Header image

john hawks weblog

paleoanthropology, genetics and evolution

Photo Credit: Contemporary human skull compared to the Kabwe cranium. John Hawks CC-BY-NC 2.0

Denisovan ancestors of the Iceland population

Last week in Nature, Laurits Skov and collaborators from Aarhus University and from Kari Stefansson’s research group in Iceland gave a high-resolution look at Neanderthal and Denisovan introgression in the Iceland population. The title of their paper is: “The nature of Neanderthal introgression revealed by 27,566 Icelandic genomes”.

I see this paper as a first step in a new phase of research. Up to now, samples of human genomes used in phylogenomic analysis have been limited in number. The 1000 Genomes Project samples include a hundred or so individuals from just a few populations. Most other studies have had much smaller sample sizes. Ancient genomes by their nature are limited in number.

Small samples do enable geneticists to answer some questions with high confidence. These are the questions where ancient people were very different from each other, and from people living today. “Did they mix at all?”, or “Assuming total isolation between them, when did their populations start to diverge?” But when you start to specify just how strict the assumptions must be to do the math, you start to see how unsatisfying the answers to such simplistic questions really are.

Truly large samples allow us to answer questions that involve small levels of difference between ancient and modern people. Ancient humans were like us in most ways. Nearly all their phenotypes overlapped with those of living populations. They were emphatically not subject to “total isolation”, they mixed repeatedly. How often did they mix? How important was that mixture to their evolution? Those questions take big samples to start to answer.

I’m going to spend some time evaluating this research and pulling out some of the new directions in separate posts. Skov and coworkers are able to identify the fine-scale similarities between haplotypes in the genomes of living people and in three high-coverage ancient genomes. In this case that includes the so-called Vindija and Altai Neanderthal genomes and the Denisova genome.

In this post, I want to focus on the Denisovan component of ancestry.

The advantages of a large-sample approach are abundantly evident when considering the very small amount of genomic DNA that Iceland people have from Denisovans. Skov and coworkers quantified the very small amount (0.1%) of DNA shared within the Iceland population from Denisovan ancestry and they discuss several scenarios for how it may have gotten there.

Finding Denisovan ancestry in western Eurasian samples is not a first-ever result, and similar small fractions of this ancestry have been found in other recent studies. Earlier this year, Anders Bergström and coworkers reviewed the variation found from whole-genome sequencing of many Human Genome Diversity Project samples, and quantified Denisovan-like haplotypes in many populations, including New Guinea populations. Their figure from the supplementary information shows the amount of each population’s genome that they infer to be Denisovan:

Denisovan ancestry in lots of populations, from Bergström et al. 2020
Figure S17B from Bergström et al. 2020, showing the amount of genomic sequence inferred to represent Denisovan ancestry in lots of populations.

Bergström and coworkers estimated that 2.8% of the ancestry of the highland New Guinea sample in their study came from Denisovans. As visible in the figure, the amount of genome identified with high confidence as Denisovan is less than the overall estimate, and amounts to around 25 million base pairs per genome in the highland New Guinea sample.

[As an aside, it is not obvious how this result reported in the Bergström et al. supplement connects to the estimated fraction of ancestry. Twenty-five million base pairs is less than one percent of a genome (and less than half a percent of a diploid genome). For my discussion here, I’m just looking at the fraction of this amount as estimated for other populations.]

If we look at western Eurasian populations here, like French or Orcadian, around 0.1 on the x-axis scale, or 1 million base pairs, are estimated to be Denisovan. Assuming the same 3:1 ratio that seems to apply to the New Guinea sample in that paper, this would suggest around 0.1 percent of the genomes of these samples are Denisovan in origin.

[As another aside, did anybody else notice that the Bergström et al. supplementary figure has more Denisovan ancestry inferred in the San sample than in French? Something with incomplete lineage sorting is surely going on that hasn’t been well quantified here. One problem is that Bergström et al. seem to have assumed near-zero introgression in African populations, and Joshua Akey’s group showed that this assumption is incorrect. ]

Bergström and coworkers (2020) is the most recent analysis of Denisovan introgression into Eurasian populations, but some earlier work from Joshua Akey’s group provide somewhat more resolution of these issues. In 2018, Sharon Browning and coworkers examined Denisovan ancestry in Eurasian populations and inferred two Denisovan source populations for introgressed haplotypes in most Asian populations. I’ll return to this paper below. Earlier this year, Lu Chen and coworkers looked at Neanderthal ancestry in sub-Saharan African samples. In the course of this analysis, they also quantified Denisovan-like genome segments in 1000 Genomes Project samples. They found a similar amount of Denisovan-like haplotypes in African and Eurasian population samples, around 1 megabase on average. Chen and coworkers interpreted this result as a baseline for incomplete lineage sorting for their purposes, which was to identify Neanderthal ancestry in sub-Saharan Africa. But it is interesting in the context of identifying true Denisovan introgression in all these populations.

So the Iceland result found by Skov and coworkers, using a vastly bigger sample of living people, seems to be pretty close to what other populations in western Eurasia have. If anybody was wondering if Denisovan ancestry in Iceland is a sign of a transoceanic migration from Melanesia to Iceland, that hypothesis isn’t necessary.

OK, let’s look more closely at what Skov and collaborators find from their enormous Iceland sample. This is a figure showing the distribution of introgressed haplotypes across all the chromosomes:

Figure 1d from Skov et al. 2020, showing parts of the genome on each chromosome that reflect Altai, Vindija, or Denisovan inferred similarity
Figure 1d shows the genomic distribution of archaic fragments and the DAV [Denisovan, Altai, or Vindija] genome that they share the most variants with: Vindija Neanderthal (50.8%), Altai Neanderthal (13.1%), Denisovan (3.3%), two or more DAV genomes (20.4%) or not shared with a DAV genome (unknown, 12.2%).

This figure is one of those nutty diagrams that shows a logogram of every chromosome split into 4 panels that each are marked up with colored blocks representing the location of introgressed haplotypes. Together these add up to around 40 percent of the genome with some evidence of introgression within the modern sampled individuals.

According to the results, the Iceland population has something like 0.1 percent Denisovan-like ancestry across their genome. Skov and coworkers went to some effort to try to understand where this Denisovan component came from.

