Camel fossils from the El Kowm Basin, Syria : diversity and evolution

Camelidae is a family of Artiodactyla which includes a depleted diversity of extant species, divided into two tribes: the Camelini consist of two domestic species (Bactrian camel, Camelus bactrianus, and dromedary, C. dromedarius; the former is also represented by a wild subspecies, C. bactrianus ferus), and the Lamini consist of two domestic (Lama glama and Vicugna pacos) and two wild species (Lama guanicoe and Vicugna vicugna). The natural distribution of Camelini is in Eurasia and Africa, while Lamini are endemic of South America and are not part of this work. However, the family has a rich fossil record, amounting to dozens of extinct species, from its origins in the middle Eocene (~45 Ma) of North America to the dispersal of Camelini into the Old World towards the end of the Miocene (~6 Ma).
After their immigration, the evolutionary history of camels in Eurasia is unclear. Several Pliocene and Pleistocene species have been named within the genera Paracamelus and Camelus, but the relationships among them are poorly understood, in particular within the last two million years. Consequently, no direct ancestor of either extant species is known. As a matter of fact, until the turn of the millennium it was not even clear if Bactrian camel and dromedary are actually different species, or only domestic forms of the same wild precursor, and osteological differences between them were hardly known. This paucity of knowledge depends in equal measure from a lack of fossils and from a lack of evolutionary investigations over these animals, which contrasts with the great historical, cultural, and economic importance that they have in the arid regions of Africa and Asia. Thus, a better understanding of Old World camels will depend both on discovery of new fossils, and on improved description of already known species, including extant ones.
A rich collection of camel fossils has been found in the oasis of El Kowm, central Syria. This locality is a 10-km wide basin where numerous artesian wells have dotted the otherwise arid plain over the span of the Pleistocene, creating as many archaeological sites. The springs were not only attractive for the steppe fauna, but also for the ancient human population which are continuously recorded since their first expansion out of Africa: the most ancient lithic assemblage, from the site Aïn al Fil, is dated to the Olduvai subchron at about 1.8 Ma, while the most recent industries grade into the Neolithic and historical periods. The El Kowm Basin has been extensively studied from an archaeological point of view, and three sites have been excavated systematically: Nadaouiyeh Aïn Askar, Umm el Tlel and Hummal. However, the stratigraphy does not record only the human presence, but also a rich macrofauna. All layers of the site present similar animal assemblages: the dominant taxa are camels, equids and bovids of different size classes, indicating an arid steppe habitat and the absence of important climatic changes. The abundance of fossils and the long, detailed stratigraphic sequence obtained by combining the major sites give to the El Kowm Basin a prominent place among Middle East paleontological localities, which are concentrated on the humid coast or the northern mountains and rarely sample faunas adapted to arid climates. More specifically, this deep and rich record of camelids is unmatched in the Middle East and in the Pleistocene of the Old World, providing a unique window through which the origins of their charismatic extant relatives can be studied.
In this doctoral thesis, I tackle the study of the El Kowm in two steps: first, I lay some necessary comparative foundations by gathering data on the osteology of extant Camelus species and describing the yet unpublished type sample of Camelus thomasi, a terminal Early Pleistocene species from Algeria which is suspected to occur in the Middle East as well. Then I proceed with the description of the camelid samples from the sites of Nadaouiyeh, Hummal, and Aïn al Fil in the El Kowm Basin.
To compare the osteology of both extant camel species, Camelus bactrianus and Camelus dromedarius, I elected to focus on simple morphometric methods. Previous observations gave rise to the suspect that qualitative traits are poorly indicated to diagnose these two closely related species, but several skeletal parts might differ in proportions between them. In order to apply the data and methods on the fossil record, I found necessary to choose simple statistical analyses which can be applied even on highly fragmentary or poorly preserved specimens. Therefore, I developed a reliable measurement system and a data transformation called Harmonic Scores, which is a combination of standardizing and scaling. The chosen methods gave satisfying result: we were able to identify and quantify several consistent interspecific differences, some of which are univocal and highly diagnostic, while others are only slightly significant and noticeable only at a population level. In addition to the descriptive results and the measurement database that were generated, some distinctive traits are suggestive of previously unknown biological adaptation: in particular, the cranial anatomy of Bactrian camels shows characters correlated with increased grazing, while its limb muscle attachments may indicate additional need for lateral stability in a heavier animal. The presence and number of humps is reflected in the vertebral column, with several differences in the lumbar region that will be helpful in the reconstruction of fossil species.
