Darwinius is an exceptionally well preserved, 47-million-year-old primate from the ancient Messel Pit in Germany whose position on the primate tree of life has been hotly contested. Now, a new study uses artificial extinction experiments to show that there is sufficient data to place the fossil accurately on the primate tree of life, and places it squarely with the lemurs and lorises.

In an unusual media blitz in 2009, Darwinius was presented as a key transitional fossil in human ancestry, but later analyses contested the claim that it was the elusive “missing link” to humans. “Darwinius is important first and foremost because it lived 47 million years ago,” says Phil Gingerich, a paleontologist at the University of Michigan who co-authored the original research article about Darwinius in 2009. “Among primates that old, Darwinius is important because it is so complete and thus preserves so much of its anatomical/skeletal morphology. And some of this morphology is primitive because Darwinius is so old.” By “primitive,” Gingerich means that Darwinius represents the ancestral state—an earlier state in the human lineage. In Gingerich’s view, subsequent evolutionary changes (“derived” traits) led to modern humans. But the study of human origins remains full of unanswered questions.

In its original description, the specimen, named Ida, was classified as a haplorhine, the same primate order as humans, other apes, and monkeys. If that placement is correct, then features of Darwinius are likely representative of the ancestral features of apes (including humans). However, larger datasets later suggested that Darwinius was actually a strepsirrhine—the primate group that contains lemurs, lorises, and galagos. If Darwinius does indeed, belong to the strepsirrhine primate group, it would not be a direct ancestor of humans, but instead the ancestor of the very distinct evolutionary branch of tooth-combed prosimians, today most diverse on the island of Madagascar.

Disagreement persists about the placement of this fossil on the primate tree of life because it is not clear whether the features that make Darwinius similar to haplorhines are indicators of shared ancestry, or whether the similarities were independently evolved in more than one branch of the primate tree of life.

Evolutionary relationships between living species are fairly straightforward to determine using the abundant information from the species’ genome (e.g., totality of DNA) and anatomical features (e.g., the shape of the animal’s teeth or the placement of holes carrying blood vessels in the bones). For extinct species, known only through fossil remains, the information available to reconstruct evolutionary relationships is more limited: DNA is often unavailable, and only a few anatomical features survive years of decay and fossilization. However, the vast majority (99.9 percent) of species that have ever lived on Earth are extinct, and the tree of living species alone is not close to representing the full diversity of life.

Fossils such as the preserved remains of animals like Darwinius are important because the origin of every major group is the result of the evolution of one or more novel features (e.g., the evolution of feathers at the origin of the lineage of birds). Although the time at which novel features evolved can sometimes be estimated using molecular data, fossils provide the only unequivocal evidence of the time at which these novel traits first evolved. “Anatomical characters are the ones that have a fossil record, and hence the ones that can be studied through time,” explains Phil Gingerich. “DNA phylogenies are all constructed in hindsight looking back from the present. Some people don’t know any better and think that the past is something we can’t really know so have to make up from the present. […] I want to know what happened in the past, and be able to know the times changes happened as well as what happened.”

In general, paleontologists argue that anatomical features discernible in fossils—the preserved remains of an animal—are not only reliable, but also provide information that is not present in DNA. But others aren’t so sure about the value of anatomical features. A 2007 study by Mark Springer and colleagues, argued that there is insufficient information in anatomical features to accurately reconstruct evolutionary relationships. Anatomical features are typically not as abundant as genomic (DNA) data, and can be more difficult to quantify. In addition, many anatomical traits are correlated with each other, which is problematic for deriving evolutionary relationships. Springer found that there was substantial disagreement between his anatomical and genomic data, and that this disagreement was not just a sample size issue. Instead, he contended that anatomical studies fail to separate homology (e.g., similarities which indicate shared ancestry) from homoplasy (e.g, similarities which were evolved independently by two or more species), thereby severely limiting their usefulness for deducing evolutionary relationships.