Denisovan ancestry may come from direct mixture of early modern people with a Denisovan group; or alternatively, Denisovan ancestry in Icelanders may have come indirectly from Denisovan mixture with Neanderthals that happened before early modern humans mixed with Neanderthals. Skov and coworkers provide a figure that shows the difference between the direct and indirect models:

Figure 2 from Skov et al. 2020 showing models for Altai Neanderthal, Vindija Neanderthal, and Denisovan introgression

Notice that in the direct model, non-African peoples get around 0.12-0.16% of their ancestry directly from a divergent Denisovan group. In the indirect model, the “introgressing Neanderthal” group gets 6-8% of its ancestry from Denisovans, and then gives 2% to the ancestors of today’s non-African peoples.

Because the Denisovan component of ancestry is so small, it is difficult to test the difference between these scenarios. The two scenarios (direct versus indirect) do predict slightly different things about the Denisova-like haplotypes. If these all came indirectly into modern humans from a Neanderthal population after Denisovan-Neanderthal introgression, then recombination should have connected many of them directly to haplotypes that otherwise resemble Neanderthals, creating a recognizable pattern. This puts a constraint on how early the Denisovan-Neanderthal mixture could have been. But if such Denisovan-Neanderthal introgression happened within the period just before modern-Neanderthal introgression, it would be very hard to tell the scenarios apart. And the sizes of these populations matter – the survival of distinct Denisovan lineages after introgression within a Neanderthal population is more likely in a larger population than a smaller one. All this means that there are a lot of ways that a different sequence of events might lead to a similar outcome.

One thing that the study concludes for certain is that this Denisovan component of Iceland genomes does not come from incomplete lineage sorting alone. These are not ancient African genetic haplotypes that merely resemble Denisovans. They really did come from Denisovans at some stage of prehistory.

A second thing that is clear is that the Denisovan-like haplotypes in Icelanders do not come directly from the Denisovan population sampled at Denisova cave in the Denisova 3 high-coverage genome. Skov and coworkers were able to examine the extent of allele sharing between the Denisova-like haplotypes in Icelanders and the Denisova 3 genome itself:

We find that the amount of derived variant sharing is compatible with a scenario where the introgressing Denisova splits from the sequenced Denisova around 300-350 kya.

That’s a fascinating conclusion. The extent of diversification within Denisovans that this would represent is as great as the greatest divergences among surviving modern human populations that survive today. Their simulations suggest greater uncertainty than reflected in the sentence I’ve quoted here. This range of uncertainty appears to extend from 270,000 years up to 400,000 years ago.

Again, this is an observation that confirms something already known from other work. The paper from Guy Jacobs and coworkers last year on Indonesian population diversity provided good evidence of deep Denisovan diversity. Many people in New Guinea today have a heritage including introgression from two different groups of Denisovans. Both of these Denisovan-like groups continued to exist until the last 40,000 years, it seems from the lengths of the introgressed haplotypes. Jacobs and coworkers denoted these Denisovan-like groups as D2 and D1. They estimated that the D2 group diverged from the population ancestral to both Denisova 3 and D1 around 360,000 years ago, and D1 diverged from the Denisova 3 population around 280,000 years ago. These findings pointed to very deep population diversity within “Denisovans”, more in fact that is present among the most diverse modern human groups.

These estimates of divergence dates are based on simplistic models of divergence with no subsequent gene flow. They are likely to be wrong. If these “Denisovan” populations behaved like modern human and Neanderthal groups, they probably shared some amount of gene flow at times after their divergence. That would make the “divergence date” estimates appear more recent than the real initial diversification of these populations.

It is not clear whether other Asian or island Southeast Asian populations may also have mixture from the “D2” population. The 2018 work from Browning and coworkers established that today’s Asian populations have Denisovan ancestry from at least two “waves” of introgression. One of those, accounting for around a third of the Denisovan-like ancestry in East Asian people today, looks to have been genetically similar to the Denisova 3 genome. The other wave was genetically quite divergent from this Denisova-3-like population. In the analysis by Browning and coworkers, they were able to find evidence for both waves in samples of living people originating from China, Japan, and Vietnam, including several ethnic groups in China. They even found a trace of the Denisova-3-like wave in Finland. But they found only the divergent wave in South Asian populations, and they did not identify either Denisovan wave in European populations other than Finland, or in samples with Native American ancestry. Browning and collaborators did not estimate a divergence date for this population, and Jacobs and coworkers did not establish whether this mainland Asian divergent wave was the same as either the D1 or D2 introgression source that they identified. So the identity of this divergent Denisova-like introgression wave at the moment is up in the air.

It is reasonable to ask whether the divergent Denisova-like wave that Browning and colleagues found in Asians might be the same as the Denisova-like ancestry that Skov and colleagues identify in Icelanders. It’s also reasonable to ask whether both of these represent the same source as the D2 population identified by Jacobs and coworkers.

I’m not convinced that they are the same, for reasons I’ll go into below. But it’s not a bad idea to start with the hypothesis that they might be the same, and think about how to test that. So I’ll propose that these signals all equate to the same divergent population, and that they originate from a Denisova-like population in South Asia.

In this hypothesis, most of the D2 component of ancestry in Sahulian populations actually reflects introgression from this divergent Denisovan group. The ancestors of today’s Sahulians would have encounted the D2 population as they transited South Asia. Later, they picked up ancestry from the different D1 population, which might come from a Wallacean or island Southeast Asian source.

Meanwhile, southwest Asian Neanderthals also had a small component of D2 ancestry. They got this from introgression or gene flow with the South Asian D2 population. That gene flow was not small, it had to be something like 6 to 8 percent of the genome of the southwest Asian Neanderthal population. The out-of-Africa wave of modern humans then picked up a small fraction of this D2 ancestry when they mixed with the southwest Asian Neanderthals. That component provides the measurable Denisovan ancestry in most of today’s people, including Icelanders. It also is present in New Guinea, but accounts for a much smaller amount of the D2 signature in this population.

This hypothesis leaves the timing of these events somewhat flexible. Today’s Sahulian groups are more divergent from other non-African populations than any of those populations are from each other, and they may have picked up their extra D2 ancestry anywhere along the road to Sahul. Jacobs and coworkers (2019) inferred the time of D2 mixture into New Guinea populations at 46,000 years ago, with an error interval going up to the time they estimate that these populations first differentiated from the stem out-of-Africa groups, around 51,000 years ago. These dates are contingent on a variety of assumptions in the article, and we cannot take them as “real” dates that we can test hypotheses with. They just give an indication of where this introgression is relative to the divergence of the New Guinea ancestral population from other populations today. Likewise, the models examined by Skov and coworkers looked at introgression among Neanderthal, Denisovan, and modern populations at or around 45,000 years ago. Their analyses provided some demonstration of how Denisovan introgression may have originated, but they could not differentiate the direct and indirect hypotheses from each other, much less provide real confidence intervals for the date of Denisovan introgression.

