestic horses since all breeds are equally represented in the signal of admixture. In short, the modern lineages of horse, wild and domestic, are but a fraction of the variation of the ancient populations.

They confirm this in a population genomic sense by looking at the enrichment for deleterious mutations. A major confound is that many lineages of horse are inbred, so they corrected for that. It turns out that all modern horses exhibit signs of being subject to a load due to accumulation of deleterious variants because of small population size (specifically, bottlenecks), as selection is less efficient at removing these mutations in small population because it is overwhelmed by random drift.

Finally, there is a lot in the paper on signals of selection around various causal alleles. It’s the typical laundry list. Many of the genes are associated with various pathologies and abnormalities, which shows you the cost of reshaping organisms and their behaviors which can occur due to domestication. Strong selection on major effect alleles often result in a cost due to antagonistic pleiotropy. If the selection benefit is high, then negative consequences be damned! Now, horses are notoriously dumb compared to donkeys, so I’d be curious if the cognitive/behavioral signals are found in humans and other mammals, and how they may have changed over time.

Songs of Ancient Mesopotamia

I guess this is taking “world music” to the next level, going back to the ancient Mesopotamians. The artist is Stef Conner, and you can read about how this reconstruction was done over at Newsweek, where there is a Soundcloud preview of her full album, The Flood. I’d actually purchase it if I could find a full digital copy, but I don’t see any out there right now (the article says it will be on iTunes next month). You can buy a physical copy at her website, but the last time I purchased a C.D. was probably in the early 2000s, so that’s not happening. Anyway, do listen the preview on Soundcloud. The drinking song above is probably not representative.

Open Thread, 12/14/2014

513367In my books list below I tried to not focus too much on evolutionary genetics and genomics. But I still feel that I was a bit narrow. Over the years my interest in science has become rather narrow because of professional focus, but when I was younger I used to read quite a bit of popular physics, such as John Gribbin’s In Search of Schrodinger’s Cat. This sort of narrowing of focus is probably inevitable, but still pretty worrying for me. Probably the last popular physics book which I read front to back was Lee Smolin’s The Trouble With Physics, nearly 8 years ago.

In other news, I’m going to try and take a break from the internet for a few days around the New Year. More precisely, I’m going to try and take a vacation with my growing family somewhere sunny and hopefully warm. We’ll see how that works out. But please don’t be concerned if I don’t post on Twitter or have a blog update for a few days +/- January 1st of 2015.

The future of publishing (does it have a future?)

Founder of the Royal Society
Founder of the Royal Society
A few years ago Joe Pickrell wrote a very influential post, The first steps towards a modern system of scientific publication. Influential because it seems have to been a reason for the development of SciReader.The developers behind PubChase also took some lessons from it. Of course we know the role that “open access” and PLOS have played in shifting the ecosystem of the consumption and production of scientific knowledge (if you don’t know what I’m talking about, an Eisen will enlighten). And Haldane’s Sieve has been instrumental in bringing preprints to biology (in particular, evolutionary biology and genomics). Finally, events like the Bay Area Population Genomics meetings fill the gap between inter-personal relationships which are critical to scientific production, and massive conferences which have an enormous overhead in implementation for the organizers and non-trivial cost of attendance for the conference-goers.

a190So is there more to say? I think so. That’s why something I wrote with Laurie Goodman and David Mittleman just dropped in Genome Biology, Dragging scientific publishing into the 21st century. It’s open access, you can read it all, and I encourage you to do so. The question that framed my thought was a simple one: how can scientific publishing become more than simply a PDF delivery platform? Using the internet to deliver PDFs is like using a gasoline engine to draw a conventional carriage designed with horses in mind. And it’s feasible because scientific publishing is a profitable field dominated by a comfortable oligopoly which captures rents from the institutional structure of modern science. Remember, high impact journals are not high impact because they provide a better experience for scientists, who are the producers and consumers of the product. They’re high impact because they are high impact, and as long as they are high impact people will need to publish in them to gain scientific credibility and prestige. Many researchers would label this a vicious circle. There’s a reason that they call Science, Nature, and Cell “glamor mags.” It’s about being seen. Ultimately, a matter of fashion, not substance.

