The palaeogenetics of cat dispersal in the ancient world

Abstract

The cat has long been important to human societies as a pest-control agent, object of symbolic value and companion animal, but little is known about its domestication process and early anthropogenic dispersal. Here we show, using ancient DNA analysis of geographically and temporally widespread archaeological cat remains, that both the Near Eastern and Egyptian populations of Felis silvestris lybica contributed to the gene pool of the domestic cat at different historical times. While the cat’s worldwide conquest began during the Neolithic period in the Near East, its dispersal gained momentum during the Classical period, when the Egyptian cat successfully spread throughout the Old World. The expansion patterns and ranges suggest dispersal along human maritime and terrestrial routes of trade and connectivity. A coat-colour variant was found at high frequency only after the Middle Ages, suggesting that directed breeding of cats occurred later than with most other domesticated animals.

Full text

1
ARTICLES
PUBLISHED: 19 JUNE 2017 | VOLUME: 1 | ARTICLE NUMBER: 0139
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
The palaeogenetics of cat dispersal in the
ancient world
Claudio Ottoni1,
2†*, Wim Van Neer3,
4, Bea De Cupere3, Julien Daligault2, Silvia Guimaraes2,
Joris Peters5,
6, Nikolai Spassov7, Mary E. Prendergast8, Nicole Boivin9, Arturo Morales-Muñiz10,
Adrian Bălăşescu11, Cornelia Becker12, Norbert Benecke13, Adina Boroneant14, Hijlke Buitenhuis15,
Jwana Chahoud
16,
17, Alison Crowther18, Laura Llorente10
†, Nina Manaseryan19, Hervé Monchot20,
Vedat Onar21, Marta Osypińska22, Olivier Putelat23, Eréndira M. Quintana Morales24,
Jacqueline Studer25, Ursula Wierer26, Ronny Decorte1, Thierry Grange2‡* and Eva-Maria Geigl2‡*
The cat has long been important to human societies as a pest-control agent, object of symbolic value and companion animal, but
little is known about its domestication process and early anthropogenic dispersal. Here we show, using ancient DNA analysis
of geographically and temporally widespread archaeological cat remains, that both the Near Eastern and Egyptian populations
of Felis silvestris lybica contributed to the gene pool of the domestic cat at different historical times. While the cat’s worldwide
conquest began during the Neolithic period in the Near East, its dispersal gained momentum during the Classical period, when
the Egyptian cat successfully spread throughout the Old World. The expansion patterns and ranges suggest dispersal along
human maritime and terrestrial routes of trade and connectivity. A coat-colour variant was found at high frequency only after
the Middle Ages, suggesting that directed breeding of cats occurred later than with most other domesticated animals.
The domestic cat is present on all continents except Antarctica,
and in the most remote regions of the world, and its evolu-
tionary success is unquestioned. While it is nowadays one of
the most cherished companion animals in the Western world, for
ancient societies barn cats, village cats and ships’ cats provided criti-
cal protection against vermin, especially rodent pests responsible
for economic loss and disease1. Owing to a paucity of cat remains
in the archaeological record, current hypotheses about early cat
domestication rely on only a few zooarchaeological case studies.
These studies suggest that ancient societies in both the Near East
and Egypt could have played key roles in cat domestication2,3.
Wildcats (Felis silvestris) are distributed all over the Old World.
Current taxonomy distinguishes five wild, geographically partitioned
subspecies: Felis silvestris silvestris, Felis silvestris lybica, Felis
silvestris ornata, Felis silvestris cafra and Felis silvestris bieti4. Modern
genetic data analyses of nuclear short tandem repeats (STR) and
16% of the mitochondrial DNA (mtDNA) genome in extant wild
and domestic cats revealed that only one of them, the north African/
southwest Asian F. s. lybica, was ultimately domesticated5.
Wildcats are solitary, territorial hunters and lack a hierarchi-
cal social structure6,7, features that make them poor candidates
for domestication8. Indeed, zooarchaeological evidence points to
a commensal relationship between cats and humans lasting thou-
sands of years before humans exerted substantial influence on their
breeding2,3,9. Throughout this period of commensal interaction,
tamed and domestic cats became feral and/or intermixed with wild
1KU Leuven—University of Leuven, Department of Imaging and Pathology, Center for Archaeological Sciences; University Hospitals Leuven, Laboratory
of Forensic Genetics and Molecular Archaeology, B-3000 Leuven, Belgium. 2Institut Jacques Monod, UMR 7592, CNRS and University Paris Diderot,
F-75013 Paris, France. 3Royal Belgian Institute of Natural Sciences, B-1000 Brussels, Belgium. 4KU Leuven—University of Leuven, Department of Biology,
Laboratory of Biodiversity and Evolutionary Genomics, Center of Archaeological Sciences, B-3000 Leuven, Belgium. 5Institute of Palaeoanatomy,
Domestication Research and the History of Veterinary Medicine, Ludwig-Maximilian University, D-80539 Munich, Germany. 6Bavarian Natural History
Collections, Bavarian State Collection of Anthropology and Palaeoanatomy, D-80333 Munich, Germany. 7National Museum of Natural History at the
Bulgarian Academy of Sciences, BG-1000 Sofia, Bulgaria. 8Radcliffe Institute for Advanced Study, Harvard University, Cambridge, 02138 Massachusetts,
USA. 9Max Planck Institute for the Science of Human History, Jena D-07743, Germany. 10Laboratorio de Arquezoología, Universidad Autónoma de
Madrid, E-28049 Madrid, Spain. 11National History Museum of Romania, RO-030026 Bucharest, Romania. 12Institute of Prehistoric Archaeology,
Free University Berlin, D-14195 Berlin, Germany. 13German Archaeological Institute, D-14195 Berlin, Germany. 14‘Vasile Pârvan’ Institute of Archaeology
of the Romanian Academy, RO-010667 Bucharest, Romania. 15Groningen Institute of Archaeology, University of Groningen, NL-9712 ER Groningen,
the Netherlands. 16Archéorient; CNRS/UMR 5133, Université Lumière Lyon II, F-69007 Lyon, France. 17Natural History Museum, Lebanese University,
LB-1107-2020 Beirut, Lebanon. 18School of Social Science, The University of Queensland, AU-4072 Brisbane, Queensland 4072, Australia. 19Institute of
Zoology, National Academy of Sciences of Armenia, AM-0019 Yerevan, Armenia. 20Labex Resmed, Université Paris IV la Sorbonne, F-75005 Paris, France.