How could we test this hypothesis? What is needed is some further large samples of other populations to test whether the small component of Denisovan ancestry can really come from a single source. 27,000 people might be overkill for testing that, but more than a few thousand will be needed.

The observation by Browning and coworkers that South Asian peoples today have more detectable ancestry from the “divergent wave” of Denisovans than European or other groups is very interesting. Bergström and coworkers this year also found more evidence of Denisovan-like ancestry within their South Asian samples than in European samples. Meanwhile, those same samples showed more Neanderthal ancestry in European groups than in South Asian groups. This set of observations was not designed to examine the question of Denisovan ancestry via introgression from Neanderthals, so I don’t want to take the results too far out of context. But on the surface this doesn’t look like a simple Denisovan-via-Neanderthal-introgression hypothesis can work for all the “divergent wave” component. The ancestors of South Asian people seem likely to have mixed directly with divergent Denisovans in addition to any divergent Denisovan ancestry they may trace to Neanderthal introgression.

OK, this is all quite complicated, because none of these studies were really designed to answer this question, and most of them are underpowered to examine the difference between direct and indirect mixture scenarios. But someone will have to get into these details with larger sample soon, because the alternative is that there were possibly many mixtures with many different divergent Denisovan groups.

For me, disproving the hypothesis that a single D2 population can account for all this Denisovan ancestry would be a welcome conclusion. After all, what reason do we have to assume that nature was parsimonious with its Denisovans? Much of what we know makes me expect to find more diversity among past peoples, not less. The divergence of the Denisovans from Neanderthals was definitely before 430,000 years ago, constrained by the fact that the Sima de los Huesos genomic samples are already on the Neanderthal population branch. According to estimates by Alan Rogers and coworkers, the Denisovan origin was likely soon after the stem Neanderthal-Denisovan common ancestors parted ways from sub-Saharan Africans, possibly close to 700,000 years ago (see my 2017 post, “How long ago did Neandertals and Denisovans part ways?”).

Four populations that only began diverging after 400,000 years ago would seem to be a paltry sampling of the Denisovan diaspora. If there weren’t other Denisovans, we’ll need to explain why.

Babies get their intestinal viruses in stages

A fascinating new paper by Guanxiang Liang and coworkers in Nature looks at how infants end up with a community of viruses in their guts: “The stepwise assembly of the neonatal virome is modulated by breastfeeding”.

Results indicate that, early after birth, pioneer bacteria colonize the infant gut and by one month prophages induced from these bacteria provide the predominant population of virus-like particles. By four months of life, identifiable viruses that replicate in human cells become more prominent. Multiple human viruses were more abundant in stool samples from babies who were exclusively fed on formula milk compared with those fed partially or fully on breast milk, paralleling reports that breast milk can be protective against viral infections. Bacteriophage populations also differed depending on whether or not the infant was breastfed. We show that the colonization of the infant gut is stepwise, first mainly by temperate bacteriophages induced from pioneer bacteria, and later by viruses that replicate in human cells; this second phase is modulated by breastfeeding.

Initially most of the viruses are bacteriophages, but as the abstract indicates, as time passes more and more of the viral community is made up of viruses that inhabit the human gut cells.

The discussion focuses upon the finding that breastfeeding reduces the level of pathogenic viruses in the infants. That finding replicates other work based on different approaches. One of the infant cohorts included in this study was from Botswana, and the findings there are more pronounced than in their U.S. (Philadelphia)-based cohorts.

In the African cohort, we not only found viruses that grow in human cells more commonly in exclusively formula-fed babies, but we also found more colonization in both feeding groups compared to US babies, emphasizing potential opportunities to intervene to reduce viral transmission to infants.

The variation of microbiomes in different human populations is a really important component of human biological variability. Studies like this one are helping to show how that variation comes about through the development process and environmental exposures.

Stone Age minds in internet time

The journalist Kenneth Miller has an article in the current Discover magazine on “How Our Ancient Brains Are Coping in the Age of Digital Distraction”. I make a brief appearance to help explain the genetic complexity of brain evolution on the human lineage.

It’s a good article and I’m going to assign it for my undergraduate class next week. My last week theme always covers what biological anthropology can tell us about the future of humanity. Usually people go overboard on the idea of “Stone Age minds”, but a look at optimality theory is rarely unrewarded.

Humans, of course, forage for data more voraciously than any other animal. And, like most foragers, we follow instinctive strategies for optimizing our search. Behavioral ecologists who study animals seeking nourishment have developed various models to predict their likely course of action. One of these, the marginal value theorem (MVT), applies to foragers in areas where food is found in patches, with resource-poor areas in between. The MVT can predict, for example, when a squirrel will quit gathering acorns in one tree and move on to the next, based on a formula assessing the costs and benefits of staying put — the number of nuts acquired per minute versus the time required for travel, and so on. Gazzaley sees the digital landscape as a similar environment, in which the patches are sources of information — a website, a smartphone, an email program. He believes an MVT-like formula may govern our online foraging: Each data patch provides diminishing returns over time as we use up information available there, or as we start to worry that better data might be available elsewhere.

Yeah, my problem is that whatever I am writing tends to look like diminishing marginal value too quickly!

Killing it: From handaxes to hafted spears, no difference in hunting or meat-eating

The change in technology from Acheulean to Middle Stone Age in Africa was a major event in human prehistory. Or was it?

If there is one generalization that we can make from the African record, with all its imperfections, it is that large cutting tools were more common in assemblages from the earlier Middle Pleistocene than later. A second generalization is that Levallois-flaked points and evidence of hafting are more common in the later Middle Pleistocene than earlier. These may seem like a trade-off, and they may have been – although I’m not aware of anyone testing with numbers whether the two are causally connected instead of merely coincidental trends.

I think most anthropologists have assumed that technological change from Acheulean handaxes to Middle Stone Age hafted spears would make some difference in how ancient hominins hunted, scavenged, and ate animals.

The biggest-ever review of this question was published last year, with data drawn from sites across Africa from 800,000 to 130,000 years ago. The paper from Geoff Smith and coworkers, published in the Journal of Human Evolution, provides a useful and detailed review of faunal exploitation across African sites from the later Acheulean and earlier Middle Stone Age: “Subsistence strategies throughout the African Middle Pleistocene: Faunal evidence for behavioral change and continuity across the Earlier to Middle Stone Age transition”.

