ArticlePDF Available

Abstract and Figures

During the Late Pleistocene, Naxos and adjacent areas, including Delos and Paros, constituted a mega-island, here referred to as palaeo-Cyclades. The extensive low-lying plains with lakes and rivers provided a suitable habitat for elephants. Due to long-term isolation from the mainland and mainland populations, these elephants evolved miniature size. The species found on Naxos had a body size of about ten percent of that of the mainland ancestor, Palaeoloxodon antiquus. During the glacial periods of the Late Pleistocene, P. antiquus may have migrated eastwards and southwards in search of better conditions and reached the islands. The dwarf species of the various Southern Aegean islands (e.g. Crete, Tilos, Rhodos, palaeo-Cyclades) are each the result of independent colonisation events. The very small size of the Naxos species respective to the dwarf elephants from Crete is explained as due to the lack of competitors. The only other elements of the contemporaneous fauna were a rock mouse (Apodemus cf. mystacinus) and a shrew (Crocidura sp.). Submergence of the area, climate change, volcanism, hunting by humans or a combination of these factors during the terminal Pleistocene may have caused the extinction of this endemic fauna.
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
A dwarf elephant and a rock mouse on Naxos (Cyclades, Greece) with
a revision of the palaeozoogeography of the Cycladic Islands (Greece)
during the Pleistocene
Alexandra A.E. van der Geer
, Hara Drinia
National and Kapodistrian University of Athens, Faculty of Geology and Geoenvironment, Department of Historical GeologyPalaeontology, 157 84 Athens, Greece
Naturalis Biodiversity Center, Department of Geology, PO Box 9517, 2300RA Leiden, The Netherlands
abstractarticle info
Article history:
Received 18 November 2013
Received in revised form 24 March 2014
Accepted 1 April 2014
Available online 12 April 2014
Body size evolution
Dwarf elephants
Island rule
New species
During the Late Pleistocene,Naxos and adjacentareas, including Delos and Paros, constituted a mega-island, here
referredto as palaeo-Cyclades. The extensivelow-lying plainswith lakes and rivers provided a suitable habitat for
elephants. Due to long-term isolation from the mainland and mainland populations, these elephants evolved
miniature size. The species found on Naxos had a body size of about ten percent of that of the mainland ancestor,
Palaeoloxodon antiquus. During the glacial periods of the Late Pleistocene, P. antiquus may have migrated
eastwards and southwards in search of better conditions and reached the islands. The dwarf species of the
various Southern Aegean islands (e.g. Crete, Tilos, Rhodos, palaeo-Cyclades) are each the result of independent
colonisation events. The very small size of the Naxos species respective to the dwarf elephants from Crete
is explained as due to the lack of competitors. The only other elements of the contemporaneous fauna were a
rock mouse (Apodemus cf. mystacinus) and a shrew (Crocidura sp.). Submergence of the area, climate change,
volcanism, hunting by humans or a combination of these factors during the terminal Pleistocene may have
caused the extinction of this endemic fauna.
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Dwarf elephants were relatively common elements in insular faunas
worldwide in the past. Thediminutive size of these fossil insular probos-
cideans started to attract the attention of naturalists in the 19th century
(e.g. Sciná, 1831; Leith Adams, 1863; Spratt, 1867). This interest
continued throughout the 20th century and is still a signicant topic
of scientic research. There are numerous recent studies on this subject
(e.g. Van den Bergh, 1999; Palombo, 2007; Ferretti, 2008; Herridge,
2010; Herridge and Lister, 2012; Liscaljet, 2012), which re-examine
old specimens, and in some cases describe recently collected ones,
and put their analyses under the perspective provided by the latest
developments in ecology, biogeography and palaeogeography.
There is, however, a set of elephant fossils, known to science since
several decades that has been ignored in these works. They are the
fossils found on the Cyclades (Southern Aegean Sea, Greece), a group
of islands (archipelago) that have recently become disconnected from
each other and the mainland by risingsea level (continental shelf islands
sensu Whittaker, 1998). The Cyclades, located between the Greek
peninsula in the west and the coast of Asia Minor in the east, consists
mainly of metamorphic and igneous rocks (Hejl et al., 2002). The sedi-
mentary deposits, including those in caves, are limited. Therefore, on
the contrary to the overwhelming amount of fossils known from other
Mediterranean islands with endemic mammals, the fossils from the Cy-
cladic archipelago are limited to a few sporadic ndings (see Section 2)
and this is the main reason why they escaped attention. Originally, the
dwarf elephant of Naxos was attributed to Elephas antiquus melitensis
(Mitzopoulos, 1961), the species from Late Pleistocene Malta. This origi-
nal nomenclature is invalid because it can by no means be conspecic
with a dwarf elephant that is endemic to Sicily at the other side of
the Mediterranean. Exchange of genetic material between the
Siculo-Maltese and Cycladic populations could not have taken place.
The consensus view is that dwarf elephants are restricted in distribution
with each (palaeo-) island harbouring its own endemic species (Azzaroli,
1982; Herridge, 2010). That said, Naxos and adjacent parts cannot have
shared the same species with any other Mediterranean island.
In addition to the description and taxonomy of the mammal fauna,
new data are here put in a broader biogeographic context. Despite the
scantiness of the material, a re-appreciation within the latest develop-
ments in palaeogeography and palaeobiogeography is most needed.
Today all endemic mammals are extinct from the Cyclades. The islands
however, continue to host an interesting fauna of endemic reptiles and
invertebrates (Wettstein, 1953; Sfenthourakis and Legakis, 2001). The
Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Corresponding author at: Naturalis Biodiversity Center, Department of Geology, PO
Box 9517, 2300RA Leiden, The Netherlands. Tel.: +31 648088216.
E-mail addresses:,
(A.A.E. van der Geer).
0031-0182/© 2014 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology
journal homepage:
Author's personal copy
effect of the original palaeogeography continues to give a strong signal
to the remaining composition of Cycladic fauna (e.g. Sfenthourakis,
1996; Fattorini, 2002;Hausdorf and Hennig, 2005)andora (e.g. Bittkau
and Comes, 2005). The now extinct mammalian fauna is a lost compo-
nent of the Cycladic ecosystem. Therefore, any contribution towards a
better understanding of these missing elements is most valuable.
The scope of this contribution is to describe in detail the Late
Pleistocene mammals from Naxos, to address their taxonomy and to
infer palaeobiogeographical and palaeoecological implications.
2. History of fossil ndings from the Cyclades
The earliest nding is from Delos, from where Cayeux (1908)
reported the discovery of an isolated incomplete third upper molar
found near the Apollo temple in a deposit of the Inopos River. Cayeux
(1908) referred the specimen tentatively to Elephas antiquus Falconer
and Cautley, 1847, although he remarked that the interplate distance
is too large for this species. Vaufrey (1929: 126) lists the specimen
under E. antiquus race mnaidriensis based on size but depicts it
(Vaufrey, 1929: Fig. 38) as E. antiquus. The second fossil was found on
Naxos in sediments of the Trypiti River. The specimen, an upper jaw of
a dwarf elephant, was described by Mitzopoulos (1961) as
Palaeoloxodon antiquus melitensis. A tip of an elephant tusk was found
exposed in an ancient articial hollow on a valley slope in northwest
Kythnos (Honea, 1975). The larger portion of the tusk was still in situ
at the base ofa 4 m tick deposit of cemented breccia associated with fos-
sil bones and quartz tools at the timeof the report. The present situation
is unknown. The tusk's dating (9160 ± 240 C14 yrs BP) is likely much
underestimated as the specimen was air-exposed and subjected to
regular wetting due to seasonal rise in ground-water level (Honea,
1975). The tusk itself was never described and its size therefore
unknown. The rest of the fossil ndings from the Cyclades are merely
anecdotal reports without proper description or depiction and for
which in some cases any evidence is missing. These are unspecied
elephant fossils from Paros (E. antiquus in Georgalas, 1929), Milos
(dwarf elephant in Papp, 1953) and Seriphos (dwarf elephant in Papp,
1953;Elephas (Palaeoloxodon)cf.melitensis in Kuss, 1973). In the
1970s also a large Apodemus was found on Naxos near or at the elephant
locality (Sondaar, 1971), later referred to as Apodemus cf. mystacinus
(Dermitzakis and Sondaar, 1978). The same sample also contained
insectivore remains (labelled as Crocidura sp.), but these were not
reported in any publication. Kuss (1973) further mentioned the presence
of endemic deer (resembling Cervus (= Candiacervus)cretensis)and
some micromammals from Amorgos. The Cycladic ndings were
and Sondaar, 1978; Sondaar and Dermitzakis, 1982; Alcover et al.,
1998; Doukas and Athanassiou, 2003), without further examination
of any of these fossils or their context.
3. Materials and methods
3.1. Fossil material
The mammalian fossils described here are a complete elephant
maxilla (left and right side; AMPG), a number of murid and insectivore
teeth and fragmentary mandibles (GIU). All material originates from
Naxos (Cyclades, Greece). The specimen from Delos was not available
to us and the depiction of Vaufrey (1929: 131) is used instead. The
whereabouts of the Paros fossil(s) is uncertain as they were only briey
mentioned without institutional information in an article on extinct
elephants in a Greek encyclopaedia. Subsequent authors only quoted
this reference without tracking the fossil or its history. Equally uncertain
and without any formal description or institutional information are the
fossils from the remainder of the Cycladic islands (Milos, Kythnos,
Serifos, Amorgos). However, these specimens are less relevant to the
present discussion as these islands have an independent history (see
Section 3.3).
The elephant maxilla described here was found in the Trypiti
River, south of Cape Moutsouna (eastern Naxos), about 150 m from
the river mouth (Fig. 2). The murid and insectivore remains also
originate from the Pleistocene deposits exposed in a dry river in the
same area. Absolute datings are not available.
Fig. 1. Map of the Mediterranean (a) withSouthern Greece (b)indicating the islandsmentioned on the text.The dashed line is the 100-misobath and approximately indicatesthe extent of
the Cyclades Plateau. (c) Map of the central Cyclades corresponding to the 100-m isobath as given in (b). The grey areas correspond to the 50-misobath. During periods of 50 m sea-level
drop, Naxos clusters together with Paros. At periods of more extensive sea-level drops (e.g. 100 m), a single mega-island is formed. 1: Peloponnesos; 2: Kythera; 3: Sounio (province of
Attika);4: Kythnos; 5: Serifos; 6: Milos;7: Santorini (Thera); 8: Euboea(Evia); 9: Andros;10: Paros; 11: Delos; 12: Naxos;13: Amorgos; 14: Astypalaia; 15:Ikaria; 16: Kalymnos; 17: Tilos;
18: Rhodos; 19: Kasos; 20: Crete.
134 A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
3.2. Comparative material
The elephant material from Naxos is compared with the following
dwarf species: Palaeoloxodon creutzburgi (Crete, Late Pleistocene),
Palaeoloxodon tiliensis (Tilos, Late Pleistocene), P. sp. (Rhodos,
Pleistocene, probably Late Pleistocene in Symeonidis et al., 1974),
Palaeoloxodon mnaidriensis(San Teodoro, Sicily, Late Pleistocene),
Palaeoloxodon melitensis(Luparello, Sicily, Middle Pleistocene),
Palaeoloxodon falconeri (Spinagallo, Sicily, Middle Pleistocene), and
Palaeoloxodon cypriotes (Imbohary, Cyprus, Pleistocene). All are repre-
sented by dental material except for the species from Rhodos which is
known by tarsal bones only.
3.3. Palaeogeographic setting
During periods of low sea level of the Pleistocene substantial parts
of the now submerged Cycladic Plateau were exposed forming clusters
of larger islands or, at periods of extensive sea-level drop, a single
mega-island (Lambeck, 1996; Kapsimalis et al., 2009; Lykousis, 2009;
see Sections 6.2 and 6.3 for detailed information on the palaeogeography)
(Fig. 1). During such periods, Naxos was part of this large island here
referred to as palaeo-Cyclades. The current islands Paros and Delos
were part of the palaeo-Cyclades as well. The smaller Cycladic islands
Milos, Kythnos, Serifos and Amorgos were not connected to the
palaeo-Cyclades and may have formed islands on their own, harbouring
their own faunas. These islands will not be further considered here as
they bear no direct relation to the much larger palaeo-Cyclades of
which Naxos forms an important part.
3.4. Size changes
Liscaljet (2012) recently divided diminutive proboscidean species,
including those from the Aegean islands, into two size classes, based
on shoulder height. Species with a shoulder height between 90 and
200 cm were listed as pygmy species, those with a shoulder height
between 200 and 250 cm as dwarfed species based on the normal
height of 250350 cm for living elephants. However, the height of the
endemic form depends not only on the degree of body size reduction,
but also on the ancestral height (which may also have exceeded
350 cm as is the case with Palaeoloxodon antiquus). Two similar-sized
dwarf species may have followed (very) different trends of body size
decrease. Furthermore, this approach splits insular proboscideans into
two articial groups. Instead, here size index Si is used, dened as
estimated body mass of individuals from insular population divided by
the estimated body mass of individuals of the mainland or ancestral
form (see Lomolino et al., 2012, 2013). Body mass is based on the aver-
age of several individualsand calculated based on linear measurements
of postcranial material where possible; if no postcranial material was
available, molar dimension was used to estimate body mass (for details
see caption to Table 3). The advantage of this approach is that in this
way insular taxa can be compared between different families and
even orders.
