PRIMARY RESEARCH PAPER
Phylogeography of Aphanius fasciatus (Osteichthyes:
Aphaniidae) in the Mediterranean Sea, with a focus on its
conservation in Cyprus
Joachim Langeneck .Chris Englezou .Matteo Di Maggio .Alberto Castelli .
Received: 4 December 2020 / Revised: 17 May 2021 / Accepted: 26 May 2021 / Published online: 7 June 2021
ÓThe Author(s) 2021
Abstract Aphanius fasciatus is a small ﬁsh occur-
ring in Mediterranean brackish environments. In
Cyprus it is known from three localities separated by
long stretches of coast. The genetic diversity of these
populations was evaluated using fragments of two
mitochondrial genes. A comparison with the other
available data showed that Cyprus populations repre-
sent a distinct lineage. The other lineages are concen-
trated in a relatively small area between the Strait of
Sicily and the Western Ionian Sea, while all other
areas include a subset of these lineages, suggesting
that the aforementioned area might have acted as a
glacial refugium. Landlocked North-African popula-
tions diverge from all other populations, suggesting
that they might have originated in the Late Pleis-
tocene, during transgression events of the Mediter-
ranean Sea in North-African inland water bodies. The
genetic diversity of A. fasciatus varied across different
Cyprus populations, with a pattern mirroring the
degree of environmental degradation, which likely
affected population genetic variability through demo-
graphic reductions. The three Cyprus populations
showed genetic uniqueness, suggesting the need of
population-based management practices; the low
genetic diversity of two populations, and the number
of threats affecting them, suggest that the species
should be considered endangered at national level and
deserves protection measures.
Keywords Aphanius fasciatus Phylogeography
Conservation Wetlands Mitochondrial DNA
The Mediterranean killiﬁsh, Aphanius fasciatus (Va-
lenciennes, 1821), is a small ﬁsh typically associated
to coastal brackish-water environments of the whole
Mediterranean region, with the exception of the
westernmost part, where it is replaced by Apricapha-
nius iberus (Valenciennes, 1846) and Apricaphanius
baeticus (Doadrio, Carmona & Ferna
2002), and possibly the Aegean Sea, where the closely
related Aphanius almiriensis Barbieri, Kottelat &
Stomboudi, 2007 may partially or totally replace it
(Valdesalici et al., 2019; Freyhof & Yog
2020). In the South-Eastern Mediterranean Sea the
Handling editor: Ce
Supplementary Information The online version contains
supplementary material available at https://doi.org/10.1007/
J. Langeneck (&)M. Di Maggio
A. Castelli F. Maltagliati
Department of Biology, University of Pisa, via Derna 1,
56126 Pisa, Italy
Freshwater Life Project, Kemp House, City Road,
London EC1V 2NX, UK
Hydrobiologia (2021) 848:4093–4114
distribution of A. fasciatus partially overlaps that of
Aphaniops dispar (Ru
¨ppell, 1829), with which it is
able to produce hybrid offspring (Villwock, 1987;
Fouda, 1995; Lotan & Ben-Tuvia, 1996). In addition,
A. fasciatus successfully colonised the Northern part
of the Suez Canal (Fouda, 1995), with a few records in
the Red Sea (Freyhof & Yog
˘lu, 2020), making
it one of the few anti-Lessepsian migrants known to
Aphanius fasciatus is characterised by habitat
ﬁdelity in its adult phase. In addition, eggs are laid
on benthic vegetation, and hatchlings show a lifestyle
similar to that of adults (Peloso, 1946; Tortonese,
1986; Maltagliati, 1998a). Despite the sporadical
presence in coastal marine waters (Torchio, 1967;
C¸ oker & Akyol, 2014), possibly related to dramatic
events, such as ﬂoods or dystrophic crises (Maltagliati,
1998a), these features contribute to a very restricted
gene ﬂow that may occur only between adjacent
populations, rendering this species a valuable model
for the reconstruction of microevolutionary processes
Although this species is listed as ‘‘Least Concern’’
(LC) by the International Union for Conservation of
Nature (IUCN) (Kottelat & Freyhof, 2007), as the
other Mediterranean species of killiﬁsh it is affected
by impacts such as habitat degradation and fragmen-
tation, pollution and competition with introduced
species (Valdesalici et al., 2015). This led to a strong
reduction of populations, with negative impacts on
their genetic diversity (Maltagliati, 2002; Cimmaruta
et al., 2003; Angeletti et al., 2010), with some
documented instances of local extinction (Changeux
& Pont, 1995; Ferrito & Tigano, 1996; Valdesalici
et al., 2015; Taybi et al., 2020). Moreover, the
fragmentation of the species in discrete populations
and the low gene ﬂow detected between them (Mal-
tagliati, 1998a,1999; Ferrito et al., 2013; Buj et al.,
2015; Cavraro et al., 2017), even at very low spatial
scale (Maltagliati et al., 2003), suggest that manage-
ment practices cannot disregard the geographical
distribution of genetic variability. The absence of
dispersal stages, and the scarce connectivity retrieved
even between adjacent populations, further suggest
that local extinction cannot be expected to recover
naturally, and that each population might deserve
speciﬁc management practices. Despite the high
number of works taking into account the distribution
of A. fasciatus across the Mediterranean Sea, however,
several populations are partially or totally unknown,
and the lack of demographic and genetic data prevents
a precise assessment of their conservation status.
The island of Cyprus represents a paradigmatic case
for A. fasciatus conservation. Despite the presence of
several coastal water bodies that might be potentially
suitable for this species, only three populations of A.
fasciatus are known, with long stretches of coast
separating them. At least one of these populations was
discovered only recently, and two of them are facing
environmental threats of anthropogenic origin (En-
glezou et al., 2018). These populations are unknown
from the demographic and genetic point of view, as the
vast majority of Eastern Mediterranean populations of
A. fasciatus. The aim of this work is the genetic
characterisation of Cyprus populations of A. fasciatus
through mitochondrial molecular markers, in order to
compare them with the other Mediterranean popula-
tions for which genetic data are available, and to
reconstruct the phylogeography of this species in the
Mediterranean Sea in a conservation context.
Materials and methods
A total of 72 individuals of Aphanius fasciatus were
sampled in November 2017 by hand net in the three
known localities in Cyprus (Englezou et al., 2018),
namely Glapsides Beach, Famagusta Bay (35.1604°N;
33.9125°E), Morphou Bay/Gu
(35.2013°N; 32.9143°E) and Akrotiri (34.6053°N;
32.9367°E). The wetland habitat at Famagusta Bay is
constituted by two coastal ponds located in Glapsides
Beach and Silver Beach, representing part of the
Pedieos River delta. After an event of mass ﬁsh
mortality that occurred in 2012 in Silver Beach pond
(Zogaris, 2017), the species has not been observed in
the latter site anymore (C. Englezou, pers. obs.);
therefore, the occurrence of A. fasciatus in Famagusta
Bay is nowadays restricted to Glapsides Pond. A.
fasciatus habitats at Morphou are part of the Serakhis
River ﬂood plain, which historically experienced
irregular, intermittent ﬂooding (Ward et al., 1958).
Several potentially suitable habitats were subjected to
extensive draining during the 1950s-1960s, and the
area is nowadays heavily impacted by agriculture,
animal farming, mine pollution and illegal industrial
4094 Hydrobiologia (2021) 848:4093–4114
waste dumping (Karaﬁstan & Gemikonakli, 2019).
Akrotiri Marsh is a natural wetland in Cyprus covering
an area of approximately 1.5 km
. It is part of a
Ramsar site, an Important Bird Area and a designated
Special Protection Area, equivalent to the EU desig-
nation, according to the mirror law (26/2007) of the
Cyprus Sovereign Base Areas (Zogaris, 2017). Since
1990s, the reduction of pond water salinity, due to
human intervention aimed at ameliorating the
stormwater drainage system in the Limassol/Amathus
area determined the proliferation of the mosquitoﬁsh
Gambusia holbrooki Girard, 1859, with an increased
competition pressure on the A. fasciatus population.
Furthermore, for molecular comparison, individu-
als of A. fasciatus from Lesina Lake (41.8860°N;
15.3640°E) and Orbetello Lagoon (42.4356°N;
11.2070°E) were sampled in April 2016 and July
2018, respectively (Table 1). Specimens were eutha-
nized in ice-cold water, ﬁxed in 96% ethanol and
preserved at -20°C until DNA extraction. Voucher
specimens were deposited in the Museo di Storia
Naturale of the University of Pisa (Italy).
DNA extraction was carried out on 5 mg of muscular
tissue using a salting-out protocol modiﬁed after
Aljanabi & Martinez (1997). Tissue lysis was obtained
by incubation at 55°C for 2 h in 290 ll of TNE 19
buffer and 20 ll of protease K; residual RNA was
degraded adding 10 ll of RNAse A and incubating for
10 min at room temperature. Lipids and proteins were
precipitated by adding 100 ll of NaCl 6 M and
centrifuging at 12,5009gfor 18 min. Absolute
ethanol at -20°C (800 ll) was added to the surnatant
containing genomic DNA; the solution was incubated
at -20°C overnight and then centrifuged at
12,5009gfor 15 min to precipitate DNA. The result-
ing DNA pellet was cleaned from impurities by adding
300 ll of 70% ethanol and centrifuging at maximum
speed for 2 min; this step was repeated thrice. The
pellet was then dried at 55°C, resuspended in 100 llof
ultrapure water, and stored at -20°C.
We ampliﬁed a fragment of the mitochondrial
control region (CR) and a fragment of the gene coding
for the subunit I of the cytochrome c oxidase (COI)
using the primer pairs CRAfF (50-
ACTATTCTTTGCCGGATTCTG-30) (Tigano et al.,
2004) and H16498 (50-
CCTGAAGTAGGAACCAGATC-30) (Meyer et al.,
1990), and FishF1 (50-TCAACCAACCACAAAGA-
CATTGGCAC-30) and FishR2 (50-ACTTCAGGGT-
GACCGAAGAATCAGAA-30) (Ward et al., 2005),
respectively. Polymerase chain reaction (PCR) ampli-
ﬁcations were carried out in 20 ll solutions using
1.5 mM of MgCl
, 0.2 mM of each dNTP, 0.1 lMof
each primer, 1 U of DreamTaq DNA polymerase
(Thermo Scientiﬁc), and *2.5 ng of template DNA.
For both markers the PCR proﬁle was set as follows:
initial denaturing step at 94°C for 3 min; 40 cycles of
denaturing at 94°C for 45 s, annealing at 45°C for
1 min, and extending at 72°C for 1 min, and a ﬁnal
extending step at 72°C for 7 min. A negative control
was included in each reaction. PCR products were
precipitated with sodium acetate and absolute ethanol
and sent to GATC Biotech (Euroﬁns Genomics) for
All available sequences for the two examined mito-
chondrial regions were obtained from GenBank
(Table 1; Supplementary Material 1; Supplementary
Material 2) and haplotype frequencies were retrieved
from the original publications whenever possible for
comparison purposes. Sequences from each gene were
aligned with ClustalX 2.1 (Larkin et al., 2007), and
alignments were edited in BIOEDIT version 7.2.5
(Hall, 1999). The program jModelTest 2.1.6 (Guindon
& Gascuel, 2003; Darriba et al., 2012), based on the
hierarchical likelihood ratio test, was used to assess
the best model of evolution for the sequences under the
Akaike Information Criterion (AIC, Akaike, 1974).
