Mitochondrial DNA, June 2011; 22(3): 55–65

FULL LENGTH RESEARCH PAPER

Divergence at Cyt-b and Co-1 mtDNA genes on different taxonomic

levels and genetics of speciation in animals

YURI PH. KARTAVTSEV

Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11

A.V. Zhirmunsky Institute of Marine Biology of the Far Eastern Branch of the Russian Academy of Sciences, Vladivostok,
Russia

(Received 12 October 2010; revised 26 October 2010; accepted 10 March 2011)

Abstract
Genetic divergence estimates using p-distances and similar measures were generated for 20,731 vertebrate and invertebrate
animal species. The results of this analysis demonstrate that the data series are realistic and interpretable when the p-distance
and its various derivates are used. The focus is on vertebrates and fish species in particular and the newest data set.
Distance data reveal increasing levels of genetic divergence of the sequences of the two genes, cytochrome b (Cyt-b) and
cytochrome c oxidase subunit 1 (Co-1), in the five groups compared: populations within species; subspecies, semi-species,
or/and sibling species; species within a genus; species from different genera within a family; and species from separate families
For personal use only.

within an order. Mean unweighted scores of p-distances (%) for these five groups are Cyt-b—1.38 ^ 0.30, 5.10 ^ 0.91,
10.31 ^ 0.93, 17.86 ^ 1.36, and 26.36 ^ 3.88, respectively; and Co-1—0.89 ^ 0.16, 3.78 ^ 1.18, 11.06 ^ 0.53,
16.60 ^ 0.69, and 20.57 ^ 0.40, respectively. The estimates show good correspondence with other analyses. These results
testify to the applicability of p-distance for most intra-species and inter-species comparisons of genetic divergence up to the
order level in animals for the two genes compared. Data reviewed provide empirical and theoretical background on
the geographic speciation mode prevalence in species origin and give a framework why per-individual species identification
(DNA barcoding) is usually successful.

Keywords: Nucleotide diversity, p-distance, speciation genetics, mitochondrial DNA, molecular evolution

Introduction family level. These genes proved to be useful for

Global initiatives such as the Consortium for
Barcoding of Life (http://www.barcoding.si.edu/), the
International Barcode of Life Project (http://www.
DNAbarcoding.org), and the Tree of Life Project
(http://tolweb.org/tree/) stimulated activity in the field
of molecular phylogenetics and evolutionary geno-
mics. Databases, especially on cytochrome oxidase
c subunit 1 (Co-1), have increased greatly. The main
objective of this study is a comparison of genetic
divergence within species and in a hierarchy of
taxonomical categories.
In the past decade, genes of mitochondrial DNA
(mtDNA) encoding cytochrome b (Cyt-b) and Co-1
proteins have been most frequently used for taxo-
nomic and phylogenetic analysis at the species and
estimating divergence in taxa up to the order level in
many animal groups (Johns and Avise 1998; Graur
and Li 1999; Hebert et al. 2002a,b; Creer et al. 2003;
Kartavtsev and Hanzawa 2007; Kartavtsev et al.
2007a,b, 2008, 2009a,b; Sasaki et al. 2007). A survey
of the evidence on intra-species divergence of
mitochondrial genes in 256 vertebrate, mostly sexually
reproducing species, indicated that 56% of them form
distinct intra-species maternal lines, which typically
are confined geographically (Avise and Walker 1999;
Avise 2000). Thus, the polytypic species or subdivi-
sion into groups of most species is well documented,
and this information is independent of ecological or
demographic data and in good agreement with
the latter. These data along with fundamental

Correspondence: Y. Kartavtsev, A.V. Zhirmunsky Institute of Marine Biology of the Far Eastern Branch of the Russian Academy of Sciences,
Vladivostok 690041, Russia. Tel: þ 7-4232-311173; fax: þ 7-4232-310900. E-mail: yuri.kartavtsev48@hotmail.com

ISSN 1940-1736 print/ISSN 1940-1744 online q 2011 Informa UK, Ltd.

