._- - --
--
Mitochondrial DNA Phylogeny and the Reconstruction of the Population
History of a Species: The Case of the European Anchovy (Engraulis
encrasicolus)
Antonios
Magoulas, * Nikolaos
*Institute of Marine Biology
University
Tsimenides,*“f
of Crete, TDepartment
and Eleftherios
of Biology, University
Zouros*~$
of Crete, and *Department
of Biology, Dalhousie
of mitochondrial
DNA restriction fragment length polymorphism
in European anchovy (Engruulis enrevealed a large number of mitotypes that form two distinct clusters (phylads). Phylad A consists of
one common mitotype and many rare secondary mitotypes that are one mutational step removed from the main
type. Nucleotide diversity and number of homoplasious
changes are low. Phylad B has a complex pattern of
mitotype connectedness,
high nucleotide diversity, and a large number of homoplasious
changes. It is suggested
that the two phylads evolved in isolation from each other and that present coexistence is the result of a secondary
contact. Moreover, phylad A has a “star” phylogeny, which suggests that it has evolved in a population that
experienced a drastic bottleneck followed by an explosion of size. Phylad A is practically the only phylad present
in the Black Sea, with its frequency dropping to 85% in the northern Aegean, and to 40% in the rest of Mediterranean and the Bay of Biscay. The Black Sea is, therefore, the most likely place of origin of phylad A.
Molecular data are consistent with a population bottleneck in the Black Sea during the last glaciation event and
a subsequent exit of phylad A with the outflow into the Aegean following the ice melting. Phylogenetic
analysis
of anchovy mtDNA provides a reconstruction
of population history in the Mediterranean,
which is consistent
with the geological information.
Analysis
crusicolus)
Introduction
“In due course they found themselves entering the
narrowest
part of the winding
straits. Rugged cliffs
hemmed them in on either side, and Argo as she advanced began to feel a swirling undercurrent.
They
moved ahead in fear, for now the clash of the Colliding
Rocks and the thunder of the surf on the shores fell
ceaselessly
on their ears” (Apollonius
of Rhodes, 3rd
Century BC).
According
to Greek Mythology,
the Colliding
Rocks at the northernmost
end of the Bosporus Straits
were immobilized
when the Argonauts
first went
through them, so that man may, henceforth, move freely between the Aegean and the Black Seas. For several
marine species, however, this remains a one-way passage. This is particularly
so for species with passive
dispersal, such as planktonic
species or species which
disperse mainly during their pelagic immature stages.
According
to Moraitou-Apostolopoulou
(1985) the
transportation
of planktonic animals from the Mediterranean to the Black Sea presents several difficulties and
this explains why planktonic
forms that are abundant
in the Aegean have not been found in the Black Sea.
The fauna of the Black Sea is much poorer than that
of the adjacent Aegean. Of the approximately
6,000
Key words: mitochondrial
chovy, Mediterranean.
DNA, phylogeography,
European
an-
Address for correspondence
and reprints: Antonios Magoulas, Institute of Marine Biology of Crete, PO. Box 2214, 71003, Iraklion,
Crete, Greece. E-mail [email protected].
Mol. Biol. Evol. 13(1):178-190.
1996
0 1996 by the Society for Molecular Biology
178
and Evolution.
ISSN: 0737-4038
species known to exist in the Aegean today, only 1,500
are also found in the Black Sea (Tolmazin 1985). Economidis and Bauchot (1976) recorded 134 fish species
in the Aegean, of which only 58 are known to occur
in the Black Sea. Though many of the Aegean species
may enter the Black Sea but fail to establish themselves
there, others simply cannot traverse the sea-way that
joins the two seas.
Because of its narrow opening to the Atlantic
Ocean through the Straits of Gibraltar and its division
into several basins with distinct geologic, hydrographic, and biotic characteristics,
the Mediterranean
has
been called a “sea of seas” (e.g., Pastor 1991, p. 20).
At the beginning
of the Messinian
(approximately
5.5
million years ago), communication
with the Atlantic
was closed and the Mediterranean
was converted into
a series of saline lakes. Very few marine species seem
to have survived this “salinity
crisis”,
which lasted
several hundred thousands years (Hsu, Cita, and Ryan
1973; Hsu et al. 1978; Por 1978, 1989; Por and Dimentman 1985; Selli 1985). The great bulk of presentday Mediterranean
species entered the region during
the early Pliocene, slightly less than 5 million years
ago, when communication
with the Atlantic was reestablished
(Por and Dimmentman
1985; Sara 1985;
Por 1989). At the end of Pliocene, a connection
was
established
between
the Black and Aegean
Seas
through the Dardanelles
and Bosporus straits (Dermintzakis 1984; Bacescu 1985; Tortonese 1985). Mediterranean waters intruded into the Black Sea and de-
mtDNA Variation
stroyed the largest part of its brackish paratethyan fauna (Ekman
1968, p. 96). In each of the glaciation
events that followed, the communication
between the
two seas was interrupted
during lowering of the sea
level at glacial maxima and re-established
during the
flooding of the interglacial
periods (e.g., Por 1978, p.
16; Sara 1985). It is generally believed that the biogeographic
consequences
of glaciations
were much
more dramatic in the Mediterranean
than in open ocean
systems (Por 1978, p. 4). For these reasons the Mediterranean
appears to provide one of the best natural
scenes to study how colonization
events, population
bottlenecks,
long term isolation, and subsequent
mixing have affected the lineage structure and geographical differentiation
of present-day populations
of marine
species.
While considerable
attention has been given to the
effects of historical and present hydrographic
complexity on the species composition
of the various basins of
the Mediterranean
Sea (e.g., Tortonese 1964; Quignard
1978), there has been little attempt to utilize information about hydrographic
complexity and the geological
history of the Mediterranean
to address issues of intraspecific phylogenies
and geographical population structure. This is now possible because the introduction
of
molecular
tools in population
genetics allows one to
combine genetic information
with information from the
geological, paleontological,
and climatic records to reconstruct the population history of a species. This molecular probing into the biogeographical
past of a species has, indeed, been used in a number of cases with
remarkable
success (for a recent review see Avise
1994).
We have attempted to do this in the European anchovy (Engruulis encrusicolus) (anchovy hereafter for
simplicity),
by using restriction fragment length polymorphism
of mitochondrial
DNA (mtDNA). The anchovy is a small pelagic fish whose distribution
includes the Mediterranean
and Black Seas, and the Atlantic coasts of north Africa and Europe as far north
as the southern North Sea and the coasts of the British
Isles (Whitehead
1984). In addition, the anchovy is a
euryhaline
species persisting in waters from low (5%0)
to high (40%0) salinities (Demir 1963). In the Black
Sea, where it is the most abundant fish species, it tolerates salinities of approximately
17%0 and, when water
temperatures
are suitable, it enters the Sea of Azof,
where salinity varies from 0 to 17%0 (Ivanov and Beverton 1985). It is a multiple spawner with pelagic eggs
and larvae. On the basis of variation in morphological
characters several authors have suggested the existence
of different subspecies or races in the Mediterranean,
but the taxonomic
status of these divisions
remains
of Anchovy
179
FIG. l.--Sample
sites and mtDNA phylad frequencies. Code, exact site, geographical
area, date of collection and size (n) of samples
are as follows: BLS: Varna, West Black Sea, Oct. 1992, 12 = 74;
KAVl: Gulf of Kavala, North Aegean, May1989, n = 57; KAV2: Gulf
of Kavala, North Aegean Sea, Oct. 1989, IZ = 155; PAG: Pagasitikos
Gulf, Central Aegean Sea, Sept. 1992, n = 20; SAR: Saronikos Gulf,
South Aegean Sea, June 1993, 12 = 72; PATl: Patraikos Gulf, East
Ionian Sea, Aug. 1989, n = 152; PAT2: Patraikos Gulf, East Ionian
Sea, Oct. 1989, IZ = 56; ADR: Ghioggia, North Adriatic Sea, Nov.