For CR the most probable evolutionary model was
HKY ?C, with C= 0.273 and the following fre-
quencies for each nucleotide base: f
= 0.376; f
= 0.165; f
= 0.158; f
= 0.301. For COI the most
probable evolutionary model was TIM3 ?I, with
I= 0.803 and the following frequencies for each
nucleotide base: f
= 0.243; f
= 0.265; f
= 0.316. Haplotype diversity (h) and nucleotide
diversity (p) were calculated using DnaSP v. 6.12.03
(Librado & Rozas, 2009). Pairwise F
localities and hierarchical analysis of molecular
variance (AMOVA, Excofﬁer et al., 1992) were
calculated with Arlequin v. 22.214.171.124. (Excofﬁer &
Lischer, 2010) considering the nucleotide substitution
model identiﬁed through jModelTest 2.1.6. Two
Hydrobiologia (2021) 848:4093–4114 4095
Table 1 Populations of Aphanius fasciatus analysed in the present study, with key references and estimates of population genetic
variability for CR and COI mitochondrial markers
References CR COI
Nh pNh p
TS Cavraro et al.
10 0 0 – – –
Orbetello TS Present study 16 0.775 ±0.088 0.0067 ±0.0010 16 0.517 ±0.132 0.0022 ±0.0009
TS Ferrito et al.
23 0.664 ±0.076 0.0135 ±0.0024 – – –
SoS Ferrito et al.
23 0.300 ±0.105 0.0020 ±0.0007 – – –
Pilo SoS Ferrito et al.
24 0.482 ±0.111 0.0019 ±0.0005 – – –
et al. (2008)
25 0.617 ±0.106 0.0047 ±0.0013 – – –
Pauli Figu SoS Pappalardo
et al. (2008)
25 0.430 ±0.124 0.0022 ±0.0008 – – –
Pauli Majori SoS Pappalardo
et al. (2008)
24 0.292 ±0.127 0.0010 ±0.0005 – – –
et al. (2008)
23 0.790 ±0.047 0.0055 ±0.0008 – – –
et al. (2008)
24 0.545 ±0.104 0.0030 ±0.0006 – – –
Santa Gilla TS Pappalardo
et al. (2008)
23 0.170 ±0.102 0.0006 ±0.0004 – – –
Cagliari TS Valdesalici
et al. (2019)
–– – 3* *
SS Ferrito et al.
23 0.688 ±0.084 0.0039 ±0.0006 – – –
Gar El Milh SS Ferrito et al.
22 0.450 ±0.112 0.0019 ±0.0005 – – –
Touggourt LNA Geiger et al.
–– – 2**-
Rio Melah LNA Ferrito et al.
20 0.626 ±0.084 0.0045 ±0.0014 – – –
Zahzah LNA Geiger et al.
–– – 6* *
Ghadira SS Ferrito et al.
22 0.247 ±0.108 0.0008 ±0.0004 – – –
Salina SS Ferrito et al.
24 0.236 ±0.109 0.0008 ±0.0004 – – –
Simar SS Ferrito et al.
24 0.486 ±0.105 0.0018 ±0.0004 – – –
Marsa Skala SS Ferrito et al.
22 0.550 ±0.095 0.0024 ±0.0005 – – –
Trapani SS Pappalardo
et al. (2008)
25 0.623 ±0.100 0.0045 ±0.0009 – – –
Marsala SS Pappalardo
et al. (2008)
24 0.616 ±0.115 0.0048 ±0.0011 – – –
SS Ferrito et al.
20 0.887 ±0.038 0.0133 ±0.0017 – – –
4096 Hydrobiologia (2021) 848:4093–4114
Table 1 continued
References CR COI
Nh pNh p
Sicily SS Landi et al.
– – – 11 0.600 ±0.154 0.0014 ±0.0005
SS Ferrito et al.
23 0.466 ±0.094 0.0068 ±0.0015 – – –
SS Ferrito et al.
22 0.247 ±0.108 0.0024 ±0.0011 – – –
Vendicari IS Ferrito et al.
23 0.166 ±0.098 0.0011 ±0.0006 – – –
IS Ferrito et al.
23 0.246 ±0.113 0.0008 ±0.0004 – – –
Ganzirri IS Ferrito et al.
23 0.087 ±0.078 0.0009 ±0.0008 – – –
Varano AS Valdesalici
et al. (2019)
–– – 4* *
Lesina AS Ferrito et al.
32 0.609 ±0.099 0.0034 ±0.0009 2 * *
AS Cavraro et al.
10 0.378 ±0.181 0.0045 ±0.0029 – – –
AS Geiger et al.
–– – 5* *
Veneto AS Geiger et al.
–– – 4* *
AS Cavraro et al.
65 0.229 ±0.068 0.0017 ±0.0007 – – –
Grado AS Ferrito et al.
23 0.087 ±0.078 0.0003 ±0.0003 – – –
Secovlje AS Buj et al.
10 0.200 ±0.154 0.0006 ±0.0005 – – –
Pag AS Buj et al.
et al. (2019)
12 0.455 ±0.170 0.0021 ±0.0009 2 * *
Dinjiska AS Buj et al.
12 0.167 ±0.134 0.0005 ±0.0004 – – –
Nin AS Buj et al.
12 0.455 ±0.170 0.0016 ±0.0007 – – –
Pantan AS Buj et al.
11 0.182 ±0.144 0.0006 ±0.0005 – – –
Ston AS Geiger et al.
et al. (2015)
11 0.182 ±0.144 0.0006 ±0.0005 4 * *
Ulcinj AS Buj et al.
11 0 0 – – –
AS Buj et al.
12 0.682 ±0.148 0.0027 ±0.0008 – – –
Hydrobiologia (2021) 848:4093–4114 4097
AMOVAs (Excofﬁer et al., 1992) were carried out to
examine the partition of genetic variance into (i) the
‘‘within locality’’ and ‘‘among locality’’ components
(2-level AMOVA) and (ii) the ‘‘within locality’’,
‘‘among localities within biogeographical sector’’ and
‘‘among geographical sectors’’ components (3-level
AMOVA). An additional AMOVA was performed in
order to check the isolation between Cyprus popula-
tions. The signiﬁcance of U-statistics parameters was
assessed by permutation tests with 10,000 replicates as
implemented in Arlequin v. 126.96.36.199. A median-joining
haplotype network was constructed by using NET-
WORK v. 10.1 (Bandelt et al., 1999).
BAPS v. 6.0 (Corander et al., 2003; Corander &
Marttinen, 2006) was employed to detect possible
hidden population substructuring by clustering genet-
ically similar sampled individuals into panmictic
groups, hereafter called ‘‘genetic clusters’’. We
applied Bayesian assignment analysis to the 317 bp
CR fragment common to ours, Pappalardo et al.’s
(2008), Ferrito et al.’s (2013), Buj et al.’s (2015) and
Cavraro et al.’s (2017) datasets, for 896 sequences
overall, and to the 491 bp COI fragment common to
ours, Geiger et al.’s (2014), Landi et al.’s (2014) and
Valdesalici et al.’s (2019) datasets, for 128 sequences
overall. BAPS adopts a Bayesian approach with a
stochastic optimization algorithm for analysing mod-
els of population structure, which greatly improves the
speed of the analysis compared to traditional MCMC-
based algorithms (Corander & Marttinen, 2006). The
only prior information given was the sampling locality
of each individual. When testing for genetic clusters,
in order to optimise computing time and check the
approximate number of clusters in our dataset, we ran
an explorative analysis with two replicates for each
value of k(the maximum number of clusters) from
k=1 to n?3, where n= number of populations
(Evanno et al., 2005). This analysis detected nine
clusters for CR and ﬁve for COI; then, we carried out a
ﬁnal run with ﬁve replicates for each value of kup to
k= 15. In addition, we used a number of reference
individuals nri = 500 and repeated the admixture
analysis 500 times per individual. Individuals with
ambiguous assignment (associated P\0.05) were not
visualised in the graphs.
Demographic history was inferred by the analysis
of the distribution of the number of site differences
between pairs of CR sequences (mismatch distribu-
tion), which was carried out on all populations using
DnaSP. Demographic analyses for COI sequences
were not carried out due to the uneven number of
individuals across populations and low number of
individuals in many populations. Expected values for a
model of constant population size were calculated and
plotted against the observed values. According to
Rogers & Harpending (1992), populations that have
experienced a rapid demographic growth in the recent
past show unimodal distributions, whereas those at
demographic equilibrium present multimodal distri-
butions. Harpending’s (1994) raggedness index (Hri;
quantifying the smoothness of the mismatch distribu-
tions and distinguishing between population expan-
sion and stability) and the sum of squared deviations
(SSD), as implemented in ARLEQUIN, were used to
test Rogers’ (1995) sudden expansion model, which
ﬁts to a unimodal mismatch distribution (Rogers &
Table 1 continued
References CR COI
Nh pNh p
Tourlida IS Ferrito et al.
23 0.814 ±0.063 0.0083 ±0.0012 – – –
Akrotiri LS Present study 22 0.636 ±0.054 0.0024 ±0.0003 24 0.620 ±0.057 0.0031 ±0.0003
Glapsides LS Present study 25 0.290 ±0.109 0.0017 ±0.0007 21 0.338 ±0.120 0.0007 ±0.0003
Morphou LS Present study 26 0.230 ±0.110 0.0010 ±0.0005 24 0.083 ±0.075 0.0002 ±0.0002
N= number of individuals; h= haplotype diversity; p= nucleotide diversity; –: data unavailable; *: sample below 10 individuals
(too small to be representative). Biogeographical sectors: TS Tyrrhenian Sea, SoS Sea of Sardinia, SS Strait of Sicily, LNA
Landlocked North Africa; IS Ionian Sea; AS Adriatic Sea; LS Levant Sea
4098 Hydrobiologia (2021) 848:4093–4114
Harpending, 1992). Moreover, population expansion
was tested through Fu’s (1997)Fs test, calculates by
using ARLEQUIN; and Ramos-Onsins & Rozas’
test calculated by means of DnaSP. Statis-
tical tests and conﬁdence intervals were based for Fs
on a coalescent simulation algorithm and for R
parametric bootstrapping with coalescence
Geographical sectors were established considering
Bianchi’s (2007) biogeographical characterisation of
the Mediterranean Sea. The sequences obtained in the
present study, together with those downloaded from
GenBank originated from ﬁve biogeographical sec-
tors, namely the Sea of Sardinia (sector 4), the
Tyrrhenian Sea (sector 3), the Ionian Sea (sector 9),
the Adriatic Sea (sectors 6, 7 and 8) and the Levant Sea
(sector 12). Although Bianchi (2007) argued that the
Adriatic Sea should be subdivided into three different
biogeographical sectors, we decided to consider it as a
single biogeographical sector, due to the post-glacial
recolonisation of the Northern and Central portions of
this sea (Bianchi et al., 2012). Moreover, we decided
to consider the Strait of Sicily as a further biogeo-
graphical sector, as it represents the transitional area
between the Western and Eastern Mediterranean
basins, even though its effect on genetic connectivity
in marine organisms varies depending on species
biological characteristics, such as developmental
mode, duration and life style of larval stages, and
adult dispersal capabilities (Bianchi, 2007; Villamor
et al., 2014). Following the results by Ferrito et al.
