A new phylogeny of the mammals

October 31, 2011 • 4:42 am

A paper in this week’s Science uses a lot of data to construct the most complete phylogeny yet of mammalian families.  Meredith et al. used 26 genes to not only construct the tree, but estimate divergence times.  Their sample comprises 97%-99% of the roughly 150 described mammalian families. Here’s the tree they get (click to enlarge; lots of detail can be seen by zooming as well):

All the nodes except for the ones denoted by solid blue circles are strongly supported.  The background shows the transition from light gray (Mesozoic) to white (Cenozoic).

A few highlights for me (and I’m not a mammalogist):

  • As Christofer M. Helgen points out in a Perspectives piece in the same issue, families prove to be good monophyletic groups: clades like bears, cats, and dogs all indeed fall into families determined earlier on morphological grounds.
  • No surprises for major groupings: monotremes (platypus and echidna) are the most distant relatives of living mammals, equally distantly related to the marsupials (as expected, a monophyletic clade) and the placentals.
  • The red panda (Ailuridae) is more closely related to procyonids and mustelids (e.g., skunks and raccoons) than to the giant panda, which is in the family Ursidae (bears; the panda’s status as a bear has been known for forty years).  This is probably not a surprise to anyone who knows mammals, but I haven’t kept up with mammalian systematics.
  • There are tons of bat families, and they’re well diverged from everything else.  Their closest relatives appear to be the families that include whales, rhinos, hippos, giraffes, cats, and pangolins, and not the rodents, which are far more distantly related.
  • All marine mammals are a monophyletic group, meaning that they shared a single common ancestor, and the closest relative of those whales, dolphins, and porpoises are the hippos.
  • The closest relatives of monkeys and apes (and us) comprise the other primates: the lemurs. No surprise there, but the next most closely related group comprises the Dermoptera (the “flying lemurs” which, as noted below, ain’t lemurs) and, after that, the rodents.
  • Sloths, armadillos and anteaters group together in a clade that is distantly related to most other mammals, but a bit more closely related to aardvarks, tenrecs, dugongs, and elephants.  And all of these groups diverged from other mammalian families a long time ago—on the order of 100 million years.
  • Finally, for connoisseurs of paleobiology, the the origin of new orders (a taxonomic level higher than families; orders comprise what most of us think of as “types” of mammals: carnivores, bats, primates, rodents, and so on) occurred after the “KPg” (Cretaceous-Paleogene; formerly called the Cretaceous-Tertiary) extinction event about 65 million years ago.  This extinction wiped out the dinosaurs, of course, but also many marine invertebrates and terrestrial plants.  The authors posit that this extinction event played an important role “in the early diversification and adaptive radiation of mammals,” perhaps by opening up  the “ecospace available for mammals.”  That’s not a new theory: we’ve all heard, for example, that the extinction of the dinosaurs allowed mammals to diversify, but it’s good to see the dates of diversification confirmed in this way.


Meredith, R. W. et al. 2011.  Impacts of the Cretaceious terrestrial revolution and KPg extinction on mammal diversification.  Science 334:521-524

96 thoughts on “A new phylogeny of the mammals

  1. What hit me straight away was the bats. There are so many species of bat – something like (correct me please) 20% of all mammals?!

    1. The same reasons there are so many songbirds. They’re small so the environment can generally support lots of them, they can fly so they can disperse easily, and they use vocal communication (changes in which can act as an impetus for speciation).

      1. Dispersal ability is a tricky thing. If creatures disperse really well, speciation will be difficult. I bet there is some intermediate dispersal rate that optimally promotes speciation in a heterogeneous, changing environment.

          1. It depends on which bird and bat taxonomists you ask. For most bird and bat taxonomists since 1920, the answer would be “One species of human”. (Ernst Mayr had a famous paper taking paleoanthropologists to task for oversplitting, and urging the number of hominid taxa be reduced. He was referring to fossil taxa, and more than one species of extant Homo was never a question.)

            Since about 1980, there’s a subset of avian taxonomists (less so in bats) that would, if applying their criteria consistently, probably recognize a handful of extant species of Homo. These taxonomists might even be a majority at the moment. However, application of this looser species concept has led to an increase of maybe 10% in the number of species. This is noticeable, but does not change the overall picture of bird (or bat) species richness.

            1. Thanks Gregory,

              When you’re not familiar with the field it’s always difficult to tell whether species richness is due to some real phenomenon or oversplitting.

    1. At long last, does this mean that we’re allowed to eat bird bats?

      You are to detest these birds. They must not be eaten because they are detestable: the eagle, the vulture, … the osprey, the stork, the various kinds of heron, the hoopoe, and the bat. —Leviticus 11:13–19

  2. I’m looking on a small screen

    In the tree, the family Felidae is closest to the family Prionodontidae

    I can’t find the latter family described anywhere

    Have I spelled it wrong?
    What’s a living example?