This controversy needed to be resolved in a scientific way. Given that an evolutionary re-run in the lab would not be possible, Robert Asher and his colleagues at the University Museum of Zoology in Cambridge, United Kingdom, turned to the next best thing: simulated extinctions using a computer model. By assessing evolutionary relationships from the simulated extinction experiments, they were able to provide statistical evidence that anatomical data in fossils, at least for primates, is sufficient to accurately place fossil species on the tree of life.

Typically, paleontologists use anatomical traits to establish evolutionary relationships between species. Teeth are particularly well preserved in the fossil record, so paleontologists often rely on features of the teeth, including number and shape, to infer evolutionary relationships. Many also rely on DNA sequences from living groups, as these are generally missing for fossils. The researchers in Asher’s group began with a comprehensive dataset of both anatomical and DNA features for the tree of living primates, which is well established. Then, to test whether the features available only in fossils contain information about evolutionary relationships, they eliminated all the features that are not available in fossils (e.g., information about DNA sequence and soft tissue anatomy). In this way, the living species were modeled in the computer as if they were fossils. Using this reduced dataset, the scientists then recalculated the primate tree of life. They did this thousands of times, sometimes focusing on fossils with only teeth, which are typically well preserved in the fossil record, and other times on fossils with features from the teeth, skull, and other bones from the rest of the body. Each time, they compared the calculated tree to the tree known to be correct.

If the reduced datasets correctly reconstructed the known tree, then the researchers would have evidence that there is sufficient information in the anatomical traits present in fossil species to infer evolutionary relationships accurately. Alternatively, if the reduced datasets produced an incorrect tree, the researchers would conclude that, alone, the traits preserved in fossils were insufficient to infer the evolutionary relationships.

They found that the reduced set of traits was largely sufficient to infer the correct evolutionary relationships. Accuracy was particularly high when information came from different parts of the body: for example, the teeth and the bones of the limbs. When available traits were concentrated in just one body part (e.g., the teeth), resulting trees were less accurate.

“Methodologically, the study is quite clever and innovative,” says Christopher Gilbert, an evolutionary anthropologist at Hunter College, City University of New York, who was not involved in the study.

Asher and his colleagues recognize that the “evolutionary signal” in fossils is not as good as that present in a living species, for which scientists have much more complete knowledge of development, soft tissues, and the totality of DNA information (e.g., the genome). Nonetheless, their artificial extinction experiments demonstrate that, at least for primates and Darwinius, anatomical features present in fossils contain adequate information for inferring evolutionary relationships.

“This is an excellent study that adds to a growing body of research suggesting paleontological data, even though often incomplete, are a crucial source of information likely to provide accurate assessments of phylogeny,” added Gilbert. “In the context of Darwinius, the study suggests that comprehensive analyses employing many fossil taxa and many characters from various anatomical regions are generally likely to be accurate. Therefore, the emerging consensus from this study and from paleoprimatologists who employ such modern methods, namely that Darwinius is a fossil strepsirrhine, is strongly supported. It really illustrates that there is very little justification for those who argue against using such methods, particularly if the goal is overall phylogenetic accuracy.”




JSTOR is a digital library for scholars, researchers, and students. JSTOR Daily readers can access the original research behind our articles for free on JSTOR.

Science News, Vol. 175, No. 13  (Jun. 20, 2009), p. 8
Society for Science & the Public
Proceedings of the National Academy of Sciences of the United States of America, Vol. 107, No. 21 (May 25, 2010), pp. 9712-9717
National Academy of Sciences
Proceedings of the National Academy of Sciences of the United States of America, Vol. 107, No. 11 (March 16, 2010), pp. 4797-4804
National Academy of Sciences
Plant Physiology, Vol. 151, No. 3  (Nov., 2009), pp. 970-977
American Society of Plant Biologists (ASPB)
Notes and Records of the Royal Society of London, Vol. 65, No. 1  (20 March 2011), pp. 83-9
The Royal Society
American Scientist, Vol. 97, No. 5  (September-October 2009), pp. 410-413
Sigma Xi, The Scientific Research Society