(2013) we decided to consider landlocked populations
of Northern Africa as belonging to a separate biogeo-
Control region (CR)
Molecular analyses allowed to obtain 317 bp CR
sequences for the 96 individuals of A. fasciatus
(GenBank accession numbers: MW147304.1-
MW147321.1). Eight haplotypes were identiﬁed in
samples from Cyprus. Interestingly, all of them were
private to the island and only the most frequent
haplotype was shared between the three populations,
whereas the remaining seven haplotypes were local-
ity-private (Supplementary Material 3). The highest
values of haplotype and nucleotide diversity were
observed in the Akrotiri population
(h=0.636 ±0.054, p= 0.0024 ±0.0003), while
the lowest values were observed in the Morphou
population (h=0.222 ±0.110,
p= 0.0010 ±0.0005), and the Glapsides population
had intermediate values for both parameters
(h= 0.290 ±0.109, p= 0.0017 ±0.0007). When
the two parameters were calculated for all populations
available in literature (Table 1), the average values of
haplotype diversity and nucleotide diversity were
h= 0.405 ±0.240 and p= 0.0029 ±0.0032,
respectively. Both parameters showed high variation
across Mediterranean populations, as testiﬁed by the
high standard deviation associated to the average
The complete dataset included 132 haplotypes,
most of which private to a single biogeographical
sector (Fig. 1). One haplotype was shared between
Strait of Sicily and Ionian Sea, two between Ionian Sea
and Adriatic Sea, and another one between Strait of
Sicily, Sea of Sardinia and Tyrrhenian Sea. More
generally, in the haplotype network Adriatic haplo-
types belonged to two distant haplogroups, all haplo-
types from Cyprus formed a haplogroup separated by
six mutations from the closest Adriatic haplotype, and
all haplotypes from internal North-African water
bodies formed a haplogroup separated by 11 mutations
from the closest haplotype from the Strait of Sicily.
Haplotypes from the Ionian Sea were distributed into
three haplogroups, two of which including Adriatic
haplotypes as well, the third one including haplotypes
from the Strait of Sicily; moreover, two haplotypes,
each shared by two individuals, were not particularly
close to any haplogroup. Lastly, no evident groupings
were observed in haplotypes sampled in the Strait of
Sicily, Sea of Sardinia and Tyrrhenian Sea (Fig. 1).
The 2-level AMOVA identiﬁed the ‘‘among-pop-
ulations’’ component as the main source of variance,
accounting for 78.8% of the variation retrieved, while
the remaining 21.2% referred to the ‘‘within popula-
tion’’ component (Table 2). The 3-level AMOVA
identiﬁed the higher source of variance in the ‘‘among
population within biogeographical sectors’’ compo-
nent (48.9%), while the ‘‘among biogeographical
sectors’’ component accounted for the 31.1%, and
the ‘‘within population’’ component for the 20.0% of
the variation retrieved (Table 2). The AMOVA car-
ried out on Cyprus populations only identiﬁed the
Hydrobiologia (2021) 848:4093–4114 4099
‘‘within population’’ component as the main source of
variance, accounting for 72.7% of the variation
retrieved, while the remaining 27.3% referred to the
‘‘among population’’ component (Supplementary
Material 4). All the associated U-statistics value were
signiﬁcantly greater than zero (Table 2).
The Bayesian assignment analysis carried out on
the complete dataset detected nine genetic clusters
with maximum-associated probability value (P=1)
Fig. 1 Median-joining haplotype network of CR sequences in Mediterranean populations of Aphanius fasciatus. Abbreviations of
biogeographical sectors are as in Table 1
Table 2 Hierarchical analyses of molecular variance derived from the cluster analysis computed from the matrix of pairwise
differences on the complete CR dataset
Source of variation df Sum of squares Variance components Percentage of variation U-statistics P
Among populations 41 3151.183 3.57249 78.75 U
= 0.788 \0.001
Within populations 854 832.086 0.96380 21.25
Total 895 3974.269 4.53529
Source of variation df Sum of
Among biogeographical sectors 5 1348.845 1.50124 31.15 U
= 0.710 \0.001
Among populations within
36 1802.338 2.35396 48.85 U
= 0.800 \0.001
Within populations 854 823.086 0.96380 20.00 U
= 0.312 \0.001
Total 895 3974.269 4.81899
Two levels were considered in the ﬁrst analysis, whereas in the second analysis the ‘biogeographical sector’ level was added.
Probability values were obtained after a permutation test with 10,000 replicates
4100 Hydrobiologia (2021) 848:4093–4114
(Fig. 2). Twenty-four sequences (2.7% of the com-
plete dataset) downloaded from GenBank were of
ambiguous assignment. The majority of populations
belonged to a single cluster, but populations in the Sea
of Sardinia, Strait of Sicily and Ionian Sea showed the
occurrence of more than one cluster. In addition,
genetic clusters showed a different repartition across
biogeographical sectors (Fig. 2). Individuals from the
Northern Tyrrhenian Sea (Diana Lagoon, Orbetello
Lagoon and Tarquinia salterns) belonged to the same
cluster, shared with some individuals of the Tunisian
populations of Gar el Milh and Lake Tunis South,
while the population of Santa Gilla, despite facing the
Southern Tyrrhenian Sea, belonged to a cluster
widespread in the Sea of Sardinia and Strait of Sicily.
Populations of the Sea of Sardinia belonged to two
clusters that were also detected in the Strait of Sicily.
Populations of the Strait of Sicily included individuals
belonging to six clusters, two of which were private to
this sector, and one restricted to the Rio Melah
population, while the Ionian Sea included three
clusters, one of which shared with the Strait of Sicily,
the other two shared with the Adriatic Sea. The
Adriatic Sea showed an interesting distribution of
genetic clusters, as all populations of the Western
coast and the north-Eastern populations (from
Secovlje to Pantan) belonged to the same cluster,
which was shared with the Ionian population of
Ganzirri Lake. On the other hand, south-Eastern
Adriatic populations (Ston, Ulcinj and Narta)
belonged to a different cluster, which was shared with
the Western Ionian population of Foce Marcellino and
the Eastern Ionian population of Tourlida. Lastly, all
individuals of the three Cyprus populations belonged
to a single private cluster.
Inference on historical demography showed that the
populations of Orbetello, Tarquinia, Casaraccio, Santa
Table 3 Harpending’s (1994) raggedness index (H
and Ramos-Onsins & Rozas’ (2002)R
calculated on CR sequences for each Aphanius fasciatus pop-
ulation. Populations from Diana lagoon and Ulcinj did not
Diana lagoon – – –
Orbetello lagoon 0.061 -1.603 0.156
Tarquinia salterns 0.464 5.053 0.159
Casaraccio pond 0.670 1.844 0.150
Pilo pond 0.111 -1.037 0.107
Pauli Figu pond 0.156 -4.670*** 0.080
Pauli maiori pond 0.634 -0.838 0.223
S’Ena Arrubia pond 0.044 -0.425 0.136
Santa Giusta 0.128 -0.949 0.085*
`pond 0.180 -1.227 0.109
Santa Gilla lagoon 0.467 -2.027** 0.141
Gar el Milh 0.106 0.171 0.147
Rio Melah 0.102 0.765 0.113
Tunis Lake 0.137 -0.430 0.152
Ghadira 0.317 0.303 0.123
Marsa Skala 0.075 0.614 0.183
Salina 0.336 -1.407 0.108
Simar 0.166 -1.306 0.111
Trapani 0.751 -0.916 0.090
Marsala 0.147 -4.070** 0.069**
Marina di Modica 0.046 -0.487 0.182
Pantano Longarini 0.496 3.657 0.171
Pantano Viruca 0.697 2.273 0.123
Vendicari 0.751 0.768 0.083
Marcellino mouth 0.321 -1.359 0.110
Ganzirri lake 0.849 0.402 0.204
Lesina lagoon 0.061 -6.513*** 0.056***
Comacchio lagoon 0.285 1.160 0.256
Venice lagoon 0.449 -1.572 0.053
Grado lagoon 0.690 -0.993 0.204
Secovlje 0.400 -0.339 0.300
Pag 0.104 -1.590* 0.134*
Dinjiska 0.472 -0.476 0.276
Nin 0.153 -2.124** 0.144*
Pantan 0.438 -0.410 0.286
Ston 0.438 -0.410 0.286
Ulcinj – – –
Narta 0.211 -3.945*** 0.099***
Tourlida 0.219 -1.075 0.140
Akrotiri 0.200 0.676 0.188
Glapsides 0.469 0.018 0.098
Table 3 continued
Morphou 0.412 -2.499** 0.092
Probability values were obtained after permutation tests with
*P\0.05, **P\0.01, ***P\0.001
Hydrobiologia (2021) 848:4093–4114 4101
`, Lake Tunis, Rio Melah, Trapani,
Marsala, Marina di Modica, Pantano Longarini, Pan-
tano Viruca, Vendicari, Ganzirri Lake, Comacchio,
Venice Lagoon, and Glapsides were characterised by
bimodal mismatch distributions, while Pilo, Pauli
Figu, Pauli Maiori, S’Ena Arrubia, Santa Gilla, Gar el
Milh, Ghadira, Salina, Simar, Marsa Skala, Foce
Marcellino, Lesina, Grado Lagoon, Secovlje, Pag,
Dinjiska, Nin, Pantan, Ston, Narta Lagoon, Tourlida,
Akrotiri and Morphou exhibited unimodal mismatch
distributions (Supplementary Material 5). The sums of
squared differences were not signiﬁcant, with the
exceptions of Tarquinia, Casaraccio, Pilo, Santa Gilla,
Ghadira, Salina, Simar, Marsa Skala, Pantano Viruca,
Vendicari, Dinjiska, Tourlida, and Glapsides (Supple-
mentary Material 5). Mismatch distributions were not
obtained for Diana Lagoon and Ulcinj because these
populations did not show polymorphism. Values of
Harpending’s (1994) raggedness varied from H
= 0.061 (Orbetello and Lesina lagoons) to
= 0.849 (Ganzirri Lake); however, none of the
values was statistically signiﬁcant (Table 3). Fu’s
neutrality test gave signiﬁcant results for
Pauli Figu, Santa Gilla, Marsala, Lesina, Pag, Nin,
Narta Lagoon and Morphou, while Ramos-Onzins &
test was signiﬁcant for Santa Giusta,
Marsala, Lesina, Pag, Nin, and Narta Lagoon
Cytochrome c oxidase - subunit I (COI)
In this study a 491 bp fragment of the COI gene was
ampliﬁed and sequenced for 85 individuals of A.
fasciatus (GenBank accession numbers:
MW138950.1-MW138959.1). We were not able to
obtain any additional sequence for Lesina population,
possibly due to a mutation in the primer annealing
region. Of the ﬁve haplotypes observed in Cyprus
individuals, one was shared by the three populations
and another one by Akrotiri and Glapsides and each
Fig. 2 Distribution of genetic clusters identiﬁed through Bayesian assignment analysis in CR sequences across Aphanius fasciatus
4102 Hydrobiologia (2021) 848:4093–4114
population had one private haplotype (Supplementary
Material 6). The highest values of haplotype and
nucleotide diversity were observed in the Akrotiri
population (h= 0.620 ±0.057,
p= 0.0031 ±0.0003), followed by Glapsides
(h=0.338 ±0.120, p= 0.0007 ±0.0003); Mor-
phou showed remarkably lower values for both
parameters (h= 0.083 ±0.075,
p= 0.0002 ±0.0002). As the majority of the local-
ities sampled in literature included 2–6 individuals,
which are insufﬁcient to have reliable estimates, a
comparison was made only with Orbetello (N=16;
h= 0.517 ±0.132, p= 0.0022 ±0.0009) and the
Sicilian population sampled by Landi et al. (2014)
(N= 11; h= 0.600 ±0.154, p= 0.0014 ±0.0005)
The complete dataset included 23 haplotypes, most
of which restricted to a single biogeographical sector.
Northern Tyrrhenian individuals (Orbetello Lagoon)
were included in a cluster close to a haplotype shared
by the Sicilian individuals and the two individuals
from Cagliari; on its part, this haplotype was separated
by only 1–2 mutations from the two haplotypes
sampled in the south-Eastern Adriatic locality of Ston
(Fig. 3). Cyprus haplotypes clustered together and
were separated by this group by only one mutation.