3.5. Measurements
Metric characters, taken from Herridge and Lister (2012) are as
follows: molar width (W), molar length (L), molar height (H), enamel
thickness (ET), plate count (PC), and hypsodonty index (HI: crown
height/width × 100). Metric characters that differ from Herridge
and Lister (2012) are lamellar frequency (LF), corresponding to DLI
Fig. 2. Location of the Trypiti River andprole of the Pleistocenedeposits near the fossil site. The thickness of the Pleistocene deposits ranges between 0.5m and 5 m (Jansen, 1973). The
outcrop consists of a reddishpaleosol (Laterite), covered by an angular conglomerate, probably of riverine origin.
Table 1
Linear measurements of Palaeoloxodon from Naxos (AMPG 999) in mm. For indices and
ratios, see Table 2.
Length Width Crown height
Left molar 122 42.7 102.6
Right molar 128 47.6 93.6
Average 125 45.2 98.1
Table 2
Indices and ratios of third molars of Mediterranean dwarf Palaeoloxodon, compared to
their ancestor, Palaeoloxodon an tiquus. Notes: ET is given in mm. Data for Naxos are
from this study. For AMPG 999 the average of the two molars is given. 1) Mean values
from Herridge and Lister (2012);P. cypriotes from the type-locality Imbohary, Cyprus,
n=4;P. falconeri from Luparello and Spinagallo Caves, Sicily, n = 17; P. antiquus from
various UK and German sites, ca 500120 ka; n = 26. 2) Herridge, 2010. Weighted
mean for P. falconeri combined from Luparello and Spinagallo Caves. Spinagallo: PC
(n = 4) 13, LF (n = 8) 9.7, and ET (n = 9) 1.7, for HI we used W (n = 8) 30.2 and CH
(n = 5) 67.5 mm. Luparello: PC (n = 6) 11.8, LF (n = 11) 11.1, and ET (n = 10) 1.4,
for HI we used W (n = 10) 31.5 and CH (n = 9) 77. P. creutzburgi (n = 2) from eastern
Crete. P. tiliensis (n = 1) from Charkadio Cave, 3) Mean values from Herridge (2010),
based on the same sample as in Herridge and Lister (2012); n = 14 for PC, n = 26 for
LF, n = 21 for ET, and n = 11 for HI. 4) Data from Theodorou (1983) for T8209 (left; the
only complete upper third molar): L 124 mm, W 47 mm, H 86 mm, ET 2.52.8 mm, and
DLI ~9.5 Data from Vaufrey (1929). The specimen is incomplete (anterior part missing).
AMPG 999 (upper) 10.5 7.1 217 2.5
P. cypriotes
(lower) 11.5 9.6 180 1.2
P. cypriotes
(upper) 11 11.8 196 1.3
P. creutzburgi
(lower) 6.8 1.7
P. falconeri
(lower) 12.5 9.4 197 1.4
P. falconeri
(upper) 12.310.52371.5
P. antiquus
(lower) 18.5 4.7 221 2.0
P. antiquus
(upper) 18.3 6.2 210 1.9
P. tiliensis
(upper) 8.1 1.6
P. tiliensis
(upper) 11.5 c. 9 2.7
P. sp.
(Delos; upper) N8c.5––
135A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
(DezimeterLamellen-Interval, or number of plates per 10 cm) as used
in German literature.
3.6. Institutional abbreviations
AMPG Museum of Palaeontology and Geology of the University of
Athens, collection Vertebrates; IvAU Faculty of Geosciences, Department
of Earth Sciences, Utrecht University, Utrecht, The Netherlands; GPM
Gemellaro Museum, Palermo, Sicily, Italy; IPH Institut de Paléontologie
Humaine, Paris, France. NBC: Naturalis Biodiversity Center, Leiden, The
4. Results
4.1. Identication
First of all, theelephant maxilla AMPG 999 from Naxos includes third
molars and not second molars as Mitzopoulos (1961) assumed, based
among others on the posterior tapering of both the crown and the
root. The bone Mitzopoulos (1961) mistook for an alveolar rim for the
subsequent molar in development likely is the palatum. Mitzopoulos
(1961) though correctly noticed that the distal end of both molars
does not show any pressure facet caused by a subsequent molar. The
wrong diagnosis of the molar automatically inuences his interpretation
of relative size because third molars are always larger than second
molars in elephants.
4.2. Genus attribution
4.2.1. Palaeoloxodon versus Elephas
In the literature, straight-tusked elephants are placed either in the
genus Elephas or in the genus Palaeoloxodon. The cladistic analysis of
Todd (2010) gives support to two separate genera within the Elephas
group, an Elephas group including the living Asian elephant, and a
Palaeoloxodon group. Palaeoloxodon was earlier shown to be a mono-
phyletic clade on its own (Shoshani et al., 2005; Ferretti, 2008), which
is conrmed by morphological differences between the two genera
(e.g. Inuzuka, 1977a, b; Shoshani and Marchant, 2001; Shoshani et al.,
2005). In this contribution, Inuzuka and Takahashi (2004) and Shoshani
and Tassy (2005) are followed in which Palaeoloxodon is treated as dis-
tinct from Elephas. The dwarf elephants (not the mammoths) of the
Mediterranean may belong to either group in the analysis of Todd
(2010), implying two separate dwarng events. Until further analysis,
Palaeoloxodon antiquus is considered ancestral to all Mediterranean
dwarf elephants, with the exception of Mammuthus creticus (=Elephas
creticus in Bate, 1905)andMammuthus lamarmorai (= Elephas
lamarmorae in Major, 1883), based on morphological grounds (for
discussion and characters, see supporting information to Herridge and
Lister, 2012). This new analysis should take secondarily derived charac-
ters and changes due to allometry into consideration (see for a discus-
sion on convergent morphologies in unrelated insular taxa Van der
Geer, 2014); in fact, this applies to the type species Palaeoloxodon
naumanni as well, which is an insular species of Japan.
4.2.2. Mediterranean Palaeoloxodon
The conspecity between Palaeoloxodon melitensisfrom Luparello
(Sicily) and from its type locality Zebbug (Malta) is considered here as
unproven, and the species name is therefore placed between brackets
for the Sicilian material. The reasoning is that in Luparello cave the
layer with P. melitensisunderlies that with Palaeoloxodon falconeri
(Imbesi, 1956), and could thus have been ancestral to the smallest
species in this cave (Palombo, 2001; Herridge, 2010). The species
name of the larger-sized dwarf species of Sicily is here put between
quotes as well, as there is no proof that is conspecicwithPalaeoloxodon
mnaidriensis from the type locality Mnajdra Gap on Malta (Herridge,
2010; van der Geer et al., 2010). A revision of the Sicilian P. mnaidriensis
is in preparation (pers. comm. Victoria Herridge). To complicate matters
even further, the difference in size between P. mnaidriensis and
P. melitensis atonesideandbetweenP. melitensis and P. falconeri at
the other side (all from Malta) may not be sufcient to warrant specic
status of the middle-sized elephant, which would imply that either
P. melitensis is a junior synonym of P. falconeri or P. mnaidriensis is
a junior synonym of P. melitensis:E. falconeri,E. melitensis and
E. mnaidriensis were described by Busk (1867),Falconer (1868),and
Leith Adams (1870) respectively. A revision of the Sicilian and Maltese
dwarf elephants is, however, beyond the scope of this research and
therefore the current taxonomy is kept consistent with the literature.
Boekschoten and Sondaar (1972) and Davies and Lister (2001)
mentioned the presence of two pygmy elephants in Cyprus: one is
Palaeoloxodon cypriotes and the second is a larger, still unnamed, form.
This form, known from the locality Xylophagou (Iliopoulos et al.,
2011), might represent a second invasion or an earlier anagenenetic
form of P. cypriotes. Pending a detailed study and revision of the Cypriot
material, the larger form is referred here as P. cypriotes.
4.2.3. The Naxos dwarf elephant
Possible candidate taxa for the Naxos specimen are Palaeoloxodon,
Elephas,Mammuthus and Loxodonta. A taxonomically informative char-
acter in elephantids is the early-wear patterns of theenamel loop visible
on the occlusal surface. Elephas, although present during the Middle and
Late Pleistocene in the Levant (Lister et al., 2013) is no option because
the initial wear pattern progresses from a row of small rings at the
apex to three subequal ellipsoid forms (Lister et al., 2013), whereas
this pattern in the Naxos molar consists of an elongated central ellipsoid
anked by two subcircular rings at the lateral and medial sides. This
pattern is known as the typical Palaeoloxodon early wear pattern
(Herridge and Lister, 2012). In Mammuthus, a sub-circular mesial ring
is present between two elongated ellipsoids (more or less the opposite
of Palaeoloxodon). The mesial expansion is either absent or in the form
of rounded loops whereas it is triangular (pointed) in Palaeoloxodon
and the Naxos molars. In addition, Loxodonta has the typical loxodont
lozenge-shaped lamellar sinus, which is lacking in the Naxos molars.
Table 3
Body mass indices (Si) of Mediterranean dwarf elephant species with P. antiquus as
ancestor. Most body mass estimations were based on postcranial elements and were
calculated using the equations of Christiansen (2004). Three indices were based on
dental elements (here indicated with an asterisk; for details see relevant species).
P. cypriotes: no complete long bone belonging to this species is known. Based on the size
and morphology of available partial femurs (Davies and Lister, 2001 and a partial femur
in NBC) and the linear dimensions of its teeth we suggest that P. cyprio tes had
approximately the same size as P. falconeri.P. leonardi: based on its humerus length
(89 cm), we estimate the body mass of the P. leonardi as 4329 kg. Rhodos P. sp.: no
complete long bone is known. The mediolateral width of the distal diaphysis of the
femur is 17 cm (Symeonidis et al., 1974). The left femur of the mounted P. mnaidriensis
in MPG has a distal diaphysis of 17 cm width and a total length of 76 cm. Based on
these data we estimate a body mass of 1500 kg for the Rhodos elephant. Delos P. sp.:
this elephant is known only from a partial M3. Its size is comparable to that of
P. mnaidriensis from Puntali, based on which we suggest that it had approximately
the same body mass as that species. Naxos P. sp. nov: we assumed a simple geometric
relationship between tooth length and body size. The index is calculated as the ratio of
the cubed linear dimensions, following Lomolino (2005). Although this method likely
overestimates body size of large-sized insular forms, it appears reliable in small-sized
phylogenetic dwarfs (see for an extensive discussion on teeth and body size reduction of
phylogenetic dwarfs in Lister (1996). Note: * = body size estimated on dental elements.
Species Si Palaeo-island Source
Palaeoloxodon creutzburgi 0.38 Crete Lomolino et al, (2013)
Palaeoloxodon cypriotes0.07 Cyprus Lomolino et al. (2013)
Palaeoloxodon cypriotes ~0.02* Cyprus This study
Palaeoloxodon falconeri 0.02 Sicily Lomolino et al. (2013)
Palaeoloxodon leonardi 0.54 Sicily This study
Palaeoloxodon mnaidriensis0.17 Sicily Lomolino et al. (2013)
Palaeoloxodon sp. 0.19 Rhodos This study
Palaeoloxodon sp. ~ 0.17* Delos This study
Palaeoloxodon sp. nov. ~0.08* Naxos This study
Palaeoloxodon tiliensis 0.09 Tilos Lomolino et al. (2013)
136 A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
Mammuthus further lacks the longitudinally grooved aspect of the ends
of the enamel loops in lateral view as seen in the Naxos molar,
Palaeoloxodon and Elephas which is considered an expression of strong
enamel folding (Lister et al., 2013). The molars of Mammuthus
meridionalis are low-crowned (HI = 115175 in Van Essen, 2003),
whereas the Naxos molars are high-crowned (HI = 217). Finally, in
Elephas the enamel of the plates tends to be more intensively folded
than in Palaeoloxodon. In sum, the occlusal wear pattern of the Naxos
molars most closely resembles that seen in Palaeoloxodon.
Palaeoloxodon antiquus was the species present during the Late
Pleistocene in Greece (Dermitzakis and Theodorou, 1980; Athanassiou,
2000; Doukas and Athanassiou, 2003) and is as such the most likely
ancestor. The species was either already widespread during the Middle
Pleistocene (Athanassiou, 2000), or occurred for the rst time in Greece
at AmbeliaGrevena around the end of the late Middle Pleistocene
(Tsoukala et al., 2011).