21Istanbul University, Osteoarchaeology Practice and Research Center, Faculty of Veterinary Medicine, TR-34320 Avcilar-Istanbul, Turkey. 22Institute of
Archaeology and Ethnology, Polish Academy of Sciences, PL-61-712 Poznań, Poland. 23Archéologie Alsace, F-67600 Sélestat, and UMR 7041, ArScan,
Nanterre, France. 24Department of Anthropology, Rice University, Houston, Texas 77005, USA. 25Natural History Museum of Geneva, CH-1208 Genève,
Switzerland. 26Soprintendenza Archeologia della Toscana, I-50121 Firenze, Italy. †Present addresses: Centre for Ecological and Evolutionary Synthesis
(CEES), Department of Biosciences, University of Oslo, NO-0316 Oslo, Norway (C.O.); BioArch, Department of Archaeology, University of York, York YO10
5NG, UK (L.L). ‡These authors contributed equally to this work. *e-mail: claudio.ottoni@ibv.uio.no; eva-maria.geigl@ijm.fr; thierry.grange@ijm.fr

2
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
ARTICLES NATURE ECOLOGY & EVOLUTION
F. s. lybica or other wild subspecies as is common today10. These
regular genetic exchanges may have contributed to the low level of
differentiation observed between modern wild and domestic cat
genome sequences11. Accordingly, the domestication process seem-
ingly has not profoundly altered the morphological, physiological,
behavioural and ecological features of cats9, in contrast to what has
been observed, for example, for dogs12.
To address questions related to the contribution of the two pur-
ported centres of cat domestication, the Near East and Egypt, and the
history of human-mediated cat dispersal, we analysed ancient and
modern cats from Europe, north and east Africa, and southwest Asia
(SWA), spanning around 9,000 years, from the Mesolithic period
to the twentieth century AD. We analysed ancient DNA (aDNA) to
explore whether a fine phylogeographic structure of maternal lin-
eages existed prior to the domestication of F. s. lybica and whether,
when and how it was reconfigured over time in response to human
intervention, thereby documenting the domestication process of the
cat. We also studied a genetically defined coat-colour marker, the
blotched tabby marking13, to monitor a phenotypic change reflect-
ing human-driven selection along the domestication pathway.
Results
Strategy for data acquisition. The mtDNA phylogeny recon-
structed from extant wild and domestic cats5 identified five geo-
graphically distinct clades (I–V, Supplementary Fig. 1), representing
the five F. silvestris subspecies. The modern domestic cat mtDNA
pool was traced back to five deeply divergent subclades (IV-A to
IV-E) of the F. s. lybica clade, representing multiple wildcat lineages
incorporated over time and space6. These subclades lack a phylo-
geographic structure, which may reflect either poor sampling of the
truly wild modern F. s. lybica, particularly in its African range, or
multiple domestication events and/or extensive gene flow between
wild and domestic populations following the dispersal of domestic
cats. In order to screen and analyse a large number of ancient sam-
ples in parallel, many of which were expected to be poorly preserved
owing to higher-temperature burial environments, we applied an
ultrasensitive high-throughput approach14 to target informative
single-nucleotide polymorphisms (SNPs) on the mtDNA that reca-
pitulate the most salient features of the previously obtained phy-
logeny (Supplementary Fig. 1). Although mtDNA alone cannot
assess possible hybridization between different populations at the
individual level, the absence of recombination and the high copy
number make it a useful genetic marker for ancient population
analyses involving a large number of poorly preserved samples. The
mtDNA phylogeny (Fig.1b) reconstructed from 286bp sequenced
in our ancient samples alongside modern data from the literature5
clearly separates the five clades of F. silvestris (posterior probabili-
ties > 0.88, Supplementary Fig. 3, Supplementary Methods) and
the five subclades of F. s. lybica (posterior probabilities > 0.77).
We examined the phylogeographic pattern and its changes across
time by grouping the mtDNA haplotypes from our study into nine
time bins (Fig.1c).
Ancient European wildcats. We found the mtDNA clade I, represen-
tative of European wildcats (F. s. silvestris), exclusively in Europe. From
the Mesolithic period to the 8th century  in Western Europe (geo-
graphic locations 1-5 in Fig.1a,c), all cats analysed (9 out of 9) carried
clade I mtDNA, whereas in southeast Europe (6-8) we observed similar
frequencies of clade I (n= 13, 42%) and clade IV (n= 18, 58%), repre-
sentative of F. s. lybica. The latter was mostly represented by one of the
lineages of subclade IV-A, hereafter IV-A1 (Supplementary Figs 1, 3),
the earliest occurrence of which, in our dataset, dates back to 7700 
in Romania (7) (Supplementary Data 1), and which is still present
today in European wild (8) and domestic cats5. The occurrence of a
F. s. lybica mitotype in pre-Neolithic southeast Europe indicates that
the native range of this subspecies extended beyond the Bosporus.
Anatolian cats from the Neolithic period to the Bronze age.