The only fossil species which has been mentioned in the Middle East is Camelus thomasi POMEL 1893, described from the Algerian locality Tighennif. Unfortunately, only few skeletal part of this species have been published by its author (a maxilla, a fragment of mandibula, and a metatarsal), and the original description was not very detailed. Additional specimens have been referred to C. thomasi on the basis of weak arguments, usually large size and geographical proximity. However, a much larger fossil sample from Tighennif has been recovered by Arambourg in 1954-56 but never published. The remains are housed at the Musée National d’Histoire Naturelle in Paris and have long been unavailable to researchers. We finally elucidate the morphology of C. thomasi by describing this collection, including a complete cranium, several mandibles and postcranial bones. Our study shows that this animal was larger than extant camels, but not as much as some remains assigned to it; that currently, no identification outside of the Maghreb can be considered reliable; and that its relationships with either extant species are not very close, unlike what has been proposed by several authors.
The sites of the El Kowm Basin which have been included in this study are Aïn al Fil, Hummal and Nadaouiyeh Aïn Askar, all excavated by the University of Basel. The combined stratigraphy starts with Aïn al Fil, which is a small site dated at 1.8 Ma. It has yielded only four camelid specimens, but two very different specimens of the same bone (the scaphoideum) give reasons to accept the existence of two unnamed species in its time span; one of them is a giant form.
The temporal sequence in Nadaouiyeh covers a time span from 0.55 to 0.15 Ma, and is bracketed between the lower and the upper sections of Hummal. The important assemblage from this site is described and assigned to a new species, named Camelus roris. A rather complete cranium is chosen as the holotype, and a left maxilla as the paratype; this form is characterized by average size, broad cranial proportions, unique orbital shape, presence of maxillar crest, posterior placement of the palatine foramina, upper dentition with relatively large M1 and small M3, and a pachyostotic mandible comparable to C. thomasi. More than hundred dental and postcranial specimens are assigned to this species, but rare instances of bones with a strongly different morphology suggest that a second species sporadically visited the locality in this period.
The stratigraphy of Hummal site starts in the late Early Pleistocene but does not have an absolute dating; the entire lower section (unit G) is estimated within 1.2 and 0.8 Ma. In this time span, abundant camelid material is found and is shown to differ from other named species, either in El Kowm or elsewhere. It also differs from the material in Aïn al Fil. Unfortunately, there are not enough well-preserved cranial specimens to warrant the definition of a new species for this assemblage.
An important hiatus divides the lower layers in Hummal from the upper section (units A-F), whose age is considered middle to late Pleistocene; unit E is possibly as old as 0.325 Ma, and the uppermost units extends into historic times. This section is subdivided into several units, corresponding to different archaeological and camelid assemblages. The largest collection is found in the Mousterian industry-bearing unit C (layer 5). Here, the material demonstrates clearly that two species existed side by side within the interval from 0.150 to 0.045 Ma, one of slightly smaller size than the extant dromedary, the other of gigantic proportions, comparable to the largest Old World camelid known. Both species could be defined on adequate material: the small camel was named Camelus concordiae, and the large one Camelus moreli.
The situation is less clear in the units D, E, and F, representing a period intermediate between Nadaouiyeh and the Mousterian layers. Our study concluded that this material cannot be divided into discrete forms, nor can it be separated from neither the older C. roris nor the younger C. concordiae. We interpret this as a period of either admixture or alternance between these two species or their close relatives; anagenetic change is not impossible but seems unlikely.
The descriptive work performed within the scope of this thesis has produced abundant data over the morphology of extant and extinct camel species, both known and new. The comparative morphometry of living Camelus species answers a century-old debate and provides a necessary reference for any further studies. The publication of a large collection of C. thomasi fossils sheds clarity over this often misunderstood species. The analysis of the El Kowm record brought to light an unexpected and vast diversity, created by a pattern of dynamic evolutionary change, with at least six species represented here: more than the number previously described worldwide.