Punctilious scientists correct Google Doodle, possibly incorrectly

November 25, 2015 • 10:45 am

I wrote yesterday about the “Lucy” Google Doodle, which looked like this:

41st-anniversary-of-the-discovery-of-lucy-5736109501841408-hp

Some of my colleagues didn’t like that Doodle, and fixed it. They couldn’t help it . . .:

Actually, I’m not sure this is scientifically accurate, as it shows Lucy (middle figure; A. afarensis) as a lineage that split off from modern humans rather than being one of our direct ancestors. We don’t know that, as Lucy’s species could have continued evolving into modern H. sapiens. The leftmost ape, if it’s a modern chimp, is correct, as they certainly branched off before hominins. But I’m not even sure, nor are the people at the Beacon Center, whether it’s a modern chimp or some ancestor of modern humans.

I therefore asked Greg Mayer, who knows more about this stuff than I do, to tell me if I was right in what I just said. He responded at some length, and even made a figure (upshot: I was right, and readers not conversant with taxonomy needn’t read on). Greg’s comments:

Yes, afarensis may have living descendants, and thus to show it as a terminal lineage isn’t quite right (or, at least, it involves additional assumptions). Some conventions for depicting fossil species in a tree with living species are needed, and it could be plausibly argued that they should be shown as terminal taxa (i.e at the tip of a branch, as in the corrected Google doodle Figure A).

Lucy phylo tree

But when a fossil species may be ancestral to a living species, we would consider the branch leading to the fossil species to have 0 length (i.e. it is at the node: Figure B).

Fossil species are usually depicted as terminal taxa, on the reasonable notion that of all the many species in an ancestral group, the probability of finding the ancestor is small. (For example, we know that early synapsids are ancestral to mammals, but the chance that Dimetrodon grandis is the ancestor of all mammals is small, so it probably is accurate to show it as a terminal taxon.) But when dealing with recent events with a rich fossil record in a geographically proscribed area, the probability of finding the ancestor does go up to the point where it’s no longer safe to make the assumption that it won’t be found.

For molecular and chromosomal data, where a complete description of the phenotype/genotype (i.e., nucleotide or banding sequence) is possible, actual ancestors can often be identified, and are often shown in trees as lying along the branches or at nodes in the tree.

Based on what we know about how speciation occurs, most ancestral species are probably paraphyletic relative to their descendants (Figure C, with afarensis zoomed in on and shown as ancestral to the lineage leading to Homo), and thus some parts of afarensis are terminal relative to Homo. If Lucy were in one of these terminal bits (the X in Figure C), then the corrected Google Doodle could be construed as accurate.

(I made the figure before checking the original Google doodle, and now realize that the leftmost ape was not identified as a chimp. If it was intended to be a chimp, then all of the above holds. However, if Google intended it to be a fossil ancestral species– which I think most likely– then it should fall along the branch, and not be a terminal taxon.)

20 thoughts on “Punctilious scientists correct Google Doodle, possibly incorrectly”

  1. Seems unlikely that any given fossil more than a few million years old is likely to be *strictly* ancestral.

    My assumption has always been “approximately ancestral” in such cases — so Lucy is (presumed?) only (at most) a short ways off the direct ancestral line.

    I do tend to prefer an indication of time in such trees, so that Lucy should be close to the root.

    Anyway, this correction seems to me to be nitpicking, and as you point out, not necessarily even correct nitpicking.

  2. Yes, but given that several species of australopithecines lived contemporaneously in the Pliocene (in both eastern and southern Africa), it almost seems more parsimonious (and conventional) to stick with tree A. Similarly, if one looks at a dinosaur tree, with everything in the tree extinct except the birds, it is conventional to but the birds as the sister group to the dromaeosaurs (or some other dinosaur lineage – I don’t know the latest on that) EVEN THOUGH a dromaeosaur lineage may have left descendants in the modern birds. This is almost always a problem when phylogenies combine living and extinct species.