DOI: 10.3109/19401736.2011.588215
 56 Y. Ph. Kartavtsev
biological generalizations made by Dobzhansky
(1955) and Mayr (1963) provide the basic ideological
framework for this paper.
In this paper, I do not consider problems related
to the construction and analysis of phylogenetic trees
and related phylogenetic issues. This is a specific
topic discussed elsewhere (Li and Zarkhih 1995;
Swofford et al. 1996; Nei and Kumar 2000; Hall
2001; Sanderson and Shaffer 2002; Felsenstein
2004). Also, a population differentiation in Fst-statistic
or similar scales is basically out of the scope of the
paper. Nucleotide diversity may be estimated in a
number of ways, for example, with measures such as
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
nucleotide diversity as the per-site measure, p (Nei
1987), and the proportion of different nucleotide cites at
a pair of randomly chosen sequences, the p-distance as
P or its estimate p (Nei and Kumar 2000, p. 33;
Kartavtsev 2009a,b, 2011). Understanding of DNA
sequence polymorphism as a result of nucleotide
substitution is of primary interest for molecular
phylogenetics. Amino acid sequence substitution rate is
also important to be estimated, but this is basically out
of the scope of this paper.
The main focus of this study is to consider the levels
of nucleotide diversity in animal populations and taxa
of different ranks and to find the possible relation of
For personal use only.
these data with some aspects of genetics of speciation.
Materials and methods
The primary nucleotide sequences of genes
(sequences for shortage) and their resemblance and
difference are the main source of data in this paper.
The conclusions are mainly based on the information
from the database on p-distances of two genes, Cyt-b
and Co-1, presented in recent reviews with necessary
details on source data and analysis (Kartavtsev and
Lee 2006; Kartavtsev 2009b, 2011).
The literature data were screened using the databases
of the Thompson Institute of Scientific Information,
the Science Citation Index SCI, the SCI database, and
other sources. Articles within 1995 – 2009 were
examined. Statistical analysis was performed using the
STATISTICA 6.0 software package (Statsoft 2001).
From this package, we employed the basic module
for calculating mean and variance parameters, as well
as those for parametric analysis of variance (ANOVA,
and the multi-dimensional version MANOVA),
canonical analysis, discriminant function analysis, and
Kruskall–Wallis non-parametric ANOVA.
Results and discussion
Genetic divergence within species and in a hierarchy
of taxonomical categories
The biological species concept (BSC) implies that a
species is an isolated reproductive unity. The molecular
data, especially pertaining to mtDNA, show that, on
the one hand, natural hybridization between species
may lead to introgression of genes from one gene pool
to the other one. On the other hand, sequences of
individual genes exemplify that the variability of DNA
markers increases with the rank of the taxon (Johns
and Avise 1998; Hebert et al. 2002a; Ward et al.
2005). Hence, I believe it is expedient to compare the
data on nucleotide divergence for several genes, from
several data sources, and, in addition, to substantiate
both the variability and distance parameters. The
latter is important for understanding the essence of
estimating a divergence at the DNA or other markers
and its connection to species identification and to
speciation process. Such complications as saturation
with mutations and an unequal rate of substitutions
among sites are able to obscure real DNA variability.
There may be other hidden factors too. In particular,
various genes may encode different functional proper-
ties of phenotype (macromolecules firstly), and this
obviously has an impact on distance estimates (Graur
and Li 1999; Kartavtsev 2009a, 2011).
If the nucleotide sequence for a particular set of loci
or alleles in a population sample is known, then