A mitotype belonging to subclade IV-A (hereafter IV-A*, see
Methods section) was predominant (12 out of 14) from around
8000 to 800 BC in Anatolia (10-13) (Fig.1a–c). Its range may have
also extended to Lebanon (15). The frequencies of IV-A1 and IV-A*
found in southeast Europe and Anatolia, respectively, are signifi-
cantly different (Fisher’s exact test; P< 0.001), suggesting a phy-
logeographic structure that mirrors the original distribution of
genetically distinct wildcat populations carrying F. s. lybica mtDNA.
The earliest occurrence of IV-A* outside the Anatolian range in our
dataset was detected in two directly radiocarbon-dated specimens
from southeast Europe, in Bulgaria (4400 BC) and Romania (3200 BC),
clearly postdating the introduction of Neolithic farming prac-
tices, and in two Late Bronze/Iron age cats (around 1200 BC) from
Greece. The range expansion of this mitotype suggests human-
mediated translocation.
Ancient Levant and Africa. Owing to very poor DNA preservation,
we could not explore the phylogeographic structure of F. s. libyca in
this area prior to the Bronze Age. Therefore, we inferred the original
distribution of the other subclades (IV-B/E) by taking into account
their temporal appearance in our dataset. We found IV-B in three
ancient cat remains dated to the 1st millennium BC from southeast
Anatolia and Jordan (13, 16, Fig.1a–c), the 6th century BC in Syria and
later in Jordan (15, 16). This clade is still found in modern wildcats from
Israel5,15. These data suggest that this subclade was mainly restricted to
a Levantine range, throughout history. Outside of this range, IV-B was
found only in Medieval Iran (17) at very low frequencies (7%).
In Africa, two lineages of IV-C (named IV-C1 and IV-C*) were
detected in five out of seven cats (including three mummies) from
Egypt with dates ranging from the 7th century BC to the 4th century
AD (20, 21, Fig.1a–c). The original range of IV-C may have extended
from Egypt along the Nile River as far south as Congo and Burundi
(27, 28), where we detected a novel sub-lineage of IV-C (IV-C2,
Fig.1) in modern wildcats that had not yet been described in the
mtDNA pool of present-day domestic cats.
Subclades IV-D and IV-E were found at low frequencies solely
in recent temporal bins of our ancient dataset (1, 9–11), most likely
as a result of human-mediated dispersal. Their basal position in
clade IV, shared with lineages found in ancient African cats (light
pink symbols in Fig.1b,c; 20, 25, 28, 29) and not detected so far in
domestic cats, may suggest an African origin.
The dispersal of Egyptian cats. Outside Africa, from the 8th cen-
tury BC to the 5th century AD, we found IV-C1 in five Classical
Antiquity period cats from Bulgaria, Jordan and Turkey (8, 11 and
16, respectively, Fig.1a–c). This range expansion is more evident
between the 5th and 13th centuries AD, when the two IV-C lineages
found in ancient Egyptian cats became substantially more frequent
both in Europe (78%; 7 out of 9) and in SWA (46%; 32 out of 70). By
contrast, none of the 41 European and 18 southwest Asian cats from
archaeological contexts predating the 8th century BC possessed IV-C
haplotypes (Fisher’s exact test; P< 0.001 in both cases).
The territorial behaviour of cats and the rapid reconfiguration
of the phylogeographic pattern observed in Europe and SWA suggest
that cats carrying IV-C haplotypes were spread by humans through-
out the eastern Mediterranean region in Classical antiquity. Further
expansion occurred during the Medieval period, whereby the IV-C1
haplotype was found at the Viking trading port of Ralswiek on the
Baltic Sea (1, Fig.1a–c) by the 7th century AD, and at the Iranian port
of Siraf by the 8th century AD (17). In the Balkans, IV-C1 persisted
throughout Medieval times up to the present (8). Translocation
of cats over even longer distances was observed by the presence
of Asian wildcat (F. s. ornata) mtDNA at the Roman–Egyptian
port of Berenike on the Red Sea (1st–2nd century AD; 23, Fig.1a–c)
and at Medieval coastal sites in Turkey (9, 10).

3
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
ARTICLES
NATURE ECOLOGY & EVOLUTION
Figure 1 | Spatio-temporal representation of cat maternal genealogies. a, Map showing the present-day distribution of Felis silvestris4 with the
geographic range of each subspecies as reported in literature5 and inferred from the data presented herein. b, Tree of mtDNA lineages observed in our
ancient samples and in modern wild and domestic cats from literature5. c, Spatio-temporal depiction of ancient cat haplotypes as depicted with symbols
from the tree in b. Rows represent the approximate geographic provenance of the samples as reported in the map in a whereas the columns pertain to
chronological periods, the limits of which were selected to separate the prehistoric and historical periods evenly, to unambiguously assign each sample to
a single bin and to take historic events into account that could have affected human–cat interactions, as indicated on the timeline above. A dot inside the
symbols indicates AMS-radiocarbon-dated samples; dashed lines inside the symbols indicate incomplete mtDNA profiles; Near Eastern modern wildcats
from literature5 are indicated by grey-shaded bins. Numbers in a and in c represent the approximate geographic locations of the sites from which the
samples are derived as reported in Supplementary Table 5.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Africa Southwest Asia Europe
2
4
a
c
b
5
3
6
8
7
9
10
11 12 13
14
15
16 17
20
212223
24
25
18 19
26
27
28 29
30
31 F. silvestris silvestris
F. silvestris lybica
F. silvestris ornata
F. silvestris cafra
F. silvestris bieti
1
F. margarita
DEC
B
A
F. s. silvestris
F. s. cafra
F. silvestris lybica
(IV)
F. s. ornata
0.002
(I)
(II)
(III)
12
**
12
F. s. bieti
(V)
Early Neolithic in
southeast Europe
Classical
Antiquity
>6500 BC
Early Bronze Age
in southwest Asia
4500 BC 3100 BC 2000 BC 800 BC AD 500 AD 1300
Roman and Byzantine empire
Modern wild
AD 1900
Ottoman Empire

4
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
ARTICLES NATURE ECOLOGY & EVOLUTION
Coat pattern. The domestication process has not markedly changed
the morphology of cats, and few traits can be used today to
identify wild or hybrid populations. Of the few traits available,
the most widely used is the tabby coat marking16. The transmem-
brane aminopeptidase Q (Taqpep) gene is responsible for the
tabby phenotypic variation in cats, with a single SNP distinguish-
ing most of the mackerel and blotched patterns that are character-
istic of the wild and domestic patterns, respectively13. To develop
a temporal framework for the emergence of a variation in coat
pattern typical of domestic cats, we investigated the three SNPs
in the Taqpep gene13. We found that the recessive allele respon-
sible for the blotched-tabby pattern in 80% of present-day cats
(W841X) occurred in our ancient dataset not earlier than the
Medieval period in SWA (3%, minimum number of total alleles,
see Methods) (Fig.2). Thereafter, its frequency increased in Europe,
SWA, and Africa (50% in total), showing late expansion of this typi-
cally domestic allele.