1993, n = 65; LION: Sete, Gulf of Lions, Dec. 1992, it = 50; BISC:
St. Jean De Luz, Bay of Biscay, March 1993, n = 48. Arrows point
to the exact site of the sample. Black: frequency of phylad A; white:
frequency of phylad B. 1 = Black Sea; 2 = Aegean Sea; 3 = Levantine Sea; 4 = Ionian Sea; 5 = Adriatic Sea; 6 = Gulf of Lions; 7
= Bay of Biscay.
doubtful (Spanakis,
references therein).
Materials
Materials
Tsimenides,
and Zouros
1989 and
and Methods
The study includes samples of anchovy from eight
different localities (fig. 1). Samples from Kavala and
Patraikos Gulfs were taken twice in order to test the
stability of the genetic composition
in time. Thus, 10
samples were analysed in total.
The Greek samples were either obtained by the
Research Vessel “Philia”
of the Institute of Marine Biology of Crete or bought fresh from the wharf. In the
former case the fish were frozen immediately
after capture and brought to the laboratory
in -30°C.
In the
latter case the fish were air-shipped to the laboratory
frozen in portable refrigerators.
The Black Sea sample
was obtained from the Institute of Fisheries of Varna
(Bulgaria) and air-shipped to our laboratory frozen in
portable refrigerators.
Livers of individual
fish from
Adriatic,
Gulf of Lions, and Bay of Biscay were
shipped in 70% ethanol. All samples were kept in
-30°C
until DNA extraction,
except the ones preserved in ethanol, which were kept in 4°C until processed.
Methods
We employed a restriction
site analysis in which
fragments were visualized after Southern blotting and
hybridization.
The hybridization
probes were obtained
180
Magoulas
et al.
as follows. Enriched mtDNA was mass-extracted
from
mature oocytes following the protocol of Lansman et
al. (198 l), but without the step of purification with cesium chloride gradient centrifugation.
For further purification of mtDNA from nuclear DNA, the extracted
DNA was digested with restriction
enzymes found to
cut the mtDNA molecule in few fragments (two such
enzymes are BamHI or BgZII). Electrophoresis
of the
digest produced clear mtDNA bands and a smear of
nuclear DNA. The mtDNA bands were recovered from
the gel by constructing
a well and lining one of its
walls by a dialysis membrane.
The DNA was blocked
onto the membrane and then recovered from the well
after a short reversion
of the polarity (Magoulas,
in
preparation).
In addition to using it as a probe, this
mtDNA was digested with 15 restriction six-cutter endonucleases
(AvaI, BamHI, BgZI, BgZII, DruI, EcoRI,
EcoRV, HindIII, KpnI, PvuII, SacII, SalI, SphI, XmnI,
XhoI) with the goal of selecting endonucleases
that produce easily and unambiguously
detected
polymorphisms. The enzymes BgZI (cutting site GCCNNNN/
NGGC),
BgZII (A/GATCT),
Hind111 (A/AGCTT),
BamHI (G/GATCC),
and EcoRI (G/AATTC)
were finally chosen. Total DNA was extracted from the liver
of scored individuals
according to Harrison, Rand, and
Wheeler (1985). Extraction from the ethanol-preserved
livers followed a standard proteinase
K/phenol/chloroform method (see Sambrook, Fritsch, and Maniatis
1989), after vacuum desiccation
of the samples. Approximately
l/20 of the total volume of each preparation was digested with each restriction
endonuclease
and the fragments were separated on 0.7% agarose gels
at 1 .O V/cm for lo-15 hours. DNA was transferred to
nylon membranes
(Sambrook,
Fritsch, and Maniatis
1989) and hybridized to approximately
100 ng of probe
DNA. As alternative probe (with no difference in the
scoring power) we used cloned fragments of mtDNA
of the American shad (Alosa supidissima). In the earlier stages of the study labelling of the probe was done
with 32P using the random priming technique (Feinberg
and Vogelstein 1983), and the mtDNA fragments were
visualised by autoradiography.
Subsequently,
the probe
was labelled with digoxigenin
and the fragments detected according
to the instructions
of the supplier
(Boehringer and Mannheim Cat. No. 1093657). Hind111
digests of lambda DNA were added in the gels as molecular weight markers. Each distinct restriction profile
produced by any of the five endonucleases
used for
routine
scoring
(in the order BgZI, BgZII, HindIII,
BamHI, EcoRI) was assigned an upper-case letter code
in alphabetical
order of detection (A, B, etc.). Thus,
each individual
was finally assigned a five-letter composite mitotype.
A restriction-site
map for the above enzymes was
constructed for all haplotypes using the method of partial and double digestions
(Sambrook,
Fritsch, and
Maniatis 1989). The map was oriented by hybridizations of single and double digests with characterized
clones of shad mtDNA. The minimum number of restriction site changes between the different mitotypes
was estimated and used to manually construct a minimum length phylogenetic
network. Maximum
parsimony phylogenetic
analysis of the mitotypes was performed using the program PAUP (Swofford 1985). Mitotypes were entered as taxa and presence/absence
of
restriction sites as characters.
Evolutionary
distances (number of nucleotide substitutions per nucleotide site; Nei and Tajima 1981; Nei
and Miller 1990) between mitotypes
were estimated
from the shared restriction
sites using the package
REAP (McElroy et al. 1991) and used for UPGMA
clustering
of mitotypes
using the NTSYS program
(Rohlf 1992). Mitotype frequencies
were used for the
estimation of mitotype diversity (the probability that a
random pair of individuals
will have different mitotypes; Nei 1987, equation 8.5) and Nei’s genetic distances (Nei 1972). The samples were clustered on the
basis of their genetic distances,
according
to FitchMargoliash
method, using the program FITCH of the
package PHYLIP (Felsenstein
1993). Nucleotide
diversities within populations
(Nei 1987, equation 10.19)
were also estimated using the program REAP
Results and Discussion
Intraspecific Phylogeny
Seven different restriction profiles were found for
BgZI (A-G), 10 for BgZII (A-J), 8 for Hind111 (A-H),
4 for BamHI (A-D), and 14 for EcoRI (A-N) in 673
fish for which the composite five-enzyme
pattern was
obtained. Typical profiles for each of these enzymes
were published in Magoulas and Zouros (1993). Mapping revealed a total of 34 restriction sites (5 for BgZI,
7 for BgZII, 8 for HindIII, 4 for BamHI, and 10 for
EcoRI) shown in figure 2. Five of the 34 sites (three
HindIII, one BamHI and one EcoRI) were invariable.
Character state (presence/absence)
combinations
at the
29 variable sites generated 46 mitotypes.