        1. So instead of “the dentally defective rat” they would have the “environmentally challenged bear”. That sounds like a sinking proposal. :-/

        2. Not this Canadian – it’s a conservative government plot along with renaming our military branches, glorifying the war of 1812, enacting “tough on crime” policies, anti-immigration laws, and anti-science (including social science) policies that is designed to create an association between militarism, “traditional values”, toughness and patriotism/nationalism as opposed to the more established Canadian values of prudence, industriousness, social justice, and cooperation. Basically, they read the research that suggests that Americans are more likely to favour Republicans after viewing the American flag and they thought “hey, maybe we could make that happen here”.

      1. If indeed, when criticizing a particularly spineless decision by Canadian politicians, Margaret Atwood was correct in claiming that the male beaver will bite off his own testicles when dangerously cornered, then maybe that change is appropriate.

        1. He will also sing the blues if he happen to be turned towards the full moon.

          What is this, Fantasy corner!? Luckily, I have that fantastic phylogram to tide me.

          1. Yes, I think what Torbjorn is saying, but no one else did, is that self-castration would pretty clearly be in logical contradiction to standard evolutionary theory; no science really needed, just pure logic. But Margaret told an amusing metaphor, which even got into the House of Parliament in terms of a few jokes about ‘free trade’ with the behemoth ‘below’.

  3. I had sort of kept up with the whales/ungulates relationship, and with the monophyly of those African groups at the outer reaches of this cladogram, but I find the bats as sister group to whales and ungulates the most surprising outcome of this study – I always thought of them as flying shrews.

    1. In some ways this is Linnean taxonomy leading our expectations astray. The early representatives and early-diverging lineages of many placental orders are “rodent-like” in that they are small little creepy guys. They aren’t rodents or shrews, cladistically, but if you looked at those ancestors today without an expert eye you might say the ancestor was more similar to one of those groups.

  4. I had heard (a decade ago) that bats were not monophyletic; that the fruit bats were only distantly related to the other bats. I guess that was a mistake (either a mistake by those earlier taxonomists or a mistake of my memory)?

    1. No, you remember correctly. That hypothesis was based on details of optic nerve anatomy (iirc) and has simply been supplanted by more and better data, primarily molecular. Science marching on.

    2. You remember correctly.

      It was never widely agreed upon, but there was a push back in the late 80’s/early 90’s towards a division of the bats. Just a couple of guys publishing a few rather quickly forgotten papers. Fruit bats were supposed to be closer to primates (visual canopy dwellers with a taste for fruit).

  5. Clarification :
    “All marine mammals are a monophyletic group,”
    All the whales/ dolphins/ porpoises etc, but NOT the seals/ walruses. They (pinnipeds etc) form a monophyletic group with the bears (Ursidae). Big surprise there!

    1. It’s at times like this that I miss having a subscription to Science. Or Nature.
      Time to get back into part-time education, and a library subscription with the course fees.

      1. Jerry: No you didn’t. You meant “cetaceans”. Sirens (dugongs and manatees) are “fully marine mammals” and they come out as sister group to elephants.

      2. It did cause me a bit of a “huh?” moment having read that statement about marine mammals and then seeing dugongs listed in a different grouping two bullet points below.

    2. “Marine mammal” is a legal term that has no consistent meaning in the scientific community. The “legal definition” of marine mammal in the U.S. (ala The Marine Mammal Protection Act” includes all critters that get their feet or fins wet in salt water (all Cetacea, all Sirenia, and various subsets of Carnivora – i. e., parts of Mustelidae, a single species of Ursidae, and all of superfamily Pinnipedea).

  6. Are all those in blue that are not primates Rodents? Starting at Tupaiidae to Erethizonitadae (hard to check spelling on my screen). 37 families?

    I wish order names were shown on this chart.

    1. Not quite. The first 2 families are treeshrews, then some lagomorphs. Rodents run from Castoridae to Erethizontidae (beavers to porcupines).

    2. The key to the main diagram states “Color-coded branches in Placentalia correspond to Laurasiatheria (green),
      Euarchontoglires (blue), Xenarthra (orange), and Afrotheria (pink).”

      “Laurasiatheria” and “Afrotheria” I recall from a study around 10 years ago attempting a phylogenetic re-grouping of the various mammals into medium-order taxons. “M. S. Springer et al., Nature 388, 61 (1997).” is probably the reference, because I don’t have the paper on my big hard drive, and that’s about the time that I started to store a selection of the stuff I had access to at the time (since then I’ve learned to be less selective, and this sort of situation is exactly why!)

      http://en.wikipedia.org/wiki/Eutheria#Subgroups gives a high level overview of the groups’ compositions.

    3. On first scan the most surprising is the Scandentia (tree shrews, Tupaiidae and Ptilocercidae) as the sistergroup of all the other Eurarchontoglires, rather than as the sistergroup to flying lemurs and primates. Another point of contention over the years has gone full circle, it seems: the perissodactyla get the artiodactyla as sistergroup, rather than the carnivores. Bats, even-toed, uneventoed, and carnivores do not split easily or convincingly.

        1. Apart from the carnivora outside of all ungulates plus bats. That is not a poorly supported node, and not standard.

          The article is deficient in discussing the poorly supported nodes, and what the statistics for alternative trees are there.

