The Loom

i-848b23240346da92123415c8e9a743d1-filename.jpgLast week I wrote about a new study that identified a fossil mammal as the closest relative to whales, helping to shed light on how whales moved from land to sea. The mammal, Indohyus, was a small four-legged creature that probably spent a fair amount of time in water and ate vegetation. The authors of the new study proposed that the ancestors of whales originally lived this way. Gradually, the whale lineage became more adapted to life in water and shifted to eating meat, as exemplified by early whales like Ambulocetus, which was something like a furry alligator.

In the comment thread, Noumenon asked this question:

I don’t understand how Indohyus and Ambulocetus, both dated to around 47 mya, can both be the ancestors of today’s whales. You say carnivory was an important transition for whales. Then Indohyus would have had to split off before Pakicetus, before whales became carnivorous.

Via email, I got a similar question from a biologist I know who is working on a book about evolution. He had read about the discovery in this article by Ian Sample in the Guardian, who declared:

Fossil hunters have discovered the remains of the earliest ancestor of the modern whale: a small deer-like animal that waded in lagoons and munched on vegetation.

So how can an ancestor be younger than its descendants?

Simple answer: It can’t.

If you head back to the original paper, you’ll see that the scientists compared the skeletons of a lot of living and dead mammals in order to draw a tree. Each species they studied is a twig on that tree. Each twig is linked to other twigs through a shared ancestry. But the scientists did not line up species in a chain of ancestors and descendants.

It is sometimes possible to find the fossil of one extinct species that evolved into another extinct species. But if scientists only studied evolution that way, they’d be ignoring a wealth of other clues to how evolution unfolded. As a lineage of mammals evolves, it acquires traits that will set all its subsequent members apart from other mammals. Roughly 150 million years ago our ancestors evolved a placenta. Placentas are found in humans, bats, elephants, whales, and thousands of other species. They are not found in platypus or echidnas. Platypus and echnidnas, thanks to their position on neighboring branches in the tree of life, tell us something about our ancestry. Like them, our own ancestors once lacked a placenta. That does not mean, however, that echidnas or platypuses are our ancestors.

By using these methods, scientists can look to Indohyus and get some ideas about what the ancestors of whales were like, even if Indohyus lived after the oldest known whales. Indohyus is the closest relative to the group of mammals called cetaceans. Only after it branched off did cetaceans gradually become more like the whales around us today. I’ve put together a simple tree showing how those traits appear to have evolved in early whales, based on what scientists know about fossil whales and how they’re related. You can find it at the end of the post.

Of course, this is a scientific hypothesis that needs to be tested. And the way to test it is to find more species like Indohyus. If paleontologists are lucky, they’ll be able to draw more branches at the base of the whale tree. And if the current hypothesis is right, a lot of the species belonging to those deep lineages will be a lot like Indohyus. They may turn out to have lived before the oldest whales, or they may have lived millions of years later. But that’s not the heart of the matter. What matters is kinship.

In the annals of misleading science reporting, this may be pretty small potatoes. But mistaking relatives for ancestors does lead to confusion, and it gets in the way of appreciating some very elegant research. And, of course, some people pretend that the fact that relatives are not direct ancestors means that evolution is false. So it’s worth getting right–not just for whales, but for humans, flowers, or any other organism.

[Update Thursday–8:30 am: changed “kangaroos” to “echidnas.” Thanks to Alan for pointing out that kangaroos have (primitive) placentas.]

i-c0c26fab7b449be7ed0352332f9dc16a-whale tree 500.jpg


  1. #1 Homie Bear
    December 27, 2007

    Yeah, it’s good to clear that up. A surprising number of people, even university-educated evolutionists, believe that chimpanzees or even monkeys are our direct ancestors. And like you say, some either deliberately or otherwise amplify that misleading knowledge to distort the science.

  2. #2 Alan Kellogg
    December 27, 2007

    Last I heard, marsupials do produce a placenta. A basic placenta that doesn’t last very long, but which serves to support the embryo until it is developed enough to make the journey to pouch or skin fold.

    BTW, marsupials are also monotremes. They’re just not Monotremes. Like Monotremes (and some placentals) marsupials have a cloaca. Which makes them monotremes. Just as all meat eating mammals are carnivores, but not all meat eating mammals are Carnivores. 🙂

  3. #3 Carl Zimmer
    December 27, 2007

    Thanks Alan. It had slipped my mind that marsupials have placentas. Maybe echidnas would be a simpler example.

  4. #4 Steviepinhead
    December 27, 2007

    Thanks, Carl, for penetrating through the unnecessarily-confusing reporting on this discovery!

    If more people would think in terms of cousins (closer and more distant) when contemplating common ancestry, rather than focusing on ancestor-descendant relationships sensu strictu, a good deal of confusion and construction of strawmen might be avoided.