Individuals sampled in the landlocked North-African
populations of Zahzah (Tunisia) and Touggourt (Al-
geria) represented a cluster separated by 4 mutations
from the closest haplotype, and also individuals from
the Western and north-Eastern Adriatic Sea (Varano
and Lesina Lakes, Po River Mouth, Veneto and Pag)
represented a clearly deﬁned haplogroup, separated by
5 mutations from the closest haplotype (Fig. 3).
The 2-level AMOVA identiﬁed the ‘‘among pop-
ulations’’ component as the main source of variance,
accounting for 83.5% of the variation retrieved, while
Fig. 3 Median-joining haplotype network of COI sequences in
Mediterranean populations of Aphanius fasciatus. Abbrevia-
tions of biogeographical sectors are as in Table 1
Table 4 Hierarchical analyses of molecular variance derived from the cluster analysis computed from the matrix of pairwise
differences on the complete COI dataset
Source of variation df Sum of squares Variance components Percentage of variation U-statistics P
Among populations 13 194.269 1.70169 83.50 U
= 0.835 \0.001
Within populations 114 38.333 0.33625 16.50
Total 127 232.602 2.03794
Source of variation df Sum of
Among biogeographical sectors 3 140.710 1.45208 59.84 U
= 0.655 \0.001
Among populations within
10 53.559 0.63834 26.31 U
= 0.861 \0.001
Within populations 114 38.333 0.33625 13.86 U
= 0.598 0.002
Total 127 232.602 2.42668
Two levels were considered in the ﬁrst analysis, whereas in the second analysis the ‘biogeographical sector’ level was added.
Probability values were obtained after a permutation test with 10,000 replicates
Hydrobiologia (2021) 848:4093–4114 4103
the remaining 16.5% referred to the ‘‘within popula-
tions’’ component (Table 4). The 3-level AMOVA
identiﬁed the higher source of variance in the ‘‘among
biogeographical sectors’’ component (59.8%), while
the ‘‘among population within biogeographical sec-
tors’’ component accounted for the 26.3%, and the
‘‘within populations’’ component for the 13.9% of the
variation retrieved (Table 4). The AMOVA carried
out on Cyprus populations identiﬁed the ‘‘within
population’’ component as the main source of vari-
ance, accounting for 68.4% of the variation retrieved,
while the remaining 31.6% referred to the ‘‘among
population’’ component (Supplementary Material 7).
All the associated U-statistics value were signiﬁcantly
greater than zero.
The Bayesian assignment analysis carried out on
the complete dataset detected ﬁve genetic clusters with
maximum-associated probability value (P=1)
(Fig. 4). Individuals were assigned to clusters with
probability values of P= 0.01 to 1; none of the
sequences showed ambiguous assignment. The major-
ity of the sampled populations belonged to a single
cluster, but individuals of the Akrotiri population were
assigned to two clusters.
Genetic clusters showed a different repartition
across biogeographical sectors if compared with the
CR diversity pattern, with the most widespread cluster
occurring in the populations of Orbetello (Northern
Tyrrhenian Sea), Cagliari (Southern Tyrrhenian Sea),
Sicily (Strait of Sicily) and Ston (South-Eastern
Adriatic Sea). One cluster was restricted to Tunisia
(Strait of Sicily) and the landlocked Algerian popu-
lation of Touggourt, while another one included all
individuals from the Western and North-Eastern
Adriatic populations. Cyprus populations were
assigned to two further clusters, one of which was
represented in all populations, being the other one
private to Akrotiri.
Fig. 4 Distribution of genetic clusters identiﬁed through Bayesian assignment analysis in COI sequences across Aphanius fasciatus
4104 Hydrobiologia (2021) 848:4093–4114
Have Pleistocenic events moulded the genetic
structure of A. fasciatus?
Molecular data on A. fasciatus obtained in the present
work together with those gathered from previous
studies allowed to give a deep insight into the pattern
of species’ genetic diversity across its geographical
range. Bayesian analyses carried out on the CR dataset
highlighted the occurrence of eight out of nine genetic
clusters in an area including the Strait of Sicily and the
Western Ionian Sea (henceforth SS-WIS). In other
terms, all lineages retrieved in the Western and Central
Mediterranean Sea occur in this area, and three of
them are private to the SS-WIS, while the lineage
occurring in Cyprus is private to the Eastern Mediter-
ranean Sea. COI sequences showed a slightly different
pattern, with one lineage widespread from the Tyrrhe-
nian Sea to the South-Eastern Adriatic Sea, one
restricted to landlocked populations of Tunisia and
Algeria, one occurring in the Western and North-
Eastern Adriatic Sea, and two lineages restricted to
Cyprus. The different degree of coverage of the two
datasets, in term of both number of populations and
individuals, could account for the differences retrieved
between the outcomes from the two mitochondrial
markers. In particular, the genetic cluster occurring in
the Western and North-Eastern Adriatic Sea has been
reported in the SS-WIS only at Ganzirri Lake, for
which COI sequences were not available. The SS-WIS
hosts a number of lineages that is much higher than
those occurring in the remaining biogeographical
sectors of the Mediterranean (one or two, none of
which is private). This genetic richness supports the
hypothesis that SS-WIS area may have acted as a sort
of refuge during Pleistocenic glaciations (Chefaoui
et al., 2017), representing nowadays a hotspot of
species’ genetic diversity. This pattern is consistent
with Petit et al.’s (2003) theoretical considerations on
the effect of glacial refugia on species genetic
structure. On the other hand, different populations of
A. fasciatus in the Strait of Sicily tend to host only one
or two lineages, suggesting that gene ﬂow between
coastal populations is very limited (Maltagliati,
1998a,b,1999), as indicated by AMOVA. Northern
Tyrrhenian populations were assigned to a genetic
cluster that occurs in coastal populations of Northern
Tunisia, while populations from Sardinia were
assigned to two genetic clusters occurring in both
Western Sicily and Malta.
The Adriatic Sea seems to have been colonised two
times independently. The Western and North-Eastern
parts of the basin host a genetic cluster that is shared
with the Ganzirri Lake population. Possibly individ-
uals of A. fasciatus colonised the Western Adriatic Sea
from the South, along the Western coast. The South-
Eastern part of the basin hosts a genetic cluster
occurring in the Eastern Ionian Sea and in the Western
Ionian population of Foce Marcellino, which likely
reﬂects colonisation of the Adriatic Sea along the
Eastern coast. An alternative hypothesis might take
into consideration an unintentional introduction of
Adriatic individuals in Ganzirri Lake, as suggested by
Rocco et al. (2007), and supported by Cavraro et al.
(2017), who argued that translocation of A. fasciatus
individuals is a possible event. At this regard, human-
mediated introduction has been considered the most
likely explanation for the occurrence of the closely
related Aphanius almiriensis Kottelat, Barbieri &
Stoumboudi, 2007 in the Palude del Capitano (South-
ern Italy) (Valdesalici et al., 2019). More speciﬁcally,
eggs or adults of A. almiriensis might have been
accidentally introduced together with ﬁshes for human
consumption from Greece during the Roman Age.
However, the Ganzirri Lake is not known to have
hosted ﬁsh-rearing facilities in the Classical Age, as
commercial aquaculture in the Cape Peloro lagoons
dates back to the beginning of the XX Century and
mostly concerns shellﬁsh (Prioli in VV.AA., 2011).
Although the introduction of some marine taxa in the
Cape Peloro lagoons has been ascribed to commercial
shellﬁsh restocking (Di Natale, 1982), the biological
cycle of A. fasciatus does not entail any dispersal stage
(Tortonese, 1986) that could be effectively transferred
with hard-bottom or soft-bottom marine molluscs.
More generally, the hypothesis of an introduced status
of the Venice Lagoon population (Nardo, 1847) and its
possible Greek origin (Cavraro et al., 2017) are not
conﬁrmed by the available molecular data, showing a
general genetic homogeneity throughout the whole
Western and North-Eastern Adriatic Sea. The only
exception to this general trend is represented by a
single individual from Comacchio lagoon, which
showed a CR haplotype belonging to the South-
Eastern Adriatic-Eastern Ionian cluster (Fig. 3). As
this is the only occurrence of this haplogroup in the
Western Adriatic, we suggest that transfer of A.
Hydrobiologia (2021) 848:4093–4114 4105
fasciatus between different localities might indeed
occur, but it is not a relevant phenomenon in shaping
the pattern of species’ genetic diversity. Moreover,
since the potentially introduced haplogroup occurs
also in the Southern Adriatic Sea, the introduction
from Greek lagoons is in our opinion a rather unlikely
explanation. In addition, the occurrence of incomplete
lineage sorting cannot be excluded. Otherwise, there is
no co-existence of the two genetic clusters in the same
population (unlike in other areas), suggesting the
existence of a biogeographical break between the
North-Eastern and the South-Eastern Adriatic Sea.
The populations of Cyprus do not belong to any of
genetic clusters occurring in the SS-WIS, although
they are not distant from them in terms of mutational
steps. This suggests that glacial refugia for these
species could have been elsewhere, possibly in an area
within the Eastern Mediterranean Sea. A more precise
reconstruction is hindered by the absence of data for
populations of the Eastern Mediterranean Sea, where
the species is locally frequent (Villwock, 1964; Lotan
& Ben-Tuvia, 1996). On the other hand, the occur-
rence of A. fasciatus in the Aegean Sea has been
debated, as it was historically confused with A.
almiriensis, which on turn has been reported only a
few times (Valdesalici et al., 2019). The hypothesis
that the latter species might largely replace A.
fasciatus in the Aegean Sea is consistent with results
by Triantafyllidis et al. (2007), who identiﬁed through
mitochondrial sequences a mean divergence of 3.45%
between Aegean populations identiﬁed in that work as
A. fasciatus and populations from the remaining parts
of the Mediterranean Sea. This divergence is close to
the 4.5% of divergence detected between A. fasciatus
and A. almiriensis using COI fragments by Valdesalici
et al. (2019), and the highest mutation rate of the COI
gene if compared to the fragments analysed by
Triantafyllidis et al. (2007) could account for the
difference. Nevertheless, a biogeographical break in
the Southern Aegean Sea has been highlighted in the
phylogeographical pattern of several marine species
(Nikula & Va
¨,2003; Domingues et al., 2005;
´rez-Losada et al., 2007; Pannacciulli et al., 2017)
and might account for the differentiation between A.
fasciatus and A. almiriensis.
A stable difference between landlocked North-
African populations and the remaining dataset is clear
from both markers, although, unfortunately, the two
datasets include different landlocked populations.
Nonetheless, landlocked populations are separated
by 10 and 5 mutational steps from the closest non-
landlocked haplotype for CR and COI, respectively.
This outcome suggests that landlocked populations
have been isolated from the coastal ones since enough
time to grant a signiﬁcant diversiﬁcation. Ferrito et al.
(2013) proposed a pre-Pleistocenic isolation of these
populations; however, this hypothesis is based on a
very conservative estimate of the mutation rate of the
control region if compared to the values retrieved in
other ﬁsh species (McMillan & Palumbi, 1997;
Burridge et al., 2008). Geological data showed that
stable connections of brackish or hyperhaline inland
water bodies of North Africa with the sea occurred
several times in the Late Pleistocene, due to sea-level
rise and consequent transgression of the sea in the
depressed areas of the continent (Richards & Vita-
Finzi, 1982; Causse et al., 1989,2003). The last event
of such transgression dates back to 55–18 kya (with a
maximum around 30 kya) (Causse et al., 2003—see
also Richards & Vita-Finzi, 1982), coarsely corre-
sponding to the A
˚lesund interstadial period, that
showed a partial reduction of the ice sheet, with
subsequent increase of the sea level (Mangerud, 1981).