4.3. Comparison with other insular Palaeoloxodon species
The Naxos elephant is readily distinguished from the mainland
species by the small size of its molars (Figs. 3 and 4). The Naxos molars
are just slightly larger than those of Palaeoloxodon cypriotes and
Palaeoloxodon falconeri and all three are a fraction of the size of their
respective ancestors. The Cretan dwarf species is the largest. The size
of the Delos specimen relative to the Tilos species is not entirely certain
because of the present unknown whereabouts of the fossil molar from
Delos. However, based on the width of 64 mm and a length of at least
155 mm, the Delos specimen would plot within the mnaidriensis
group. The Rhodos species is not known by dental remains.
The lamellar frequency of the Naxos specimen is comparable to that
of Palaeoloxodon falconeri, based on data given by Herridge and Lister
(2012, Fig. 2). However, the number of lamellae in a 10 cm length of
tooth (LF) is inversely related to crown width in mainland elephants
(Lister et al., 2013) and was shown to be of limited taxonomic value in
dwarf Palaeoloxodon taxa (Herridge and Lister, 2012; see also Table 2).
This also applies to PC. Scoring this type of variable is thus nothing
more than another approach to size. These data will therefore be lacking
for the here considered dwarf Palaeoloxodon taxa in scatter plots with
molar size and body size index. The only taxonomically informative
characters are limited to body size and perhaps hypsodonty index
and enamel thickness. Enamel thickness seems to be increased in the
Aegean taxa instead of decreased as in the Siculo-Maltese taxa.
When the Si of the Naxos specimen is compared to that of other
Mediterranean dwarf elephants of the genus Palaeoloxodon (Table 3),
the material from Naxos differs from Palaeoloxodon creutzburgi,
Palaeoloxodon mnaidriensisand P. sp. from Delos in its greater size
reduction (= lower Si). The Naxos specimen further differs from
Palaeoloxodon cypriotes and Palaeoloxodon falconeri (Spinagallo, Sicily)
in its lesser size reduction (=higher Si). It is comparable in size reduc-
tion to Palaeoloxodon tiliensis and the larger individuals of P. falconeri
(Luparello, Sicily).
4.4. A new species for the palaeo-Cyclades
The current view on the distribution of insular dwarf elephants is
that every island harbours its own endemic species (Doukas and
Athanassiou, 2003) as already suggested earlier (Sondaar, 1977;
Dermitzakis and Sondaar, 1978; Theodorou, 1983; Theodorou et al.,
2007). Species can thus not be shared by two islands unless these
islands were connected to each other previously during periods of low
sea-level and the period of isolation since the break-up of the islands
was insufcient for speciation. The reasoning behind this is that
dwarfed elephants are believed to have had a reduced their overseas
long-distance dispersal abilities. Due to their size reduction, the maxi-
mum distance they could have crossed up and down is reduced accord-
ingly, in comparison to mainland elephants. One-way long-distance
chance dispersals of size-reduced elephantids may have taken place,
giving rise to yet another speciation event, as probably in the case of
Stegodon orensis if it is derived from Stegodon sp. B from southern
Sulawesi (Van den Bergh et al., 2001) after crossing perhaps 100 km
(Van den Bergh, 1999). The limitations of overseas return possibilities
inevitably mean that even if two dwarf species are similar in size and
morphology and are derived from the same ancestor but they evolved
independently on two unrelated islands, they must be assigned to two
Fig. 3. Elephant maxillae from Naxos, Peloponnesus (continental P. antiquus,AMPG),and
Delos (drawing from Vaufrey, 1929). All drawn to the same scale.
Fig. 4. Sizes of upper third molars (M3) of insular dwarf elephants. The length of M3, as
depicted here, depends on the number of lamellae of each molar. As a consequence, the
molars that had lost some of their anterior lamellae appear to be shorter and relatively
broader than they should. This however, does not considerably affect the overall look
of the diagram, as it can still provide a very good indication of the size of each species.
Symbols: s: P. falconeri from Spinagalo, Sicily; l: P. melitensisfrom Luparello, Sicily; N:
P. lomolinoi nov. sp.; t: P. tiliensis from Tilos; p: P. mnaidriensisfrom Puntali (own data
from GPM); d: P. mnaidriensis from Ghar Dalam; a: mainland P. antiquus. Data sources:
Ambrosetti (1968) for P. falconeri,Theodorou (1983) for P. tiliensis,Herridge (2010) for
P. mnaidriensis (Ghar Dalam) and P. antiquus. Own data: P. melitensis, Luparello (IPH),
P. mnaidriensis,Puntali(GPM).
137A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
differentspecies as any genetic contact was unlikely to have taken place.
Based on similar body size of the dwarf elephants, the relevant islands
for Naxos hereare Tilos and Sicily, and the distance to be crossed greater
than 150 km, which is well beyond the swimming capacity of modern
Asian elephants (Johnson, 1978). For that reason a new species for the
dwarf elephant of Naxos is erected here.
5. Systematic palaeontology
5.1. The dwarf elephant from Naxos
Order PROBOSCIDEA Illiger, 1811
Family ELEPHANTIDAE Gray, 1821
Genus Palaeoloxodon Matsumoto, 1924
Type species Palaeoloxodon naumanni (Makiyama, 1924)
Palaeoloxodon lomolinoi sp. nov.
Holotype: AMPG 999 (Fig. 5), an isolated maxilla preserving both
third molars (M3). Curated at the Museum of Palaeontology and Geology
of the National and Kapodistrian University of Athens in Greece.
Synonymy:Elephas melitensis in Mitzopoulos, 1961: 334, textg. 1;
Palaeoloxodon antiquus melitensis in Mitzopoulos, 1961: 336; P. antiquus
cf. melitensis in Mitzopoulos, 1961: 340; a dwarf elephant in Sondaar,
1971:66;P. antiquus melitensis in Marinos and Symeonidis, 1976:354;
P. antiquus melitensis (lapsus calami) in Dermitzakis and Sondaar,
1978: 827; Elephas antiquus melitensis in Kotsakis, 1979: 36; elephant
similar in size to Elephas melitensis in Shoshani and Tassy, 1996: 238;
Elephas unnamedspeciesBinAlcover et al., 1998: 19; elephant probably
belonging to the palaeoloxodontine line in Palombo, 2001: 488; an
endemic species smaller than E. mnaidriensis in Palombo, 2004:
366367; Elephas sp. in Van der Geer et al., 2010:Plate6.
Type Locality: Trypiti river, Naxos, Greece.
Distribution and age: Late Pleistocene, Naxos, Greece.
Etymology: the species is named after Mark Lomolino, in honour of
his contribution to the knowledge of island biogeography.
Referred specimens: Holotype only.
Measurements:seeTables 1 and 2.
Diagnosis: a dwarf palaeoloxontine elephant from Naxos.
Differential diagnosis:Palaeoloxodon lomolinoi is about ten percent
the body mass of Palaeoloxodon antiquus, has a higher lamellar frequency
and thicker enamel. It differs from Palaeoloxodon creutzburgi,
Palaeoloxodon mnaidriensisand P. sp. from Delos in its greater size
reduction. It differs from Palaeoloxodon cypriotes and Palaeoloxodon
falconeri (Spinagallo, Sicily) in its lesser size reduction. Palaeoloxodon
lomolinoi has thicker enamel (ranging between 2 and 3 mm) than all
other Mediterranean dwarf Palaeoloxodon, except for Palaeoloxodon
Description of the holotype: (adapted from Mitzopoulos, 1961)an
exceptionally well-preserved partial upper jaw, consisting of the
maxillar bones and the complete palatum of an adult individual. Both
sides contain a fully erupted third molar in wear. The number of plates
is 10.5 of which two are not in wear. The most anterior plate is oval. The
following complete plate is merged with a worn central median pillar
that lies between the two anterior plates to form a Y-shaped structure.
The rest of the plates in wear have a rhomboid, lamellar form. The
width decreases with plate number. The enamel loops are simple,
cigar-shaped, scarcely folded with small mesial expansions. The occlu-
sal surface of the plates at the rst stage of wear is broken up into three
connected oval islets, somewhat obscured due to thepresence of matrix
on the specimen. The enamel is thick. The occlusal surface is oblong
with the proximal part wider than the distal part.
Remarks: originally, the material of Palaeoloxodon lomolinoi was
attributed to the Maltese species Palaeoloxodon antiquus melitensis by
Mitzopoulos (1961). However, since dwarf elephants are considered
to have had reduced overseas dispersal abilities (see discussion above)
and can thus not have reached the Cyclades from Sicily or Malta, the
Naxos dwarf elephant can therefore not have been co-specicwith
any Siculo-Maltese dwarf species. Palaeoloxodon lomolinoi has about
thesamesizeasPalaeoloxodon tiliensis and P. melitensis(Luparello,
Sicily). Palaeoloxodon lomolinoi is smaller than P. sp. (Delos), which
might be an older chronospecies. The individual age of the specimen
seems rather high, as about 80% of the plates are in wear, translating to
an age of above 50 Asian Elephant years (following Roth and Shoshani,
1988) assuming that wear proceeded with the same rate in insular
Palaeoloxodon as in extant Elephas.
5.2. The rock mouse from Naxos
Order RODENTIA Bowdich, 1821
Family MURIDAE Illiger, 1815
Genus Apodemus Kaup, 1829
Type species Apodemus agrarius (Pallas, 1771)
Apodemus cf. mystacinus (Danford & Alston, 1877) (Fig. 6)
Holotype: MCZ 14887, male, skin and skull of a male. Curated at the
Museum of Comparative Zoology of Harvard University of Cambridge
in the USA.
Type locality: Turkey, Adana Province, Bulgar Dagh Mt, Zebil.
Referred material: Thirteen isolated upper rst molars (NX1NX9,
NX40), four isolated upper third molars (NX41NX44), four isolated
lower rst molars (NX5154), two mandibles with rst and second
molars (NX56, NX58) (Fig. 6). All specimens are stored at IvAU.
Synonymy:largeApodemus in Sondaar, 1971: 66; big Apodemus sp. in
Kotsakis, 1979:36.
Dimensions: See Table 4.
Distribution and age: Late Pleistocene, Naxos, Greece.
Fig. 5. Occlusal and labial views of the holotype AMPG 999 of Palaeoloxodo n lomolinoi
sp. nov.
138 A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
Description: Upper dentition: on the M1, the t1 is positioned slightly
more posterior than the t3. Because of the strong connection betweent4
and t7, the posterior cusps form a continuous ridge from t6 to t9.A t12 is
clearly present as a spur of the t8 in all specimens but one, which is
relatively worn. In most M2, the t12 is a ridge-shaped extension of the
t8, but its development is more variable than in the M1. One M2 has a
clearly cusp-shaped t12, in two others it seems to be missing. The t1 is
only slightly larger than the t3.
Lower dentition: the m1 is mainly characterised by its strong labial
cingulum. A c1 is invariably present, as well as up to three accessory
cusplets, which, however, can also be incorporated in a ridge. The two
anterolophids are of similar size, the anterocentral cusp is small. The
anterolingual cusp of the m2 is strongly developed. Some of the m2
have a very strong labial cingulum, but in most it is reduced to some slight
patches. In two specimens, the cingulum bears to accessory cusplets.
Remarks:Sondaar (1971) and Kotsakis (1979) already noted the
presence of an Apodemus in the Naxos fauna. In the same sample
remains of a crocidurine shrew (Fig. 6) and a bat are present, which
will be published separately.
The Naxos Apodemus is a large species. In size, it agrees well with
measurements of Apodemus mystacinus as given by Van de Weerd
(1973),Mayhew (1978) and Niethammer (1978). Compared to the
latter two, the m1 from Naxos seems relatively short. In that respect
the material from Naxos metrically ts best with the sample form
Varkiza 1, Greece (Middle or Late Pleistocene; Van de Weerd, 1973).
The strong labial cingulum on the m1 and m2 also point to
A. mystacinus, although the development of accessory cusplets is less
than suggested by Niethammer (1978, Fig. 63).Moreover,Niethammer
(1978) diagnosed the species as having at least on accessory cuspule
on the m2, which is only present in part of the specimens in the Naxos
assemblage. This could be a difference between subspecies, as the typical
mystacinus and Apodemus m. epimelas do show some differences in
the dentition. According to Storch (1975), the t12 is incorporated
in the ridge formed by t9t8t7 in A. m. mystacinus, but is separate in
A. m. epimelas. This implies that the Naxos Apodemus can be placed in
the latter subspecies.
According to Masseti (2012),Apodemus mystacinus still occurs on
Naxos. However, neither Niethammer (1978),norStorch (2004)
includes the island in the distribution of the species. Storch (2004)
noted that the subspecies show a clear separation in the Aegean,
A. m. mystacinus occurring on the eastern islands, whereas Apodemus
m. epimelas is restricted to the west. Storch (2004) explained this differ-
ence by pointing out that A. mystacinus is a rock dweller, which would
be well adapted to cross rocky land bridges during sea-level lows.