The review is a meta-analysis of published faunal data from site reports and articles. Looking at sites from Morocco across to Ethiopia and down through Kenya and Tanzania to South Africa, Smith and coworkers compiled faunal lists and other information from a varied array of more than 40 sites, some of which had faunal evidence from multiple time periods. They conclude that what seems like a big change in technology made no difference at all to which prey animals hominins hunted.

The currently available faunal data do not support a broadening of the hominin dietary niche during the Middle Pleistocene. While smaller-sized bovids, such as gazelles, grysbok, southern reedbuck, and springbok, are preserved throughout the Middle Pleistocene sample, these species never illustrate significant increases. Similarly, dangerous game (e.g., Cape buffalo and long-horned buffalo; see SOM Table S2) are recorded in small numbers throughout the faunal dataset. Increased proportions of dangerous game within MSA assemblages have been argued to reflect improvements in projectile technology, which provides the ability to hunt game at greater distances (Klein et al., 2007). While the proportion of these species increases from the early (0.03%) to late Middle Pleistocene (4%), they never predominate. Results from this metastudy suggest that such a change neither occurred on a broad scale during the Middle Pleistocene nor in parallel with the appearance of MSA technologies.

The data are pretty clear about two things. First, there is an enormous difference between open air and cave sites in faunal evidence.

Figure 6 from Smith et al 2019, showing representation of different size prey animals in sites from Africa
Figure 6 from Smith et al. 2019. The left panel shows Acheulean-era sites, with prey species in different size classes (size class 1 is smallest, 6 is elephants). It's striking the difference between cave sites and open air sites, especially in the lack of large mammal remains.

Hominins did not move the bones of elephants and hippos to caves, and they rarely moved large bovid remains. This is a known bias in zooarchaeology, but when you see a figure like this, it really reinforces just how much the hominin behavioral record is biased by the places that archaeologists prefer to dig. Some of the bias is that there are many cave sites in the MSA time period. Another bias is that archaeologists have been more likely to investigate open air sites that have obvious preservation of large mammal remains. A large fraction of hominin foraging behavior may be entirely missing from the record.

The other obvious feature of the results is that there really is no difference between Acheulean and MSA faunal size distributions other than the greater number of MSA sites that are caves.

A third important generalization from the data is that later Middle Pleistocene sites have more evidence for hominin processing of carcasses than earlier sites. The earlier sites have greater evidence for carnivore involvement, while later sites have more cutmarks and evidence for transport of selected body parts. However, that greater intensity of hominin involvement may reflect the bias toward cave sites in the later Middle Pleistocene.

There is too little data on mortality profiles across these sites for Smith and coworkers to do a comparison of selectivity of prime age adults in prey species. Still, there are many Oldowan-era sites that have a clear bias toward selection of prime age adults by hominins, at least for some prey size classes. It is not probably going to be very informative to look at more detailed aspects of prey selection without first knowing that the hominin contribution is much higher than carnivore contribution to faunal assemblages.

The meta-analysis approach obviously has limits. The quality of data is constrained by the data that has been reported. This varies greatly in depth and detail because some sites were excavated more than 50 years ago.

Also, the chronological information about sites is of varied quality. A good example is Kabwe, included here as a Sangoan/early MSA site. The recent paper by Rainer Grün and coworkers that provides some dates for the hominin material makes it clear that the contextual relationship of artifacts, faunal remains, and hominin fossils in the Broken Hill mine is basically unknown. It may be that the faunal material is of later Middle Pleistocene age, but it’s not obvious what (if any) hominin involvement may have been involved in accumulating animal bones. Kabwe is one of the rare sites that have both hominin fossil material, archaeological material, and fossil material. The unfortunate reality is that being from the same site does not provide evidence of association.

Other sites have similar difficulties. It may not seem objectionable acknowledge the uncertainty about geological age and simply lump sites as early or late. Still, the earlier open-air sites may be quite short duration accumulations that have very large uncertainty of geological age, while later cave sites may sample thousands of years with a more precise knowledge of age. Much rides on whether they are classified as “Acheulean” or “MSA”, yet that classification also depends in some cases on a small artifact assemblage. This kind of classification may serve the purpose of testing whether these coarse groupings have any correlation to faunal exploitation. But we cannot assume that such groupings are meaningful.

An enormous hole in this kind of project is the lack of knowledge of which (if any) hominins are responsible for the faunal remains. To their credit, Smith and coworkers are agnostic about which species of hominins are represented by the behavioral patterns across these many sites:

Currently, associated hominin fossils are limited and do not allow for direct correlations between various African Middle Pleistocene hominin species (e.g., Homo heidelbergensis, Homo rhodesiensis, Homo naledi and Homo sapiens) and a specific lithic technology or subsistence strategy. In general, this period is recognized by a mosaic of lithic entities and hominin species and the broad scale trends discussed here are complementary with more local and regional scale variation in terms of exact timings and causal mechanisms of behavioral change and continuity.

That’s an appropriately cautious statement. We do not know which hominin species made the artifact assemblages sampled at these sites. There were likely other species or divergent populations in addition to the ones listed. All of them were tool users and ate hunted animals.

With that caution in place, it remains interesting that there is no pattern over time across faunal assemblages in prey size selection. I think that’s striking evidence that earlier Middle Pleistocene hominins had achieved the ability to efficiently hunt prey species in the proportions that made ecological sense for them. The fact that later hominins selected the same prey sizes suggests that this aspect of hominin hunting was near equilibrium regardless of the precise toolkits they used.

If you ask me who was doing the hunting, I’d say it was every hominin that existed in Africa across this time period. I think it is likely that they all had similar niches with respect to carnivory.

What ancient DNA is telling us about the prehistory and history of the horse

Ludovic Orlando has a great review of recent research into the origins and evolution of domesticated horses: “Ancient Genomes Reveal Unexpected Horse Domestication and Management Dynamics”. The review is open access in BioEssays.

Here’s a cool fact:

With nearly 300 ancient genomes sequenced at above onefold coverage, horses have provided the largest genomic time series characterized to date after humans.

I’m a bit surprised that dogs aren’t higher than horses, but horse bones are quite a bit more common in some archaeological settings.