Discussion
Zooarchaeological and iconographic evidence for early cat
domestication. Owing to the paucity of cat remains in the archeo-
logical record and the lack of established osteometric features dis-
tinguishing remains from wild and domestic F. s. lybica2, current
hypotheses about early cat domestication are grounded in scanty
evidence when compared to other domesticated animals. A com-
plete skeleton found in Cyprus in association with a human burial
dated to around 7500 BC suggests that cats were probably tamed
by early Neolithic sedentary communities that had been growing
cereals in SWA, concomitant with the emergence of commensal
rodents3. Similarly, the skeletons of six cats in an elite Predynastic
cemetery in Egypt, around 3700 BC, may suggest a close cat–human
relationship in early ancient Egypt2.
The iconography of Pharaonic Egypt constitutes a key source of
information about the species’ relationship with humans, and has
motivated the traditional belief that cat domestication took place in
Egypt1,17. Numerous depictions in Egyptian art from the 2nd mil-
lennium BC document a progressive tightening of the relationship
between human and cat, as illustrated in particular by the popu-
larization of the motif of the ‘cat under the chair’ of women after
around 1500 BC1,17.
Here, we show that mitochondrial lineages corresponding to
these two purported domestication centres contributed at different
times to the gene pool of modern domestic cats. We deduced this
by establishing the ancestral phylogeography of wild cats in the Old
World and by observing its reconfiguration through time, which
reveals the spread of cats through human agency following ancient
land and maritime trade routes.
Distribution of wildcats. Our aDNA data (Fig.1a and Supplementary
Fig. 4) show a clear phylogeographic structure. F. s. silvestris was
confined to Europe, whereas F. s. lybica was found in SWA and
southeast Europe. A clear understanding of the present distribu-
tion of wild F. s. lybica in Anatolia has proven elusive until now
owing to a lack of genetic data. It has commonly been assumed
that the native range of the modern European wildcat includes
Anatolia5,18,19. Our phylogeographic reconstruction demonstrates
that mtDNA clade IV, corresponding to F. s. lybica, was predomi-
nant in Anatolia for many millennia beginning in the Neolithic
period at the latest. Not a single instance of clade I, corresponding to
F. s. silvestris, was detected in our samples from SWA (Fig. 1).
Nevertheless, we cannot exclude its presence in the wilds of
Anatolia, in particular in the forest and mountain refuges of north-
ern Anatolia and the Caucasus.
We found two distinct IV-A mitotypes on either side of the
Bosporus. In Anatolia, from around 8000 BC to 800 BC, almost all
cats (12 out of 14) belonged to the IV-A* mitotype. By contrast,
cats carrying a distinct mitotype, IV-A1, were present in southeast
Europe by the beginning of the 8th millennium BC. This suggests that
F. s. lybica was distributed across Anatolia from the early Holocene
epoch at the latest, prior to the formation of the present-day exten-
sion of the Black Sea, and that it made its way to southeast Europe
before the onset of farming in the Neolithic period. A split in an
ancestral Anatolian cat population in the late Pleistocene epoch,
presumably during the Last Glacial Maximum, followed by local
differentiation and/or drift and founder effect, might have been
responsible for the distribution of distinct clade IV mtDNA lineages
in Anatolia and southeast Europe. F. s. silvestris and F. s. lybica occur
across different biotopes that include, respectively, temperate wood-
land and open bushland4. The expansion of open bushlands during
the Late Pleistocene epoch might have attracted F. s. lybica into the
Africa Southwest Asia Europe
Mackerel
(TaM/TaM or TaM/Tab)
>6500 BC
4500 BC
3100 BC
2000 BC
800 BC
AD 500
AD 1300
AD 1900
Blotched
(Tab/Tab)
Ta
M
, wild type Ta
b
, blotched (W841X)
Figure 2 | Spatio-temporal representation of the alleles determining the phenotypic variation in the shape of tabby patterns, mackerel (TaM)
and blotched (Tab). To overcome issues of potential allelic drop-out, each individual is defined by at least one observed allele, except for the
few instances in which both alleles were detected. The image shows a ‘cat under the chair’ with a tabby mackerel marking, typical of F. silvestris lybica
(Anna (Nina) Macpherson Davies, Copy of Wall Painting from Private Tomb 52 of Nakht, Thebes (I, 1, 99–102) Cat Eating Fish. Photo: © Ashmolean
museum, Oxford, UK).

5
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
ARTICLES
NATURE ECOLOGY & EVOLUTION
Balkans when the Bosporus was a land bridge and the Balkans rep-
resented a refuge for warm-adapted species20,21.