At present, a more precise identiﬁcation of a Late
Pleistocene event that might have caused the isolation
of North-African landlocked populations of A. fascia-
tus is impossible. However, we would discard the most
recent dates, as they would imply an exceedingly high
mutation rate. Instead, transgression events that
occurred approximately 200 kya (Causse et al.,
2003) would be compatible with the known range of
mutation rates for ﬁsh control region. The number and
extent of these events, and in particular of the most
recent one, has been largely debated (Richard & Vita-
Finzi, 1982; Causse et al., 1989; Vita-Finzi et al.,
1991; Causse et al., 2003). Results from this study
suggest that isolation of A. fasciatus in North-African
inland water bodies dates back to the oldest of such
events, supporting Causse et al.’s (1989,2003)
hypothesis of a more limited transgression occurred
A unique phylogeographical pattern
among Mediterranean ﬁshes
A comparison of the phylogeographical pattern
retrieved with other ﬁsh species with a similar
distribution shows clear differences, possibly related
4106 Hydrobiologia (2021) 848:4093–4114
to the peculiar life cycle traits of A. fasciatus. The
majority of studies carried out on Mediterranean
marine demersal ﬁsh species retrieved a high connec-
tivity across the basin (Vin
˜as et al., 2004; Bargelloni
et al., 2005; Domingues et al., 2005; Barbieri et al.,
2014). In these species divergent lineages are often co-
occurring and show similar frequencies, possibly
testifying events of past vicariance, but also conﬁrm-
ing the occurrence of secondary contact events
(Angiulli et al., 2016; Barros-Garcı
´a et al., 2020). A
similar pattern can be identiﬁed in benthic species with
pelagic larvae (Suzuki et al., 2004; Francisco et al.,
2014). A break between the Aegean Sea (including
any population in the Black Sea) and the remaining
Mediterranean Sea has been identiﬁed in some dem-
ersal species (Suzuki et al., 2004; Domingues et al.,
2005), but in these cases the degree of genetic
divergence is much lower than that identiﬁed between
A. fasciatus and A. almiriensis, suggesting a less
pronounced effect of the Southern Aegean as a
phylogeographical barrier. Conversely, although A.
fasciatus might thrive in internal water bodies, its
ability to inhabit coastal environments and to occa-
sionally use marine environments for dispersal (Tor-
chio, 1967;C¸ oker & Akyol, 2014) makes its
phylogeographical pattern completely different from
those identiﬁed in exclusive freshwater species, that
are often restricted to a single river drainage (Perea &
Doadrio, 2015; Lorenzoni et al., 2021), and usually
show the occurrence of several inland glacial refugia
(Costedoat & Gilles, 2009).
A comparison with ﬁsh species typically associated
to brackish-water environments, and with an appar-
ently similar ecology, provides more suitable compar-
isons. A strong structuring between Western and
Eastern Mediterranean, and among different lagoon
systems, has been revealed for the sand gobies
Pomatoschistus marmoratus (Risso, 1810) and Po-
matoschistus tortonesei Miller, 1969 (Mejri et al.,
2009,2011). The divergence between P. marmoratus
populations from different lagoon systems might be
enough to consider them as different species (Mejri
et al., 2011). In general, deep divergence between
separate populations and even cryptic speciation
events are relatively frequent in euryhaline Gobiidae
(Stefanni & Thorley, 2003; Neilson & Stepien, 2009).
This is probably due to the absence of dispersal stages
combined with the exclusively benthic adults; con-
versely, the potential for migration across stretches of
sea allows a shallower structuring in A. fasciatus.A
similar pattern of substantial isolation between pop-
ulations occurring in different brackish-water bodies
was retrieved in the pipeﬁsh Syngnathus abaster
Risso, 1827, a species occurring mostly in brackish-
water environments across the Mediterranean Sea and
characterised by egg brooding in a ventral pouch and
hatching of benthic juveniles (Sanna et al., 2013).
However, unlike A. fasciatus,S. abaster shows as well
a further structuring into three deeply divergent clades
occurring in different areas of the Mediterranean Sea
(Western Mediterranean, Tyrrhenian ?Adriatic Sea
and Strait of Sicily), suggesting that this strictly
sedentary benthic species is in fact unable to effec-
tively disperse by rafting, and that its dispersal
capabilities are closer to those of euryhaline gobies
than to those of A. fasciatus. The three-spined
stickleback Gasterosteus aculeatus Linnaeus, 1758 is
characterised by a high tolerance towards salinity
variations, although it is not as pronounced as in A.
fasciatus. It is, however, a species with boreal afﬁnity,
showing a higher capability for migration across
marine environments, and a rather scattered distribu-
tion in the Mediterranean basin, where it occurs
mostly in freshwater environments. As a consequence,
stickleback populations occurring in coastal Mediter-
ranean environments are completely isolated from
each other and originated independently from an
ancient, panmictic European population (Ma
¨,2008). Finally, the Iberian killiﬁsh Apricapha-
nius iberus is very similar to A. fasciatus as regards its
ecology and life history traits, despite showing a much
narrower distribution. This species and A. fasciatus
show similar levels of genetic divergence between
populations, despite the very different spatial scale
(Pappalardo et al., 2015). In particular, Pappalardo
et al. (2015) retrieved the same CR haplotype shared
between the populations of S’Ena Arrubia (Sardinia)
and Trapani (Sicily); by comparison, all the ﬁve
populations of A. iberus assayed, coming from a
relatively restricted area in South-Western Spain,
included only private haplotypes.
In conclusion, the phylogeographical pattern
retrieved in A. fasciatus is completely different from
both marine demersal ﬁshes and freshwater Mediter-
ranean ﬁshes; the effect of the South Aegean break
might recall a pattern identiﬁed in some marine
species, but it is far more pronounced in the A.
fasciatus group, where it actually leads to a
Hydrobiologia (2021) 848:4093–4114 4107
differentiation at species level. However, the phylo-
geographical pattern retrieved in A. fasciatus does not
fully match that retrieved in other small species
typically associated to brackish-water environments,
usually showing a more pronounced structuring and a
lower connectivity even with smaller distances
between populations. This difference is possibly due
to the capability of A. fasciatus adults to sporadically
cross marine environments (Torchio, 1967;C¸ oker &
Akyol, 2014) and thus move between coastal ponds.
On the other hand, the comparison with the phylo-
geographic structure of the stickleback, a species that
frequently uses marine environments for dispersal,
and has been demonstrated to be able to perform trans-
oceanic migrations, shows that migration events are
nevertheless sporadic enough to ensure population
Genetic diversity and conservation of A. fasciatus
Based on our molecular data, populations of A.
fasciatus in Cyprus are genetically divergent from
the other Mediterranean populations. The group of
haplotypes from Cyprus is separated by ﬁve (CR) or
two (COI) mutational steps from the closest extra-
Cyprus haplotype, respectively (Figs. 2,4). The
Bayesian assignment analysis identiﬁed Cyprus hap-
lotypes as separate genetic clusters for both mito-
chondrial regions, underlining the high degree of
uniqueness of these populations. Moreover, the two
mitochondrial regions were consistent in showing
remarkable genetic divergence among the Cyprus
populations. At the within-population level, the sam-
ple from Akrotiri marshes was characterised by the
highest values of genetic diversity parameters; the
population from Morphou Bay/Gu
showed the lowest values of genetic diversity, while
the population from Glapsides exhibited intermediate
values. Comparing these values with literature data for
CR, both Glapsides and Morphou have haplotype
diversity values below the average, with only a few
populations with values below those of Morphou. All
Cyprus populations showed nucleotide diversity val-
ues below the average for CR. In the case of COI,
however, the Akrotiri population exhibited the highest
values for both haplotype and nucleotide diversity,
although it should be noted that information for this
gene is available for only ﬁve populations. The pattern
of within-population genetic diversity in Cyprus
mirrors the levels of environmental stress affecting
biotopes at the three localities. Although the impact of
drainages and the redirection of freshwater through the
Akrotiri area might have fostered the success of the
invasive mosquitoﬁsh Gambusia holbrooki, with neg-
ative impact on the distribution of A. fasciatus,
Akrotiri marsh is a protected area. Therefore, the high
within-population genetic diversity of Akrotiri popu-
lation may be a consequence of the protection
measures adopted. Moreover, this A. fasciatus popu-
lation exhibited very high values of haplotype diver-
sity, even if compared with other Mediterranean
populations. The brackish pond in Glapsides Beach
is located on the delta fan of the ephemeral Pedieos
River, and together with the pond in Silver Beach
represents the only suitable habitat for A. fasciatus in
Famagusta Bay. The species, however, disappeared
from Silver Beach after a mass mortality event in 2012
(Zogaris, 2017), possibly due to heavy habitat alter-
ations related to urbanisation of the area (Seffer et al.,
2011). Further samplings did not show a recolonisa-
tion of this area (C. Englezou, pers. obs.). The
intermediate genetic diversity retrieved in this popu-
lation is compatible with a moderately disturbed
environment, where local mortality events can be the
cause of population decline, with consequent genetic
loss. The Morphou area was characterised by the
presence of multiple environmental stressors, most of
them of anthropogenic origin (Englezou et al., 2018),
and unsurprisingly, this population showed the lowest
values of genetic diversity, suggesting that it may have
experienced relevant demographic decline(s) with
subsequent genetic loss. Nevertheless, the highly
signiﬁcant value of Fu’s F
might suggest a recent
population expansion, but this is not conﬁrmed by
Harpending’s raggedness index and Ramos-Onzins &
. A reduction of genetic diversity due to
mass mortality events has already been reported in A.
fasciatus (Ferrito & Tigano, 1996; Maltagliati, 2002),
although several allegedly extinct populations showed
clues of recovery (Lo Duca & Marrone, 2009;
Valdesalici et al., 2015). However, genetic loss
resulting from population decline might impair the
ability of the population to withstand further environ-
mental stress, thus increasing the likelihood of its
extinction (Markert et al., 2010).
Although the three populations of A. fasciatus from
Cyprus may be seen as a statistically supported single
4108 Hydrobiologia (2021) 848:4093–4114
genetic cluster, each of them shows characteristics of
uniqueness. In fact, the presence of locality-private
haplotypes, together with results of the AMOVA,
which identiﬁed the among-population component as
the highest source of variation, suggests that gene ﬂow
among populations in Cyprus is restricted, or even
absent, and that they have been isolated for long time.
In fact, although Cyprus coastline hosts several
habitats that might be potentially suitable for A.
fasciatus, extensive sampling failed in ﬁnding addi-
tional populations, and the three known biotopes in
which A. fasciatus is present are separated by at least
170 km of coastline (280 km in the case of Glapsides-
From a strict conservationist point of view, our
results suggest that each population of A. fasciatus in
Cyprus should be treated as a single unit of conser-
vation, in view of its genetic uniqueness, suggesting
that recolonisation from one of the other localities
should clearly be regarded as a highly unlikely event.