Although, in contrast to Storch (2004), we do not want to rule out
rafting as a mechanism for dispersal for this species, his argument ts
well with nding the Greek subspecies on Naxos. After all, during sea-
level lows the island would be far easier accessible from the west than
from the east.
6. Discussion
6.1. Fossil insular elephants from the Southern Aegean Sea
The Cyclades does not form the only island complex in the Southern
Aegean Sea with insular proboscideans. A fossil dwarf mammoth lived
on Crete (Mammuthus creticus, earlyMiddle Pleistocene), and fossils
of endemic palaeoloxodontine elephants have been found on Crete
(Palaeoloxodon creutzburgi, Late Pleistocene), and the Dodecanese
islands, namely Tilos (Palaeoloxodon tiliensis, Late Pleistocene), Rhodos
(unnamed species; originally reported as Palaeoloxodon antiquus
mnaidriensis by Marinos and Symeonidis, 1973), Astypalaia (unnamed
dwarf elephant, reported as personal data in Doukas and Athanassiou,
2003) and Kasos (unnamed elephant, reported as personal data in
Masseti, 2012) (for overviews see Doukas and Athanassiou, 2003;
Masseti, 2009). Apart from these fossils, a number of elephant fossils
from southern Aegean islands belong to mainland P. antiquus.These
islands were connected to the mainland and include Kalymnos
(Masseti, 2002, 2006; specimen gured in Masseti, 2009,Fig.6),Kythera
(Petrochilos, 1938; Dermitzakis et al., 1982), Ikaria (Masseti, 2006,
2009), which are not considered here. Gökceada (= Imbros, Masseti,
2009) is part of the northern Aegean Sea and was connected to Asia
With the exception of the Cretan mammoth, all the other Aegean
proboscideans are palaeoloxodine elephants (pending a full description
of the larger elephant from Xylophagou, Cyprus). There are two hypo-
thetical phylogenies available for these elephants. The rst is that they
all result from independent colonisations from various mainlands and
evolved in parallel into dwarf forms(convergent evolution). The second
is that the islands of the Southern Aegean Sea acted as a speciation
region for elephants. After all, archipelagos have been characterised as
Fig. 6. Smallmammals from Naxos.Apodemus cf. mystacinus.1:NX2,leftupperrst molar
(M1 sin.); 2: NX 37, right upper second molar (M2 dext.; inverse); 3: NX 41, left upper
third molar (M3 sin.); 4: NX 56, mandible with rst and second mo lars (m1, m2).
Crocidura sp. 5: NX 137, left mandible with molar series (m1m3 sin.); 6: NX 87, left
maxillary with fourth premolar to second molar (P4M2 sin.). All materials from IvAU.
Table 4
Measurements of the occlusal surface of themolars of Apodemuscf. mystacinusfrom Naxos
(average between brackets), in mm.
Length Width n
M1 2.282.53 (2.43) 1.541.68 (1.61) 13
M2 1.451.68 (1.55) 1.451.59 (1.53) 8
M3 1.031.21 (1.10) 1.051.14 (1.10) 4
m1 2.112.30 (2.25) 1.401.47 (1.42) 6
m2 1.411.46 (1.44) 1.391.40 (1.40) 2
Lower toothrow 5.76
139A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
speciation machines(sensu Rosenzweig, 1995; see also Whittaker and
Fernández-Palacios, 2007). In this scenario a single mainland ancestor
colonises an island of the archipelago from where it successively
disperses to the other islands where further radiation takes place.
Here the rst scenario is adopted because in the case of the Southern
Aegean Sea we deal with a set of islands that is surrounded by vast
mainlands (Fig. 7). The distances between each (palaeo-)island and
the respective nearest mainland are a fraction of the inter-island dis-
tance and colonisation for the mainland is thus more likely than from
any other island. In addition, although the dispersal abilities of living
elephants are well known (Johnson, 1978), the same unlikely applies
to insular elephants. Due to reduced endurance as a result of their size
reduction, it is likely to assume that they could cover a fraction of the
distance their full-grown ancestors could cover. In addition, the loss of
a strong pneumatisation of the dwarf elephant skull (Palombo, 2001;
Van der Geer, 2014) and absolutely shorter trunk may have reduced
their swimming capacities even further. In all likelihood, the various
Southern Aegean dwarf species of Palaeoloxodon are the result of
independent colonisation events from the mainland.
6.2. Geological and palaeogeographical setting
The Cyclades Plateau is part of the AtticCycladic com plex, consisting
mainly of metamorphic and igneous rocks (Hejl et al., 2002), and divides
the northern Aegean Sea from the southern Aegean Sea (or Cretan Sea).
The numerous outcropping islands on the plateau today are the result
of its complex geomorphology (Kapsimalis et al., 2009). The Cyclades
Plateau was a practically aseismic region during the Late Pleistocene
and still is today (Papazachos, 1990). So for our purposes sea-level
changes are the only relevant factor in assessing its palaeogeography,
at least as far as the Late Pleistocene is concerned. The maximal depth
of the Cyclades Plateau is less than 250 m. During periods of low sea
level substantial parts of the now submerged Plateau were exposed
forming clusters of larger islands or, at periods of more extensive sea-
level drop, a single mega-island (Lambeck, 1996; Kapsimalis et al.,
2009; Lykousis, 2009) of about 10,000 km
(Kapsimalis et al., 2009).
Lykousis (2009) suggests that during the Middle Pleistocene, larger
areas were exposed subaerial, and, the Cyclades Plateau was connected
to the mainland of Eurasia during major glaciations. Lykousis (2009),
based on seismic data, argues that it was connected to both the Greek
mainland and Asia Minor during oxygen isotopic stages 10 and 12
(480350 kyrs BP) and stage 8 (300250 kyrs BP) and that the area
subsided progressively, implying that the connection remained intact
during isotopic stage 9 but with less land exposed above sea level.
During isotopic stage 6 (180140 kyrs BP), the land mass was separated
from Asia Minor but still connected to the mainland of Greece. Finally,
the Plateau formed a single mega-island at oxygen isotopic stage 2
(3018 kyrs BP) (Lykousis, 2009). These isotope stages correspond to
glacial periods and low sea levels. In between, sea level was (much)
higher and the emerged area (much) smaller. The breaking up of
the single island into the present numerous islands started around the
onset of the Holocene (ca 12 kyrs BP; Kapsimalis et al., 2009). The
model of Lykousis (2009) is in conict with the view that the seaways
between EuboeaAndros and KeaSounion were already open during
the early Pleistocene (Anastasakis et al., 2006), resulting in a signicantly
more insular condition of the Cycladic Plateau throughout the
Since the current knowledge of the palaeogeography of the Cyclades
cannot help us further, the maximum stratigraphic range of the fossil
taxa from Naxos is here used as a starting point. Apodemus mystacinus
is present in Europe and Asia Minor since the early Pleistocene
(Masseti, 2012) while Palaeoloxodon antiquus was present during the
Middle and Late Pleistocene (Athanassiou, 2000). This would imply a
Middle or Late Pleistocene age for the Naxos fauna (this also ts the
metrical values of the Naxos Apodemus, which indicate a Middle or Late
Pleistocene age as well). Most of the palaeoloxodontine phylogenetic
dwarfs are of Late Pleistocene age (e.g. Tilos, Crete, Rhodos and the
large-sized elephants of Sicily and Malta), there are, however, some
that have been attributed to the Middle Pleistocene (the smallest-sized
elephants of Sicily and Malta).
6.3. Colonisation window and subsequent evolution
Kurtén (1968) suggested that during glacial stages of the Pleistocene
the South European peninsulas served as refuges for Palaeoloxodon
antiquus, which is adapted to the temperate climates of the interglacial
stages. During such a colder stage sea-level drops are expected and the
palaeo-Cyclades was not only much closer to the mainland but also
much larger (Fig. 1). At excessive sea-level drops, the palaeo-Cyclades
was connected to the mainland of Greece via Euboea (Anastasakis and
Fig. 7. Palaeogeographic map of the Southern Aegean when the sea level was 100 m below present day sea-level. The map also indicates the possible dispersal routes of four insular
elephants and the overseas distances they wouldhave had to cross. In all cases the mainland was signicantly closer to each palaeo-island than the palaeo-islands from each other.
140 A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
Dermitzakis, 1990). At such moments, P. antiquus could have easily
reached the palaeo-Cyclades overland where it found sufcient area to
survive as the submerged shelf areas of today formed at plains then.
After submerging of the land connection, colonisation was still possible
but this time overseas by swimming as in the case of several other insu-
lar elephants (Van den Bergh, 1999; Van der Geer et al., 2010). Perhaps
independently, the murid and the insectivore reached the palaeo-island
supposedly by rafting. The nature of the ecological assemblage,
consisting solely of an elephant, a mouse and a shrew, strongly favours
overseas chance dispersal. When the climate became warmer again and
sea level rose, the degree of isolation increased and may have either led
or started a progressive body size reduction. Eventually the island may
have become too small to sustain a population of large herbivores
leading to the extinction of Palaeoloxodon lomolinoi. Other causes
which could have led to the extinction of thedwarf elephant are climate
change, volcanic activity e.g. of nearby Santorini (Thera), hunting
pressure exerted by humans, disease or a combination of factors, all of
which may be independent of sea-level changes. As to volcanic activity,
Theodorou (1988) suggested that the Minoan eruption of Santorini was
one of the factors that led the Tilos dwarf elephant to extinction. If this
applies to Naxos as well, it's impossible to pinpoint a precise point in
time, because the Santorini volcanic eld has had twelve major and
numerous minor eruptions over the last 200 ka (Druitt et al., 1989),
the effect of which on Naxos is unknown (no ash layers were found).
On the other hand, the Cyclades Plateau is unlike any other island in
the Mediterranean. Sea-level changes can dramatically alter its size
and topography. At low sea levels it forms a large island with extensive
at areas, but a modest sea-level rise leads to a fragmented landscape
consisting of many smaller islands with rather rough terrain.
Several possible episodes of isolation, dispersal, evolution and
extinction can be deduced from Lykousis' (2009) interpretation of
seismic data, where extinction can have been caused by (re)connection
to the mainland or by substantial submergence of habitat. Lykousis
(2009) does not specify how much of the Plateau around Naxos was
exposed during interglacials, so the most plausible colonisation window
has to be narrowed down by elimination.
First of all, should the palaeogeographic reconstruction of Lykousis
(2009) turn out to be incorrect and the palaeo-Cyclades was separated
from the mainland throughout the entire Pleistocene, then Palaeoloxodon
lomolinoi could have lived on Naxos during the colder stages of the
Middle Pleistocene as well. A low sea level is important for two reasons.
First, it minimises the distance between mainland and island, and thus
increases the chances of a successful overseas crossing. Second, it
increases the size of the palaeo-Cyclades by adding large territories
with minimal inclination that form suitable habitats for elephants.
This scenario would add isotope stages 16 (approx. 650 kyrs BP), 12
and 10 as possible colonisation and evolution window for P. lomolinoi.
In the case that the model of Lykousis (2009) is correct, the Middle
Pleistocene glacials have to be excluded because of their land connec-
tions. Furthermore, isotopic stage 9 is excluded on the ground of its
short duration (around 50 kyrs) during which the complete sequence
of isolation, successful sweepstakes colonisation and subsequent evolu-
tion of dwarsm before land connection was re-established should have
taken place. The data in Lykousis (2009) allow for at most a very short
period, if at all, of isolation. Isotopic stage 7 is somewhat longer (about
70 kyrs) but in our opinion still unlikely short. In addition, Lykousis
(2009) data indicate that the land connection with mainland Greece
through Attika and Euboea remained intact throughout.
Taking all evidence together, the only relevant period for evolution
of the Naxos fauna is a period when emerged landmasses formed a
large island, the palaeo-Cyclades, and where submerged areas of today
formed large plains and provided sufcient habitat to sustain an
elephant population. This could be approx. 650 kyrs BP, 480 krys BP or
between 140 and 30 kyrs BP. Available palynological data from the
study area narrow the colonisation window signicantly. The relevant
deposits are quite rich in pollen and represent at least 20 taxa
(Dalongeville and Renault-Miskovsky, 1993). The deposits cannot be
pinpointed in time but can most likely be attributed to the period
between 116 and 11.5 kyrs BP (Würm) and represent the different
stages of alternating open forest and arboreal steppes, or, in the case
of one particular deposit, perhaps between 238 and 128 kyrs BP (Riss;
Dalongeville and Renault-Miskovsky, 1993). A Late Pleistocene age, or
perhaps a latest Middle Pleistocene age, is in line with the geological
age of other Palaeoloxodon dwarfs from the Aegean and is not
contradicted by the micromammal data.