Botai horses indeed did not show close genetic affinities to modern domestic breeds. They clustered instead together with the Przewalski's horse, a horse discovered in the late 1870s roaming wild in Mongolia, and considered since as the only truly wild horse living on the planet. In short, the earliest domestic horses known in the archaeological record appeared to be the direct ancestors of the only modern horse that was supposed to have never been domesticated. It then became obvious that current models of horse evolution required serious rethinking.

Instead, the ancestors of today’s domesticated horses appear in the archaeological record around 4100 years ago in Hungary. The date and subsequent widespread occurrence of this single lineage of horses is striking as a parallel to the spread of steppe ancestry in humans inhabiting the same regions. It would seem that the spread of Bronze Age steppe peoples brought a lineage of horses everywhere the people went.

The expansion of this population of domesticated horses is matched by the equally striking extinction of many diverse lineages of horses throughout Eurasia. The review suggests that additional lineages remain to be found, with hints of their existence given by the rare occurrence of very divergent mitochondrial DNA haplotypes in a few ancient specimens.

The review goes on to discuss several other issues related to horse genetics. One important topic is the introduction of progressively stronger inbreeding in the last 2000 years. Some societies selected for particular male patrilines, while others were more ecumenical.

Ancient DNA can cast an invaluable light into ancient phenotypes that are invisible from skeletal evidence. The most widely reported of these phenotypes so far relate to pigmentation. That’s not only because pigmentation phenotypes are highly visible, it’s also because they are genetically simple enough to enable reliable phenotype prediction from sparse SNP data.

With horses, there is a lot to go on for pigmentation analysis. Many well-known color variations exist today, and they differ within and between horse breeds. The genetic variants that correlate with many of them are known. The element related to pigmentation that Orlando includes in his review relates to a cultural pattern:

It is noteworthy that this information cannot only reveal the traits that past breeders most likely selected, but can also help document past funerary traditions. For instance, the analysis of coat coloration loci in the 13 complete horse skeletons found in the funerary monument of Berel (Kazakhstan) revealed that ≈2500 years ago already, Scythian Pazyryk Iron Age nomads herded the full diversity of horse coat colors present in the region today. Since those horses were specifically killed for the funerals of Pazyryk elite members, the genetic data also showed that sacrifices were not targeted toward particular family groups or coat colors.

Orlando here cites his own 2017 work in a paper led by Pablo Librado.

Overall it is a useful and interesting review. We are relatively advanced in knowledge of horse domestication now compared to many other domesticates, at least with respect to the major ancestors of today’s breeds. We have to keep in mind how much we don’t know about ancient variation that has not been sampled, and the discovery of presently-unknown populations should not surprise us.

The value of reading the literature

Probably the most common thing for fellow researchers to ask me about my blog is, “How do you have time?”

In a new post, the Drugmonkey “Reading the literature and writing”

I think a lot of scientists really don’t like to read deeply into the literature. At best, perhaps they weren’t ever trained to read deeply.
So I’m not talking about reading the literature in a “keep up with the latest TOCs for all the relevant journals” kind of way. I’m talking about focused reading when you want to answer a question for yourself and to put it into some sort of scholarly setting, like a Discussion.

For me, blogging is part of how I engage with the literature in my field. I take notes, I cite sources, and I look into the way that people have developed their ideas. When it makes sense, I share that with others. And I am much more likely to think deeply when I am thinking about how to convey something for other people.

I do get the feeling that this kind of deep reading and engagement is actually quite rare, and that many researchers have a more tactical approach to reading. “Tactical” just means that people try to maximize their short-term benefits from reading and writing. I think it’s better to be strategic in professional reading, and sometimes that means reading with deep engagement.

Quote: Loren Eiseley on the march of progress

In 1948, Scientific American published an article by the anthropologist Loren Eiseley, titled, “Antiquity of modern man”. In it, Eiseley conveyed some details of a skull discovery from Fontèchevade, France, which had been made the previous year by the French archaeologist Germaine Henri-Martin.

The story of the Fontèchevade cranial remains is a long and complicated one, with a simple ending. What seemed in 1948 like evidence for an ancient population of modern people living during the last interglacial turned out to be only around 35,000 years old. The two Fontèchevade individuals look modern but they weren’t older than Neanderthals at all.

Still, at the time they were discovered, the Fontèchevade remains helped keep alive the idea of a parallel population of modern ancestors in Europe even after the Piltdown hoax was exposed. They were perceived as the most well-documented evidence of a “presapiens” population, which helped sideline Neandertals as possible human ancestors. For that reason they are very important to the historical development of paleoanthropology.

Eiseley leapt right onto the Fontèchevade story as proof of a deep history of modern humans—hence, the article’s title, “Antiquity of modern man.”

I sort of love Eiseley’s take on the usual, “You wouldn’t notice him on the subway” line:

It is a small skull within the size range of living females. There is nothing Neanderthaloid about it. This woman could have sat across from you on the subway yesterday and you would not have screamed. You might even have smiled.

But the reason I wanted to share this article is for Eiseley’s meditation on the on the idea of the “March of Progress”. This passage follows as he tells the history of early discoveries of fossil humans in England, beginning with Kent’s Cavern. Eiseley notes that the discovery came at a time when the great age of the Earth was unknown, and thus had to be fit within a short timescale.

In Eiseley’s telling, the huge change in scientific view that followed was not enough to shift ideas to the radical idea that our own evolution had formed a tree.

It took the rest of the century and the long thought of a biological genius, Charles Darwin, before the idea of eons of time became acceptable. and the bodies of men and animals were seen to melt and flow and change from age to age like the hills they moved upon.
Even then, perhaps, the vision was still beyond us. The human mind always tends to erect new dogma to shelter itself in hastily erected systems against what is not known or what proves at last to be unknowable. The forms of paleoanthropic, big-browed fossil men began to be discovered. Though their numbers were few, scientists fitted them into a system--a single line of ascent leading to modern man. A form like Pithecanthropus, for example, led on in the following age to Neanderthal man. and the latter was regarded as our own direct ancestor. At the other end of the succession. the beginning, was an ape generally conceived of as differing little from a modern chimpanzee.
The sequence was thought of as short and very direct. The time scale was still being underestimated, and western Europe, actually marginal to the Asiatic land mass, was unconsciously overemphasized as an evolutionary center for mankind. In addition, certain preconceptions were making it difficult to survey the problem of the origin of modern man in an unprejudiced light.
The most obvious of these preconceptions was, of course, the idea that since the remains of Neanderthal man had been found in European deposits immediately underlying our own species, we must be a later breed. Thus there could be no valid remains of Homo sapiens that were as old as Neanderthal man in Europe.