What can not continue, will not continue. There isn’t a need to talk about creative destruction today as if it’s a novel concept, we’ve seen “smartphones” swallow the functionality of whole industries (e.g., watches and cameras), and it seems inevitable that ride-sharing will radically transform the nature of the taxi industry in the United States. I hold that the dominance and profitability of scientific publishing firms today is in large part a function of norms within modern science which enable and perpetuate a coordination problem. Once the norm starts shifting, it will change very fast, because many of the people who are publishing in the glamour magazines only do so begrudgingly because they feel they have to.

So is there a future for organizations such as the Nature Publishing Group? I think there is. The key is to take more to heart the idea that scientists are their customers. I don’t think the sector will be as awash in money in the future, so it needs to be leaner and more efficient. Publishers need to really start innovating so that scientists don’t just focus on something like “impact factor,” but also questions such as “is this journal going to package and present my results in a way that communicate well with my colleagues?” In other words, one needs to focus on the substance of what scientific publishing is supposed to be about, beyond obtaining a tenure track position, the furtherance of mutual understanding! Second, the journals can also invest in sharpening their style so that they always maintain some value-add over spare operations such as preprint servers. The ecosystem of scientific communication will remain vast, but it will evolve. Scientific publishers need to reposition themselves into a smaller but more specialized niche soon, because the market is likely to shift underneath their feet before they know it.

Razib’s 20 Holiday Books, 2014

41PHSZN6AELMy friend Randall Parker sent me an email where he suggested I should put up a post relating to books for the holidays. This makes sense, since I’m a book nerd. Over the years I’ve started to realize time is precious, and have offloaded a lot of the hard work of figuring out things to others who have domain specific specialties (e.g., I have friends who are into nutrition or exercise, and rely on them to give me appropriate pointers and direction). As implied by the title I’ll probably try and do this every year now. Also, I’ll avoid textbooks in the following list, and will attempt to be more diverse in my disciplinary focus than usual….

The Genetical Theory of Natural Selection. The variorum edition from 1958 with J. H. Bennett is what you want. R. A. Fisher is dense, but this isn’t a textbook. If you understand 10%, that’s a lot of understanding.

The Isles. Norman Davies’ magisterial narrative history of the British Isles.

In Gods We Trust. You won’t look at religion the same way after.

kwonhardcoverKnowledge and the Wealth of Nations. All about endogenous growth theory and its origins. More interesting than it sounds. As important as it sounds.

From Plato to NATO. This book has had many lesser copycats.

The Truth About Everything. A history of western philosophy. It has illustrations.

Prehistory of the Mind. Underrated hybrid of evolutionary psychology and paleoanthropology.

The Fall of Rome. A materialist take on the decline and fall of the Roman Empire. I’ve read this book three times.

The Number Sense. Like The Language Instinct for numeracy.

What_Hath_God_Wrought_-_The_Transformation_of_AmericaWhat Hath God Wrought. An anti-Jacksonian history of early America.

When Genius Failed. The template for “too big to fail.”

The Imitation Factor. Great short read on behavioral ecology.

Mutants. Armand Leroi can write beautifully even about the grotesque.

Descartes’ Baby. The child is the father of the man.

downloadCalculus Made Easy. This is an old and chatty book. It’s not a text.

A Beautiful Math. Game theory and John Nash’s science.

Grand New Party. Not the Tea Party.

Genome. Compulsively readable.

The Human Web. One of William H. McNeill’s later books.

From Dawn to Decadence. Jacques Barzun’s magnum opus.

Readers are invited to offer their list of 20 in the comments. Randall has been challenged to put one up at his blog. Though I hope people will try and make the books at least somewhat accessible and relevant to a general intelligent audience (e.g., no books on Design Patterns or The Art of Computer Programming).