Currently, IV-A1 is found in the European wildcat population
and also in modern domestic cats5 (Fig. 1). Our data imply that
admixture episodes potentially occurring through time between
overlapping populations of wild F. s. silvestris and F. s. lybica could
be in part responsible for F. s. lybica mtDNA introgression in pres-
ent-day European wildcat populations. Conservation programs
should also take into account past natural admixture when aiming
at neutering and removing hybrids that are believed to have a role in
cryptic extirpations of wild F. s. silvestris populations4.
Origin and dispersal of domestic cats. Our data show that mito-
type IV-A* had a wide distribution stretching across Anatolia from
west to east throughout the Neolithic, Bronze age and Iron age.
Its range may have extended as far south as the Levant, where we
inferred the presence of subclade IV-B. These findings suggest that
in the Fertile Crescent, cats that developed a commensal relation-
ship with early farming communities during the Neolithic period
carried at least mitotypes IV-A* and IV-B. Mitotype IV-A* later
spread to most of the Old World, representing the Near Eastern
contribution to the mtDNA pool of present-day domestic cats.
This spread may have started as early as around 4400 BC into
southeast Europe, the date of the first appearance of IV-A* in our
European dataset, and therefore subsequent to the neolithisation
of Europe. This suggests that the human-mediated translocation of
cats began in prehistoric times, corroborating the interpretation of
the finding of a cat buried around 7500 BC in Cyprus3. We also found
IV-A* in cat remains from the Roman–Egyptian port of Berenike
on the Red Sea and in one Egyptian mummy (Fig.1a–c), which may
hint at an introduction of cats from SWA to Egypt.
Our data provide the first evidence for an African origin for one
of the mitochondrial lineages of present-day domestic cats, namely
clade IV-C. Indeed, we found the lineages C1 and C* in the majority
of Egyptian cat mummies. These cats were worshipped and, dur-
ing the Greco–Roman period, kept in temple precincts to be mum-
mified17. We show that, despite a local ban on cat trading being
imposed in Egypt as early as 1700 BC22, cats carrying IV-C mtDNA
spread to most of the Old World. The increasing popularity of cats
among Mediterranean cultures and particularly their usefulness on
ships infested with rodents and other pests presumably triggered
their dispersal across the Mediterranean22. Indeed, depictions of cats
in domestic contexts, already frequent during the New Kingdom
in Egypt around 1500 BC (‘cat under the chair’, Fig. 2), are found
on Greek artifacts from as early as the end of the 6th century BC
(Supplementary Methods). The Egyptian cat must have been very
popular, as IV-C1 and C* represented more than half of the mater-
nal lineages in Western Anatolia during the 1st millennium AD,
and occurred twice as frequently as the local mitotype IV-A*. This
suggests that the Egyptian cat had properties that made it attrac-
tive to humans, presumably acquired during the tightening of the
human–cat relationship that developed during the Middle and New
Kingdoms and became even stronger afterwards1,17. As the most
pronounced genetic changes that distinguish wild and domestic cats
are apparently linked to behaviour11, it is tempting to speculate that
the success of the Egyptian cat is underlain by changes in its socia-
bility and tameness.
North of the Alps, domestic cats appeared soon after the Roman
conquest, yet remained absent outside the Roman territory until
Late Antiquity23. In medieval times it was compulsory for seafar-
ers to have cats onboard their ships24, leading to their dispersal
across routes of trade and warfare. This evidence explains, for
example, the presence of the Egyptian lineage IV-C1 at the Viking
port of Ralswiek (7–11th century AD)24. The expansion of the domes-
tic cat may have been fostered by a diversification in their cultural
usage, which in Medieval Europe included the trade of domestic
cat pelts as cloth items25. Spread of the black rat (Rattus rattus)
and house mouse (Mus musculus) by sea routes as early as the
Iron age, documented by zooarchaeological and genetic data26,
probably also encouraged cat dispersal for the control of these
new pests.
Increased translocation as a result of long-distance trade is also
witnessed by the finding of Asian F. s. ornata mtDNA in cats from
the Roman Red Sea port of Berenike (1st–2nd century AD) and from
Turkey in the 6–7th century AD. This was probably the result of
increasingly intensive and direct trade connections between south
Asia and the Mediterranean basin via the Indian Ocean and Red
Sea27, but possibly also via the Silk Road connecting central Asia
with Anatolia28. Long-distance maritime routes29, as described for
instance in the 1st century AD Periplus of the Erythraean Sea, prob-
ably explain the occurrence of IV-A*, typical of SWA, as far south as
East Africa (30, Fig.1a–c).
Upon arrival in these various new locations, introduced cats
reconfigured the phylogeographic landscape of the species through
admixture with local tame or wild cats, leading to a transfer of deeply
divergent mitochondrial lineages in the domestic pool (IV-D/E and
possibly III-F. s. ornata although these lineages are found only at
low frequency in modern domestic cats5). Modern genetic data
have shown that admixture with domestic cats still occurs today
in European wildcat populations10,16, and intensive conservation
programs have been implemented to preserve the integrity of
F. s. silvestris4,30.
Evolution of the tabby pattern. Our study also sheds light on
the late emergence in domestic cats of a key phenotypic trait,
the blotched coat marking caused by a SNP in the Taqpep gene13.