I appreciate that line, “new dogma to shelter itself in hastily erected systems”.

Quote: Arthur Keith on Weidenreich and Piltdown

In 1944, after receiving Franz Weidenreich’s monograph on the fossil sample from Choukoutien, China (now spelled as Zhoukoudian), Arthur Keith wrote a letter to the editor of Nature to express his thoughts on Weidenreich’s ideas: “Evolution of modern man (Homo sapiens).

Keith, who held many sensible views even earlier in his career, really softened toward the light of facts in the later 1930s and 1940s, and wrote several notes and letters complimenting the ideas of scientists whom he had criticized in his earlier work. Dart is probably the most famous example, but Weidenreich comes in for praise in this Nature correspondence.

Still, there is the bone of contention:

I have mentioned that as, regards the origin of modern races of mankind, Dr. Weidenreich and I have reached a large measure of agreement, all save in the case of that most ancient of Englishmen, Piltdown man (Eoanthropus). Dr. Weidenreich is of the belief that all surviving races of mankind have passed through a "Neanderthaloid" stage in their evolution, a stage which was apparently omitted in the case of Piltdown man. He is therefore removed by Dr. Weidenreich from the list of authentic fossil men, his skull being assigned to a modern type of man, while his lower jaw is given to a fossil anthropoid akin to the orang. Virchow solved the mixed simian characters of Pithecanthropus in a similar way, assigning the skull to an ape and the femur to a man. In England we find it hard to believe that there lived in the Weald of Sussex, in earliest pleistocene times, a modern type of man and a rather human-like ape, and that by some strange chance the bones of these two became mingled in the Piltdown gravel bed. Not only was the Piltdown race alive in England when the rest of Europe seems to have been occupied by human stock of the Neanderthal breed, but also this ancient race appears to have come down to mid-pleistocene times; at least it is on such a supposition we can best explain the characters of the Swanscombe and London fossil skulls.

People often wonder why Piltdown comes up so much today, when it has been known to be fake for seventy years—much longer in fact than it was thought to be real. One reason is that it carried such weight at the time when it was accepted, and really stymied the development of ideas in paleoanthropology at a time when evolution was undergoing a synthesis. Its proponents (at first) and defenders (in later years) saw the fossil as objective evidence for unstated or unconscious assumptions about the advancement of European, and specifically English, peoples in prehistoric times.

Quote: Fay Cooper-Cole on bones and races

In 1945, American Naturalist published a lecture by Fay Cooper-Cole on “Some problems of human racial development and migration”. Cooper-Cole had been a student of Frans Boas and helped to establish Anthropology at the University of Chicago. He had also been an expert witness at the Scopes trial in 1925. At the time this lecture was published he was 64 years old.

The lecture is interesting from a historical point of view, although overall it reads like Cooper-Cole was speaking off the cuff with little systematic preparation. The history includes things like Piltdown that were wrong, and others like the placement of the Swanscombe and Steinheim crania that today do not carry the significance that they once did.

But it’s fascinating to read the paragraphs near the beginning and end of the lecture. In them he examined more enduring issues.

At the beginning:

My difficulties begin when I use the term human. Where is the dividing line between prehunian and human? Just about the time of the Dayton trial came the discovery of Australopithecus. A London journal came out with an illustrated article calling this the oldest find of fossil man in Africa or perhaps in the world. American papers copied and one of our British colleagues was so fearful I might use the material in my testimony that he sent me a long message saying the find, was not man but an ape far advanced toward man.
How human does a "man-ape" have to be before we class him as human? Does the distinction lie in cranial capacity? If so, the six-year-old Australopithecus had already passed out of and beyond the range of Anthropoid apes of like age.

And the end, as he runs out of time to talk about modern human differences:

Those of you who deal with other animals don't realize how fortunate you are. You can put your subjects into a cage; you can control their mating and you can watch results over several generations. No such happy situation faces the student of man. He is a slow breeder; he refuses to conform in his mating habits; and you can't lock him up for long. Paleontologists may say "if these bones could only speak," but if we judge by living man, most information on racial matters that the talking bones would give would be misinformation.

That last line is uncharacteristic of the lecture as a whole. Cooper-Cole’s view on races as expressed in the lecture mostly relates a view of human variation like that of his contemporary Earnest Hooton. But the limits on that understanding come out in this last paragraph, and I think the last sentence is very quotable.

Chicken ancient DNA and natural selection during medieval times

Three years ago, Liisa Loog and coworkers published a fascinating paper quantifying natural selection from ancient DNA data in chickens: “Inferring Allele Frequency Trajectories from Ancient DNA Indicates That Selection on a Chicken Gene Coincided with Changes in Medieval Husbandry Practices”.

Ancient DNA approaches are more and more being applied to zooarchaeological remains, especially in the study of domestication. Our commensal domesticated animals have some great opportunities for examining evolutionary questions. Geneticists have invested a lot in understanding their genetics today, including very good whole-genome drafts for all of them, making it easier to look at ancient genomes. We have a good idea of what some SNP polymorphisms do in these populations today, so that gives an opportunity to study the evolution of functional variation.

Finally, because of a history of strong human selection, the genetic structure of some phenotypic variation is strongly biased toward large-effect alleles. That’s pretty ideal when we are considering ancient DNA data, because small sample size makes it pretty much impossible to see very slight changes in allele frequencies over time. So if you’re studying a trait that is affected by hundreds of loci with small effect sizes, you are unlikely to be able to statistically demonstrate evolutionary change, even it if really happened.

But there are some problems using ancient DNA data to look at selection all the same. The authors discuss these at some length. This is an important concept to understand:

Ancient DNA data can provide direct information on changes in allele frequencies through time, and as such allows us to directly link past selection to contemporaneous ecological factors. However, ancient DNA sample sizes are typically small, and samples tend to be sparsely and unevenly distributed in space and time. Several methods exist for studying selection using ancient DNA (reviewed in Malaspinas 2016), but typically they lack the ability to model the confounding effects of gene flow—a process that could lead to overestimation of selection coefficients (Mathieson et al. 2015).

The confounding effects of gene flow are what slowed some of my work on Holocene natural selection around 10 years ago. When we look at the skeletons of post-Neolithic human populations in Europe, it is clear that people changed substantially in many skeletal phenotypes over time. But it is hard to tell how much of that phenotypic evolution was the result of natural selection, because there was also large-scale migration into Europe from other parts of the world. Ancient DNA has made that recent migration even more apparent, with some of today’s populations deriving much more of their ancestry from Bronze Age central Asia and the Near East than from Neolithic people in Europe.