Moreover, genetic data, together with results obtained
in previous surveys by Zogaris (2017) and Englezou
et al. (2018), highlighted that two out of three
populations show traces of demographic and genetic
deterioration, possibly related to the high degree of
anthropogenic environmental stress. Although our
results suggested that the Glapsides Beach population
is in a genetically better condition than that of
Morphou Bay, the local extinction in the close site
of Silver Beach advises that its status should be
carefully monitored. Recolonisation of the brackish
environment of Silver Beach could occur naturally
either through migration across Famagusta Bay, where
A. fasciatus has already been observed in marine
environments (C¸ oker & Akyol, 2014), or through
seasonal ﬂooding of the Pedieos River delta. However,
recent surveys showed that the brackish environment
at Silver Beach is currently deteriorated, with a
reduction of the area and an almost complete replace-
ment of macrophytes, playing a relevant role in A.
fasciatus biological cycle, by algal ﬁlms (C. Englezou,
pers. obs.). Therefore, although technically feasible,
the reintroduction in the Silver Beach environment
would most likely be inconsequential. The grim
situation of the Morphou Bay population, affected
by multiple anthropogenic stressors, might be partially
relieved by the inclusion of this area in the proposed
Natura 2000 site Akdeniz Special Environmental
Protected Area (S.E.P.A.), as suggested by Englezou
et al. (2018), or by the designation of a new, smaller
protected area corresponding to biotopes where A.
The genetic uniqueness of Cyprus populations and
the low level of connectivity between them, together
with the fragmentation of their habitat and the high
number of anthropogenic and environmental threats
affecting this species suggest that A. fasciatus deserves
high-priority conservation interest in Cyprus. How-
ever, it should be noted that, according to the
International Union for Conservation of Nature
(IUCN) Red List categories and criteria, A. fasciatus
is included in the ‘‘Least Concern’’ (LC) category
(Kottelat & Freyhof, 2007). Discrepancies between
the IUCN Red List and national red lists are not
uncommon (Brito et al., 2010). Although the majority
of such cases is due to missed taxonomic update or
absence of data for a speciﬁc area, there are several
instances where a species listed as LC by the IUCN
can be considered relevant for conservation at the
national level. This discrepancy has been observed
mostly (1) for species at the edge of their natural range,
where populations are by deﬁnition scarce and scat-
tered (Brito et al., 2010; Helfman, 2013), (2) for
exploited species that are object of strict management
practices in order to ensure sustainability of their
harvesting (Helfman, 2013), and (3) for species that in
a speciﬁc geographical area are affected by a severe
habitat degradation, that impairs survival of their
populations. A case of discrepancy between IUCN and
national red lists is that of another European vertebrate
associated with wetlands, the spade-footed toad Pelo-
bates fuscus (Laurenti, 1768), which is listed as LC by
the IUCN (Dufresnes et al., 2019), but it is considered
‘‘Endangered’’ (EN) in Italy (Andreone et al., 2004;
Crottini & Andreone, 2007). The more conservative
measures adopted in Italy were necessary due to the
limited number of known populations, the relevant
habitat reduction that occurred in the last two centuries
and the presence of several threats to the survival of
this species on the national territory. It should be
stressed that the IUCN Red List and national red lists
have two fundamentally different purposes, with the
former list aiming at assessing the risk of a complete
extinction of a species, focusing on the global pattern
of threats, while national red lists are aimed at
preserving the highest level of biodiversity within
the national borders, focusing on speciﬁc geographical
areas with their threats and processes. In our opinion,
Hydrobiologia (2021) 848:4093–4114 4109
both approaches somehow neglect the biogeographi-
cal component and its effects on the genetic diversity
of organisms. Commonly, the IUCN does not consider
the intraspeciﬁc patterns of diversity, including the
possibility of genetic loss across its geographical
range, unless it is sanctioned by the taxonomy (e.g.
endemic subspecies). National red lists tend to assign
to the same category species that could sustain
abundant populations but are affected by extrinsic
processes, as well as species at the edge of their
distributional range, where populations are by neces-
sity scanty, again without taking into account genetic
diversity patterns (Helfman, 2013). The case of A.
fasciatus in Cyprus recalls that of P. fuscus in Italy
from several points of view: (1) both species occur in
populations that are fragmented and separated by large
streaks of habitat made unsuitable by human activities;
(2) both species are affected by multiple anthro-
pogenic and environmental stressors across their
national distribution; (3) national populations of both
species show genetic uniqueness and distinctness from
other populations, although the divergence is too low
to justify a taxonomic recognition (Crottini et al.,
2007; Litvinchuk et al., 2013; present data); (4) the
conservation of both species strictly depends on the
conservation of their habitats. The listing of these
species in national red lists and the subsequent
enforcement of protection measures would therefore
represent a further step towards the protection of
wetlands and the maintaining of their ecosystem
services, where these taxa can be considered as
umbrella species (Andreone et al., 2004; Valdesalici
et al., 2015).
More generally, the naturally fragmented distribu-
tion of A. fasciatus, the low degree of gene ﬂow
between populations, and the subsequent possibility of
small-scale genetic differentiation retrieved in some of
them (Maltagliati et al., 2003) suggest that the
management of this species should take into account
the conservation status of single populations, as local
extinction might lead to signiﬁcant diversity losses for
the species. However, even though this approach is
recommendable in Cyprus, it is not possible to strictly
protect all A. fasciatus populations in the Mediter-
ranean area. Therefore, further studies on the distri-
bution and diversity of the less-known genetic
lineages should be carried out, in order to design a
conservation approach aimed at maximising protec-
tion of A. fasciatus genetic diversity.
Acknowledgments We are grateful to S. Gu
Parmatsias, V. Michael and the team at the Akrotiri
Environmental Centre as well as the Freshwater Life Project
team for their support in the sampling of populations of A.
fasciatus in Cyprus, to M. Marcelli and M. Renzi for their help in
sampling A. fasciatus from Orbetello Lagoon and, lastly, to two
anonymous reviewers whose comments greatly contributed to
improving our manuscript.
Funding Open access funding provided by Universita
within the CRUI-CARE Agreement. Sampling activities were
funded by Freshwater Life Project (U.K. Charity No. 1172393),
while laboratory work was funded by Fondi di Ateneo (FA) of
the University of Pisa.
Data availability Voucher specimens have been deposited in
the Museum of Natural History of the University of Pisa. Gene
sequences have been made available in GenBank.
Code availability Not applicable.
Conﬂict of interest The authors do not have any conﬂict of
interest to declare.
Ethical approval All applicable international, national, and/
or institutional guidelines for the care and use of animals were
followed by the authors.
Open Access This article is licensed under a Creative Com-
mons Attribution 4.0 International License, which permits use,
sharing, adaptation, distribution and reproduction in any med-
ium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative
Commons licence, and indicate if changes were made. The
images or other third party material in this article are included in
the article’s Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons licence and your
intended use is not permitted by statutory regulation or exceeds
the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this licence, visit
Akaike, H., 1974. A new look at the statistical model identiﬁ-
cation. IEEE Transactions on Automatic Control 19:
Aljanabi, S. M. & I. Martinez, 1997. Universal and rapid salt-
extraction of high quality genomic DNA for PCR-based
techniques. Nucleic Acids Research 25: 4692–4693.
Andreone, F., P. E. Bergo
`, S. Bovero & E. Gazzaniga, 2004. On
the edge of extinction? The spadefoot Pelobates fuscus
insubricus in the Po Plain, and a glimpse at its conservation
biology. Italian Journal of Zoology 71: 61–72.
4110 Hydrobiologia (2021) 848:4093–4114
Angeletti, D., R. Cimmaruta & G. Nascetti, 2010. Genetic
diversity of the killiﬁsh Aphanius fasciatus paralleling the
environmental changes of Tarquinia salterns habitat.
Genetica 138: 1011–1021.
Angiulli, E., L. Sola, G. Ardizzone, C. Fassatoui & A. R. Rossi,
2016. Phylogeography of the common pandora Pagellus
erythrinus in the central Mediterranean Sea: sympatric
mitochondrial lineages and genetic homogeneity. Marine
Biology Research 12: 4–15.
Bandelt, H. J., P. Forster & A. Ro
¨hl, 1999. Median-joining
networks for inferring intraspeciﬁc phylogenies. Molecular
Biology and Evolution 16: 37–48.
Barbieri, M., F. Maltagliati, M. I. Rolda
´n & A. Castelli, 2014.
Molecular contribution to stock identiﬁcation in the small-
spotted catshark, Scyliorhinus canicula (Chondrichthyes,
Scyliorhinidae). Fisheries Research 154: 11–16.
Bargelloni, L., J. A. Alarcon, M. C. Alvarez, E. Penzo, A.
Magoulas, J. Palma & T. Patarnello, 2005. The Atlantic-
Mediterranean transition: discordant genetic patterns in
two seabream species, Diplodus puntazzo (Cetti) and Di-
plodus sargus (L.). Molecular Phylogenetics and Evolution
´a, D., E. Froufe, R. Ban
´n, J. C. Arronte, F. Baldo
& A. de Carlos, 2020. Phylogeography highlights two
different Atlantic/Mediterranean lineages and a phenotypic
latitudinal gradient for the deep-sea morid codling Lepid-
ion lepidion (Gadiformes: Moridae). Deep Sea Research
Part I: Oceanographic Research Papers 157: 103212.
Bianchi, C. N., 2007. Biodiversity issues for the forthcoming
tropical Mediterranean Sea. Hydrobiologia 580: 7–21.
Bianchi, C. N., C. Morri, M. Chiantore, M. Montefalcone, V.
Parravicini & A. Rovere, 2012. Mediterranean Sea biodi-
versity between the legacy from the past and a future of
change. In Stambler, N. (ed.), Life in the Mediterranean
Sea: a look at habitat changes. Nova Science Publishers,
New York: 1–55.
Brito, D., R. G. Ambal, T. Brooks, N. De Silva, M. Foster, W.
Hao, C. Hilton-Taylor, A. Paglia, J. P. Rodrı
´guez & J.
´guez, 2010. How similar are national red lists and
the IUCN Red List? Biological Conservation 143:
Buj, I., J. Mioc
´, Z. Marc
´, P. Mustaﬁc
ˇ, D. Zanella, M.
´, T. Mihiniac
´aleta, 2015. Population
genetic structure and demographic history of Aphanius
fasciatus (Cyprinodontidae: Cyprinodontiformes) from
hypersaline habitats in the eastern Adriatic. Scientia Mar-
ina 79: 399–408.
Burridge, C. P., D. Craw, D. Fletcher & J. M. Waters, 2008.
Geological dates and molecular rates: ﬁsh DNA sheds light
on time dependency. Molecular Biology and Evolution 25:
Causse, C., R. Coque, J. C. Fontes, F. Gasse, E. Gibert, H. Ben
Ouezdou & K. Zouari, 1989. Two high levels of continental
waters in the southern Tunisian chotts at about 90 and 150
ka. Geology 17: 922–925.
Causse, C., B. Ghaleb, N. Chkir, K. Zouari, H. Ben Ouezdou &
A. Mamou, 2003. Humidity changes in southern Tunisia
during the Late Pleistocene inferred from U-Th dating of
mollusc shells. Applied Geochemistry 18: 1691–1703.
Cavraro, F., S. Malavasi, P. Torricelli, C. Gkenas, V. Liousia, I.
Leonardos, I. Kappas, T. J. Abatzopoulos & A.
Triantafyllidis, 2017. Genetic structure of the South
European toothcarp Aphanius fasciatus (Actinopterygii:
Cyprinodontidae) populations in the Mediterranean basin
with a focus on the Venice lagoon. The European Zoo-
logical Journal 84: 153–166.
Changeux, T. & D. Pont, 1995. Current status of the riverine
ﬁshes of the French Mediterranean basin. Biological
Conservation 72: 137–158.
Chefaoui, R. M., C. M. Duarte & E. A. Serra
˜o, 2017. Palaeo-
climatic conditions in the Mediterranean explain genetic
diversity of Posidonia oceanica seagrass meadows. Sci-
entiﬁc Reports 7: 2732.