Wildcats exhibit a mackerel-like coat pattern, whereas the blotched
pattern is common in many modern domestic breeds13. In our data-
set, the first occurrence of the recessive allele W841X, which is asso-
ciated with the blotched markings, dates to the Ottoman Empire in
SWA and later increases in frequency in Europe, SWA and Africa
(Fig.2). This result is in agreement with the iconography from the
Egyptian New Kingdom through the European Middle Ages, where
cats’ coats were mainly depicted as striped, corresponding to the
mackerel-tabby pattern of the wild F. s. lybica1,17 (Fig.2). It was only
in the 18th century AD that the blotched markings were common
enough to be associated with the domestic cat by Linnaeus13, and
physical traits started to be selected only in the 19th century AD for
the production of fancy breeds15. Thus, both our data and recent
genomic data11 suggest that cat domestication in its early stages may
have affected mainly some behavioural features, and distinctive
physical and aesthetic traits may have been selected for only recently.
A similar pattern of late emergence of other phenotypic traits
has been observed in chicken31, but contrasts with what has been
observed in horses, where coat-colour differentiation appeared at
an early stage of domestication32.
Conclusive remarks. The comprehensive aDNA genetic study
of cats across time and space that we present, provides answers
to longstanding questions concerning the domestication pro-
cess of the cat and contributes to a better understanding of how
humans have reshaped global biodiversity through species trans-
locations23,26,33. By revealing the original phylogeographic dis-
tribution of wildcats and its profound modification through
human-mediated dispersal of tamed cats through time, we show
that both Near Eastern and Egyptian cat lineages contributed
at different times to the maternal genetic pool of domestic cats,
with one or other present in the vast majority of present-day cat
breeds. Cat domestication was a complex, long-term process
featuring extensive translocations that allowed admixture events
between geographically separated cat populations at different
points in time.

6
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
ARTICLES NATURE ECOLOGY & EVOLUTION
Methods
Ancient DNA analyses. Ancient DNA analysis was performed in dedicated
aDNA facilities in Paris and Leuven from bone, teeth, skin and hair samples
(the last two when available in Egyptian mummies) of 352 ancient cats. e ages
of the archaeological remains were determined using direct accelerator mass
spectrometry (AMS) radiocarbon dating (KIK-IRPA, Belgium), stratigraphic
associations with AMS dates, and contextual archaeological evidence
(Supplementary Data 1). All dates in the text are reported in calibrated radiocarbon
years BC. DNA was also extracted from claws and skin samples of 28 modern
wildcats from Bulgaria and east Africa (Supplementary Methods).
Amplification of nine mtDNA and three nuclear DNA fragments in the
Taqpep gene was preceded by the elimination of carry-over contamination based
on the dUTP/UNG system34 and carried out in three separate multiplex PCRs.
Phylogenetically informative SNPs in the mtDNA were selected following the
most up-to-date worldwide cat phylogeny5 (Supplementary Fig. 1). We targeted
42 informative SNPs in nine short regions distributed across the mitochondrial
ND5, ND6 and CytB genes that recapitulate the most salient features of the
phylogeny obtained with longer portions of these genes (Supplementary Fig. 1).
The diagnostic SNPs were screened with Pyrosequencing assays (Biotage, Qiagen)
and sequencing on a PGM Ion Torrent platform (Institut Jacques Monod, Paris)
of amplicon libraries following the ‘aMPlex Torrent’ workflow and downstream
sequence analysis with a bash script described elsewhere14 (Supplementary Code).
The aMPlex Torrent approach combines the sensitivity of multiplex PCR with
the power and throughput of next-generation sequencing14 and made it possible
to screen and analyse a large number of poorly preserved ancient samples in
parallel. We obtained mtDNA sequences from 209 out of 352 ancient cats (59%;
Supplementary Data 1, 2), with expectedly lower success rates for old samples from
hot environments (Supplementary Fig. 2). The mtDNA profiles ranged from 286 to
449 bp, 12 of which were incomplete profiles generated from two to seven mtDNA
fragments. The authentication criteria adopted rely on: (i) strict contamination
prevention controls including physical containment as well as material and reagent
decontamination34–36; (ii) extensive replications performed through independent
PCRs (at least two, but up to eight with one up to three independent DNA
extracts). For samples where the low DNA content decreased data reliability,
we increased the number of replicates and used different PCR assays, multiplex
and simplex PCR, pyrosequencing and the aMPlex Torrent method performed
in independent laboratories (Paris and Leuven) so that samples with different
preservation levels could be genotyped with similar reliability. More details about
DNA extraction, amplification, sequencing and the authentication criteria can be
found in the Supplementary Methods.
Phylogeographic analyses. Each specimen was assigned to an mtDNA clade using
the terminology previously proposed5, including specimens with an incomplete
profile (shapes with an inner dashed line in Fig.1c). Owing to our streamlined
sequencing assay, some of the subclades and lineages of IV-A and IV-C observed
in the 2007 study were collapsed into a single haplotype, which we named
IV-A* and IV-C*, respectively (Supplementary Fig. 1). The ancient and modern
sequences generated here were aligned to 159 sequences from Driscoll’s study5.
A Bayesian tree of 66 unique 286 bp-long haplotypes (Supplementary Fig. 3)
was constructed as described in detail in the Supplementary Methods. An ML
tree, obtained as described in the Supplementary Methods, had the same topology.
Frequencies of haplotypes A* and A1 in Anatolia and southeast Europe, and
of clade C in before and after the 8th century AD in SWA, were tested using a
Fisher’s exact test.
Nuclear markers. We typed allelic variations within the Taqpep gene associated
with coat-colour pattern differences—W841X, D228N and T139N13. The results
presented here are intended to be indicative of allele frequencies. Given the low
level of independent replications of our assay and the risk of allelic dropout,
especially in ancient degraded samples, we could not ascertain genotypes,
except for a few heterozygous samples showing a fairly high number of reads
in at least two independent amplifications (Fig.2, Supplementary Methods
and Supplementary Data 2). Assuming that none of the alleles is amplified
preferentially, and adopting a conservative strategy that accounted for the
minimum number of alleles observed, our data across the spatial and temporal
framework showed that 7 out of 67 successfully amplified cat samples possessed
at least one mutant Tabby-W841X allele, of which two were heterozygotes (BMT2
and MET9). In 88 cats we could screen the allele D228N and in all instances
we observed the wild-type. Among 63 cats successfully screened for T139N,
we detected the mutant allele (C to A) in three specimens.