The large number of individuals
examined and of
mitotypes detected in this study allows us to make inferences about the evolution of mtDNA and invites the
application of tests to evaluate the effects of significant
historical events on present-day
structure of mtDNA
lineages.
Recent work (Rand, Dorfsman,
and Kann
1994; Ballard and Kreitman 1994) has challenged the
generally held view that mtDNA variation in natural
populations is governed solely by the random processes
mtDNA Variation
B-
B-
@
B-B
B-
FIG. 2.-Restriction
site map of Engraulis encrasicolus mtDNA.
Total length is approximately
17,200 bp. The orientation of the molecule is shown inside the circle. Invariable sites are given in bold.
Sites whose exact locations are not determined are indicated with asterisk. The evolutionary
changes of each site are also indicated. @
denotes a site change in the path connecting the two phylads, A denotes a change in phylad A, and B denotes a change in phylad B.
Superscripts + and - denote site gain and loss, respectively.
of neutral evolution. We will present our results on the
assumption of neutrality, while recognizing
that the nature of our data cannot provide direct support for one
or the other view.
Attempts to connect the 46 mitotypes with a single
mutational
step (loss or gain of a restriction
site) revealed the existence of two distinct groups of mitotypes, henceforth
called phylad A and phylad B (fig.
3). Within phylad A, mitotype AAAAA is connected
to all but 3 of the 22 mitotypes by a single step. The
remaining three types are connected to AAAAA by two
steps through an existing intermediary.
An ambiguity
arises
with EAADA
which can be connected
to
AAAAA through EAAAA or AAADA. The first alternative was chosen because both EAADA and EAAAA
were found in the same population (Black Sea), whereas AAADA was found once in the Gulf of Lions. The
high frequency of AAAAA and the fact that it occupies
the centre of a star-type
phylogeny
suggests
that
AAAAA is the most likely ancestral mitotype in this
phylad. Using the infinite allele model under the neutral coalescent theory, Crandall and Templeton (1993)
arrived at three criteria for rooting an intra-specific
phylogeny. Mitotype AAAAA meets all these criteria:
of Anchovy
181
a) It is by far the most frequent mitotype (mean intraphylad frequency 0.81), b) it has 19 connections
in the
mutational network, whereas mitotypes with the second
degree of connectedness
have only two connections,
and c) it is located in the midpoint of the cladogram.
Thus, under the assumptions
of neutral coalescent theory, AAAAA is the most likely ancestral mitotype of
phylad A.
Phylad B is more complex. There is no single mitotype with which the other 22 types can be connected
by one or two steps and there are many networks that
can be produced using the criterion of minimal number
of steps for joining all 23 types. The network shown
in figure 3 was chosen because it joins with single steps
the most common type, BBBBB (intraphylad frequency
of 0.378), to the next two most frequent ones, BBCBB
and BBBBC (frequencies
of 0.274 and 0.135), and,
then, joins the fourth most common type, BBCBC (frequency of 0.078), to the second most frequent type
(BBCBB), again with a single step. We have examined
26 other equally parsimonious
networks. In all of them
mitotype BBBBB has the highest connectedness
and in
all but one occupied a position close to midpoint. Given that it is also the most common mitotype in phylad
B, BBBBB appears as the most likely ancestral mitotype in this phylad under the assumptions
of neutral
coalescent theory.
Using again the assumption
of coalescent theory,
Casteloe and Templeton
(1994) proposed a heuristic
method to estimate the probability
that a haplotype is
the oldest in a network of haplotypes.
When this argument was applied to phylad A, it gave a root probability of 0.280 for mitotype AAAAA, 0.235 for mitotype AAABA,
0.233 for AAAAG,
and 0.230 for
EAAAA. When applied to phylad B, it gave a probability of 0.283 for BBBBB, 0.227 for BBCBB, 0.186
for BBBBC, 0.134 for BCBBB, and 0.122 for BBCBC.
These results
agree with the characterization
of
AAAAA and BBBBB as the most likely ancestral mitotypes of phylads A and B respectively.
The approach
of Casteloe and Templeton (1994) takes into account
whether the haplotype lies in the interior or at the tip
of the network, but ignores the connectedness
of the
haplotype.
This apparently
explains
why mitotypes
AAABA, AAAAG, and EAAAA were assigned probabilities comparable to that of AAAAA, despite the fact
that they are much less frequent and have much lower
connectedness
(2 versus 19). The same applies to mitotype BCBBB in phylad B.
There are several ways to join the two phylads
into one composite network. If we assume that types
AAAAA and BBBBB are the most likely ancestral
types in each phylad, we may join them directly. This
option would require a six-step inter-phylad
gap and
182
Magoulas
et al.
Phylad A
AAFAA
f‘ :’
t”
418
tz1
tn.
BFBBB
I
Ftylad B
BDCBF
FIG. 3.-Parsimony
mutational network of the 46 mitotypes found in anchovy. Common mitotypes (frequency > 0.03) are shown in
ellipses. The number along the line connecting two mitotypes indicates the restriction site in which a change (gain or loss) occured (numbering
of sites as in fig. 2). Arrows indicate the direction of site loss (not the direction of evolution). The three lines connecting the two phylads
represent possible evolutionary
pathways (solid line: according to the coalescent theory; broken lines: according to the maximum parsimony
criterion).
would consider the intermediate
types AABAA, GAABA, BABBB, and BBBBA as homoplasious
derivatives of one or the other central type. The other option
is to minimize the gap separating the two phylads, by
assuming that existing rare intermediate
mitotypes are
older than one or the other of the central types (see fig.
3). The network joining all 46 mitotypes involves either fifty steps (if AAAAA
is directly
joined
to
BBBBB) or 47 steps (if AAAAA is joined to BBBBB
through existing minor mitotypes).
The manual network described above is in agreement with the results of a phylogenetic
analysis using
the PAUP program. More than 100 equally parsimo-
nious trees were produced, all of which connected the
mitotypes of phylad A into one group and the mitotypes of phylad B into another. Mitotypes within each
phylad were connected
by single steps and the two
phylads were connected to each other by three steps,
as expected from the criterion of maximum parsimony.
The consistency
index of these trees was 0.617, which
indicates that 18 of the 47 changes were homoplasious.
Separate phylogenetic
analysis for each phylad showed
that four homoplasious
changes occurred in phylad B
and one in phylad A.
The UPGMA phenogram
of mitotypes
derived
from their evolutionary
distance matrix is shown in fig-
mtDNA Variation
A’
4
5
8
10
14
15
17
3
6
20
19
18
11
12
13
of Anchovy
183
PAG BLS
BISC AAR
B
.‘
,.+l~yladA
N. AEGEAN
-q&
FIG. 4.-UPGMA
phenogram of the mitotypes, based on their
evolutionary
distance. The two phylads are connected by a node at a
distance of approximately
0.037. Mitotypes are numbered as in Appendix 1.
ure 4. The main feature of this figure is that the interphylad divergence was three times as large as the intraphylad divergence.
Specifically,
the mean within phylad divergence
was 0.012 and 0.016 for phylad A and
B respectively,
whereas the mean sequence divergence
between the two phylads was 0.037, which is among
the highest degrees of within-species
mtDNA divergence reported in marine fish species (see Billington
and Hebert 1991, table 5).