Cimmaruta, R., F. Scialanca, F. Luccioli & G. Nascetti, 2003.
Genetic diversity and environmental stress in Italian pop-
ulations of the cyprinodont ﬁsh Aphanius fasciatus.
Oceanologica Acta 26: 101–110.
C¸ oker, T. & O. Akyol, 2014. An overview on the ﬁsh diversity in
the coasts of Turkish Republic of Northern Cyprus
(Mediterranean). Ege Journal of Fisheries and Aquatic
Science 31: 113–118.
Corander, J. & P. Marttinen, 2006. Bayesian identiﬁcation of
admixture events using multilocus molecular markers.
Molecular Ecology 15: 2833–2843.
Corander, J., P. Waldmann & M. J. Sillanpa
¨, 2003. Bayesian
analysis of genetic differentiation between populations.
Genetics 163: 367–374.
Costedoat, C. & A. Gilles, 2009. Quaternary pattern of fresh-
water ﬁshes in Europe: comparative phylogeography and
conservation perspective. The Open Conservation Biology
Journal 3: 36–48.
Crottini, A. & F. Andreone, 2007. Conservazione di un anﬁbio
iconico: lo status di Pelobates fuscus in Italia e linee guida
d’azione. Quaderni della Stazione Ecologica del civico
Museo di Storia naturale di Ferrara 17: 67–76.
Crottini, A., F. Andreone, J. Kosuch, L. J. Borkin, S.
N. Litvinchuk, C. Eggert & M. Veith, 2007. Fossorial but
widespread: the phylogeography of the common spadefoot
toad (Pelobates fuscus), and the role of the Po Valley as a
major source of genetic variability. Molecular Ecology 16:
Darriba, D., G. L. Taboada, R. Doallo & D. Posada, 2012.
jModelTest 2: more models, new heuristics and parallel
computing. Nature Methods 9: 772.
Di Natale, A., 1982. Extra-Mediterranean species of Mollusca
along the Southern Italian coasts. Malacologia 22:
Domingues, V. S., G. Bucciarelli, V. C. Almada & G. Bernardi,
2005. Historical colonization and demography of the
Mediterranean damselﬁsh, Chromis chromis. Molecular
Ecology 14: 4051–4063.
Dufresnes, C., I. Strachinis, E. Tzoras, S. N. Litvinchuk & M.
¨l, 2019. Call a spade a spade: taxonomy and distri-
bution of Pelobates, with description of a new Balkan
endemic. Zookeys 859: 131–158.
Englezou, C., S. Gu
¨cel & S. Zogaris, 2018. A new record of
Aphanius fasciatus (Valenciennes, 1821) on Cyprus:
insights for conservation. Cahiers de Biologie Marine 59:
Evanno, G., S. Regnaut & J. Goudet, 2005. Detecting the
number of clusters of individuals using the software
Hydrobiologia (2021) 848:4093–4114 4111
STRUCTURE: a simulation study. Molecular Ecology 14:
Excofﬁer, L. & H. E. L. Lischer, 2010. Arlequin suite ver 3.5: a
new series of programs to perform population genetics
analyses under Linux and Windows. Molecular Ecology
Resources 10: 564–567.
Excofﬁer, L., P. E. Smouse & J. M. Quattro, 1992. Analysis of
molecular variance inferred from metric distances among
DNA haplotypes: application to human mitochondrial
DNA restriction data. Genetics 131: 479–491.
Ferrito, V. & C. Tigano, 1996. Decline of Aphanius fasciatus
(Cyprinodontidae) and Salaria ﬂuviatilis (Blenniidae)
populations in freshwater of eastern Sicily. Ichthyological
Exploration of Freshwaters 7: 181–184.
Ferrito, V., F. Maltagliati, A. Mauceri, A. Adorno & C. Tigano,
2003. Morphological and genetics variation in four Italian
populations of Lebias fasciata (Teleostei, Cypridontidae).
Italian Journal of Zoology 70: 115–121.
Ferrito, V., A. M. Pappalardo, A. Canapa, M. Barucca, I.
Doadrio, E. Olmo & C. Tigano, 2013. Mitochondrial
phylogeography of the killiﬁsh Aphanius fasciatus (Tele-
ostei, Cyprinodontidae) reveals highly divergent Mediter-
ranean populations. Marine Biology 160: 3193–3208.
Fouda, M. M., 1995. Life history strategies of four small-size
ﬁshes in the Suez Canal, Egypt. Journal of Fish Biology 46:
Francisco, S. M., V. C. Almada, C. Faria, E. M. Velasco & J.
I. Robalo, 2014. Phylogeographic pattern and glacial
refugia of a rocky shore species with limited dispersal
capability: the case of Montagu’s blenny (Coryphoblennius
galerita, Blenniidae). Marine Biology 161: 2509–2520.
Freyhof, J. & B. Yog
˘lu, 2020. A proposal for a new
generic structure of the killiﬁsh family Aphaniidae, with
the description of Aphaniops teimorii (Teleostei: Cyprin-
odontiformes). Zootaxa 4810: 421–451.
Fu, Y. X., 1997. Statistical tests of neutrality of mutations
against population growth, hitchhiking and background
selection. Genetics 147: 915–925.
Geiger, M. F., F. Herder, M. T. Monaghan, V. Almada, R.
Barbieri, M. Bariche, P. Berrebi, J. Bohlen, M. Casal-
Lopez, G. B. Delmastro, G. P. J. Denys, A. Dettai, I.
Doadrio, E. Kalogianni, H. Ka
¨rst, M. Kottelat, M. Kovac
M. Laporte, M. Lorenzoni, Z. Marc
Perdices, S. Perea, H. Persat, S. Porcellotti, C. Puzzi, J.
Robalo, R. S
ˇanda, M. Schneider, V. S
Stoumboudi, S. Walter & J. Freyhof, 2014. Spatial
heterogeneity in the Mediterranean Biodiversity Hotspot
affects barcoding accuracy of its freshwater ﬁshes.
Molecular Ecology Resources 14: 1210–1221.
Guindon, S. & O. Gascuel, 2003. A simple, fast and accurate
method to estimate large phylogenies by maximum-like-
lihood. Systematic Biology 52: 696–704.
Hall, T.A., 1999. BioEdit: a user-friendly biological sequence
alignment editor and analysis program for windows 95/98/
NT. Nucleic Acids Symposium Series 41: 95–98.
Harpending, H. C., 1994. Signature of ancient population
growth in a low-resolution mitochondrial DNA mismatch
distribution. Human Biology 66: 591–600.
Helfman, G. S., 2013. National ‘‘versus’’ global red lists of
imperiled ﬁshes: why the discord? Environmental Biology
of Fishes 96: 1159–1168.
Karaﬁstan, A. & E. Gemikonakli, 2019. Contaminant evaluation
in ﬁsh from the mining impacted Morphou Bay, Cyprus,
using statistical and artiﬁcial neural network analysis. Mine
Water and the Environment 38: 178–186.
Kottelat, M. & J. Freyhof, 2007. Handbook of European
Freshwater Fishes. Published by the authors. 646 pp.
Landi, M., M. Dimech, M. Arculeo, G. Biondo, R. Martins, M.
Carneiro, G. R. Carvalho, S. Lo Brutto & F. O. Costa, 2014.
DNA barcoding for species assignment: the case of
Mediterranean marine ﬁshes. PLoS ONE 9: e106135.
Larkin, M. A., G. Blackshields, N. P. Brown, R. Chenna, P.
A. McGettigan, H. McWilliam, F. Valentin, I. M. Wallace,
A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson & D.
G. Higgins, 2007. Clustal W and Clustal X version 2.0.
Bioinformatics 23: 2947–2948.
Librado, P. & J. Rozas, 2009. DnaSP v5: a software for com-
prehensive analysis of DNA polymorphism data. Bioin-
formatics 25: 1451–1452.
Litvinchuk, S. N., A. Crottini, S. Federici, P. De Pous, D.
Donaire, F. Andreone, M. L. Kalezic
A. Lada, L. J. Borkin & J. M. Rosanov, 2013. Phylogeo-
graphic patterns of genetic diversity in the common
spadefoot toad, Pelobates fuscus (Anura: Pelobatidae),
reveals evolutionary history, postglacial range expansion
and secondary contact. Organisms Diversity and Evolution
Lo Duca, R. & F. Marrone, 2009. Conferma della presenza di
Aphanius fasciatus Valenciennes, 1821) (Cyprinodontif-
ormes Cyprinodontidae) nel bacino idrograﬁco del ﬁume
Imera meridionale (Sicilia). Naturalista Siciliano, S. IV 33:
Lorenzoni, M., A. Carosi, S. Quadroni, V. De Santis, I. Vanetti,
G. B. Delmastro & S. Zaccara, 2021. Cryptic diversity
within endemic Italian barbels: revalidation and descrip-
tion of new Barbus species (Teleostei: Cyprinidae). Journal
of Fish Biology 98: 1433–1449.
Lotan, R. & A. Ben Tuvia, 1996. Distribution and reproduction
of killiﬁsh Aphanius dispar and A. fasciatus and their
hybrids in the Bardawil Lagoon on the Mediterranean coast
of Sinai (Egypt). Israel Journal of Zoology 42: 203–213.
¨kinen, H. S. & J. Merila
¨, 2008. Mitochondrial DNA phylo-
geography of the three-spined stickleback (Gasterosteus
aculeatus) in Europe – Evidence for multiple glacial
refugia. Molecular Phylogenetics and Evolution 46:
Maltagliati, F., 1998a. Does the Mediterranean killiﬁsh Apha-
nius fasciatus (Teleostei: Cyprinodontidae) ﬁt the one-di-
mensional stepping-stone model of gene ﬂow?
Environmental Biology of Fishes 53: 385–392.
Maltagliati, F., 1998b. A preliminary investigation of allozyme
genetic variation and population geographical structure in
Aphanius fasciatus from Italian brackish-water habitats.
Journal of Fish Biology 52: 1130–1140.
Maltagliati, F., 1999. Genetic divergence in natural populations
of the Mediterranean brackish-water killiﬁsh Aphanius
fasciatus. Marine Ecology Progress Series 179: 155–162.
Maltagliati, F., 2002. Genetic monitoring of brackish-water
population: the Mediterranean toothcarp Aphanius fascia-
tus (Cyprinodontidae) as a model. Marine Ecology Pro-
gress Series 235: 257–262.
4112 Hydrobiologia (2021) 848:4093–4114
Maltagliati, F., P. Domenici, C. Franch Fosch, P. Cossu, M.
Casu & A. Castelli, 2003. Small-scale morphological and
genetic differentiation in the Mediterranean killiﬁsh:
Aphanius fasciatus (Cyprinodontidae) from a coastal
brackish-water pond and an adjacent pool in northern
Sardinia. Oceanologica Acta 26: 111–119.
Mangerud, J., 1981. The early and middle Weichselian in
Norway: a review. Boreas 10: 381–393.
Markert, J. A., D. M. Champlin, R. Gutjahr-Gobell, J. S. Grear,
A. Kuhn, McGreevy, Jr., T. J., Roth, A., Bagley, M. J., & D.
E. Nacci, 2010. Population genetic diversity and ﬁtness in
multiple environments. BMC Evolutionary Biology 10:
McMillan, W. O. & S. R. Palumbi, 1997. Rapid rate of control-
region evolution in Paciﬁc butterﬂyﬁshes (Chaetodonti-
dae). Journal of Molecular Evolution 45: 473–484.
Mejri, R., S. Lo Brutto, O. K. Ben Hassine & M. Arculeo, 2009.
