Yesterday I blogged about the ~27 million year old elephantimorph Eritreum, a creature that stood only about four feet high at the shoulder, but there were once even smaller proboscideans. About sixty million years ago in what is now Morocco there lived a rabbit-sized (~5 kg) hoofed mammal that is one of the earliest known relatives of the modern behemoths of Africa and Asia. Called Eritherium azzouzorum, it was a small mammal that could have a major influence on our understanding of elephant evolution.
Up until now the oldest known proboscidean (that's the group containing modern elephants and all their extinct relatives) was Phosphatherium, a ~56 million year old creature found in the Eocene sediments of Morocco. Eritherium, from the Paleocene of the same country, appears to be even older. It's exact age is a little fuzzy at the moment, but if it is really about 60 million years old then it documents part of the major radiation of placental mammals that occurred after the extinction of the non-avian dinosaurs.
The remains of Eritherium presently known are fragmentary. There are about 15 bits of skull that have been recovered as well as some teeth and a portion of the maxillae (upper jaws) preserving premolars 3-4 and molars 1-3. This might not seem like much, but as paleontologists are well aware the teeth of a mammal can tell you a lot. As such the teeth of Eritherium identify it as one of the earliest proboscideans. It is similar to Phosphatherium, but it also retaining features seen in more generalized hoofed mammals (often referred to as "condylarths")* like the retention of a canine tooth, three incisors, and a first premolar in juveniles (but lost in adults).
*[There is some debate as to whether the condylarths are a natural evolutionary group. Since the details of this issue are still being ironed out, however, I will refer to the ancestral stock from which the earliest proboscideans probably sprang as "condylarths" in order to keep things simple.]
Without more of the skeleton it is somewhat difficult to tell what Eritherium might have looked like (did it have a tiny trunk?), but it nevertheless has important implications for the debate over the timing of the radiation of placental mammals. According to the headline-making mammal "supertree" published in 2007, for example, the first proboscideans appeared about 100 million years ago, well before the extinction of the non-avian dinosaurs. The discovery of Eritherium refutes this hypothesis. Indeed, the hypothesis derived from the genetic data appears to have put the divergence of proboscideans about 40 million years too early!
It is unlikely that Eritherium was the last common ancestor of all proboscideans, but it does represent one of the first distinguishable members of the group and shows a lot of similarities to earlier hoofed mammals. This means that the earliest proboscideans appear to have diverged from other mammals just before or soon after the end-Cretaceous mass extinction and not tens of millions of years further back into the Cretaceous. This is not to say that evolutionary trees based on molecular or genetic data are not trustworthy, but rather that they must always be tested by what is found in the fossil record. The converse is also true, especially since molecular and genetic data have been key to discoveries like our shared common ancestry with chimpanzees somewhere between 8-5 million years ago and the identification of whales as highly specialized artiodactyls. (In fact, any evolutionary tree you see is a hypothesis that has to be tested by further research and discovery.)
Combined with other early proboscideans like Phosphatherium and Numidotherium, Eritherium gives paleontologists a better look at the early evolution of a group represented only by Asian and African elephants today. Eritherium is not as imposing as some of its later relatives, like the mastodons and mammoths, but it has helped researchers better define the timing and placement of when the earliest distinguishable ancestors of elephants appeared. Hopefully future discoveries in northern Africa will continue to fill in our understanding of early proboscidean evolution, and I will be interested to see how this find influences the persistent "fossils vs. genes" debate.
Gheerbrant, E. (2009). Paleocene emergence of elephant relatives and the rapid radiation of African ungulates Proceedings of the National Academy of Sciences, 106 (26), 10717-10721 DOI: 10.1073/pnas.0900251106
Brian, better put your helmet and flak vest on - you have made enemies ;-) http://microecos.wordpress.com/2008/02/07/afrotheres-of-the-world-unite/
Damn, that's right I still owe Christopher Taylor a t-shirt.
Just to play geneticist's advocate for a moment (weird...), I'm not sure that Eritherium necessarily refutes the Cretaceous divergence of proboscideans.
The group *could* have become genetically differentiated while retaining a plesiomorphic appearance for a considerable amount of time. 40 million years might seem like a stretch, but just look at Sphenodon which appears morphologically rather similar to its Mesozoic relatives whereas genetic evidence suggests that it has undergone very rapid and dramatic recent evolution on the gene level--purportedly faster than any other vertebrate studied.
Alternatively, one might make the argument the the early elephant fossil record is poor and the new taxon merely represents a late-surviving member of the plesiomorphic stem lineage. I don't personally find such arguments very compelling, but it's hard to disprove gaps in the fossil record. To fall back on cliche (again): the absence of evidence of Cretaceous proboscideans isn't really evidence of their absence.
All that being said, I think it's clear that something of tremendous evolutionary significance was occurring in the Paleocene/Eocene--even if the "fuse" had been lit well before. At any rate (zing), I am in absolute agreement with you that phylogenetic hypotheses based on genetic data must be continuously tested against fossil evidence and vice versa. I think the tuatara example is a good demonstration that neither morphology nor genes alone can provide a complete picture of evolution.
Neil; Thanks for the thoughtful reply. Maybe, like you said, there is a longer fuse, but then Eritherium would be one hell of a persistent type. More fossils are needed, and hypotheses will no doubt change, but at present it seems more likely to me that the divergence of proboscideans was closer to the K/T boundary than the "supertree" study inferred based upon genetic analysis of living elephants. Either way we need more data.