Geographic
Variation
Appendix
1 gives the distribution
of the 46 mitotypes in the ten samples. The Monte-Carlo
simulation
method of Roff and Bentzen (1989) showed that the
samples are highly heterogeneous.
Out of 1,000 trials
none produced a chi-square value higher than that produced by the observed distribution.
The frequencies of
Appendix
1 were used to calculate Nei’s pairwise genetic distances and construct a Fitch-Margoliash
phenogram among populations
(fig. 5A). It can be seen from
the phenogram that the Black Sea and the three North
Aegean samples are grouped together. Adriatic appears
to be distinct from the other samples due to its low
frequency of phylad A, but more samples are needed
to substantiate
this difference.
The recognition
of two distinct mitotype phylads
raises the question of how these phylads are distributed
among samples. Table 1 summarises
Appendix
1 in
pAT2
I
Pbylad B
LION
FIG. 5.-Fitch-Margoliash
phenograms of the samples, based on
Nei’s genetic distances calculated in different ways. A: Genetic distances calculated using the mitotype proportions (as shown in Appendix 1). B: Genetic distances obtained when each phylad is considered
separately. The phenogram of phylad A is depicted in the same scale
as the phenogram for phylad B (small ellipse) for the sake of comparison, and in a larger scale (large ellipse). The Black Sea was excluded from the phenogram of phylad B, because it contains practically
no type B, and the three North Aegean samples were pooled into one
to minimize the effect of low number of B mitotypes in each individual
sample.
terms of the most common mitotypes in each phylad.
For 54 individuals,
not included in the appendix, the
complete five-fold restriction profile could not be obtained, but the phylad could be safely inferred from the
scored enzymes. These individuals are included in table
1 (part B). The frequencies
of phylads are shown in
figure 1. Nei’s distances were also calculated considering the mitotypes of each phylad separately and the
resulting Fitch-Margoliash
phenograms
are shown in
figure 5B. It can be seen from figure 5B that there is
considerable
intra-phylad
heterogeneity
among samples for phylad B, but practically none for phylad A.
The Evolutionary
History of Anchovy
mtDNA
The marked genetic differentiation
among geographic populations
of anchovy that we report here is
not common among marine fishes with pelagic eggs
and larvae (Avise, Reeb, and Saunders 1987). Also un-
184
Magoulas
et al.
Table 1
Distribution
of Mitotypes
BLS
A. Animals
B. Animals
KAVl
with all five enzymes
AAAAA . .
Other.....
Total A . .
BBBBB . .
BBCBB . .
BBBBC . .
BBCBC . .
Other B . . .
TotalB...
Total
Partitioned
58
11
69
0
0
0
0
1
1
. . . . . 70
KAV2
. 72
.. 1
Total . . . . . 73
PAG
SAR
PAT1
PAT2
30
6
36
30
28
8
6
10
82
17
4
21
10
9
3
5
6
33
ADR
LION
BISC
Total
5
2
7
21
16
9
2
2
50
17
4
21
14
6
2
4
3
29
16
3
19
11
2
10
1
4
28
304
73
377
112
81
40
23
40
296
scored
90
31
121
8
8
1
1
2
20
40
6
46
2
6
0
2
1
11
14
3
17
1
0
1
0
1
3
17
3
20
15
6
6
2
10
39
141
57
20
59
118
54
57
50
47
673
with five or fewer enzymes
Phylad A
Phylad B
in Two Phylads
scored
126
21
46
11
17
3
22
49
43
99
21
33
9
56
21
29
19
29
396
331
147
57
20
71
142
54
65
50
48
727
Noms.-Within
phylad A the common mitotype (AAAAA) is treated separately and all the others are pooled in one
class. Within phylad B the four most common mitotypes (BBBBB, BBCBB, BBBBC, BBCBC) are treated separately and
the rest pooled.
usual is the deep division of mitotypes in two phylads
and the heterogeneous
geographic distribution
of these
phylads. Avise et al. (1987) have distinguished
five
phylogeographic
categories on the basis of molecular
divergence
of mitotypes
and their geographic
distribution. Anchovy fits category II, which is characterized
by the co-existence
of discontinuous
and highly diverged mitotypes in several geographic regions. This
pattern implies an allopatric evolution of the diverged
mitotypes and a secondary contact within the areas of
coexistence.
The evidence
summarized
in table 2
strongly suggests that phylad A has evolved in a population that maintained
a much lower effective size
than was the case for phylad B. The present-day
disTable 2
Values of Various
Mitotypes
Measurements
tribution of mitotypes and the geologic record of the
region converge on the hypothesis that the Black Sea
is the place of origin of phylad A. The strongest evidence comes from the geographical
isolation of the
Black Sea from the Mediterranean
and the fact that the
present population
in this sea consists almost exclusively of individuals
of phylad A. The mean mtDNA
sequence divergence (Nei 1987) among individuals
of
the Black Sea population (excluding the single type B
animal) is 0.0017. If we use the conventional
mtDNA
evolutionary
rate of 2% per million years (e.g., Moritz,
Dowling, and Brown 1987), this value translates into
an expected average time to common ancestry for the
mitotypes of ca. 80,000 years. Since anchovies have
for Phylad A, Phylad B, and the Whole Collection of
Phylad A
Number of haplotypes
..... ...........
Branching pattern . . . . . . . . . . . . . . . . . . . .
Number of homoplasies
...............
Number of steps the most distant mitotype is removed from the presumed ancestral mitotype . . . . . . . . . . . . . . . . . . . .
Mitotype diversity . . . . . . . . . . . . . . . . . . . .
Nucleotide diversity . . . . . . . . . . . . . . . . . .
Tajima’s D . . . . . . . . . . . . . . . . . . . . . . . . . .
Tajima’s D corrected for homoplasies . . . .
23
Radial
1
2
0.347
0.0017
-2.203 (P < O.Ol)a
-2.194 (P < 0.01)
Phylad B
23
Complex
4
4
0.756
0.005 1
- 1.289 (NS)b
- 1.106 (NS)
Total
46
21
0.748
0.0171
0.012 (NS)
0.074 (NS)
a P = probability that the value is nonsignificant.
b NS = nonsignificant.
Noms.-For
the calculation of mitotype and nucleotide diversities for phylads all individuals belonging to the same
phylad were pooled into one hypothetical population, assuming that the collection of our samples is representative of the
global distribution of mitotypes within phylads.
mtDNA Variation
roughly one generation
per year (see Magoulas
and
Zouros 1993), the approximate
female effective population size can be estimated at 80,000 (the mean number of generations
to common
ancestry;
see Avise
1989). If the rate of evolution
is only a fifth of the
above (Rand 1994, and references therein), the effective population
size will be 400,000. Even the latter
estimate is incompatible
with the present huge sizes of
anchovy populations in the Black Sea (the usual annual
catch is approximately
500,000 mt [GFCM statistical
bulletin no. 8, FAO, Rome, 1991, p. 2011, which corresponds roughly to 20 X lo9 individuals).
A population bottleneck in the Black Sea could explain this reduction of the effective population size. The large number of mitotypes in phylad A, all of which are recent
derivatives of AAAAA, suggests that this was the mitotype that survived the bottleneck
and gave rise to
other mitotypes, perhaps during an explosive recovery
of the population following the bottleneck.
