The mummy returns… and sheds new light on old questions


M. T. P. Gilbert, E-mail:


Whether as the ancient Egyptian crocodile-god Sobek, a terrifying predator of African waterways, or simply as a premium handbag leather, the Nile crocodile (Crocodylus niloticus) has long held the fascination of mankind. Despite 200 years of study, however, uncertainty remains as to its taxonomy. While resolving such issues are key to understanding the origins and biogeography of the so-called true crocodiles of genus Crocodylus, given widespread ongoing range contraction, such issues are paramount for design of future conservation strategies. In this issue of Molecular Ecology, Hekkala et al. (2011) apply analysis of modern, historic and ancient DNA (aDNA) to the questions, with far-reaching implications. First they demonstrate that, as currently described, the Nile crocodile is paraphyletic, with individuals from the east and western clades separated by a number of New World crocodile species. The consequences of this finding are as important for conservation efforts as for their impact on crocodile taxonomy. Furthermore, they strike at the heart of the long-standing debate over whether aDNA analysis of ancient Egyptian mummies is scientifically sound.

The taxonomic history of the Nile crocodile has had its ups and down. Following initial description as ‘Crocodylus niloticus’ by the Austrian naturalist Joseph Nicolai Laurenti in his influential 1768 treatise on the poisonous function of reptiles and amphibians, subsequent study of geographically dispersed and morphologically diverse samples rapidly expanded the number of species and subspecies throughout the 18th and 19th century. However, as the potential for geographic and morphological plasticity within single species became clear during the 20th century, the accepted number of species collapsed. As a consequence, it is today managed for conservation purposes as a single entity (Ross 1998).

While convenient, whether this approach is sensible is a different question. Thanks to decreases in genetic analysis costs, expanded sampling efforts and development of increasingly sophisticated analytical tools, the concept that many large, geographically distributed animal species can be treated as a single panmitic population for management purposes is rapidly being brought into question. For example, the killer whale (Orcinus orca) was for many years treated as a single species and conservation unit (Rice 1998). Recent studies, however (e.g. Morin et al. 2010), have demonstrated that it consists of genetically and ecologically well-differentiated groups.

Inspired by previous smaller studies that (i) hinted that genetic differentiation exists between Nile crocodile populations, that (ii) suggest the species may even be paraphyletic and (iii) concerned in general by its range collapse, in particular, in the western part of the continent, Hekkala et al. (2011) undertook a three-pronged study to clarify matters. Using a conventional approach as a base, they first sequenced over 5000 bp of DNA (mitochondrial and nuclear, from nine loci) from individuals sampled across wild African and Malagasay populations, as well as several related New World species. Then, calibrating their data with the fossil record for the order Crocodilia, they subsequently estimated divergence times for key points in the genus’ history. The results confirm without doubt that the Nile crocodile is subdivided into two distinct clades, one predominantly found in the west of the continent and the second in the east (Fig. 1) that likely diverged from each other ca. 8.13 MYA. Additionally, the Nile crocodile is paraphyletic, with four New world species monophyletic with the eastern clade, the Cuban (Crocodylus rhombifer), Morelet’s/Mexican (Crocodylus moreletii), the American (Crocodylus acutus) and the Orinoco (Crocodylus intermedius) crocodiles (Fig. 1). Furthermore, Nile crocodile inter- and intra-clade genetic diversity is consistent with levels observed in other accepted species.

Figure 1.

 The evolutionary history of genus Crocodylus showing species discussed in this piece and their karyotypes. Both DNA sequencing and karyotype analysis indicates that the single species traditionally known as the Nile crocodile, C. niloticus is paraphyletic, with a predominantly Eastern African clade monophyletic with the New World Crocodylus species (represented here by C. rhombifer). The identity of the Western African clade is shown to match the type specimen for C. suchus (Fig. 1 inset), a description previously believed to be synonymous with C. niloticus, but that now warrants full species recognition. Photo credits: Kent Vliet (C.n., C.p., C.r., C.siam.), Matthew Shirley (C.such.), John Thorbjarnarson (M.c.) and Evon Hekkala (C.s. type).

Although two genetically distinct clades, given limited resources available for conservation efforts, that alone may be insufficient grounds for treating them as independent units. Arguably, even if different, should the 2 clades be reproductively compatible and play an ecologically similar role, then it may be economically reasonable to treat them as a single unit. In this regard, Hekkala et al. (2011) second approach is interesting; through karyotyping, they demonstrate that chromosome numbers vary between clades, with 2n = 32 in the eastern and 2n = 34 in the western clade (Fig. 1). How significant this is for reproductive isolation (thus conservation implications) is currently unclear. While small differences in chromosome number in closely related species often preclude generation of fertile hybrids—a classic example being donkey and horse, with 2n = 31 and 32 chromosomes, respectively—this is not always the case. Perhaps pertinent to the Nile crocodile is the fact that other members of genus Crocodylia, in particular, the saltwater (C. porosus) and Siamese crocodile (C. siamensis), have been observed to successfully hybridize in captivity (Chawananikun et al. 1999), despite chromosome differences (2n = 34 and 30, respectively, Fig. 1, although there is some disagreement about the karyotype of C. siamensis [c.f. Kawagoshi et al. 2008] with suggestions that C. siamensis itself may exist as two cryptic lineages, with 2n = 30 and 34 karyotypes respectively).