The specimen from Delos might represent an earlier stage of
isolation than the specimen from Naxos and in that case would consti-
tute a chronospecies of Palaeoloxodon lomolinoi. To rule out the option
of an earlier, independent colonisation followed by an independent
evolutionary process further study of the Delos specimen and its
context is needed. The species on Rhodos, Crete and Tilos on the other
hand are doubtlessly the result of independent colonisations by
Palaeoloxodon antiquus from different regions. Rhodos and Tilos
were colonised from Asia Minor, Crete from Greece through Kythera
and Antikythera, which were connected to each other and to the
The palaeo-Cyclades was not far removed from the mainland. A
moderate degree of isolation is suggestive of a high colonisation rate,
while a relatively large area can accommodate a high number of species
(Lomolino, 2000). Yet, the recorded biodiversity of the palaeo-Cyclades
is lower than that for Crete, an island with similar size and only slightly
higher degree of isolation during periods with low sea level during the
Late Pleistocene. Apart from a dwarf elephant, a large mouse and a
shrew, Crete also harboured several deer species and an otter during
the Late Pleistocene. This difference might simply be an artefact of the
lack of fossils since there are thousands of fossils from Crete but only a
handful from the Cyclades. A complete picture of the palaeobiodiversity
and biostratigraphy of the palaeo-Cyclades is thus currently impossible to
build and will perhaps never be. The available bits and pieces, however,
give some clues about biogeography. First of all, the palaeo-Cyclades
supported an elephant of about a quarter the body mass of that of
Crete during the same period. This small relative size may be related
to the absence of any competitor or predator as Lomolino et al. (2012)
found that body size evolution of mammals on islands is contextual,
and the absence or presence of competitors and predators is a main
factor. However, this is by no means certain when taking the fragmen-
tary fossil record of the palaeo-Cyclades into account. The smaller size
may also be related to island area, but such a relation has never been
proven for elephantids and is in conict with the fact that Tilos, which
is much smaller, harboured a similar-sized elephant. Secondly, the
meagre fauna consisting of a dwarf elephant, shrews and a eld
mouse puts the island on a par with relatively remote islands like
Middle Pleistocene Malta and Flores. As with the previous issue, this
may prove untrue in case more taxa are found.
6.4. Remarks on palaeo-dietary ecology and habitat
Limited ecological information, at least for the Late Pleistocene, can
be inferred from the fauna. The rock mouse is a species of primarily
deciduous forests and Mediterranean woods but prefers drier habitats
than most other Apodemus (Tchernov, 1986). It has been reported
from rocky areas of dry, open environments (Mayhew, 1978; Montuire
et al., 1994). Part of the palaeo-Cyclades was likely covered with decid-
uous forest and woodlands with, as can be inferred from the geology,
barren rocky areas as well. As to the large herbivore spectrum of the
fauna, the only component was Palaeoloxodon lomolinoi which might
have deviated ecologically from its ancestor Palaeoloxodon antiquus.
Palombo et al. (2005) and Palombo and Iacumin (2010) demonstrated
dietary plasticity for the latter species. Its microwear patterns indicate
wide dietary breadth and the ability to exploit feeding resources in a
large diversity of habitats (Rivals et al., 2012). Apart from that, the
European populations of this species show a trend towards an increased
141A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
amount of browse in its diet from the Middle Pleistocene (graze-
dominated mixed feeder) to the Late Pleistocene (Eemian interglacial,
MIS 5e; leaf browser) irrespective of latitude (Rivals et al., 2012). If
the founder population of the palaeo-Cyclades belonged to a Late
Pleistocene meta-population (see Section 6.3) and it would have been
adapted predominantly to leaf browsing or browse-dominated mixed
feeding. A higher amount of dietary plasticity is an advantage for a
species on an island simply because of the ability of niche broadening
without much need for adaptation. The slightly thicker enamel observed
in molars of dwarf species such as P. lomolinoi might be related to diet
but this relation has never been investigated systematically. Actually,
this might very well be linked to increased dietary breadth rather than
to niche narrowing and specialisation in parallel with the case of
increased tooth crown height (hypsodonty) in ungulate lineages
(Rivals et al., 2010).
Palynological data indicate a Late Pleistocene climate that was
warmer than the present with a Mediterranean character and either
drier or wetter than the present dependent on the stage (Dalongeville
and Renault-Miskovsky, 1993). The arboreal component differs per
deposit and varies between 40% (warm, wet) and 20% (warm, dry); in
one case 12% in combination with a high percentage of aquatic plants,
which is a local phenomenon.
The palaeo-Cyclades had a varied landscape with mountains and
valleys. The present-day islands were the highlands, but in between
there were low, large and relatively at plains with an average slope
of just 1.58° (Kapsimalis et al., 2009). Two major palaeo-plains can be
recognised on the palaeo-Cyclades: a central plain located between
the highlands of what are now SyrosDelosMykonos and Antiparos
ParosNaxos, and a southern plain between the highlands of Antiparos
ParosNaxos and FolegandrosSikoosIos (Kapsimalis et al., 2009).
These plains with their freshwater resources likely formed suitable
habitats for elephants.
7. Conclusions
1. During the Late Pleistocene, Naxos and adjacent areas formed a large
mega-island, referred to here as the palaeo-Cyclades.
2. Areas that are at present submerged formed extensive low-lying
plains with lakes and rivers, providing suitable habitats for large
3. Fossils of a dwarf elephant, formally named here Palaeoloxodon
lomolinoi, a rock mouse (Apodemus cf. mystacinus), shrews and bats
were found on Naxos in Late Pleistocene sediments.
4. The dwarf elephant from Naxos evolved miniature size due to long-
term isolation. It had a body mass of about 10% of that of its mainland
5. The various Aegean insular dwarf species each result from indepen-
dent colonisations.
We thank Hans de Bruijn and Olaf Schuiling for their help with the
collection history of the small mammal remains from Naxos. We further
wish to thank Pascal Tassy and Isabelle Rouget for their search for
the cast of the Delos specimen in the Paris collections, Athanassios
Athanassiou for providing details concerning the specimen from Paros
and the occurrence of P. antiquus in Greece, George Anastasakis for the
fruitful discussions on the palaeogeography of Cyclades, Hans de Bruijn
for his measurements of Apodemus, George Iliopoulos for information
on the Cypriot elephants and Marco Masseti, Victoria Herridge and
Maria Rita Palombo for sharing their views on the evolution of dwarf
elephants with us. We are grateful to Astrid Scharlau and Nikos
Mandilaras for indicating to us the fossiliferous locality and to them
and Anna Heijstee for their assistance during eldwork. We thank, in al-
phabetical order of institutional abbreviation, George Theodorou
(AMPG), Wilma Wessels (GIU), Carolina di Patti (GPM) and Thomas
Ingicco (IPH), and Reinier van Zelst (NBC) for access to collections in
their care. Finally, we thank two anonymous reviewers and the editor
for their constructive comments which greatly improved the manu-
script. This research has been co-nanced by the European Union
(European Social Fund ESF) and Greek national funds through the Op-
erational Program Education and Lifelong Learningof the National
Strategic Reference Framework (NSRF) Research Funding Program:
THALIS UOA-Island biodiversity and cultural evolution: examples
from the Eastern Mediterranean, Madagascar, Mauritius and
Philippines during the past 800,000 years (MIS375910, KA:70/3/
Alcover, J.A., Campillo, X., Macias, M., Sans, A., 1998. Mammal species of the world:
additional data on insular mammals. Am. Mus. Novit. 3248, 129.
Ambrosetti, P., 1968. The Pleistocene dwarf elephants of Spinagallo (Siracusa, South-
Eastern Sicily). Geol. Romana 1, 217229.
Anastasakis, G.C., Dermitzakis, M., 1990. Post middle Miocene paleogeographic evolution
of the central Aegean Sea and detailed Quaternary reconstruction of the region: its
possible inuence on the distribution of the Quaternary mammals of the Cyclades
Islands. Neues Jb. Geol. Paläontol. Monat. 1990, 116.
Anastasakis, G., Piper, D.J.W., Dermitzakis, M.D., Karakitsios, V., 2006. Upper Cenozoic
stratigraphy and paleogeographic evolution of Myrtoon and adjacent basins, Aegean
Sea, Greece. Mar. Pet. Geol. 23, 353369.
Athanassiou, A., 2000. Presence of fossil elephants in the area of Penios Valley (NW
Peloponnesus, Greece). Ann. Géol. Pays Hellén. 38, 6376.
Azzaroli, A., 1982. Insularity and its effects on terrestrial vertebrates: evolutionary and
biogeographic aspects. In: Montanaro Gallitelli, E. (Ed.), Palaeontology: Essential of
Historical Geology. S.T.E.M., Mucchi, Modena, pp. 193215.
Bate, D.M.A., 1905. Four and a half months in Crete in search of Pleistocene mammalian
remains. Geol. Mag. 2 (5), 193202.
Bittkau, C., Comes, H.P., 2005. Evolutionary processes in a continental island system:
molecular phylogeography of the Aegean Nigella arvensis alliance (Ranunculaceae)
inferred from chloroplast DNA. Mol. Ecol. 14, 40654083.
Boekschoten, G.J., Sondaar, P.Y., 1972. On the fossil mamalia of Cyprus, I and II. Proc. K.
Ned. Akad. Wet. B 75 (4), 306338.
Busk, G., 1867.Description of the remains of three extinctspecies of elephant, collected by
Capt. Spratt, C.B.R.N., in the ossiferous cavern of Zebbug, in the island of Malta. Trans.
Zool. Soc. Lond. 6, 227306.
Cayeux, L., 1908. Découverte de l'Elephas antiquus à l'île de Délos (Cyclades). C.R. Hebd.
Seances Acad. Sci. 147, 10891090.
Christiansen, P., 2004. Body size in proboscideans, with notes on elephant metabolism.
Zool. J. Linnean Soc. 140, 523549.
Dalongeville, R., Renault-Miskovsky, J., 1993. Paysages passés et actuels de l'île de Naxos.
Recherches dans les Cyclades, Résultat des travaux de la RCP 583, Collection de la
Maison de l'Orient méditerranéen, Série archéologique. Maison de l'Orient et de la
Méditerranée Jean Pouilloux, Lyon, pp. 957.
Davies, P., Lister, A.M., 2001. Palaeoloxodon cypriotes, the dwarf elephant of Cyprus: size and
scaling comparisons with P. falconeri (SicilyMalta) and mainland P. antiquus. In:
Cavarretta, G., Gioia, P., Mussi, M., Palombo, M.R. (Eds.), The World of Elephants, Pro-
ceedings of the 1st International Congress, Roma, 1620 Ottobre 2001, pp. 479480.
Dermitzakis, M.D., Sondaar, P.Y., 1978. The importance of fossil mammals in reconstruc-
tion paleogeography with special reference to the Pleistocene Aegean Archipelago.
Ann. Géol. Pays Hellén. 46, 808840.
Dermitzakis, M.,Theodorou, G.,1980. Map of themain fossiliferous localities of Proboscidea
in Aegean area. Editions of the Laboratory of Geology and PalaeontologyUniversity of
Athens, Athens.
Dermitzakis, M.D., Symeonidis, N.K., De Boer, L.E.M., Sondaar, P.Y., 1982. The Evolution
of Elephants. Edition of the Laboratory of Geology and Palaeontology. University of
Athens, Athens (in Greek).
Doukas, C.S., Athanassiou, A., 2003. Review of the Pliocene and Pleistocene Proboscidea
(Mammalia) from Greece. Deinsea 9, 97110.
Druitt, T.H., Mellors, R.A., Pyle, D.M., Sparks, R.S.J., 1989. Explosive volcanism on Santorini,
Greece. Geol. Mag. 126 (2), 95126.
Falconer, H., 1868. On the fossil remains of E. melitensis. An extinct pygmy species of
elephant and other Mammalia from the ossiferous caves of Malta. Palaeontol. Mem.
Notes 2, 292307.
Fattorini, S., 2002. Biogeography of the tenebrionid beetles (Coleoptera, Tenebrionidae)
on the Aegean Islands (Greece). J. Biogeogr. 29, 4967.
Ferretti, M., 2008. The dwarf elephant Palaeoloxodon mnaidriensis from Puntali Cave,
Carini (Sicily; Late Middle Pleistocene): anatomy, systematics and phylogenetic
relationships. Quat. Int. 182, 90108.
Georgalas, G.C., 1929. Elephas (palaeozoology). Megali Helliniki Egkyklopaedia 9, 945946
(In Greek).
Hausdorf, B., Hennig, C., 2005. The inuence of recent geography, palaeogeography and
climate on the composition of the fauna of the central Aegean Islands. Biol. J. Linn.
Soc. 84, 785795.
Hejl, E., Riedl, H., Weingartner, H., 2002. Post-plutonic unroong and morphogenesis of
the AtticCycladic complex (Aegean, Greece). Tectonophysics 349, 3756.
Herridge, V., 2010. Dwarf Elephants on Mediterranean Islands: A Natural Experiment in
Parallel Evolution. (PhD Thesis) University College, London.