Support for this hypothesis comes from the application of Tajima’s
test for neutrality
(Tajima 1989),
which compares the observed average number of pairwise nucleotide
differences
(k) between haplotypes in
a sample to the number (A4) expected from the number
of segregating
sites. Under the infinite site mutation
model the quantity D, defined as k - M, is zero. A
positive D indicates balancing
selection or population
subdivision.
A negative D indicates directional
selection, recent population
bottleneck,
or background
selection of slightly deleterious
alleles (see also Rand,
Dorfsman,
and Kann 1994). The results from the application of Tajima’s test are shown in table 2. For phylad B the test gives a negative but not significant D.
For phylad A it gives a negative and significant D. The
significance
remains after correction for the incidence
of homoplasy (by estimating the number of restriction
site differences
between the mitotypes
from the network of fig. 3 rather than from the matrix of presence/
absence of restriction sites). The negative value of phylad A provides further support for a recent bottleneck
in the Black Sea, although it could also be explained
by a recent selective sweep.
The Population History of the Anchovy
in the Mediterranean
The existence of two discontinuous
mitotype phylads and the pattern of geographic
distribution
of mitotypes in anchovy (figs. 1, 3, and 5) can be understood
in terms of past geological
events and current hydrographic patterns in the Mediterranean.
Specifically, we
will provide evidence to support the following points:
(a) The anchovy entered the Mediterranean
about 5
million years ago, when the present opening of Gibraltar was established.
(b) Phylad A originated in the
of Anchovy
185
Black Sea, which was colonized through the Mediterranean at the end of the Pliocene. (c) The current high
frequency of phylad A in the Mediterranean
and the
Atlantic is due to a massive outflow from the Black
Sea. (d) Current hydrographic
forces contribute to the
dynamic maintenance
of the geographical
structure of
mitotypic diversity that originated from these geologic
events.
Colonization
of the Mediterranean
and the Black Sea
Anchovy is not included in the list of fish species
cited by Por and Dimentman
(1985) as possible Messinian relicts in the Mediterranean,
i.e., species that
survived the Messinian
salinity crisis in situ. It must
have entered the Mediterranean
from the Atlantic during the Trubi transgression,
at the beginning
of the
Pliocene, when the communication
between Atlantic
and Mediterranean
was opened again and marine conditions were re-established
permanently
in the Mediterranean.
It is known that 80% of the Black Sea fish species
are of Mediterranean
origin, the rest tracing their origin
from Paratethys, the body of water which during Oligocene and Miocene covered present-day
central Europe and extended beyond the Black and Caspian Seas
(Ekman 1968; Quignard 1978). The view that the Black
Sea anchovy is a relict of the Tertiary Period was entertained in the past, but is no longer accepted (Zenkevitch 1963, p. 435). This view is also supported by
the fact that the implied long separation of the AtlantoMediterranean
and Paratethyan
stocks (as far back as
15-20 million years) would have caused a higher divergence between phylads A and B, even under the
assumption of a slow rate of mtDNA evolution for fish.
We conclude that anchovy has been one of the many
species that invaded the Black Sea during the late Pliocene, when this sea was connected to the Mediterranean.
As mentioned,
several lines of evidence suggest
that the Black Sea is the place of origin of phylad A
and that a population bottleneck resulted in the “star”
phylogeny and low diversity of this phylad. An explanation for this presumptive
bottleneck can be found in
the glaciation history of the Black Sea in the Pleistocene. During glacial maxima, the anchovy populations
must have suffered drastic reductions
in the isolated
Black Sea, as this species cannot reproduce in temperatures below 13°C (Demir 1963; Palomera 1992). The
most recent bottleneck in the Black Sea may have occurred 20,000-30,000
years ago, during the height of
the last Glacial (Por 1989, pp. 24 and 27).
186
Magoulas et al.
Massive Exit of Anchovy
Mediterranean
from
the Black Sea into the
Phylad A is now present in all localities sampled.
Several observations
suggest that this is the result of a
recent exit of anchovy from the Black Sea into the
Eastern Mediterranean
rather than of gradual diffusion
via the Aegean Sea. These include the relatively young
age of phylad A, the high homogeneity
of intra-phylad
A composition
of samples (fig. 5B), the fact that type
AAAAA is by far the most common mitotype of phylad A in all samples, and that the minor A types found
in all samples are recent derivatives
of this type. The
sharp drop of the phylad’s frequency over the relatively
small distance separating the North from the Central
Aegean suggests that, under the present regime of hydrographic conditions
in the Aegean, gene flow from
the North Aegean into the main Mediterranean
is not
free.
Massive water spills from the Black Sea into the
Mediterranean
during the interglacial
periods and especially the deglaciation
events (or terminals)
are accepted facts in physical oceanography.
According
to
Herman (1988), “following
warming and subsequent
massive deglaciation,
sea level rose; when the sea stand
reached the Bosporus sill (-36 m) the connection
between the Mediterranean
and the Black Sea was reestablished and the low salinity Black Sea water spilled
over into the Mediterranean”,
resulting in the flooding
of the eastern Mediterranean
by large volumes of low
salinity water. Other authors (Por 1989, and references
therein) write that by the end of the last Glacial and
especially
during its termination
event (9,000-12,000
years ago), the Levantine
basin was hydrographically
separated from the Ionian basin by a tongue of diluted
and cold (13°C) water penetrating
from the direction
of the Aegean Sea and reaching the African coast during the winter. We note that the population
recovery
from the bottleneck
in the Black Sea and the outflow
into the Mediterranean
are not very distant in terms of
geologic time, as both must have occurred during the
warming period after the last glacial maximum.
Current Factors Affecting the Phylogeography
Anchovy in the Mediterranean
of
Phylad B is practically absent from the Black Sea,
while it has a frequency
of 15% in the neighbouring
North Aegean. This can be explained
by postulating
either selection against phylad B in the Black Sea or
that there is a blockade in the anchovy gene flow from
the Aegean to the Black Sea. Everything
that we know
about the hydrography
of the passage joining the two
seas and the biology of the species is consistent with
a view of unidirectional
gene flow. The elongated (ca.
31 km), narrow (no more than 1.5 km wide in most of
its length and 550 m at its narrowest point), and shallow (ca. 30 m at its shallowest point) Straits of Bosporus heavily restricts free water circulation
and is to
a large extent responsible
for the preservation
of the
physicochemical
and biotic differences that distinguish
the Black Sea from the Mediterranean
Sea. As an example, salinity is as high as 38-39%0 in the Aegean
Sea, but only 17-18%0 in the Black Sea. In the Straits
of Bosporus there is a surface current of cold, lowsalinity (17%~~) water flowing from the Black Sea towards the Marmara Sea and a subsurface
current of
warmer, high-salinity
(38.5%0) water flowing in the opposite direction. In volume, the former is twice as large
(Vergnaud-Grazzini
1985). Eggs and larvae of anchovy
are primarily distributed
in shallow layers, mainly in
the upper 10 m (e.g., Demir 1963; Palomera 1991). The
anchovy spawns in most parts of the Black Sea, but to
a greater extent in the southern regions (Nierman et al.