    Why add unnecessary assumptions to what you’d like to portray? Tree A simply depicts the unassailable conclusion that afarensis shares common ancestry with humans much more recently than chimps do, even if Tree B is how things really happened. So I guess I still like the revised doodle, and that actually holds whether the creature on the left is meant to be a modern chimp OR an earlier hominin that shares common ancestry with us further back in time than afarensis.

    Without sequences from these extinct species, and the use of clock-like genes, there seems little justification for putting some extinct organism in an arbitrary location “down” the tree.

    1. The last point is just meant to say that in such trees the branch lengths are meaningless and arbitrary, the branching order is what is informative.

      1. We have dating on the fossil, and we know the branch must be prior to that and after the branch with chimps, so we don’t have to have sequencing & clock to get a more meaninful tree, vis-a-vis time.

        One could even use “fuzzy” lines, and or put circles or something to serve as “error bars” on the branching points, to reflect uncertainty.

        Of course one can do a tree whose only purpose is to show common ancestry, and such are fine. But why waste the infomation potential of the vertical dimension?

        1. Good point. But even after we make the branch lengths meaningful with respect to time, we are still left with the arbitary decision as to whether you put afarensis on a line directly leading to humans, on a side branch that ends “in mid-air” (before reaching the present), or as portrayed in Tree A. Other phylogenies that mix extinct and living organisms often simply make all lines reach the top of the tree in order to emphasize genealogical relationships and branching order, even if those lines really “end” further down the tree for the extinct organisms..

  3. Since we are being punctilious, I assume the left-most creature represents not a modern chimp but an ancestor of Lucy and Homo Sapiens. We assume it must be chimp-like, but we should remember that the Pan line and the Homo line have had equal time (6 to 13 million years) to diverge from the most recent common ancestor.

    1. I can go along with that. If we assume that the ape is our ancestor and not a chimp then that part of the linear google doodle is correct. Further, if we assume that the early hominin is not necessarily A. afarensis but is instead an early hominin ancestor to humans (and it could</I. be A. afarensis, as far as I know), then that part of the doodle is also correct.

  6. I second Greg Mayer’s comments. There is an anathema against ever letting a fossil species be ancestral to a modern species or modern group. For clams, that may be an excellent convention, but for hominids not so much. For example, for Homo ergaster, is there likely to any sister species of it in Africa that was ancestral to early humans? If not, doesn’t that suggest that it really is our ancestor?

    The observations of “several species of australopithecine” may be correct, but a note of caution should be sounded. Paleoanthropologists are under a lot of pressure to split rather than lump, so that they can announce that they have found a new species of hominid.

    1. Re your last sentence.

      I’d never thought about what might underlie the lumper/splitter divide but the moment I read your last sentence, I had an “ah ha!” reaction. Makes perfect sense. thank you.

  7. The left most seems to be knuckle-walking, and that would mean chimp rather than common ancestor of chimp and human.

    1. Gorillas knuckle walk too, so the common ancestor of humans and chimps could well have been a knuckle walker. Either knuckle walking evolved independently twice, or it was lost in humans– either way it’s two evolutionary events, so there’s nothing to choose from on parsimony grounds between the chimp-human common ancestor being a knuckle walker or not. Some people have argued that knuckle walking in gorillas is fundamentally different from knuckle walking in chimps (in the way bat wings differ from bird wings), so that knuckle walking in the two apes is analogous and therefore independently evolved, and some have argued the exact opposite; I am not in a position to judge.

      GCM

  8. I like Greg’s hand-drawn diagrams, but shouldn’t he have “I think” written above them?

  9. Could be even more complicated when you consider that recent hominids such as H. sapiens, H. neanderthalensis and Denisovans interbred and behaved like a complex ring species, possibly even with H. erectus. There is no reason to suppose that the African Australopiths and archaic hominids couldn’t have done the same.

    The problem comes from trying to represent evolution as a tree, which makes it look like speciation was an event rather than a slow process during which barriers to hybridization might have been slow to develop.

    1. I suppose that’s one way of saying that the minutiae of human evolution are a bit of a Charlie Foxtrot.

Comments are closed.