142 A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
Herridge, V.L., Lister, A.M., 2012. Extreme insular dwarsm evolved in a mammoth. Proc.
R. Soc. B 279 (1741), 31933200. (22).
Honea, K., 1975. Prehistoric remains on the island of Kythnos. Am. J. Archaeol. 79,
Iliopoulos, G., Athanassiou, A., Konstantinou, G., 2011. New dwarf elephant material from
the Pleistocene of Cyprus. In: Van der Geer, A., Athanassiou, A. (Eds.), Program and
Abstracts, European Association of Vertebrate Palaeontologists 9th Annual Meeting,
Heraklion, Crete, Greece 1419 June, 2011. Natural History Museum of Cret e,
Heraklion, p. 30.
Imbesi, M.,1956. Sugli elefanti nanidella grotta di Luparello, Palermo.In: Blanc, G.A. (Ed.),
Actes du 4. Congres International du Quaternaire, Rome-Pise, Aout-Septembre 1953.
Istituto Italiano di Paleontología Umana, Rome, pp. 443447.
Inuzuka, N., 1977a. On a fossil skull of Palaeoloxodo n naumanni from Saruyama, Shimosa-
machi, Chiba Prefecture, central Japan. J. Geol. Soc. Jpn. 83, 523536.
Inuzuka, N., 1977b. On the origin of Palaeoloxodon naumanni a comparative osteology
of the cranium. J. Geol. Soc. Jpn. 83, 639655.
Inuzuka, N., Takahashi, K., 2004. Discrimination between the genera Palaeoloxodon and
Elephas and the independent taxonomical position of Palaeoloxo don (Mammalia:
Proboscidea). In: Baquedano Pérez, E., Rubio Jara, S. (Eds.), Miscelánea en homenaje
a Emiliano Aguirre II. Museo Arqueologico Regional, Madrid, pp. 234244.
Jansen, J.B.H., 1973. Geological map of Naxos(1/50 000). NationalInstitute of Geologyand
Mining Research, Athens, 1 sheet.
Johnson, D.L., 1978. The origin of island mammoths and the Quaternary land bridge
history of the Northern Channel Islands, California. Quat. Res. 10, 204225.
Kapsimalis, V., Pavlopoulos, K., Panagiotopoulos, I., Drakopoulou, P., Vandarakis, D.,
Sakelariou, D., Anagnostou, C., 2009. Geoarchaeological challenges in the Cyclades
continental shelf (Aegean Sea). Z. Geomorphol. N.F. 53, 169190.
Kotsakis, T., 1979. The Quaternary vertebrates on the Aegean islands: palaeobiogeographical
implications. Ann. Géol. Pays Hellén. 30, 3164.
Kurtén, B., 1968. Pleistocene Mammals of Europe. Aldine Publishing Company, Chicago,
Kuss, S.E., 1973. Die pleistozänen Säugetierfaunen der ostmediterranen Inseln. Ber.
Naturforsch. Ges. Freibg. Br. 93, 4971.
Lambeck, K., 1996. Sea-level change and shoreline evolution in Aegean Greece since
Upper Palaeolithic times. Antiquity 70, 588611.
Leith Adams, A., 1863. Observations on the fossiliferous caves of Malta. J. R. Soc. 4 (2),
Leith Adams, A., 1870. Notes of a Naturalist in the Nile Valley and Malta. Edmonton and
Douglas, Edinburgh.
Liscaljet, N., 2012. Napakaliit trompa: new pygmy proboscidean remains from the
Cagayan Valley (Philippines). Quat. Int. 276277, 278286.
Lister, A.M., 1996. Dwarng in island elephants and deer: processes in relation to time of
isolation. Symp. Zool. Soc. Lond. 69, 277292.
Lister, A.M., Dirks, W., Assaf, A., Chazan, M., Goldberg, P., Applbaum, Y.H., Greenbaum, N.,
Horwitz, L.K., 2013. New fossil remains of Elephas from the southern Levant: implica-
tions for the evolutionary history of the Asian elephant. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 386, 119130.
Lomolino, M.V., 2000. Ecology's most general, yet protean pattern: the speciesarea
relationship. J. Biogeogr. 27, 1726.
Lomolino, M.V., 2005. Body size evolution in insular vertebrates: generality of the island
rule. J. Biogeogr. 32, 16831699.
Lomolino, M.V., Sax, D.F., Palombo, M.R., van der Geer, A.A.E., 2012. Of mice and
mammoths: evaluations of causal explanations for body size evolution in insular
mammals. J. Biogeogr. 39, 842854.
Lomolino, M.V., van der Geer, A.A., Lyras , G.A., Palombo, M.R., Sax, D.F., Rozzi, R., 2013. Of mice
and mammoths: generality and antiquity of the island rule. J. Biogeogr. 40, 14271439.
Lykousis, V., 2009. Sea level changes and shelfbreak prograding sequences during the last
400 ka in the Aegean margins: subsidence rates and palaeogeographic implications.
Cont. Shelf Res. 29, 20372044.
Major, C.I.F., 1883. Die Tyrrhenis: Studien über geographischeVerbreitung von Tieren und
Panzen im westlich Mittelmeergebiet. Kosmos 13 (117), 81106.
Makiyama, J., 1924. Notes on a fossil elephant from Sahamma, Tôtômi. Mem. Coll. Sci.
Kyoto Imp. Univ. B 1 (2), 255264.
Marinos, G., Symeonidis, N., 1973. Erstmalige Funde von Zwergelefanten auf dem Insel
Rhodos. Anz. Math. Naturwiss. Kl. Österr. Akad. Wiss. 10, 129131.
Marinos, G., Symeonidis, N., 1976. Island populations of dwarf mammals on the Aegean
archipelago during Quaternary. Ann. Géol. Pays Hellén. 28, 352367 (In Greek).
Masseti,M., 2002. A glimpse of palaeontology. In: Masseti, M. (Ed.), Island of Deer. Natural
History of the Fallow Deer of Rhodes and of the Vertebrates of the Dodecanese
(Greece). City of Rhodes, Environmental Organization, Rhodes, pp. 3540.
Masseti, M., 2006. Deer and dwarf elephants of the Dodecanese. In: Karageorghis, V.,
Giannikouri, A. (Eds.), The Holocene redenition of the faunal ecosystems of Rhodes
and other nearby islands. Conservation and Preservation of the Cultural and Natural
Heritageof the Large Islands of the Mediterranean. Ministryof Culture, Archaeological
Institute of Aegean Studies, Foundation Anastasios G. Leventis, Athens, pp. 113122.
Masseti, M., 2009. Mammals of the Mediterranean islands: homogenisation and the loss
of biodiversity. Mammalia 73, 169202.
Masseti, M., 2012. Atlas of Terrestrial Mammals of the Ionian and Aegean Islands. Walter
de Gruyter GmbH, Berlin/Boston.
Mayhew, D.E., 1978. Late Pleistocene small mammals from Arnissa (Macedonia, Greece).
Proc. K. Ned. Akad. Wet. B. 81 (3), 302321.
Mitzopoulos, M.K., 1961. Über einen pleistozänen Zwergelefanten von der Insel Naxos
(Kykladen). Prak. Akad. Athenon 36, 332340.
Montuire, S., Sen, S., Michaux, J., 1994. The Middle Pleistocene mammalian fauna from
Emirkaya-2, Central Anatolia (Turkey): systematics and paleoenvironment. N. Jb.
Geol. Paläont. 193, 107144.
Niethammer, J., 1978. Apodemus mystacinus (Danford and Alston, 1877) Felsenmaus.
In: Niethammer, J., Krapp, F. (Eds.), Handbuch der Säugetiere Europas, Band 1.
Akademische Verlagsgesellschaft, Wiesbaden, pp. 306315.
Pallas, P.S., 1771. Reise durch verschiedene Provinzen des Russischen Reichs, vol. I. Saint
Petersburg p. 454.
Palombo, M.R., 2001. Paedomorphic features and allometric growth in the skullof Elephas
falconeri from Spinagallo (Middle Pleistocene, Sicily). In: Cavaretta, G., Gioia, P.,
Mussi, M., Palombo, M.R. (Eds.), The World of Elephants. Proceedings of the 1st
International Congress, Roma, 1620 October 2001. Consiglio Nazionale delle
Ricerche, Rome, pp. 492496.
Palombo, M.R., 2004. Dwarng in insular mammals: the endemic elephants of
Mediterranean islands. In: Baquedano Pérez, E., Rubio Jara, S. (Eds.), Miscelánea en
homenaje a Emiliano Aguirre II. Museo Arqueologico Regional, Madrid, pp. 354371.
Palombo, M.R., 2007. How can endemic proboscideans help us understand the island
rule? A case study of Mediterranean islands. Quat. Int. 169170, 105124.
Palombo, M.R., Iacumin, P., 2010. What did straight-tusked elephants eat? Quatern. HS 3,
Palombo, M.R., Filippi, M.L., Iacumin, P., Longinelli, A., Barbieri, M., Maras, A., 2005.
Coupling tooth Microwear and stable isotope analysis for palaeodiet reconstruction:
the case study of Late Middle Pleistocene Elephas (Pa laeoloxodon)antiquus teeth
from Central Italy (Rome area). Quat. Int. 126128, 153170.
Papazachos, B.C., 1990. Seismicity of the Aegean and surrounding area. Tectonophysics
178, 287308.
Papp, A., 1953. Die paläogeologische Entstehung der Agäis nach dem derzeitigen Stand
unserer Erkentnisse. In: Wettstein, O. (Ed.), Herpetologia Aegaea. Österreichische
Akademie der Wissenschaften, Vienna, pp. 815818.
Petrochilos, J., 1938. Découverte de l'Elephas antiquus dans l'île de Cythère et l'âge de sa
séparation du continent. CR Somm. Soc. Géol. Fr. 1938, 5960.
Rivals, F., Semprebon, G.M., Solounias, N., 2010. Advances in ungulate dental wear
techniques reveal new patterns of niche breadth and expansion throughout the
Cenozoic. J. Vertebr. Paleontol. 2010, 151A152A (SVP Program and Abstracts Book).
Rivals, F., Semprebon, G., Lister, A., 2012. An examination of dietary diversity patterns in
Pleistocene proboscideans (Mammuthus,Palaeoloxodon,andMammut)fromEurope
and North America as revealed by dental microwear. Quat. Int. 255, 188195.
Rosenzweig, M.L., 1995. SpeciesDiversity in Space and Time. Cambridge University Press,
Roth, V.L., Shoshani, J., 1988. Dental identication and age determination in Elephas
Sciná, D., 1831. Rapporto sulle ossa fossili di Mardolce e degli altri contorni di Palermo.
Reale Tipograa di Guerra, Palermo.
Sfenthourakis, S., 1996. A biogeographic analysis of terrestrial isopods (Isopoda,
Oniscidea) from central Aegean islands (Greece). J. Biogeogr. 23, 687698.
Sfenthourakis, S., Legakis, A., 2001. Hotspots of endemic terrestrial invertebrates in
southern Greece. Biodivers. Conserv. 10, 13871417.
Shoshani, J., Marchant, G.H., 2001. Hyoid apparatus: a little known complex of bones and
its contributionto proboscidean evolution. In: Cavaretta, G., Gioia, P., Mussi, M.,
Palombo, M.R. (Eds.), The world of elephants. Proceedings of the 1st International
Congress, Roma, 1620 October 2001. Consiglio Nazionale delle Ricerche, Rome,
pp. 668675.
Shoshani, J., Tassy, P. (Eds.), 1996. The Proboscidea: Evolution and Palaeoecology of
Elephants and Their Relatives. Oxford University Press, Oxford, New York, Tokyo.
Shoshani, J., Tassy, P., 2005. Advances in proboscidean taxonomy& classication,anatomy
& physiology, and ecology & behavior. Quat. Int. 126128, 520.
Shoshani, J., Ferretti, M., Lister, A.M., Saegusa, H., Agenbroad, L.D., Mol, D., Takahashi, K.,
2005. On the relationships within the Elephantinae using hyoid characteristics. In:
Agenbroad, L.D., Symington, R.L. (Eds.), The world of elephants. Mammoth Site
Scientic Papers, 4. Mammoth Site, Hot Springs, S. Dakota, pp. 160165.
Sondaar, P.Y., 1971. Palaezoogeography of the Pleistocene mammals from the Aegean. In:
Strid, A. (Ed.), Evolution in the Aegean. Opera Botanica, 30, pp. 6570.
Sondaar,P.Y., 1977. Insularityand its effect on mammal evolution.In: Hecht, M.N., Goody,
P.L., Hecht, B.M. (Eds.), Major Patterns in Vertebrate Evolution. Plenum, New York,
pp. 671707.
Sondaar, P.Y., Dermitzakis, M.D., 1982. Relation migrationland vertebrates, paleogeography
and tectonics. In: Pichon, X.L., Augustidis, S.S., Mascle, J. (Eds.), International
Symposium on the Hellenic Arc and Trench (H.E.A.T.), April 810 1981. Athens, pp.