1994). In the mouth of Bosporus in the Black Sea the
planktonic
stages of the anchovy must be almost exclusively
at the upper part of the water column that
moves swiftly southward. The lower northward-moving current contains no immature stages and, besides,
upon entering the northern end of the straits, it sinks
precipitously
into the hostile oxygen-deprived
deep
water of the Black Sea (Tolmazin 1985). Demir (1963)
suggests that adult anchovy exit the Black Sea into the
Sea of Marmara in the fall and move back in the Black
Sea in the spring, but provides no evidence for this
claim. In contrast, Ivanov and Beverton (1985) provide
detailed information
about anchovy migration within
the Black Sea but do not include the anchovy among
the species that migrate as adults through the Bosporus
straits. Kocatas et al. (1993) list a number of species
that move from the Sea of Marmara to the Black Sea,
and, again, anchovy is not included among them. It
seems, therefore, that the unidirectional
flow of immature stages from Black Sea to the Aegean is not
counterbalanced
by active movement of adults in the
opposite direction.
The northern Aegean is the Mediterranean’s
portal
area of Black Sea waters. Three samples of anchovy
from this area, taken from two different stations at two
seasons 3 years apart, produced a consistently
high frequency (85%) of phylad A (fig. 1). To a large extent
this is the result of gene flow from the Black Sea into
the Aegean. The large drop of this frequency
in the
Central Aegean is suggestive of a second barrier to free
gene flow, one that geographically
coincides with the
boundary
of the northern and southern basins of the
Aegean Sea. The subdivision
of the Aegean into two
basins has been long recognized. The northern basin is
characterized
by a continental
shelf favouring neritic
mtDNA Variation
FIG. 6.-Surface
geostrophic currents in the Aegean and Adriatic
Seas. The gyres in northern Aegean and southern Adriatic are shown
(modified from Ovchinnikov
1966).
and meroplanktonic
forms (Moraitou-Apostolopoulou
1985). This basin is under the influence of cool, lowsalinity Black Sea waters which are entrained into a
cyclonic circulation affecting the northern and western
parts of the Aegean and causing an ecological isolation
of the northern basin from the southern basin (Theocharis et al. 1993) (see fig. 6). In the southern basin
the continental
shelf is very limited and the area has
the characteristics
of a pelagic zone with a fauna that
resembles the Western and Eastern Mediterranean
(Moraitou-Apostolopoulou
1985).
A similar north-south
subdivision
is known to exist in the Adriatic Sea. According
to Artegiani et al.
(1993), the northern part is under the strong influence
of fresh water inflow, mainly from the PO River, whereas a topographically
controlled cyclonic gyre, situated
in the South Adriatic pit, partially isolates the northern
Adriatic from the influence of the Mediterranean
(fig.
6). This isolation is also reflected by the anchovy distribution within the Adriatic. In the southern Adriatic
anchovy is 80 times less abundant than in the northern
Adriatic (Casavola et al. 1988). The low frequency of
phylad A in the northern Adriatic (14%) can be seen
as a result of a historic event that prevented its entrance
when this phylad became abundant in the rest of the
Mediterranean
and of current conditions
that prevent
the free movement between the South and North Adriatic.
It is known that the Straits of Gibraltar does not
represent a zoogeographic
barrier for most marine species, with the result that the Mediterranean
and the adjacent Atlantic coasts support very similar faunas and
form together what is recognized
as the Atlanto-Mediterranean
province
(e.g., Ekman
1968; Tortonese
1964). The homogeneity
of the sample from the Bay
of Biscay with those from the Mediterranean
in terms
of phylad A and B proportions
is consistent
with the
of Anchovy
187
free movement of anchovy through the Straits of Gibraltar. The difference of this sample from the others in
terms of intra-phylad
B composition
remains unexplained. This is, however, a minor difference, since all
four major B mitotypes of the Mediterranean
samples
were found in high numbers in the Bay of Biscay and
BBBBB was also the most common type in this sample.
An electrophoretic
study of populations
of anchovy from the North Aegean and Ionian Seas (Spanakis,
Tsimenides,
and Zouros 1989), produced results that
varied among loci. All samples were statistically
homogeneous
for one of the four polymorphic
enzyme
loci (Est). The samples from the Aegean were highly
heterogeneous
at two loci (Ldh and Pgm), whereas the
samples from the Ionian were not. All the samples from
the Aegean were homogeneous
for the fourth locus
(Pgi) and so were the samples from the Ionian, but the
two seas were highly heterogeneous
at this locus.
These results are consistent
with recent suggestions
that geographic patterns of allozyme variation may be
influenced
more by locus specific selection than by
forces of gene flow and random drift and may, thus, be
less useful in reconstructing
the phylogenetic
history
of a species (Karl and Avise 1992). However, the fact
that the only enzyme locus that joined the samples
from the same sea into a homogeneous
group also recognized the two seas as genetically heterogeneous
populations is consistent with our present finding of a large
mtDNA differentiation
between North Aegean and Ionian samples.
The two anchovy phylads were found to coexist
over a large and ecologically
diverse area extending
from the North Aegean to the Bay of Biscay. This suggests that the phylads A and B are not coincident with
geographical
races or subspecies, which has been suggested by several authors in the past (Spanakis, Tsimenides, and Zouros 1989). Both phylads were also
collected from the same schools in all these localities
(except the Black Sea). This apparent lack of ecological or ethological
differentiation
between animals belonging to different phylads suggests that the molecular
discontinuity
and the marked geographical
differences
of the various mitotypes that we have described here
are not in themselves evidence for the existence of two
incipient taxa. At present, we have no evidence to suggest that the history of the anchovy that resulted in this
rich and informative
pattern of mitotype diversity has
led to the development
of any form of reproductive
isolation between local populations
or between mitotypic variants occurring within the same population.
Acknowledgments
We are grateful to a number of people who helped
us with the collection of the samples: Prof. F? Kolarov,
- _-
188
Magoulas et al.
Dr. I. Dobrovolov,
K. Michailov, Dr. G. Kotoulas, Dr.
T. Patarnello,
Dr. G. Goulielmos,
Dr. A. Machias, and
A. Argyrokastritis.
We thank K. Ekonomaki and V. Terzoglou for technical assistance, Dr. G. Kotoulas for discussion and comments during the whole course of the
work, and Dr. I? Bentzen for providing the clones of
shad mtDNA. A.M. was supported by a graduate scholarship from the Institute of Marine Biology of Crete
and the Institute of Molecular
Biology and Biotechnology (Foundation for Research and Technology-Hellas). This research was supported by a contract from
the Commission
of the European Communities.
APPENDIX 1
Distribution of Mitotype Frequencies in the Samples. Several Individuals were Heteroplasmic. Heteroplasmy in Anchovy Has Been Discussed Elsewhere (Magoulas and Zouros 1993). For the Purpose of the Present Analysis Only the “Stronger” of the Two
Mitotypes Found in Heteroplasmy
Was Taken into Account.
BLS
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
AAAAA . . .
AAAAG . . .
AAAAH . . .
AAAAI . . . .
AAAAJ....
AAAAL . . .
AAABA . . .
AAACA . . .
AAADA . . .
AABAA . . .
AAEAA . . .
AAFAA . . .
AAGAA . . .
AAHAA . . .
ABAAA . . .
AGAAG . . .
AIAAA . . . .
AJAAA . . . .
CAAAA . . .
DAAAA . . .
EAAAA . . .
EAADA . . .
GAABA . . .
BABBB . . .
BABBC . . .
BBABC . . .
BBBBA . . .
BBBBB.. . .