DNA polymorphism can be assessed in several ways.
The best measures of DNA sequence divergence are
p and p or its derivates (see details in Nei 1987;
Nei and Kumar 2000; Kartavtsev 2009a, 2011).
Numerical simulations showed that when p-distances
are small (, 20%), different substitution models give
similar scores of in-time diversity (Nei and Kumar
2000, p. 41). Useful to remember as well is that
because of heterogeneity of substitution rates along
sequences and different parts of genes, an important
correction of the p-distance is gamma correction
(e.g. Nei and Kumar 2000; Felsenstein 2004).
Divergence of DNA nucleotide sequences on the intra-
species and inter-species levels. As measures for
comparison, one may employ uncorrected
p-distances, distances of the two-parameter Kimura
model (K2P), or other indices (GTR, General Time
Reversible model, TrN, Tamura-Nei model etc.), used
in the literature for the genes Cyt-b, Co-1, and others
(Kartavtsev 2009a, 2011). The possibility of their use
follows from theory and from numerical simulation, as
noted in Nei and Kumar (2000, p. 41) and outlined
earlier (Kartavtsev 2009a, 2011).
Variation rows of pairwise K2P comparisons for
sequences of the Cyt-b gene—presented, for instance, in
a review of data on vertebrate animals—show a far from
normal distribution (Johns and Avise 1998). This
creates additional problems of analyzing this and other
genes, in which the distance distributions also seem to
deviate from the normality. The analysis of their
distribution is based on the data table including
20,731 species (Kartavtsev 2011) for the five groups
 Sequence divergence at CO-1 and CYT-B and genetics of speciation
of comparison (Groups 1–5) and for genes Cyt-b and
Co-1, respectively. Five groups of comparison are as
follows: Group 1, populations within species; Group 2,
subspecies, semi-species, and sibling species; Group 3,
morphologically distinct species within genera;
Group 4, genera within a family; and Group 5, families
within an order. Indeed, original data showed great
variability and different patterns of distributions for
both the two genes and groups of comparison (Figure 1).
In such cases, means of the estimates generally provide
more satisfactory variation in row distributions as was
indeed obtained (Kartavtsev 2009a, 2011).
A one-way ANOVA (model with random effects for
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
groups of the same size) showed that the mean
distances in the five groups analyzed were significantly
different for the two genes: Cyt-b, F ¼ 2048.60,
degrees of freedom [d.f.] ¼ 4, 3138; p , 0.0001;
Co-1, F ¼ 9876.80, d.f. ¼ 4, 19,089; p , 0.0001.
Pooling data in a two-way MANOVA (see scheme
For personal use only.
Figure 1. Cyt-b (top) and Co-1 (lower) genetic distance frequency distribution plotted against different species compiled in the data table
for five groups of comparison, Groups 1– 5 (from Kartavtsev 2011).
57
below) for the two genes produced a statistically
significant increase in the p-distances in the hierarchy
of the comparison groups: F ¼ 124.15, d.f. ¼ 4, 279;
p , 0.000001 (Figure 2, top). Interaction of factors in
this data set is not statistically significant: F ¼ 1.82,
d.f. ¼ 4, 279; p ¼ 0.1258. However, this pooling is
not quite correct for all of the mtDNA sequences
compared because it includes heterogeneous groups of
different sizes. Consequently, categorized represen-
tation of the mean values with weighting an individual
score on a sample size (n) for each gene is more correct
(Figure 2, bottom). However, both approaches
showed that the distance for the two genes increases
with the rank. Mean unweighted distances (%) for
the five groups were as follows: Cyt-b—(Group 1)
1.38 ^ 0.30, (Group 2) 5.10 ^ 0.91, (Group 3)
10.31 ^ 0.93, (Group 4) 17.86 ^ 1.36, and
(Group 5) 26.36 ^ 3.88; and Co-1—(Group 1)
0.89 ^ 0.16, (Group 2) 3.78 ^ 1.18, (Group 3)
 58 Y. Ph. Kartavtsev
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