Regardless, it is clear from the modern genetic data that current taxonomy of the Nile crocodile is unsatisfactory. At this point, Hekkala et al. (2011) took a further step that was bold, if not controversial. Given their clear evidence that Nile crocodiles are likely distinct, predominantly geographically separated, species, and inspired by the intriguing written record presented by Herodotus in book II (Euterpe) of his ‘Histories’, that ancient Egyptian priests were aware of two forms of crocodiles in the Nile (Herodotus, in Hekkala et al. 2011), the authors wished to investigate whether there was evidence of past cohabitation of the clades. To do this, they took an approach many would view as dangerous as that of tourist naïvely bathing in the murky waters of a crocodile infested lake—they attempted to add a historical dimension to their study through analysis of mitochondrial DNA extracted from historical museum collections (dating to within the last 150 years), and more controversially, ancient mummified crocodiles estimated to date as far back as 3000 YBP.

While genetic analyses of historical and ancient material are becoming a routine tool, the question of whether authentic aDNA can be recovered from mummified Egyptian remains is a sure-fire way to stir up heated passions within the aDNA community, often to a degree that mystifies external observers (c.f. Marchant 2011). But why should this be? Egyptian mummies are famous for their morphological preservation and abundance, thus seem ideal for genetic study. Furthermore, the study of mummified Egyptian aDNA, in particular, from humans, has had a long, colourful history, attracting much media interest. One of the first aDNA analyses was Svante Pääbo’s 1985 report of successfully cloning a 3.4 kb piece of DNA from a 2400-year-old mummified child (Pääbo 1985). Subsequent studies used similar material to investigate diverse questions including the relatedness of Egyptian royalty (e.g. Hawass et al. 2010) to what pathogenic diseases were present in antiquity (e.g. Nerlich et al. 1997). Thus, given the number of publications that apparently successfully utilize such material, it seems ideal for application to any question that will benefit from a historical perspective. And yet despite this, a small number of authors (this author included) have not always been willing to accept the results of such studies, either at best advocating caution with regard to interpretation of the findings, or at worst flatly disbelieving that the results are real.

Two problems underlie the dispute. The first is the well-documented relationship between rate of DNA degradation and temperature. As described best by the Arrhenius equation, as temperature goes up, the rate of DNA degradation increases exponentially (c.f. Lindahl 1993). While the absolute rate is modified to some degree by factors including pH and water availability, calculations based on degradation rates of DNA in free water have been used to demonstrate that at temperatures similar to those that might be expected in Egyptian tombs, PCR amplifiable DNA may at best survive for hundreds of years (e.g. Gilbert et al. 2005). This time span is similar to experimental observation based on Egyptian papyrus DNA (Marota et al. 2002), but considerably less than the thousands of years required by many published studies. In short, therefore, we critics have argued there is no logical reason why DNA would survive that long in such conditions to enable PCR-based analysis. Thus, the second problem is that the published results may be derived from modern human contaminant DNA on the ancient samples (e.g. Gilbert et al. 2005). In riposte, many who have published such studies argue that our analyses are too simplistic, for example, because mummification practices such as cavity evisceration and natron-based embalming that rapidly desiccates specimens will slow the rate of DNA degradation or that within tomb conditions are cooler and dryer than our models assume (e.g. Zink & Nerlich 2005). While this author remains sceptical about the latter argument—the temperature within many Egyptian tombs is warmer and damper than might be expected because of flooding and breaking of their environmental isolation through grave robbing (c.f. Gilbert et al. 2005), indeed sufficiently warm and damp that early accounts by some of the most famous tomb excavators such as Carter and Belzoni explicitly mentioned their surprise at the fact—the former argument is intriguing, and curiously there is little published data that can be used to address the issue. Thus, the importance of Hekkala et al.'s (2011) data that are in many ways unique. The authors report successful PCR amplification and sequencing of short mtDNA fragments from many of the historical samples, but more importantly, 8/22 mummified crocodile hatchlings estimated to date between 200BC and 200AD, well beyond the DNA survival range believed plausible by critics. More importantly, by using a non-human species and analysing it under conditions far stricter than used in previous studies on such material, it is very difficult to explain away the results as contamination. Thus, not only are the biological results significant—the data demonstrates both that crocodiles from today’s western clade were once found in the Nile, now home to the eastern clade, and that the western clade corresponds to one of the now discounted species, C. suchus (Geoffroy 1807, in Hekkala et al. 2011) thus warranting its resurrection—but in a stroke, the data reopens the debate about what could be feasible with mummified Egyptian material. There are some caveats, notably (i) the oldest (pre-dynastic) material attempted did not yield results, (ii) unlike humans, crocodiles have nucleated red blood cells and a thick keratinized skin layer, thus more DNA overall in their tissue, that is, better protected from contamination and (iii) if the DNA survival is a feature of mummification, mummification techniques changed continually through Egyptian history (c.f. Gilbert et al. 2005), and thus, how generally applicable the results are remain unknown. Despite this, however, the results are sufficient to persuade at least one hardened sceptic to open his mind to a potential future in which mummified Egyptian remains play a new role in questions, whether of biological or anthropological interest.

M.T.P.G. is interested in the application of degraded materials to the study of anthropology, ecology, and evolution.