Code availability. A bash script and accessory fasta and gff files for data
analysis of the aMPlex Torrent data are provided as Supplementary Code.
Data availability. Sequence data that support the findings of this study have been
deposited in Dryad (http://dx.doi.org/10.5061/dryad.g4p30).
Received 10 October 2016; accepted 10 March 2017;
published 19 June 2017
References
1. Engels, D. W. Classical Cats: e Rise and Fall of the Sacred Cat
(Routledge, 2001).
2. Van Neer, W., Linseele, V., Friedman, R. & De Cupere, B. More evidence for
cat taming at the Predynastic elite cemetery of Hierakonpolis (Upper Egypt).
J. Arch. Sci. 45, 103–111 (2014).
3. Vigne, J.-D., Guilaine, J., Debue, K., Haye, L. & Gérard, P. Early taming of the
cat in Cyprus. Science 304, 259 (2004).
4. Yamaguchi, N., Kitchener, A. C., Driscoll, C. & Nussberger, B.
Felis silvestris. IUCN http://dx.doi.org/10.2305/IUCN.UK.2015-2.RLTS.
T60354712A50652361.en (2015).
5. Driscoll, C. A. et al. e Near Eastern origin of cat domestication. Science
317, 519–523 (2007).
6. Driscoll, C. A., Macdonald, D. W. & O’Brien, S. J. From wild animals to
domestic pets, an evolutionary view of domestication. Proc. Natl Acad. Sci.
USA 106, 9971–9978 (2009).
7. Price, E. O. Animal Domestication and Behavior (CABI Publishing, 2002).
8. Wozencra, C. in Mammal Species of the World 3rd edn (eds Wilson, D. E. &
Reeder, D. M.) http://go.nature.com/2p5vsxb (Johns Hopkins Univ. Press, 2005).
9. Zeder, M. A. in Biodiversity in Agriculture (eds Gepts, P. et al.) 227–259
(Cambridge Univ. Press, 2012).
10. Mattucci, F., Oliveira, R., Lyons, L. A., Alves, P. C. & Randi, E. European
wildcat populations are subdivided into ve main biogeographic groups:
consequences of Pleistocene climate changes or recent anthropogenic
fragmentation? Ecol. Evol. 6, 3–22 (2015).
11. Montague, M. J. et al. Comparative analysis of the domestic cat genome
reveals genetic signatures underlying feline biology and domestication.
Proc. Natl Acad. Sci. USA 111, 17230–17235 (2014).
12. Axelsson, E. et al. e genomic signature of dog domestication reveals
adaptation to a starch-rich diet. Nature 495, 360–364 (2013).
13. Kaelin, C. B. et al. Specifying and sustaining pigmentation patterns in
domestic and wild cats. Science 337, 1536–1541 (2012).
14. Guimaraes, S. et al. A cost-eective high-throughput metabarcoding
approach powerful enough to genotype ~44 000 year-old rodent remains
from Northern Africa. Mol. Ecol. Resour. http://dx.doi.org/10.1111/1755-
0998.12565 (2016).
15. Driscoll, C. A., Clutton-Brock, J., Kitchener, A. C. & O’Brien, S. J. e taming
of the cat. Genetic and archaeological ndings hint that wildcats became
housecats earlier—and in a dierent place—than previously thought. Sci. Am.
300, 68–75 (2009).
16. Pierpaoli, M. et al. Genetic distinction of wildcat (Felis silvestris) populations
in Europe, and hybridization with domestic cats in Hungary. Mol. Ecol. 12,
2585–2598 (2003).
17. Málek, J. e Cat in Ancient Egypt, revised edn 144 (British Museum Press,
1993–2006).
18. Spassov, N., Simeonovski, V. & Spiridonov, G. e wild cat (Felis silvestris
Schr.) and the feral domestic cat: problems of the morphology, taxonomy,
identication of the hybrids and purity of the wild population.
Hist. Nat. Bulg. 8, 101–120 (1997).
19. Kitchener, A. C. & Rees, E. E. Modelling the dynamic biogeography of the
wildcat: implications for taxonomy and conservation. J. Zool. 279,
144–155 (2009).
20. Ivanova, S. et al. Magura Cave, Bulgaria: a multidisciplinary study of Late
Pleistocene human palaeoenvironment in the Balkans. Quat. Int. 415,
86–108 (2016).
21. Dubey, S. et al. Molecular evidence of Pleistocene bidirectional faunal
exchange between Europe and the Near East: the case of the bicoloured
shrew (Crocidura leucodon, Soricidae). J. Evol. Biol. 20, 1799–1808 (2007).
22. Faure, E. & Kitchener, A. C. An archaeological and historical review of the
relationships between felids and people. Anthrozoos 22, 221–238 (2009).
23. Peters, J. Römische Tierhaltung und Tierzucht. Eine Synthese aus
archäozoologischer Untersuchung und schrilich-bildlicher Überlieferung 444
(Verlag Marie Leidorf, 1998).
24. Johansson, F. & Hüster, H. Untersuchungen an Skelettresten von Katzen aus
Haithabu (Ausgrabung 1966 – 1969) 86 (Wachholtz, 1987).
25. Ewing, E. Fur in dress 168 (Batsford, 1981).
26. Jones, E. P., Eager, H. M., Gabriel, S. I., Johannesdottir, F. & Searle, J. B.
Genetic tracking of mice and other bioproxies to infer human history.
Trends Genet. 29, 298–308 (2013).