As long as you only have one population to look at, it will be hard to tell whether change was caused by in situ natural selection. Change might instead have resulted from immigration from populations that you haven’t sampled.

This isn’t a unique problem to ancient DNA. It is true of the skeleton, of teeth, of anything where you want to study change over time. You need to know something about the whole spectrum of populations that have contributed genes into later people, to be confident about the relative roles of gene flow and selection.

(Of course, these are not mutually exclusive forces. Change that looks like it comes from migration may be caused by natural selection increasing the survival and reproduction of immigrants. Or natural selection upon one population may increase its fitness to the extent that individuals begin to emigrate to other populations. Demography and selection are intertwined.)

If anything, ancient DNA should make it somewhat easier to untangle the effects of selection from those of migration. The reason is that your sample of DNA can rapidly grow larger than your sample of morphology. All you need to determine a genome is a small part of a bone, while examining a morphological character requires that exact part to have been preserved and recovered.

At the moment, the situation may seem like the opposite. But that’s only because ancient DNA right now is relatively expensive. Once it becomes practical to get genomes from tiny bone fragments at minimal expense, we will know much more about ancient allele frequencies than we know about ancient morphological traits.

Anyway, back to the chickens. Loog and coworkers focused on two genes previously implicated in domestication, as they describe in their background:

The domestication of chickens (Gallus gallus domesticus) from wild junglefowl took place over a relatively short evolutionary timeframe (Rubin et al. 2010). In addition, the genetic architectures underlying many of the phenotypic traits that distinguish domestic populations are well understood (Eriksson et al. 2008; Rubin et al. 2010). As such, chickens present an ideal opportunity to correlate the timing of allele frequencies shifts with changes in human imposed selective pressures over the past several thousand years. To do so, we focused on two loci: thyroid-stimulating hormone receptor (TSHR) and β-carotene dioxygenase 2 (BCDO2), both previously identified as targets of selection in domesticated chickens (Rubin et al. 2010).

Having already identified candidate loci, they focused on these in the ancient samples. But they faced the problem that Asian chickens have different frequencies of selected alleles, and Asia is a possible source population for the chickens. So they had to straighten out whether in situ selection or gene flow from Asia was a better explanation for any frequency changes they found.

The authors compared the time course of frequency change in European chickens to their inferred time course of gene flow from Asia. For both genes, they infer that the frequency changes were initiated before substantial gene flow from Asian chickens was underway.

Summary of chicken allele frequencies over time from Loog 2017
Figure 1 from Loog et al. 2017. Original caption: "Observed allele counts through time for the TSHR locus (a) and the BCDO2 locus (b). Wild type alleles in ancient samples are represented by yellow dots and derived alleles by red dots. Modern allele proportions are shown as solid bars to the right of each panel."

I don’t think the approach they took is especially satisfying. It would be better to test whether their loci of interest (TSHR and BCDO2) have different dynamics than the genome-wide pattern.

Still, the genes themselves are pretty interesting in the way they impact chicken biology. They give a good idea of how zooarchaeological information from many sites might be integrated into a coevolutionary picture of domesticate evolution and human cultural practices.

Quote: Sweeping out the cruft for cladistics in human origins

Eric Delson, Niles Eldredge, and Ian Tattersall in 1977 published on one of the first cladistic analyses of humans and our close relatives: “Reconstruction of Homlnid Phylogeny: A Testable Framework Based on Cladlstic Analysis”.

I was reading through this paper the other day, and found the introductory paragraph provocative, and certainly worth a re-reading.

The controversy that has raged to this date in hominid phylogenetics will continue to bedevil us, if we persist in approactfing the problem in the usual fashion. As things stand, neither new data, nor new techniques for their manipulation, will now alter the level of controversy one way or the other---only the number of people active in the field will determine how many different opinions are held. This situation stems from the form in which hypotheses are conventionally proposed---and statements of phylogenetic relationship are very much hypotheses. The simple truth is that the sundry hominid "phylogenies" available in the literature are scenarios amalgams of statements of ancestor-descendant relationship thoroughly admixed with, and largely (although not always consciously) based on, a priori models of the evolutionary process, on interpretations of the significance of the stratigraphic and geographic occurrence of fossils, and reconstructions of the palaeoenvironment and functional anatomy--hence adaptations, of these creatures. Statements as complex as these, as far removed from the data base as they frequently are, are difficult to test rigorously and leave us only with a vague feeling that one perhaps seems more "sensible" or more plausible probabilistically than another. We may celebrate the controversy as healthy, but it is a bit disconcerting to realize that phylogenies stated in these terms are by their very nature incapable of rigorous comparison, testing and rejection.

This has a markedly ironic tone today, in the best Alanis Morissette sense. The passage introduces a paper that argues that the “scientific” approach of cladistics should clear up the confusion of hominid systematics. It didn’t turn out that way.

The advent of cladistic methodology in human evolution studies only brought new forms of disagreement. As a younger generation of scientists applied cladistic approaches, they introduced a much broader diversity of views about hominin relationships than had existed in the 1960s and 1970s.

Many of the players writing about hominin relationships were aggressive in their papers and at scientific conferences. They irritated many others by acting like they were the “real” scientists, using “objective” methods.

Some other paleoanthropologists established themselves as the anti-cladists, publishing about how cladistic approaches “atomize” traits that should really be considered as part of integrated functional or developmental packages. The cladists would often assert that such hypotheses are untested, and therefore should not be considered in phylogenetic analysis.

This would be a great area for a historian of science to examine. It was a transition in methods, a generational turnover, and many of today’s active paleoanthropologists built their careers by publishing in this area.

Reading a review of 'The Idea of the Brain'

This week I read a review in Nature of Matthew Cobb’s forthcoming book, The Idea of the Brain: The Past and Future of Neuroscience. The review by Stephen Casper is here: “Neuroscience needs some new ideas”.

As recounted by Casper, the book covers the history of metaphors for the brain, looking at the ways that these metaphors have shaped research agendas over the decades, and sometimes impeded understanding.