BBBBC.. . .
BBBBD . . .
BBBBE....
BBCBA . . .
BBCBB.. . .
BBCBC.. . .
BBCBF . . . .
BBCBK . . .
BBCBM . . .
BBCBN . . .
BCBBB.. . .
BCDBB . . .
BDCBF....
BEBBB . . . .
BEBBC . . . .
BFBBB . . . .
BHBBC . . .
FBBBC . . . .
58
1
5
0
0
0
1
1
0
0
0
0
0
0
0
2
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Total
70
......
KAVl
KAV2
PAG
SAR
PAT1
PAT2
ADR
LION
BISC
Total
90
2
2
0
0
4
4
5
0
3
1
1
2
0
1
0
0
0
3
1
1
1
0
0
0
0
0
8
1
0
0
0
8
1
1
0
0
0
0
1
0
0
0
0
0
0
40
0
0
0
1
2
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
2
0
0
0
0
6
2
0
0
0
0
0
0
0
0
0
0
0
0
14
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
17
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
1
0
0
3
15
6
0
0
1
6
2
0
0
0
1
1
0
0
0
1
0
1
1
30
0
0
0
0
3
0
0
0
0
0
1
0
1
0
0
0
0
0
1
0
0
0
4
0
1
0
30
8
0
0
0
28
6
0
1
1
0
1
0
1
1
0
0
0
0
17
0
0
0
0
0
0
0
0
1
0
0
0
0
2
0
0
0
1
0
0
0
0
1
0
0
0
10
3
1
0
0
9
5
0
1
0
0
0
0
1
0
0
1
0
1
5
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
21
9
0
0
0
16
2
0
0
0
0
0
0
0
0
0
0
0
0
17
0
1
1
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
14
2
0
1
0
6
4
0
0
0
0
0
0
0
0
0
0
0
0
16
1
0
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
2
0
0
0
11
10
0
0
0
2
1
0
0
0
0
0
0
0
0
0
0
0
2
304
4
9
1
1
11
8
8
1
4
2
3
3
1
3
2
1
1
4
3
1
1
1
13
1
1
3
112
40
1
2
1
81
23
1
2
1
1
2
1
2
1
1
1
1
4
54
57
50
47
673
141
57
20
59
118
mtDNA Variation of Anchovy
LITERATURE CITED
APOLLONIUS OF RHODES. 3rd century BC. The Argonautica.
English translation by E.V. Rieu: The voyage of Argo, 2nd
ed. 1971. Penguin Books, Harmondsworth,
Middlesex, England.
ARTEGIANI, A., M. GACIC, A. MICHELATO, V. KOVACEVIC, A.
Russo, E. PASCHINI, I? SCARAZZATO,and A. SMIRCIC. 1993.
The Adriatic Sea hydrography and circulation in spring and
autumn (1985-1987). Deep-Sea Res. II 40: 1143-l 180.
AVISE, J. C. 1989. Gene trees and organismal histories: a phylogenetic approach to population biology. Evolution 43:
1192-1208.
-.
1992. Molecular population structure and the biogeographic history of a regional fauna: a case history with lessons for conservation biology. Oikos 63:62-76.
-.
1994. Molecular markers, natural history and evolution. Chapman 8z Hall, New York.
AVISE, J. C., J. ARNOLD, R. M. BALL, E. BERMINGHAM, T.
LAMB, G. NEIGEL, C. A. REEB, and N. C. SAUNDERS. 1987.
Intraspecific
phylogeography:
The mitochondrial
DNA
bridge between population genetics and systematics. Annu.
Rev. Ecol. Syst. l&489-522.
AVISE, J. C., C. A. REEB, and N. C. SAUNDERS. 1987. Geographic population structure and species differences in mitochondrial DNA of mouthbrooding
marine catfishes (Ariidae) and demersal spawning toadfishes (Batrachoididae).
Evolution 41:991-1002.
BACESCU, M. 1985. The effects of geological and physicochemical factors on the distribution of marine plants and
animals in the Mediterranean.
Pp. 195-212 in M. MORAITOU-APOSTOLOPOULOUand V. KIORTSIS, eds. Mediterranean
marine ecosystems. Plenum Press, New York and London.
BALLARD, W. J., and M. KREITMAN. 1994. Unraveling selection
in the mitochondrial genome of Drosophila. Genetics 138:
757-772.
BILLINGTON,N., and D. N. HEBERT. 1991. Mitochondrial DNA
diversity in fishes and its implications
for introductions.
Can. J. Fish. Aquat. Sci. 48(Suppl. 1):80-94.
CASAVOLA, N., G. MARANO, L. DE MARTINO, and C. SARAGINO. 1988. Preliminary evaluation of anchovy and sardine
stocks in the lower Adriatic. FAO Fisheries Report 394:8&
90.
CASTELOE, J., and A. R. TEMPLETON. 1994. Root probabilities
for intraspecific gene trees under neutral coalescent theory.
Mol. Phylogenet. Evol. 3: 102-l 13.
CRANDALL,K. A., and A. R. TEMPLETON. 1993. Empirical tests
of some predictions from the coalescent theory with applications to intraspecific phylogeny reconstruction.
Genetics
134:959-969.
DEMIR, N. 1963. Synopsis of biological data on anchovy Engruulis encrusicolus (Linnaeus) 1758 (Mediterranean
and
adjacent seas). FAO Fisheries Synopsis 26. Rome.
DERMITZAKIS, M. 1984. A guide to the geology of Greece.
University of Athens. Athens.
ECONOMIDIS,I?, and M. L. BAUCHOT. 1976. Sur une collection
de poissons de mers helleniques (mers Agee et Ionienne)
deposee au Museum national d’Histoire
naturelle. Bull.
Mus. Nat. Hist. Nat. 392:871-903.
189
ELAN, S. 1968. Zoogeography of the sea. Sidgwick and Jackson, London.
FEINBERG, A. I?, and B. VOGELSTEIN. 1983. A technique for
radiolabeling DNA restriction endonucleases
fragments to
high specific activity. Anal. Biochem. 1326-13.
FELSENSTEIN,J. 1993. Phylogeny Inference Package, version
3.5~. University of Washington.
HARRISON, R. G., D. M. RAND, and W. C. WHEELER. 1985.
Mitochondrial DNA size variation within individual crickets. Science 228: 1446-1448.
HERMAN, Y. 1988. Paleoenvironments
and episodic stagnation
of the eastern Mediterranean,
during the Brunhes Epoch.
Rapp. Comm. Int. Mer. Medit. 31,2:106.
Hsu, K. J., M. B. CITA, and W. B. E RYAN. 1973. The origin
of Mediterranean
evaporites. Init. Repts. Deep Sea Drill.
Proj. 13:1203-1231.
Hsu, K. J., L. MONDATERT, D. BERNOULLI, M. B. CITA, A.
ERICKSON, R. E. GARRISON, R. B. KIDD, I? MELIERES, C.
MULLER, and R. WRIGHT. 1978. History of the Mediterranean salinity crisis. Init. Repts. Deep Sea Drill. Proj. 42:
1053-1078.
IVANOV, L., and R. J. H. BEVERTON. 1985. The fisheries resources of the Mediterranean.
Part two: Black Sea. Gen.
Fish. Count. Medit., Studies and Reviews 60. FAO, Rome.