  5. #5 Mike from Ottawa
    December 27, 2007

    Carl: Have you considered doing an updated edition of At Water’s Edge?

  6. #6 Jeanette Garcia
    December 30, 2007

    Thank you for a very informative article. It’s irritating how otherwise intelligent people fail in the critical thinking department when it comes to knowing the difference between a common ancestor and a distant relative.

    The Waters Edge blew me away when I read it, especially the evolution of whales. The idea that a four legged land mammal would return to the sea so thoroughly that they could never, ever, again return to land boggles the mind. As I live near the ocean and have seen pods of whales swimming in the cold Pacific ocean I am filled with wonder when I think of their early beginnings, running along a warm, inland sea. I would love to read an update of this book.

  7. #7 chris y
    January 1, 2008

    Excellently lucid explanation. But a question occurred to me as I read it which you may be able to answer.

    Most of the coverage of Indohyus that I’ve seen seems to assume that the line leading to it became vegetarian, and that the primitive condition in the common ancestor of Indohyus and Pakicetus was a plant eater, as in your diagram. But what is the evidence, if any, that precludes a carnivorous ancestor with the Indohyid line developing vegetarian habits, as with pandas? Or is it simply that a dietry shift in one direction is far more usual than the other, and is assumed?

  8. #8 Carl Zimmer
    January 1, 2008

    Chris [7]: Good question. If you were to put the tree I showed here into the bigger tree of mammals, you’d see that most of the closest relatives of whales and Indohyus were plant-eaters (like hippos, camels, cows, etc.) So the simplest explanation would be that the common ancestor of hippos, camels, cows, whales, and Indohyus, etc. were plant eaters, and that there was a transition to meat eating in whales. The hypothesis that the reverse happened would require lots of transitions from meat eating to plant eating, with no evidence that such transitions actually occurred. But there’s a mystery–one lineage of these mammals, called mesonychids, had teeth for eating meat. They may or may not be close to the whale line. So did mesonychids and whales turn to meat independently, or did their ancestor make the switch? More fossils and more research are required.

  9. #9 Rick Gebethner
    January 4, 2008

    Your discussion of relations versus ancestors brings to mind a couple of items which I ran across but cannot properly attribute to sources. The first is was of a u-shaped atoll with a certain kind of spider which had diversified to the point where the spiders at the two ends of the atoll had become separate species. The two could not viably breed though any given spider population could breed with its neighbours. This and another similar study have been used to demonstrate the limitation of thinking in terms of distinct species where gradations exist. An interesting by-product is that genes could – through intermediaries – be passed on from one species of the spider to the other. In effect relatives could also be ancestors. The other item – that I also cannot properly cite – is the apparent passing of genes from chimpanzees to humans several millions of years after the lineages had split. Again, relations can become ancestors. And it is not out of the realm of the possible that the chimps got those genes from a species even further distinct from us. This is all wonderful stuff; but, lineages as branches on a tree is a concept that’s a little old and doesn’t do justice to the complexities of life.

  10. #10 Nick
    January 7, 2008

    I was searching for this kind of a blog for months now. Actually lost the hope of finding one, but here i am Thanks for the great articles! Looking forward for a little read after lunch.Ill Bookmark this blog.Thanks again.

  11. #11 Jason Goodbody
    January 8, 2008

    The probability of a fossil find actually representing a population that is an ancestor to a living species is absurdly small.

    In all likelyhood it is on a branch that shares a common ancester with said species. The closer it is to the split, the more excited scientists are that they’ve found a “link” along the chain. What they have found is a species that certainly left no currently living descendents.

    I think it is safe to say that any fossil find can only shed light on what a metaphorical cousin of our ancester looks like and almost never of the ancestor itself.

  12. #12 Jason Goodbody
    January 8, 2008

    I said “scientists” when I should have said “science writers”. Oops

  13. #13 Jim Thomerson
    February 2, 2008

    This whole discussion sounds like it took place in the 1950’s, before the cladistic revolution. Modern thinking about relationships (at least among the people I know) is cladistic. We do not accept that one can recognize an ancestor, thus all ancestors are hypothetical. We do not understand the concepts of transition species or missing links. Sister groups is where it is at! Paleontologists, I think, particularly paleoanthropologists, have been fairly slow to embrace cladistics. Well, I knew about Hennig in 1966, but did not publish my first cladistic analysis until 1997, so I should not be too critical of those who are taking their time to carefully consider the matter.

  14. #14 Joe G
    February 5, 2008

    Is there genetic any data which would demonstrate that the transformations required, if cetaceans evolved from land mammals, are even possible?

    IOW is there any way to test te premise (that cetaceans evolved from land mammals) without first assuming it?

New comments have been disabled.