Spratt, T.A.B., 1867. On the bone-caves near Crendi, Zebbug, and Melliha, in the Island of
Malta. Q. J. Geol. Soc. 23, 283297.
Storch, G., 1975. Eine mittelpleist ozäne Nager-Fauna von der Insel Chios, Ägäis
(Mammalia, Rodentia). Senckenberg. Biol. 56, 165189.
Storch, G., 2004. Late Pleistocene Rodent dispersal in the Balkans. In: Grifths, H.I.,
Krystufek, B., Reed, J.M. (Eds.), Balkan Biodiversity: Pattern and Process in the
European Hotspot. Kluwer Academic Publishers, Dordrecht, pp. 135145.
Symeonidis, N.K., Bachmayer, F., Zapfe, H., 1974. Entdeckung von Zwergelefanten auf der
Insel Rhodos (Ausgrabungen 1973). Ann. Naturhist. Mus. Wien 78, 193202.
Tchernov, E., 1986. The rodents and lagomorphs from UbeidiyaFormation: systematics,
paleoecology and biogeography. In: Tchernov, E. (Ed.), The Lower Pleistocene
Mammals of Ubeidiya(Jordan Valley). Association Paleorient, Paris, pp. 235350.
Theodorou, G.E., 1983. The Dwarf Elephants of the Charkadio Cave on the Island of Tilos
(Dodekanese, Greece). (PhD thesis) University of Athens (In Greek).
Theodorou, G.E., 1988. Environmental factors affecting the evolution of island endemics:
the Tilos example from Greece. Mod. Geol. 13, 183188.
Theodorou, G., Symeonidis, N., Stathopoulou, E., 2007. Elephas tiliensis n. sp. from Tilos
island (Dodecanese, Greece). Hell. J. Geosci. 42, 1932.
Todd, N.E., 2010. New phylogenetic analysis of the family Elephantidae based on cranial
dental morphology. Anat. Rec. 293, 7490.
143A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
Author's personal copy
Tsoukala, E., Mol, D., Pappa, S., Vlachos,E., van Logchem, W., Vaxevanopoulos, M., Reumer,
J., 2011. Elephas antiquus in Greece: new nds and a reappraisal of older material
(Mammalia, Proboscidea, Elephantidae). Quat. Int. 245, 339349.
Van de Weerd, A., 1973. Rodentia from two Pleistocene ssuresllings near Athens. Proc.
Ned. Akad. Wet. B 76, 148166.
Van den Bergh, G.D., 1999. The Late Neogene elephantoid-bearing faunas of Indonesia
and their palaeozoogeographic implications; a study of the terrestrial faunal succes-
sion of Sulawesi, Flores and Java, including evidence for early hominid dispersal
east of Wallace's line. Scr. Geol. 117, 1419.
Van den Bergh, G.D., de Vos, J., Sondaar, P.Y., 2001. The late quaternary palaeogeography
of mammal evolution in the Indonesian Archipelago. Paleontol. Palaeoclimatol.
Palaeoecol. 171, 385408.
Van der Geer, A.A.E., 2014. Parallel patterns and trends in functional structures inextinct
island mammals. Integr. Zool. 9, 165180.
Van der Geer, A., Lyras, G., de Vos,J., Dermitzakis, M., 2010. Evolution of Island Mammals:
Adaptation and Extinction of Placental Mammals on Islands. Wiley-Blackwell, Oxford.
Van Essen, H., 2003. Tooth morphology of Mammuthus meridionalis from the southern bight
of the North Sea and from several localities in the Netherlands. In: Reumer, J.W.F., de
Vos, J., Mol, D. (Eds.), Advances in Mammoth Research (Proceedings of the Second Inter-
national Mammoth Conference, Rotterdam, May 1620 1999). Deinsea, 9, p p. 453511.
Vaufrey, R., 1929. Les éléphants nains des îles méditerranéennes et la question des
isthmes pléistocènes. Arch. Inst. Paleon. Hum. Mem. 6, 1220.
Wettstein, O., 1953. Herpetologia Aegaea. Österreichische Akademie der Wissenschaften,
Whittaker, R.J., 1998. Island Biogeography: Ecology, Evolution and Conservation. Oxford
University Press, Oxford.
Whittaker, R.J., Fernández-Palacios, J.M., 2007. Island Biogeography: Ecology, Evolution,
and Conservation, 2nd edn. Oxford University Press, Oxford.
144 A.A.E. van der Geer et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 404 (2014) 133144
... Remains of more or less dwarfed Pleistocene elephants are known from the following Aegean islands: Astypalea, Crete, Dilos, Kalymnos, Kassos, Kos, Kythera, Kythnos, Milos, Naxos, Paros, Rhodos, Serifos and Tilos (Doukas and Athanassiou 2003, Tsoukala et al. 2011, Masseti 2012, Sen et al. 2014, Van der Geer et al. 2014. These islands are all in the southern Aegean, and some of them are part of the Hellenic arc (Crete, Kassos, Kythera and Rhodos) whereas others islands belong to the Cycladic arc. ...
... These islands are all in the southern Aegean, and some of them are part of the Hellenic arc (Crete, Kassos, Kythera and Rhodos) whereas others islands belong to the Cycladic arc. The presence of Pleistocene elephants in the islands of Astypalea, Kythnos, Milos, Paros and Serifos is mentioned in the literature (see in particular Kotsakis 1990, Doukas andAthanassiou 2003, and references therein) but without any description or illustration (see a recent review in Van der Geer et al. 2014). The elephant fossils from the other islands are briefly described herein in alphabetic order. ...
... Fieldwork carried out during the last two decades did not provide any additional fossil material, except for "a part of deciduous molar of a dwarf elephantid" from the Koutalas Cave (Chania) that Iliopoulos et al. (2010: 5) referred to Elephas sp. The systematic status of this large elephant from Crete was not discussed in later studies dealing with the Crete elephants, such as by Van der Geer and Lyras (2011), Herridge and Lister (2012), Lomolino et al. (2013) and Van der Geer et al. (2014, 2016. Consequently, we suggest to maintain the species distinction of "E. ...
Full-text available
This paper provides a synthesis of the present knowledge on dwarfed endemic elephants from the Pleistocene of the south Aegean islands. Pleistocene elephants are quite well documented from Crete and Tilos, but with scarce remains on other islands. The systematics and affinities of these elephants are discussed here in the light of recent knowledge on their dispersal history and morphological features. There were apparently three different species on Crete, an older species of Early Pleistocene age and related to Mammuthus , and two others of Middle or Late Pleistocene age, namely Palaeoloxodon creutzburgi and P. chaniensis . The unique m3 from Kassos is similar in size and morphology to P. creutzburgi . From the other islands, the most famous and particularly well-documented species is P. tiliensis from Tilos. It was a dwarfed form estimated as being 1.8 m at the shoulders. Other important records are from the islands of Rhodos, Naxos, Dilos, Kalymnos and Kythera. These islands yielded palaeoloxodontine elephant fossils presumably of the Middle-Late Pleistocene age. The pattern of their dentition and the character of the limb bones, when known, resemble those of the European straight-tusked elephant P. antiquus , and the general opinion is that they were derived from this species. The main question discussed in the present study is the relationship between the elephant occurrences and palaeogeographic evolution of the Aegean domain. It appears that elephants populated Crete at least twice at different times using sweepstake roots. On other islands, elephants probably became isolated because of the subsidence of the Aegean domain and the sea level rise during the Late Pleistocene, which reduced land surfaces and food resources. Hence different degrees of dwarfism existed in these elephants and varied from one island to another.
... Yet those prehistoric proboscideans weren't always colossuses. On Cyprus, Crete, and numerous other Mediterranean islands, paleontologists have discovered the remains of ''dwarf'' elephants (Symeonides et al 2001;Poulakakis et al 2002;Poulakakis et al 2006;Theodorou et al 2007;Van der Geer et al 2014;Sen et al 2014). We also know that Sardinia once had diminutive mammoths. ...
Full-text available
The one-eyed giant has dominated popular imagination for thousands of years. The Cyclops combines two monstrous elements into one: a humanoid of colossal physique and a physical abnormality that only rarely occurs in real life. The most famous Cyclops, the cannibalistic shepherd of the ancient Greek Homeric myth isn't the only one. He is one of many that appear not just in Greek mythology, but in other cultures throughout the ancient world. In this work we shall discuss the possible origin behind the one-eyed monstrous giant, using both mythology and paleontology.
... In any case, the acquisition of new material and basic systematic study will give us the opportunity to carry out more targeted research on this relatively unexplored topic (i.e., phylogenetic analysis) that will allow us to extent our knowledge on the chiropteran evolution in the Balkan Peninsula. Mayhew (1977), 9 Kuss (1970), 10 Kotsakis et al. (1976), 11 van de Weerd (1973), 12 Horacek and Poulianos (1988), 13 Storch (1975), 14 Kuss (1973), 15 Reumer and Doukas (1985), 16 Schmidt-Kittler et al. (1995), 17 Vasileiadou et al. (2003), 18 Revilliod (1922), 19 Hulva et al. (2007), 20 Vasileiadou and Koufos (2005), 21 Vasileiadou and Zouros (2012) a This study P. Piskoulis and K. Chatzopoulou ...
Bats (Mammalia: Chiroptera) have an almost worldwide distribution since the Early Eocene till present. The review of the fossil record of this group in Greece revealed the presence of five families (Rhinolophidae, Vespertilionidae, Miniopteridae, Rhinopomatidae, Megadermatidae) and more than 20 species of Rhinolophus, Samonycteris, Myotis, Nyctalus, Pipistrellus, Hypsugo, Vespertilio, Eptesicus, Barbastella, Plecotus, Miniopterus, and Rhinopoma including some taxa that cannot be identified beyond the family level—Rhinolophus and Myotis are by far the most diverse bat genera in Greece. Bat fossils are discovered in at least 23 localities, all of them East of Pindus Mountain Range (Macedonia, North Aegean Islands, Central Greece, Peloponnese, Cyclades, Dodecanese Islands and Crete), distributed from the Early Miocene until the Early/Middle Holocene. Herein, we also present for the first time some preliminary information for the chiropteran fauna from the Late Pleistocene locality Loutra Almopias Cave, which seems to be the richest and most diverse up to date in Greece.
... On the other hand, the available finds are inadequate for drawing well-supported conclusions regarding its endemicity. Remarks Van der Geer et al. (2014a) depicted a left mandible and a left maxillary of Crocidura sp., but did not provide any description. ...
Several Aegean islands are known for their Pleistocene endemic mammal species. In isolation from the mainland, these mammals adapted to insular environments evolving unique characters. Only flying animals (birds, bats, insects) and animals that are capable of long-distance overseas traveling—by swimming, rafting, or floating—managed to successfully colonize the islands. Crete, the largest Greek island, had the richest endemic fauna. It has been isolated from the mainland since roughly five million years ago. Most mammals endemic to Crete did not survive into the Holocene, but went extinct before. Only the Cretan shrew is a surviving relict of the Pleistocene fauna. All native Cretan mammals today are feral descendants of introduced species. The review here of the fossil record of Crete shows the presence of 17 species of endemic mammals, consisting of dwarf elephants and mammoths, dwarf hippos, several species of deer (varying in size from dwarf deer to giant deer), giant mice, an otter, and the Cretan shrew. These taxa are distributed from the Early Pleistocene until the late Late Pleistocene, and their fossils mainly originate from coastal localities. The most biodiverse of these localities is Liko Cave, west of Réthymnon on the northern coast. The peak in taxonomic diversity is noted during the Late Pleistocene, when the large herbivores consisted of at least eight species of deer and a dwarf elephant. The majority of Pleistocene Cretan species follow the “island rule,” with the Cretan dwarf mammoth (Mammuthus creticus) at the extreme end of the trend toward dwarfism, weighing a mere 4% of the body mass of its mainland relative, M. meridionalis, and the Cretan rat (Kritimys catreus) at the extreme end toward gigantism, being about 6.7 times larger than its mainland relative, Praomys. The notorious exception is shown by the Cretan deer, Candiacervus, which instead evolved a spectacular adaptive radiation into eight species. Close to Crete are Kárpathos and Kassos. These islands were united during large periods of the Middle and Late Pleistocene, and harbored four endemic species: one elephant, two deer, and a mouse. The elephant (Palaeoloxodon aff. creutzburgi) evolved a size reduction similar to that of the Cretan elephant, in accordance with the “island rule.” The Karpathos deer belong to two species and are probably the result of in situ speciation. On Tilos, an older deer fauna was replaced by a younger elephant fauna. These elephants (P. tiliensis) show a significant degree of dwarfism (approx. 9% of the body mass of its mainland relative). Fossils of elephants have also been discovered on Rhodes, Naxos, Astypálaea, and Delos, which are all dwarf forms of the mainland species P. antiquus. The presence of fossil elephants has further been reported from Kýthnos, Sériphos, Milos, and Paros. The presence of fossil deer finally has been also reported from Amorgós and Rhodes.