          There seems to be a visual trick going on: some of the blue circles on bluish background seem grey.

  7. No surprise there, but the next most closely related group comprises the rodents.

    That’s a surprise? Apparently, you haven’t met my brother-in-law.

  8. Great phylogeny. I love this shit.
    Winners of the least-closely-rlated-to-anybody-else lineage-survival sweepstakes (eutherian division):
    the aardvark

      1. It’s a single-species Order.
        All that’s left of its 75-myo lineage.
        Everybody else died, I guess.

  9. Question about orders and families and whatnot: Aren’t taxonomic “levels” arbitrary? What’s the difference between one monophyletic group and any other?

    1. Yep, “levels” are arbitrary. They shouldn’t even be called “levels”, since one “level” doesn’t even include comparable things — each order doesn’t have the same age, diversity, or anything else.

      Difference between one monophyletic group and others? Objectively, nothing but age and diversity, really. Subjectively, we humans find certain clades more interesting than others, and give them names, but there are so many clades once you have a complete phylogeny, there’s not much point in naming/learning all the clades.

  10. I haven’t kept up with Fossa phylogeny, but last I heard the Malagasy carnivores (…or, rather, Carnivores) were partly with Viverrids and partly with Herpestids, but it looks like this new phylogeny is sticking them all with the Herps. I’m unsure what this means as to the question of how many colonization events there were, but it would seem to point to just one. Pretty interesting.

    1. Since 2003 ,
      A.D. Yoder, M.M. Burns, S. Zehr, T. Delefosse, G. Veron, S.M. Goodman and J.J. Flynn, 2003. Single origin of Malagasy Carnivora from an African ancestor.
      Nature 421: 734-737.

  11. Interesting study, but there are still some weird bits. The non-cetacean (i.e. non-whale) marine mammals are not, as was well known, related to cetaceans. The seals, sea lions, and walruses form a monophyletic group. This is not unexpected, but there had been a long debate as to whether seals and sea lions were two separate origins of the marine habit from carnivores. (Seals and sea lions swim in very different ways– seals with their hind legs, sea lions with their front legs.) Manatees and dugongs are also a separate return to the sea for mammals (as well as, presumably, sea otters).

    The closest relatives of primates are flying lemurs– a not unexpected but nonetheless exciting result. (Flying lemurs are not lemurs.)

    The sister group to rodents+rabbits are the tree shrews– this I find surprising. But the node is not well supported, and the next node over is the primates+flying lemurs, and tree shrews have long been associated with primates, so this would be interesting.

    The clearest gaffe is that canids (dogs and foxes) are actually nested within the bats– this can’t possibly be true!

    (I posted this on the picture-only page by mistake a few minutes ago. There are some other orphaned comments there as well.)

    1. The clearest gaffe is that canids (dogs and foxes) are actually nested within the bats– this can’t possibly be true!

      No? What about flying foxes, then, huh? 😉


    2. I’m not sure what you mean about the canids being nested in the bats.

      In the vertical layout, the canids are at the bottom of a big clade, and while the first bat (myzopodidae) is immediately below that, it’s at the top of a different big clade. (The one that includes the other pictured bat, and apparently several things below that that I assume are all bats, down to rhinopomatidae.)

    3. The last one I don’t see from the diagram. Canidae branches off the Felidae et cetera clade (if that is the term), all the bats unless I missed some is in the next nesting over.

      Maybe you could make a case that horses, whales et cetera is “in” the bat nesting instead (they branch from a common ancestor with bats)?

    4. The seals, sea lions, and walruses form a monophyletic group. This is not unexpected, but there had been a long debate as to whether seals and sea lions were two separate origins of the marine habit from carnivores.

      This could, of course, still be true. Phocids and otariids + walrus are each others’ closest living relatives, but it’s possible that their marine habits were independendently derived from different terrestrial relatives that are all extinct.
      Seems unlikely, but possible. I don’t know if the pre-seal fossil record is any good.

    5. Canidae are not within the bats, but the lay-out puts then one line up. They are in the agreed position.
      The bears are the sistergroup of the sea-carnivores (Phocidae otariidae Odobenidae), istead of the sistergroup of sea-carnivores + weasel + red panda+ skunk + washbears. That is against previous findings.

    1. Actually, probably not. Tufte takes a rather dim view of phylogenetic trees. See his discussion of them in Beautiful Evidence. Before he published this, several discussants on his discussion board (including Joe Felsenstein and myself) tried to convince him otherwise, but unsuccessfully.

      1. How does he propose to visually represent biological relationships then? I think an even worse alternative than a hard to understand graphic is a complete absence of a graphic.

        1. I’d have to check the final version of Beautiful Evidence to see what he finally said, but basically he thought phylogenetic studies were data-thin, self-referential, and not at all rigorous. He likened phylogenetics to post-modernism, and found trees wanting by the standards of Feynman diagrams. He probably wouldn’t think phylogenetic relationships are worth representing at all.

  12. Thanks for the breakdown and illustration for those of us who don’t have ready access to Science. There’s something beautiful about connected species shown alongside one another.

Leave a Reply