This past Friday, our journal club took on “The affinities of Homo floresiensis based on phylogenetic analyses of cranial, dental, and postcranial characters” (Argue et al. 2017). Essentially, Argue and colleagues attempted to figure out what other hominin species H. floresiensis (often called the Hobbit) was most closely related to, using statistical tree-building methods.
Since it was published in 2004 by Brown et al., H. floresiensis has been a bit of a mystery. Much like Homo naledi, there’s been a lot of discussion about where it belongs in the human family tree because its anatomy was A) weird and B) totally unexpected for its age (somewhere between 100-60 thousand years old) and its geographic location (on Flores, a small Indonesian island). The Hobbit was very short in stature, with a very small brain (in the range of orangutans, chimpanzees, and the much-older australopithecines), large teeth for its size, primitive-looking wrist bones, and disproportionately large feet relative to its height and leg length (hence its nickname of the Hobbit). Its discovery on Flores was a surprise because the other hominins that have been found in Indonesia (like Homo erectus from Java) were older and had larger brains (and we generally think brain size in the human lineage has increased over time – but last week’s chat about H. naledi also brought this up as a problematic assumption).
In their new article, Argue et al. set out to test two hypotheses: either the Hobbit is a late survivor from an earlier primitive hominin lineage, or it is a dwarfed descendent of H. erectus. They also commented on another controversial claim that’s been made about the Hobbit – that it is simply a modern human with a genetic/developmental pathology. They tested their hypotheses by applying two tree-building methods to a large sample of characters (particular features or measurements of the skeleton) from the skull, teeth, and postcranial (below the head) skeleton. One method (parsimony) attempts to build the shortest possible tree (one that requires the fewest changes in traits to get from species to species), while the other method uses probability to figure out which trees are most likely to occur (given a particular model of evolution).
When you set out to do a project like this, you’re forced to make some choices as a paleoanthropologist. If you have isolated postcranial bones from a hominin site where you’ve previously found fossils of more than one hominin species from the same time, how do you decide which body belongs with which head? You also confront the issue that not all researchers agree on which specimens belong in which species. And, as always, the fossil record is incomplete; you don’t have all of the characters for all of the species. To account for these potential problems, Argue et al. tested their two hypotheses with several different data sets – for example, they did one test where they considered all of the potential postcranial skeletal material that’s been called Homo habilis to actually be H. habilis and another where they excluded the questionable material.
What Argue et al. found was that their two different hypothesis testing methods and various different data sets produced broadly similar results in support of the first hypothesis: the Hobbit either shared a common ancestor with Homo habilis or is the sister group to the grouping of Homo habilis/Homo erectus/Homo ergaster/Homo sapiens. They are able to reject the hypothesis that the Hobbit is a dwarfed H. erectus (and reject a number of other species as possible close relatives). What this suggests is that (as was proposed for Homo naledi in last week’s papers) there is a long ghost lineage (unknown ancestors) for the Hobbit dating back more than 1.75 million years that is still waiting to be found. Ghost lineages – so hot right now.
Read on for some BS&M discussion bits!