... The Naxos toads are further differentiated by a deep mitochondrial divergence, dated to at least two million years (Fig. 2B), which coincides with the beginning of the Pleistocene. This period was marked by sea level subsidence that could have shortly opened land bridges with the mainland, contributing to vicariance events in the Cyclades (Sfenthourakis & Legakis, 2001;Valakos et al., 2008;van der Geer et al., 2014). Post-Messinian connections were hypothesized between Africa and Sicily at the same epoch for the African green toad Bufotes boulengeri (as well as other herpetofauna, Stöck et al., 2016), which subsequently evolved into the Sicilian endemic B. b. siculus. ...
Full-text available
The dynamic biogeography of glacial refugia may cause complex patterns of genetic admixture between parapatric taxa, which in turn can mislead their systematics, diversity, and distributions. We investigated this issue for green toads (Bufotes) inhabiting the circum-Aegean region, a biodiversity hotspot of the Eastern Mediterranean. A previous phylogeographic study based on mitochondrial and microsatellite loci identified the hybrid zone between the European (viridis) and Anatolian (sitibundus) lineages of B. viridis all over the Balkan Peninsula, but subsequent range-wide genomic analyses (>1000 SNPs) located this transition in Turkey, a thousand kilometers eastwards. In order to clarify the diversity and taxonomy of the circum-Aegean populations, we reconciled these conflicting findings by integrating previous data with pure sitibundus individuals. Our results confirmed that the viridis/sitibundus hybrid zone extends in western Anatolia, but that southeastern European populations feature cytonuclear discordances and a high and structured microsatellite diversity. This remarkable situation may stem from a massive geographic displacement of the hybrid zone during the last glacial fluctuations, an underappreciated yet seemingly common feature among the herpetofauna of the region. Our study thus contributes to the rising view that mitochondrial DNA can be a poor predictor of current phylogeographic structure, hence the need for genomic data, especially for narrowly-distributed taxa. Finally, the analyses unambiguously support the distinction of a micro-endemic clade of green toads unique to some Cyclades islands, for which we provide a formal taxonomic description.
... Rhodes, Carpathos, Tilos, Crete, Cyprus) ( Van der Geer et al., 2014;Koufos, Kostopoulos and Vlachou, 2005: 188). This is in good agreement with the suggestion made by Lykousis (2009) that the gradual submersion of the northern and central 'dry Aegean' did not happen before MIS 9 (i.e. ...
This thesis explores possibilities for hominin movement and occupation over the exposed dry land landscapes of the Aegean region during the Early and Middle Pleistocene (focusing more on the Middle Pleistocene ca. 0.8- 0.2 Mya). The point of departure and inspiration is the recent palaeogeographical reconstructions from the study area. Geological evidence reveals the existence of extended terrestrial landscapes, with attractive environments, connecting western Anatolia to Europe via the Greek mainland, during the glacial lowstands of the Middle Pleistocene, and possibly during certain interglacials. These lands are now lost, lying underwater, but, in spatial terms, a completely new spectrum of possibilities opens up for hominins moving across or settling over this part of Eurasia, affecting the wider narrative regarding the early settlements out of Africa. Yet, the research potential of the submerged landscapes of the Aegean has not been fully integrated in the way(s) we study and interpret the Lower Palaeolithic evidence from this region. The discussion about the early colonisation of Europe has been long focused on the western part of the continent due to the abundance of available evidence. The wider Aegean region was excluded, until recently, as a ‘cul de sac’ that blocked movement and dispersal towards the west, representing a gap in the European Lower Palaeolithic archive, with very little to contribute in terms of material culture or hominin fossil evidence. Advances in palaeogeography and geoarchaeology and exciting new finds urging now for a reconsideration. Could the Aegean exposed lands provide land bridges for movement and favourable niches for occupation, offering perhaps an eastern gateway to Europe during the Early and Middle Pleistocene? In order to answer these questions I drew information from archaeology and palaeoanthropology, palaeozoology and palaeoenvironments, and geology and palaeogeography. These multiple lines of evidence have been synthesised within an affordance-based GIS framework, which centres on the relationship between the hominins and their ‘affording’ world. The new methodological scheme developed here led to new hypotheses and scenarios of movement and occupation, predicting areas in the Aegean, onshore and offshore, with increased research potential for the Lower Palaeolithic, based on the level of suitability for the hominin survival, subsistence and dispersal. The findings of my study suggest that despite the serious methodological challenges imposed by landscape dynamics, temporal limitations and extensive discontinuities in the archaeological record, a cross - and inter - disciplinary approach can help us gain valuable insights into the nature of the past landscapes and land use by hominins. In this respect, the complex topography concept and the concept of affordances constitute the backbone of my approach. The first, by setting out the background against which suitability was built, and the second, by attributing a lived and experienced element into the past landscape. The contribution of this study is twofold: (a) offers a framing heuristic, to the newly founded discipline of the continental shelf prehistoric research, for testing further ideas on hominin movement and occupation in dynamic environments; and (b) proposes trans-Aegean corridors of opportunity for dispersal and occupation areas, complementing the current Lower Palaeolithic narrative with a potential eastern gateway to Europe.
Rodents (Mammalia: Rodentia) constitute today the mammalian order with the highest species diversity and the largest geographical distribution, since they live in almost all terrestrial habitats worldwide in great numbers. The fossil record shows that they appeared in the Paleocene, possibly in Asia, and diversified and spread mostly during the Eocene and Neogene. The present review of the Greek rodent fossil record reveals the presence of 23 subfamilies in 8 families (Sciuridae, Gliridae, Castoridae, Dipodidae, Muridae, Eomyidae, Ctenodactylidae, Hystricidae) from 108 Upper Oligocene to Upper Pleistocene localities. Fifty-three (53) species and one subspecies have been named from Greek type localities. The vast majority of these species (42) come from Miocene localities, with 11 of them from the Lower Miocene locality Aliveri, 11 more from the Upper Miocene locality Maritsa 1 and five from the Miocene/Pliocene boundary locality Maramena.
Full-text available
Insectivores (Mammalia: Eulipotyphla) are today a rather successful mammalian order, living almost worldwide in a large variety of habitats. The present chapter focuses on the review of the fossil record of Eulipotyphla in Greece, which starts in the late Oligocene. The fossil remains are isolated dental elements (more rarely postcranial elements), revealed in 47 localities, belonging to the three extant eulipotyphlan families present in Greece also today, Erinaceidae (five genera), Talpidae (six genera) and Soricidae (13 genera), and also the extinct families Dimylidae (one genus), Heterosoricidae (two genera), and Plesiosoricidae (one genus). Eight species which are still recognized have been described from Greek Neogene localities (the erinaceids Galerix symeonidisi, Parasorex kostakii, Schizogalerix macedonica, the talpids Myxomygale engesseri, Desmanella dubia, and the soricids Heterosorex ruemkeae, Amblycoptus jessiae, Deinsdorfia kerkhoffi), while three others are now considered junior synonyms of taxa described from elsewhere (Desmanodon meuleni, Galerix atticus, Dibolia dekkersi).
The aim of the present Thesis is the examination of the chiropteran collection that has been retrieved from the two chronologically different fossiliferous assemblages of Loutra Almopias Cave A (Pella, Macedonia, Greece), which will contribute to the knowledge of the Quaternary bats of the Greek region and the Balkan Peninsula. The specimens retrieved from the cave’s floor sediments (LAC) are of Late Pleistocene age, whereas the specimens retrieved from the elevated chamber LAC Ia are of latest Pleistocene age. Attempts to date fossils failed due to their lack of collagen. The purpose was served by the first comprehensive systematic taxonomy and phenetic analysis of a fossil chiropteran fauna from the Greek region, correlation of the studied specimens with the two chronologically different faunal assemblages of the cave site and other modern and fossil assemblages from the Greek region and the broader region of the Balkan Peninsula, taphonomic analysis and a palaeoclimatological/palaeoecological approach. These were based on the determination of the 9004 chiropteran specimens according to their morphological and metrical characteristics, which resulted in the identification of 17 species from LAC and 20 from LAC Ia from three families (Rhinolophidae, Vespertilionidae, Miniopteridae) and nine genera (Rhinolophus, Myotis, Nyctalus, Pipistrellus, Vespertilio, Eptesicus, Plecotus, Barbastella, Miniopterus). Eight species described from Cave A are the first known records in Greece and one species is the first known record of a Late Pleistocene locality from the Balkan Peninsula. Nineteen species refer to the southernmost appearance of the Late Pleistocene of the Balkan Peninsula. Cave A served primarily as a nursery roost for many bats and secondarily as a warm refuge during colder periods. Predation from the European Eagle Owl, Bubo bubo, was opportunistic, indicating a mixed assemblage. Cave A acted as a refugium during the Pleistocene glacial oscillations and as a starting point for the repopulation of Central and Northern Europe after the Last Glacial Maximum. An increase in cold-adapted species is observed during the latest Pleistocene LAC Ia, which roughly coincides with the onset of Younger Dryas. The majority of the identified species are cave dwellers and have a preference for warm climatic conditions. A variety of different landscapes are used for foraging with a preference in mixed and/or forested areas with the presence of water bodies, being a must for a significant proportion of the identified taxa. The morphological characteristics of the chiropteran specimens from Cave A are similar to those of their extant representatives indicating only minor alterations to their body size apart from the body size of the Greater Horseshoe Bat, Rhinolophus ferrumequinum, which decreases from Late to latest Pleistocene, indicating its dependency on climate. The phenetic analysis of the chiropteran species from Cave A is the first ever conducted for Chiroptera from the Greek region and it confirms the reliability of the method for species discrimination of European bats. In conclusion, the Late Pleistocene bat fauna from Cave A is, up to date, the richest and most diverse not only from the Greek region, but from the Balkan Peninsula, too.
Of all known insular mammals, hippos and elephants present the extremes of body size decrease, reducing to 4 and a mere 2% of their ancestral mainland size, respectively. Despite the numerous studies on these taxa, what happens to their relative brain size during phyletic dwarfing is not well known, and results are sometimes conflicting. For example, relative brain size increase has been noted in the Sicilian dwarf elephant, Palaeoloxodon falconeri, whereas relative brain size decrease has been postulated for Malagasy dwarf hippos. Here, I perform an analysis of brain, skull, and body size of 3 insular elephants (Palaeoloxodon “mnaidriensis,” P. tiliensis, and P. falconeri) and 3 insular hippos (Hippopotamus madagascariensis, H. lemerlei, and H. minor) to address this issue and to test whether relative brain size in phyletic dwarf species can be predicted. The results presented here show that the encephalization of all insular elephants and hippos is higher than that of their continental relatives. P. falconeri in particular has an enormous encephalization increase, which has so far not been reported in any other insular mammal. Insular brain size cannot be reliably predicted using either static allometric or ontogenetic scaling models. The results of this study indicate that insular dwarf species follow brain-body allometric relationships different from the expected patterns seen for their mainland relatives.
Mediterranean dwarf elephants represent some of the most striking examples of phyletic bodysize change observed in mammals and are emblematic of the ‘island rule’, where small mammals become larger and large mammals dwarf on islands. The repeated dwarfing of mainland elephant taxa (Palaeoloxodon antiquus and Mammuthus meridionalis) on Mediterranean islands provide a ‘natural experiment’ in parallel evolution, and a unique opportunity to investigate the causes, correlates and mechanisms of island evolution and body-size change. This thesis provides the first pan-Mediterranean study that incorporates taxonomic and allometric approaches to the evolution of dwarf elephants, establishing a framework for the investigation of parallel evolution and key morphological correlates of insular dwarfism. I show that insular dwarfism has evolved independently in Mediterranean elephants at least six times, resulting in at least seven dwarf species. These species group into three, broad size-classes: ‘smallsized’ (P. falconeri, P. cypriotes and M. creticus), ‘medium-sized’ (P. mnaidriensis and P. tiliensis) and ‘large-sized’ (Palaeoloxodon sp. nov. and ‘P. antiquus’ from Crete). Size-shape similarities between independent lineages from the east and central Mediterranean indicate that homoplasy is likely among similar-sized taxa, with implications for the existence of meta-taxa. These homoplasies appear to result from the exploitation of ontogenetic trajectories common to the Elephantidae, underpinning the evolution of small size. Interspecific allometry between dwarf and full-sized species can be seen to result from these common, but grade-shifted ontogenetic trajectories, and this may also be true of broader macroevolutionary trends in the Proboscidea. These size-related grade-shifts suggest that similar, but increasingly extreme, modifications of pre-natal development underpin the evolution of insular dwarfism in elephants. By incorporating research into the morphology and ontogeny of teeth and post-crania in fullsized extant and extinct elephants, this thesis provides new insights into insular dwarfism, elephant systematics and elephant functional morphology and adaptation.