A study on Pomatoschistus tortonesei Miller, 1968 (Per-
ciformes, Gobiidae) reveals the Siculo-Tunisian Strait
(STS) as a breakpoint to gene ﬂow in the Mediterranean
basin. Molecular Phylogenetics and Evolution 53:
Mejri, R., M. Arculeo, O. K. Ben Hassine & S. Lo Brutto, 2011.
Genetic architecture of the marbled goby Pomatoschistus
marmoratus (Perciformes, Gobiidae) in the Mediterranean
Sea. Molecular Phylogenetics and Evolution 58: 395–403.
Meyer, A., T. D. Kocher, P. Basasibwaki & A. C. Wilson, 1990.
Monophyletic origin of Lake Victoria ﬁshes suggested by
mitochondrial DNA sequences. Nature 347: 550–553.
Neilson, M. E. & C. A. Stepien, 2009. Evolution and phylo-
geography of the tubenose goby genus Proterorhinus
(Gobiidae: Teleostei): evidence for new cryptic species.
Biological Journal of the Linnean Society 96: 664–684.
Nikula, R. & R. Va
¨, 2003. Phylogeography of Cerasto-
derma glaucum (Bivalvia: Cardiidae) across Europe: a
major break in the Eastern Mediterranean. Marine Biology
Nardo, G., 1847. Prospetto della fauna marina volgare del
Veneto Estuario. G. Antonelli, Venezia: 44.
Pannacciulli, F. G., F. Maltagliati, C. de Guttry & Y. Achituv,
2017. Phylogeography on the rocks: the contribution of
current and historical factors in shaping the genetic struc-
ture of Chthamalus montagui (Crustacea, Cirripedia).
PLoS ONE 12: e0178287.
Pappalardo, A. M., V. Ferrito, A. Messina, F. Guarino, T.
Patarnello, V. De Pinto & C. Tigano, 2008. Genetic
structure of the killiﬁsh Aphanius fasciatus, Nardo 1827
(Teleostei, Cyprinodontidae), results of mitochondrial
DNA analysis. Journal of Fish Biology 72: 1154–1173.
Pappalardo, A. M., E. G. Gonzalez, C. Tigano, I. Doadrio & V.
Ferrito, 2015. Comparative pattern of genetic structure in
two Mediterranean killiﬁshes Aphanius fasciatus and
Aphanius iberus inferred from both mitochondrial and
nuclear data. Journal of Fish Biology 87: 69–87.
Peloso, A. M., 1946. Osservazioni sulla gametogenesi e sul ciclo
sessuale del Cyprinodon calaritanus. Archivio di Zoologia
Perea, S. & I. Doadrio, 2015. Phylogeography, historical
demography and habitat suitability modelling of freshwa-
ter ﬁshes inhabiting seasonally ﬂuctuating Mediterranean
river systems: a case study using the Iberian cyprinid
Squalius valentinus. Molecular Ecology 24: 3706–3722.
´rez-Losada, M., M. J. Nolte, K. A. Crandall & P. W. Shaw,
2007. Testing hypotheses of population structuring in the
Northeast Atlantic Ocean and Mediterranean Sea using the
common cuttleﬁsh Sepia ofﬁcinalis. Molecular Ecology
Petit, R. J., I. Aguinagalde, J.-L. de Beaulieu, C. Bittkau, S.
Brewer & R. Cheddadi, 2003. Glacial refugia: hotspots but
not melting pots of genetic diversity. Science 300:
Ramos-Onsins, S. E. & J. Rozas, 2002. Statistical properties of
new neutrality tests against population growth. Molecular
Biology and Evolution 19: 2092–2100.
Richards, G. W. & C. Vita-Finzi, 1982. Marine deposits
35000–25000 years old in the Chott el Djerid, southern
Tunisia. Nature 295: 54–55.
Rocco, L., V. Ferrito, D. Costagliola, A. Marsilio, A. M. Pap-
palardo, V. Stingo & C. Tigano, 2007. Genetic divergence
among and within four Italian populations of Aphanius
fasciatus (Teleostei, Cyprinodontiformes). Italian Journal
of Zoology 74: 371–379.
Rogers, A. R., 1995. Genetic evidence for a Pleistocene popu-
lation explosion. Evolution 49: 608–615.
Rogers, A. R. & H. Harpending, 1992. Population growth makes
waves in the distribution of pairwise genetic differences.
Molecular Biology and Evolution 9: 552–559.
Sanna, D., F. Biagi, H. Ben Alaya, F. Maltagliati, A. Addis, A.
Romero, J. De Juan, J.-P. Quignard, A. Castelli, P. Franzoi,
P. Torricelli, M. Casu, M. Carcupino & P. Francalacci,
2013. Mitochondrial DNA variability of the pipeﬁsh Syn-
gnathus abaster. Journal of Fish Biology 82: 856–876.
Seffer, J., G. K. Yalinca, W. Fuller, I. K. Tuncok, V. Sefferova
Stanova, O. Ozden & G. Eroglu, 2011. Management plan
for Famagusta Wetlands SEPA. Unpublished ﬁnal report,
Nicosia, Cyprus: 50.
Stefanni, S. & J. L. Thorley, 2003. Mitochondrial DNA phylo-
geography reveals the existence of an Evolutionarily Sig-
niﬁcant Unit of the sand goby Pomatoschistus minutus in
the Adriatic (Eastern Mediterranean). Molecular Phylo-
genetics and Evolution 28: 601–609.
Suzuki, N., M. Nishida, K. Yoseda, C. U
˘,T.S¸ ahin & K.
Amaoka, 2004. Phylogeographic relationships within the
Mediterranean turbot inferred by mitochondrial DNA
haplotype variation. Journal of Fish Biology 65: 580–585.
Taybi, A. F., Y. Mabrouki & I. Doadrio, 2020. The occurrence,
distribution and biology of invasive ﬁsh species in fresh
and brackish water bodies of NE Morocco. Arxius de
`gica 18: 59–73.
Tigano, C., A. Canapa, V. Ferrito, M. Barucca, I. Arcidiacono &
E. Olmo, 2004. Osteological and molecular analysis of
three Sicilian populations of Aphanius fasciatus (Teleostei,
Cyprinodontidae). Italian Journal of Zoology 71: 107–113.
Tigano, C., A. Canapa, F. Ferrito, M. Barucca, I. Arcidiacono,
A. Deidun, P. J. Schembri & E. Olmo, 2006. A study of
osteological and molecular differences in populations of
Aphanius fasciatus Nardo 1827, from the central
Mediterranean (Teleostei, Cyprinodontidae). Marine
Biology 149: 1539–1550.
Torchio, M., 1967. Osservazioni e considerazioni sulla presenza
in acque mediterranee costiere di ciprinidi, ciprinodontidi e
Hydrobiologia (2021) 848:4093–4114 4113
gasterosteidi. Natura: Rivista di Scienze Naturali 5:
Tortonese, E., 1986. Cyprinodontidae. In Whitehead, P. J. P., M.
L. Bauchot, J.-C. Hureau, J. Nielsen & E. Tortonese (eds),
Fishes of the North-castern Atlantic and the Mediterranean,
Vol. 2. UNESCO, Bungay: 623–626.
Triantafyllidis, A., I. Leonardos, I. Bista, I. D. Kyriazis, M.
T. Stoumboudi, I. Kappas, F. Amat & T. J. Abatzopoulos,
2007. Phylogeography and genetic structure of the
Mediterranean killiﬁsh Aphanius fasciatus (Cyprinodonti-
dae). Marine Biology 152: 1159–1167.
Valdesalici, S., A. Brahimi & J. Freyhof, 2019. First record of
Aphanius almiriensis from Italy and notes on the distri-
bution of Aphanius fasciatus (Teleostei: Aphaniidae).
Journal of Applied Ichthyology 35: 541–550.
Valdesalici, S., J. Langeneck, M. Barbieri, A. Castelli & F.
Maltagliati, 2015. Distribution of natural populations of the
killiﬁsh Aphanius fasciatus (Valenciennes, 1821) (Tele-
ostei: Cyprinodontidae) in Italy: past and current status,
and future trends. Italian Journal of Zoology 82: 212–223.
Villamor, A., F. Costantini & M. Abbiati, 2014. Genetic struc-
turing across marine biogeographic boundaries in rocky
shore invertebrates. PLoS ONE 9: e101135.
Villwock, W., 1964. Genetische Untersuchungen an altweltli-
chen Zahnkarpfen der Tribus Aphaniini (Pisces, Cyprin-
odontidae) nach Gesichtpunkten der Neuen Systematik.
Journal of Zoological Systematics and Evolutionary
Research 2: 267–382.
Villwock, W., 1987. Further contributions on natural hybrids
between two valid species of Aphanius,Aphanius dispar
¨ppell) and Aphanius fasciatus (Valenciennes)
(Pisces: Cyprinodontidae) from the Bardawil-Lagoon,
North Sinai, and al-Quatanir, West of the Suez Canal,
¨binger Atlas des Vorderen Orients A, Beihefte
˜as, J., J. Alvarado Bremer & C. Pla, 2004. Phylogeography
of the Atlantic bonito (Sarda sarda) in the northern
Mediterranean: the combined effects of historical vicari-
ance, population expansion, secondary invasion, and iso-
lation by distance. Molecular Phylogenetics and Evolution
Vita-Finzi, C., G. W. Richards, C. Causse, R. Coque, J.
C. Fontes, F. Gasse, E. Gibert, H. Ben Ouezdou & K.
Zouari, 1991. Comment and Reply on ‘‘Two high levels of
continental waters in the southern Tunisian chotts at about
90 and 150 ka’’. Geology 19: 94–96.
VV.AA, 2011. Lo stato della pesca e dell’acquacoltura nei mari
italiani. Cataudella, S. & M. Spagnolo (eds), Ministero
delle politiche agricole alimentari e forestali, Rome, Italy,
Ward, I. L., D. P. McGregor & M. Grehan, 1958. Hydrology and
water development in Cyprus. Proceedings of the Institu-
tion of Civil Engineers 9: 233–258.
Ward, R. D., T. S. Zemlak, B. H. Innes, P. R. Last & P. D. N.
Hebert, 2005. Dna barcoding Australia’s ﬁsh species.
Philosophical Transactions of the Royal Society of Lon-
don, series B, Biological Sciences 360: 1847–1857.
Zogaris, S., 2017. Conservation study of the Mediterranean
Killiﬁsh Aphanius fasciatus in Akrotiri Marsh, (Akrotiri
SBA, Cyprus)-Final Report. Darwin Project DPLUS034
‘‘Akrotiri Marsh Restoration: a ﬂagship wetland in the
Cyprus SBAs BirdLife Cyprus’’. Nicosia, Cyprus.
Unpublished ﬁnal report, 64 pp.
Publisher’s Note Springer Nature remains neutral with
regard to jurisdictional claims in published maps and
4114 Hydrobiologia (2021) 848:4093–4114
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”),
for small-scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are
(“Terms”). For these purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or
a personal subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or
a personal subscription (to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the
Creative Commons license used will apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data
internally within ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking,
analysis and reporting. We will not otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that
Users may not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to
circumvent access control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil
liability, or is otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by
Springer Nature in writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer
Nature journal content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates
revenue, royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain.
Springer Nature journal content cannot be used for inter-library loans and librarians may not upload Springer Nature journal
content on a large scale into their, or any other, institutional repository.
information or content on this website and may remove it or features or functionality at our sole discretion, at any time with or
without notice. Springer Nature may revoke this licence to you at any time and remove access to any copies of the Springer Nature
journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express
or implied with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or
warranties imposed by law, including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be
licensed from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other
manner not expressly permitted by these Terms, please contact Springer Nature at