27. Sidebotham, S. E. Berenike and the Ancient Maritime Spice Route
(Univ. California Press, 2011).
28. Ball, W. Rome in the East: e Transformation of an Empire 523
(Routledge, 2000).
29. Boivin, N., Crowther, A., Helm, R. & Fuller, D. East Africa and Madagascar
in the Indian Ocean world. J. World Prehist. 26, 213–281 (2013).
30. Witzenberger, K. A. & Hochkirch, A. e genetic integrity of the ex situ
population of the European wildcat (Felis silvestris silvestris) is seriously
threatened by introgression from domestic cats (Felis silvestris catus).
PLoS ONE 9, e106083 (2014).

7
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol
ARTICLES
NATURE ECOLOGY & EVOLUTION
École Normale Supérieure, Paris, for granting access to the pyrosequencer;
M. Larmuseau and A. Van Geystelen for discussions and assistance with nuclear
SNP analyses; K. Knaepen, M. Coomans, and A. Giucca for support in laboratory
procedures in Leuven; and J. Nackaerts of the veterinary hospital Kruisbos
(Wezemaal, Belgium) for providing cat blood samples. We also thank the curators
of the following collections for facilitating access to the material under their care
and the permission to take tissue samples: the Royal Museum for Central Africa
(Tervuren, Belgium), the Muséum National d’Histoire Naturelle and Musée du
Louvre (Paris, France), the British Museum and Natural History Museum
(London, UK) and the Bavarian State Collection of Anthropology and Palaeoanatomy
(Munich, Germany).
Author contributions
The project was initiated by W.V.N., E.-M.G., C.O., T.G. and R.D. The ancient DNA study
was conceived and designed by T.G., E.-M.G. and C.O. C.O. carried out the molecular
laboratory work, with support of S.G. and analysed the data. J.D. generated the aMPlex
Torrent data. The archaeological bone samples were provided by W.V.N., B.D.C., J.P.,
N.S., M.E.P., N.Bo., A.M.-M., A.Bă., C.B., N.Be., A.Bo., H.B., J.C., A.C., L.L., N.M., H.M.,
V.O., M.O., M.O., O.P., E.M.Q.M., J.S., U.W., and W.V.N. and B.D.C. were responsible for
their curation and archaeozoological recording. The authors’ list from A.Ba. to U.W. is
in alphabetical order. C.O., E.M.G. and T.G. wrote the paper. W.V.N., B.D.C., J.P., N.S.,
M.E.P., N.Bo., A.M.-M. contributed to further discussion about the interpretation of the
data and the outline of the paper. N.Bo. and M.E.P. revised the English. All the authors
gave final approval for publication.
Additional information
Supplementary information is available for this paper.
Reprints and permissions information is available at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to C.O., T.G. and E.-M.G.
How to cite this article: Ottoni, C. et al. The palaeogenetics of cat dispersal in the
ancient world. Nat. Ecol. Evol. 1, 0139 (2017).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Competing interests
The authors declare no competing financial interests.
31. Girdland Flink, L. et al. Establishing the validity of domestication genes
using DNA from ancient chickens. Proc. Natl Acad. Sci. USA 111,
6184–6189 (2014).
32. Ludwig, A. et al. Coat color variation at the beginning of horse
domestication. Science 324, 485 (2009).
33. Boivin, N. L. et al. Ecological consequences of human niche construction:
examining long-term anthropogenic shaping of global species distributions.
Proc. Natl Acad. Sci. USA 113, 6388–6396 (2016).
34. Pruvost, M., Grange, T. & Geigl, E.-M. Minimizing DNA-contamination
by using UNG-coupled quantitative real-time PCR (UQPCR) on degraded
DNA samples: application to ancient DNA studies. Biotechniques 38,
569–575 (2005).
35. Champlot, S. et al. An ecient multistrategy DNA decontamination
procedure of PCR reagents for hypersensitive PCR applications. PLoS ONE 5,
e13042 (2010).
36. Bennett, E. A. et al. Library construction for ancient genomics: single strand
or double strand? Biotechniques 56, 289–300 (2014).
Acknowledgements
This research has been funded by the IAP program (BELSPO), the KU Leuven
BOF Centre of Excellence Financing on CAS, and the CNRS (T.G. and E.-M.G.).
The high-containment laboratory of the Institut Jacques Monod, Paris was supported by
a grant to E.-M.G. from the University Paris Diderot, ARS 2016-2018. The sequencing
facility of the Institut Jacques Monod, Paris, and J.D., were supported by grants to
T.G. from the University Paris Diderot, the ‘Fondation pour la Recherche Médicale’
(DGE20111123014), and the ‘Région Ile-de-France’ (grant 11015901). C.O. was
supported by the FWO mobility program (V4.519.11N, K2.197.14N, K2.057.14N).
Faunal research carried out by J.P. and team in Anatolia received funding by the
German Research Foundation (DFG PE424/10-1,2). Research by N.Bo. M.E.P., and
A.C. was supported by an ERC grant (206148) and UK NERC Radiocarbon Facility
grant (NF/2012/2/4). The archaeological and archaeozoological research conducted
by A.Bă. and A. Bo. was supported by the Romanian National Authority for Scientific
Research, UEFISCDI (PN-II-ID-PCE-2011-3-1015 and PN-II-RU-TE-2014-4-0519).
Research at Songo Mnara was directed by S. Wynne-Jones and J. Fleisher with support
from the National Science Foundation (BCS1123091) and the Arts and Humanities
Research Council (AH/J502716/1). We thank G. Larson and E. A. Bennett for critical
reading of the manuscript; the Ufficio Beni Archeologici della Provincia Autonoma di
Bolzano for granting access to the archaeological material of Galgenbühel/Dos de la
Forca and J. Crezzini for help in sampling; M.-A. Félix, Institut Jacques Monod and