The late-nineteenth-century discovery of neurons led to a clash of rival images. Reformers imagined separate components, comparable to the wires and signals of the nascent telecommunications infrastructure. Conservatives cast the nervous system as a continuous network (or reticulum) akin to the blood circulation, feeling that this explained how volition and mind might work; to them, discrete signalling implied heterodox notions of mind, perhaps even of the soul.
The post-1940 proliferation of references to enchanted looms, ghosts in machines, logical circuits, reptile brains, parallel processors and uploaded minds grew from those foundations. Cobb notes, but only in passing, that we need new images to make sense of research developments ranging from artificial intelligence to mini-brains grown in the laboratory to brain implants. He doesn’t try to invent examples.

I’m looking forward to this book coming out later in the month. I put a lot of history into my course on evolutionary neuroscience, Biology of Mind. The history is important to enable students who are not themselves neuroscience students to understand how the science has shaped other fields, as well as to think through the limits on current knowledge.

Hippos in South America

The Washington Post has an opinion piece by Robert Gebelhoff asking the tough questions about a South American herd of African herbivores: “The great conundrum of Pablo Escobar’s hippos”:

Today, Escobar’s herd has grown to upward of 100 strong. To residents, they are a threatening menace, but among scientists their presence is the source of spirited debate. Are Escobar’s hippos “invasive”? Or are they “introduced”? Are they threatening the local ecological community? Or are they helping to “rewild” the area? The answer is far from clear, but the debate could change the way we think about preserving habitats.

He raises several interesting points, especially the fact that large notoungulates existed in South America in a hippo-like niche that is now empty, probably because of human-induced extinctions. I figure that hippos in South America are a fair exchange considering that African waterways are increasingly choked by invasive South American water hyacinths.

In the long run, of course, all large herbivores will be managed, whether they are introduced, invasive, or native. Already in large parts of the world, it is human aesthetic taste that shapes the management of large mammals of all kinds. Even the most “natural” environments are shaped by people, with conservation priorities that may look quite different from the same environments 10,000 years ago.

A recent paper in Ecology by Jonathan Shurin and coworkers looked at the way that hippos in Columbia are affecting their environment: “Ecosystem effects of the world’s largest invasive animal”.

Hippos in Africa fertilize lakes and rivers by grazing on land and excreting wastes in the water. Stable isotopes indicate that terrestrial sources contribute more carbon in Colombian lakes containing hippo populations, and daily dissolved oxygen cycles suggest that their presence stimulates ecosystem metabolism. Phytoplankton communities were more dominated by cyanobacteria in lakes with hippos, and bacteria, zooplankton, and benthic invertebrate communities were similar regardless of hippo presence. Our results suggest that hippos recapitulate their role as ecosystem engineers in Colombia, importing terrestrial organic matter and nutrients with detectable impacts on ecosystem metabolism and community structure in the early stages of invasion. Ongoing range expansion may pose a threat to water resources.

The unavoidable reality is that people will shape their understanding of nature in ways that are convenient for themselves. One day society may forget the effects of wild megafauna entirely. Or they may engineer environments using the remaining megafauna in ways that are convenient to humans.

Mask fiasco

This week, the New York Times published a piece by information scientist Zeynep Tufekci, “Why Telling People They Don’t Need Masks Backfired”. The main idea is that the message (ordinary people will not be protected by wearing masks during a pandemic) was obviously contradicting the very real need for masks by medical providers. The essay covers several of the reasons why mask-wearing is helpful, even if it is imperfect.

I appreciated the tone and the ending of the piece:

Research shows that during disasters, people can show strikingly altruistic behavior, but interventions by authorities can backfire if they fuel mistrust or treat the public as an adversary rather than people who will step up if treated with respect.

Quote: Eileen Whitehead Erlanson defining taxonomic inflation

I was reading today to find the origin of the term “taxonomic inflation”. This is a common idea today from people who criticize an overzealous attention to defining species. The term “taxonomic inflation” is especially used by detractors of the phylogenetic species concept, on the logic that this species concept results in naming species from insufficient data.

In a discussion of species concepts, I wanted to provide this perspective and cite its origin, and I found that nobody who today uses the term ever cites its origin.

I found the earliest use of the term in a 1934 paper by the botanist Eileen Whitehead Erlanson. She uses it in a discussion of the problems of recognizing varieties below the species level:

The taxonomist must find some way of describing all the multitude of natural forms, and in Rosa the problem of how to treat the units smaller than a species is perplexing. GREGOR (23) pointed out that if a varietal name is appended to every specimen which differs in some minor point from the general description of the group, then "every genotype and every clonal modification of an individual would ultimately deserve varietal or sometimes specific rank." To name all the possible combinations of minor characteristics leads to such a state of taxonomic inflation that the categories below the genus lose all practical value. Nor can the morphological criteria alone be accepted as supplying a completely reliable system of classification (23), for morphologically similar forms may differ markedly physiologically, as has been found in R. acicglaris, R. woodsii Lindl., and R. arkansana Porter. Probably the most promising method for dealing with the smaller units is that of distinguishing minor variations by numbers or letters, which makes it possible to express dif- ferent character combinations clearly and succinctly without degrading scientific nomenclature. This has been advocated by VAVILOV (42), HALL and CLEMENTS (26), CLAUSEN (7), and HALL (25). It was recommended for the varieties of rose species by MATTHEWS (31) and is a method that I also would endorse.

The paper is “Experimental Data for a Revision of the North American Wild Roses” in the Botanical Gazette, which later became the International Journal of Plant Sciences.

The next person to use the term “taxonomic inflation” was the paleontologist William K. Gregory in a 1936 paper, On the Meaning and Limits of Irreversibility of Evolution”. I mention this because it is the only other early use of the term, but Gregory used it in a very different way that is not similar to the present usage of the term. Gregory was concerned that paleontologists were overzealous in their application of the idea that specializations cannot be reversed in evolution, and that this caused them to accept a much higher antiquity for some species lineages than they would if they could accept that specialized characters sometimes reverse to an ancestral state. Hence, the number of evolutionary lineages is “inflated” by the assumption that they must have separated before any specialized traits evolved.

The use of the term, taxonomic inflation, really took off in the 1960s after the president of the Linnaean Society of London, one T. M. Harris, devoted a speech to “The Inflation of Taxonomy”. This was a wholesale critique of oversplitting at every level of taxonomy, and was clearly influential upon later authors. However, as a speech, Harris seems to have felt no need to cite earlier authors. Nor did many later authors cite Harris’ work.

It seems that “taxonomic inflation” was an idea that many wanted to use, and few sought for the term’s source. I’m going to credit Erlanson.