KARL, A., and J. C. AVISE. 1992. Balancing selection at allozyme loci in oysters: implications from nuclear RFLPs. Science 256: 100-102.
KOCATAS, A., T. KORAY, M. KAYA, and 0. E KARA. 1993.
Review of the fishery resources and their environment in
the Sea of Marmara. FAO Stud. Rev. 64:87-143.
LANSMAN, R. A., R. 0. SHADE, J. E SHAPIRA, and J. C. AVISE.
1981. The use of restriction endonucleases to measure mitochondrial DNA sequence relatedness in natural populations. III. Techniques and potential applications. J. Mol.
Evol. 17:214-226.
MAGOULAS, A., and E. ZOUROS. 1993. Restriction-site
heteroplasmy in anchovy (Engraulis encrusicolus) indicates incidental biparental inheritance of mitochondrial DNA. Mol.
Biol. Evol. 10:3 19-325.
MCELROY, D, I? MORAN, E. BERMINGHAM,and I. KORNFIELD.
1991. The Restriction Enzyme Analysis Package, version
4.1. Smithsonian Tropical Research Institute, Balboa, Panama.
MORAITOU-APOSTOLOPOULOU, M. 1985. The zooplankton
communities of the Eastern Mediterranean
(Levantine basin, Aegean Sea); influence of man-made factors. Pp. 303331 in M. MORAITOU-APOSTOLOPOULOUand V. KIORTSIS,
eds. Mediterranean marine ecosystems. Plenum Press, New
York, London.
MORITZ, C., T. DOWLING, and W. M. BROWN. 1987. Evolution
of animal mitochondrial DNA: relevance for population biology and systematics. Annu. Rev. Ecol. Syst. l&269-292.
NEI, M. 1972. Genetic distance between populations- Am. Natur. 106:283-292.
-.
1987. Molecular evolutionary
genetics. Columbia
University Press, New York.
NEI, M., and J. C. MILLER. 1990. A simple method for estimating average number of nucleotide substitutions within
190 Magoulas et al.
and between populations from restriction data. Genetics
125:873-879.
NEI, M., and E TAJIMA. 1981. DNA polymorphism detectable
by restriction endonucleases. Genetics 97: 145-163.
NIERMAN,U., E BINGEL,A. GORBAN,A. D. GORDINA,A. C.
Gucu, A. E. KIDEYS,A. KONSULOV,G. RADU, A. A. SUBBOTIN,and V. E. ZAIKA. 1994. Distribution of anchovy eggs
and larvae (Engruulis encrusicolus) in the Black Sea in
1991-1992. ICES J. Mar. Sci. 51:395406.
OVCHINNIKOV,
I. M. 1966. Circulation in the surface and intermediate waters of the Mediterranean Sea. Oceanology 6:
48-59.
PALOMERA,I. 1991. Vertical distribution of eggs and larvae of
Engruulis encrusicolus in stratified waters of the western
Mediterranean. Mar. Biol. 111:37++.
. 1992. Spawning of anchovy Engruulis encrusicolus in
the Northwestern Mediterranean relative to hydrographic
features in the region. Mar. Ecol. Prog. Ser. 79:215-223.
PASTOR,X., ed. 1991. The Mediterranean. Greenpeace Books,
Collins and Brown, London.
POR, E D. 1978. Lessepsian migration. Springer-Verlag, BerlinHeidelberg.
-.
1989. The legacy of Tethys. An aquatic biogeography
of the Levant. Kluwer Academic Publishers, Dodrecht.
POR, E D., and C. DIMENTMAN.1985. Continuity of Messinian
biota in the Mediterranean basin. Pp. 545-557 in D. J.
STANLEYand E C. WEZEL,eds. Geological evolution of the
Mediterranean basin. Springer-Verlag, Berlin.
QUIGNARD,J. l? 1978. Le Mediterranee creuset ichthyologique.
Bull. Zool. 45(Suppl. 11):23-36.
RAND,D. M. 1994. Thermal habit, metabolic rate and the evolution of mitochondrial DNA. TREE 9: 125-131.
RAND, D. M., M. DORFSMAN,and L. KANN. 1994. Neutral and
non-neutral evolution of Drosophila mitochondrial DNA.
Genetics 138:741-756.
Row, D. A., and l? BENTZEN.1989. The statistical analysis of
mitochondrial DNA polymorphisms: x2 and the problem of
small samples. Mol. Biol. Evol. 6:539-545.
ROHLF, E J. 1992. Numerical Taxonomy and Multivariate
Analysis System, version 1.70. Exeter Software, New York.
SAMBROOK,J., E. E FRITSCH,and T. MANIATIS.1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
SARA, M. 1985. Ecological factors and their biogeographic
consequences in the Mediterranean ecosystems. Pp. 1-17 in
M. MORAITOU-APOSTOLOPOULOU
and V. KIORTSIS, eds.
Mediterranean marine ecosystems. Plenum Press, New York
and London.
SELLI,R. 1985. Evolution of the Tyrrhenian Sea. Pp. 131-151
in D. J. STANLEYand E C. WEZEL, eds. Geological evolution of the Mediterranean basin. Springer-Verlag, Berlin.
SPANAKIS,E., N. TSIMENIDES,and E. ZOUROS. 1989. Genetic
differences between populations of sardine, Surdinu pilchurdus, and anchovy, Engruulis encrusicolus, in the Aegean and Ionian Seas. J. Fish. Biol. 35417-437.
SWOFFORD,D. L. 1985. Phylogenetic analysis using parsimony, version 2.4. Illinois Natural History Survey, Campaign,
Ill.
TAJIMA,E 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123:
585-595.
THEOCHARIS,
A., D. GEORGOPOULOS,
A. LASCARATOS,
and K.
Nrrr~s. 1993. Water masses and circulation in the central
region of the Eastern Mediterranean: Eastern Ionian, South
Aegean, and Northwest Levantine 1986-1987. Deep-Sea
Res. II 40:1121-l 142.
TOLMAZIN,D. 1985. Changing of the coastal oceanography of
the Black Sea II: Mediterranean effluent. Prog. Oceanogr.
15:277-3 16.
TORTONESE,E. 1964. The main biogeographical features and
problems of the Mediterranean fish fauna. Copeia 1964:98107.
TORTONESE,E. 1985. Distribution and ecology of endemic elements in the Mediterranean fauna (fishes and echinoderms). Pp. 57-83 in M. MORAITOU-APOSTOLOPOULOU
and
V. KIORTSIS,eds. Mediterranean marine ecosystems. Plenum Press, New York and London.
VERGNAUD-GRAZZINI,
C. 1985. Late Cenozoic stable isotope
record. Pp. 414451 in D. J. STANLEYand E C. WEZEL,
eds. Geological evolution of the Mediterranean basin.
Springer-Verlag, Berlin.
WHITEHEAD,I? J. I? 1984. Engraulidae. Pp. 282-283 in l? J. l?
WHITEHEAD,M. L. BOUCHOT,J. C. HURGAN,J. NIELSEN,
and E. TORTONESE,eds. Fishes of the northeastern Atlantic
and the Mediterranean. UNESCO.
ZENKEVITCH,
L. 1963. Biology of the seas of the U.S.S.R. G.
Allen and Unwin, London.
RICHARD G. HARRISON, reviewing
Accepted
September
13, 1995
editor