For personal use only.
Figure 2. Resulting graphs of two-way MANOVA and mean
p-distance values at five levels of differentiation in animal species
(from Kartavtsev 2011). Top: variation among comparison groups
without weighting the distance scores on the species number (n).
The main effect (i.e. the differences among five comparison
groups) is exemplified: F ¼ 124.15, d.f. ¼ 4, 178; p , 0.000001.
Statistically significant is also the difference in mean distance score
for the two genes; interaction of two factors is non-significant
(see text). Bottom: variation among five comparison groups with
weighting the distance scores on n. Significant are the effects on the
group of comparison and the interaction between gene and taxa.
Interaction effect (i.e. the differences among the distance scores at
two genes in five comparison groups) is exemplified: F ¼ 101.05,
d.f. ¼ 4, 22,227; p , 0.0001. Bars are confidence intervals for
the mean (95%). All p-distances and means for five comparison
groups at each of two genes Cyt-b and Co-1 are presented in
the original paper (Kartavtsev 2011). Comparison groups: Group 1,
within species, among individuals of the same species; Group 2,
within sibling species (plus semi-species, and subspecies),
Group 3, within genus, among morphologically distinct species of
the same genus; Group 4, within family, among genera of the same
family; and Group 5, within order, families of a certain order.
11.06 ^ 0.53, (Group 4) 16.60 ^ 0.69, and
(Group 5) 20.57 ^ 0.40 (Kartavtsev 2011).
Taking into account variation in sample size (n)
for each ith distance measure in comparison
groups (Kartavtsev 2011), we performed a two-way
MANOVA with p-distances weighted by n (factor 1,
comparison Groups 1 –5: see above; factor 2, genes:
Cyt-b and Co-1; also, a model with random effect
of factors was applied; Figure 2, bottom). In this
MANOVA, the effect of factor 1 (i.e. group of
comparison) was significant F ¼ 4964.01,
d.f. ¼ 4, 22,227; p , 0.000001. The effect of factor
2 (mean p-distance differences for two genes) proved
to be non-significant: F ¼ 1.15, d.f. ¼ 1, 22,227;
p ¼ 0.2842. The interaction between factors 1 and 2
was significant too: F ¼ 101.05, d.f. ¼ 4, 22,227;
p , 0.000001. The categorized graph of the distri-
bution of mean weighted p-distance values supported
the earlier conclusion on the increase in distances
with the rank of the groups compared. Fit of the
bivariate distribution (the taxa rank, “taxa” against
the distance score, “p”) showed that there is
accordance with the linear regression model, although
factorial impact is moderate (67–70%): for Cyt-b,
taxa ¼ 1.9002 þ 0.1201 £ p (rp ¼ 0.84; R 2 ¼ 0.7091;
t ¼ 97.93, d.f. ¼ 3141; p , 0.0001); for Co-1,
taxa ¼ 1.4332 þ 0.1471 £ p (rp ¼ 0.82; R 2 ¼ 0.6717;
t ¼ 198.96, d.f. ¼ 19,092; p , 0.0001). Thus, it is
possible to conclude that there is little, if any, impact
of saturation at both the genes up to the order level
in our data set. The lower graph in Figure 2 clearly
shows the meaning of the factor interaction: the
p-distance values or its derivates of these two genes
differ among some of the five comparison groups;
that is, the substitution rates are different for Cyt-b and
Co-1 at least in some of the groups of animal
taxa compared. This conclusion, on an extended
data set, validates the same conclusion made earlier
for these two genes (Kartavtsev and Lee 2006;
Kartavtsev 2009a).
The data presented in Figures 1 and 2 demonstrate
that both the genes show a trend of increasing mean
p-distances with an increasing rank of the groups
compared from populations to orders. Because of the
importance of this conclusion, the data presented
in Figures 1 and 2 were additionally tested using non-
parametric Kruskall – Wallis ANOVA. In this case,
unweighted scores were used to have more conserva-
tive estimation. For gene Cyt-b, H ¼ 57.01, d.f. ¼ 4,
n ¼ 85; p ¼ 0.0001. For gene Co-1, H ¼ 74.05,
d.f. ¼ 4, n ¼ 103; p ¼ 0.0001. Thus, the comparative
