The latest edition of the Journal of Vertebrate Paleontology contains not one, but two papers on the resolutely unglamorous topic of growth series. This was a sufficiently momentous occasion that I felt a blog post was warranted. No, really, don't stop reading! Because I want to try and convince you that constructing growth series is a critical part of systematic and evolutionary studies that the aspiring paleontologist neglects at their peril.
Although it may seem obvious, it's important to realize that many (if not most) organisms change their morphology dramatically over the course of their lifespan. This causes problems for systematic biologists, who want to compare character states in different types of organisms in order to reconstruct phylogeny. You have to compare like with like. In other words, it's no good comparing the juvenile of one species with an adult of another if the character you're studying changes with age.
Still think this is obvious? Well OK smarty-pants, let's take a look at a published research paper, namely Flannery, Archer, and Maynes (1987), which proposed a radical shake-up of the phylogeny of phalangerid marsupials. One of the critical morphological characters used by the authors was the extent of the exposure of the ectotympanic (the bony tube that forms the floor of the external ear canal in mammals) on the front face of the postglenoid process (a vertical bony "buffer" at the back of the jaw joint that means your lower jaw doesn't end up wrapped round your ears when your temporalis muscle contracts).
The problem is that the extent of the ectotympanic on the postglenoid process is age-related. In young phalangerids, the bone makes up more than half of the process, but as they age it gets rudely shoved aside by another bone, the squamosal; the squamosal grows more than the ectotympanic, which gradually contributes less and less to the postglenoid process, until in the end it is barely visible. This was a problem, because a number of the phalangerid species that were included in the study were very rare in museum collections and were represented by only juvenile or very aged individuals. As a results, some species in the study appeared to have a postglenoid process with an ectotympanic exposure, while others didn't. Based on these apparent differences, the authors tried to define a group of phalangerid possums that were united by the fact that they had no ectotympanic exposure on the postglenoid process. Unfortunately, all phalangerids eventually lack an ectotympanic exposure if they live long enough, so this royally mucked-up their phylogeny.
As a graduate student working on phalangerids, I spent a lot of time reading the Flannery et al paper, and it soon became apparent that I wasn't seeing the same characters in specimens that they were. In the end, the way that I teased apart the mystery of the postglenoid process (not very Hardy Boys, I know) was to build a growth series of phalangerid skulls. I was fortunate that I had a very large museum collection - the Mammal Department of London's Natural History Museum - just down the road from me, with hundreds of specimens covering nearly all the known species of the group. So I could select specimens of different ages from the same species, same sex, and even the same geographic population, thus minimizing the chance that other types of variation could creep in and mess things up for me (this, by the way, is one of the reasons we bother to collect so many specimens of the same animal).
In the end, I was able to untangle this problem because I had access to a larger collection of specimens than the Australian group. The challenge for paleontologists is that we're rarely this lucky. The chances of an animal being preserved as a fossil are almost vanishingly small and the temporal resolution of the fossil record is so low that even when two animals are found in the same site, they may come from populations separated by hundreds of thousands, if not millions of years. So it takes a rare set of circumstances to preserve a fossil population at the same level of detail that can be obtained from collecting Recent specimens.
The first of the two JVP papers describes just such a situation. Karen Black, Mike Archer, Suzanne Hand, and Henk Godthelp describe a collection of fossils from the Riversleigh area of Queensland . Over the last 30 or so years Riversleigh has yielded a host of spectacular fossils from the Oligo-Miocene of Australia; the ones described by Black et al belong to an extinct species of herbivorous marsupial called Nimbadon lavarackorum that lived around 15 million years ago. Nimbadon belongs to an extinct group called the diprotodontids, who were the heavyweights of the ancient Australian mammal fauna; some species were the size of rhinos, although Nimbadon was considerably smaller, around the size of a sheep.
What makes the Riversleigh find so important is the nature of the site where the fossils were found, AL90. 15 million years ago, this was a cave with a vertically-positioned entrance, down which from time to time unfortunate individuals of Nimbadon would blunder. In effect, the cave acted like a natural pitfall trap, sampling the local Nimbadon population. As a result of this, Black and her co-workers were able to construct a genuine growth series for this long-vanished animal, which demonstrated that the early growth patterns of the skull are mirror those seen in living marsupials, and which also gave insights into the development of the large air sinuses in the cranium that are a distinctive characteristic of diprotodontids.
The important thing about this paper is that the nature of the site means Black et al can be confident that they are actually sampling the same population. I'm not convinced that the same can be said of the second paper, by John Scannella and Jack Horner . This attracted a fair amount of attention in the national press, and quite a lot more attention here at Yale, by proposing that the ceratopsian dinosaur Torosaurus (very much an icon for the Peabody Museum) is actually the adult form of the much better known Triceratops. By studying a growth series of Triceratops, Scannella and Horner argue, first, that bone histology of skulls of Triceratops originally classified as "adult" suggests that the animals are not fully grown and, second, that the anatomical changes seen in the skull of Triceratops as it ages, if projected forward into an "adult" animal, would produce something that looks suspiciously like Torosaurus. There are no known juvenile specimens of Torosaurus.
Given that Torosaurus is such an iconic presence at the Peabody (a life-size bronze reconstruction of it towers over Whitney Avenue, outside the Museum) you might be expecting me to launch into a searing denunciation of Scanella and Horner's work. If so, I fear you will be disappointed. This sort of reexamination of taxonomic hypotheses is exactly what science is all about and using growth series is exactly how one should go about doing so. The slightly raised eyebrow is because I'm mildly skeptical that the growth series used by Scanella and Horner (which was originally proposed by Horner and Goodwin in 2006 ) actually represents a single population of Triceratops - it seems unlikely that specimens drawn from a variety of different localities, of different ages, could constitute a growth series of the consistency seen in Black et al's study. Whether this makes any difference or not remains to be seen.
Both these papers are important because they once again remind us, should any reminder be needed, that fossils were once living animals, and the morphology that today is literally "set in stone" was once plastic and changeable. As we try and piece together evolutionary patterns, the importance of the unglamorous growth series should not be underestimated.
[Accompanying figure is taken from Black et al, 2010. For those of you that want references, here they are:
 Flannery, T.F., M. Archer, and G. Maynes. 1987. The phylogenetic relationships of living phalangerids (Phalangeroidea: Marsupialia) with a suggested new taxonomy. pp477-506 in Archer M. (ed) Possums and Opposums: Studies in Evolution. Sydney, Surrey Beatty & Sons and the Royal Zoological Society of New South Wales.
 Black, K.H., M. Archer, S.J. Hand, and H. Godthelp. 2010. First comprehensive analysis of cranial ontogeny in a fossil marsupial—from a 15-million-year-old cave deposit in Northern Australia. Journal of Vertebrate Paleontology 30(4):993-1011.
 Scanella, J.B., and J.R. Horner. 2010. Torosaurus Marsh, 1891, is Triceratops Marsh, 1889 (Ceratopsidae: Chasmosaurinae): synonymy through ontogeny. Journal of Vertebrate Paleontology 30(4):1157-1168.
 Horner, J. R., and M. B. Goodwin. 2006. Major cranial changes during Triceratops ontogeny. Proceedings of the Royal Society of London B 273:2757–2761.]