analysis of the data for nucleotide sequences of genes
Cyt-b and Co-1, performed for groups with increasing
the rank for each of the genes separately, demonstrates
(with a probability of error p , 0.0001) that, in
animals, genetic divergence increases with the taxon
rank. Heterogeneity of the gene evolution rate, also
significant in our data for the two genes (Figure 2,
bottom), is widely known in the literature (e.g. Li
1997; Machordom and Macpherson 2004), which
was noted previously. An applied outcome has come
from the difference between levels 1 and 3; that is, an
order of magnitude difference that exists in an average
 Sequence divergence at CO-1 and CYT-B and genetics of speciation
p-distance value between the intra-species and
inter-species levels suggests that most species may be
easily discriminated at these mtDNA markers using
only few specimens. In other words, the vast data
reviewed provide theoretical and good empirical
background for per-individual species identification
or DNA barcoding.
The differences in p-distance estimates between the
two genes can have the following interpretations.
Firstly, the substitution rate may in fact be different
in the two genes but hidden somehow. For instance,
the data on taxonomic groups from the most
representative sources (Johns and Avise 1998; Hebert
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
et al. 2002a,b), which can differ in divergence level,
may be differently represented in our database.
Actually, heterogeneity of K2P values at the Cyt-b
gene was found for the vertebrate groups examined:
amphibians and reptiles have the highest variability,
and birds the lowest variability (Johns and Avise
1998). Significant heterogeneity of the nucleotide
diversity was obtained for Co-1 among flatfish genera
(Figure 3; Kartavtsev et al. 2008). Inter-species
heterogeneity of nucleotide diversity estimates at
Cyt-b can be found even within a single fish genus
(Garcia-Machado et al. 2004). Secondly, in the two
most representative works on Co-1, several different
For personal use only.
measures were used (Hebert et al. 2002a,b). In
addition, instead of K2P and other similar measures
(expected distance), the non-corrected p-distance
(observed distance) was employed in many studies.
In general, the shortcoming of analysis of such data
array is high biological heterogeneity of the material
and the presence of some unknown or not identifiable
components of the estimates (some of them were
mentioned above). For instance, p-distances and other
distance measures can differentially represent one
and the same group of comparison. However, non-
weighted p-distances in the most numerous compari-
son group (morphologically distinct species within a
genus) did not statistically significantly differ between
the two groups: the p-distance and other distance
Figure 3. Plot of p-distances for Co-1 gene sequence data within
flatfish genera (from Kartavtsev et al. 2008). On the x-axis are
three flatfish groups: 1, Pseudopleuronectes; 2, Verasper þ
Hippoglossoides; 3, Cynoglossus. On the y-axis are the p-distance
scores for intra-group comparisons.
59
estimates (K2P, GTR, TrN, etc.; Kartavtsev and
Lee 2006; Kartavtsev 2009a). However, unmodified
p-distance must undergo homoplasy faster, that is,
be smaller than the expected values of K2P, GTR,
TrN, and so forth (Nei and Kumar 2000, p. 41).
The differences between these groups are also non-
significant, when n is used as a covariance in ANOVA
of the distance scores. However, the differences
between the groups are significant, if the distance
scores are weighted by n: for Cyt-b, F ¼ 231.38;
d.f. ¼ 1, 943; p , 0.01; for Co-1, F ¼ 207.60;
d.f. ¼ 1, 13,888; p , 0.01 (Kartavtsev 2009a). The
latter differences apparently are caused by unequal
representation of taxa in compared groups and also
their different numeric representations. This effect is
still obscure; for example, there was no correlation
detected between the distance score and n:
r p ¼ 2 0.0122, n ¼ 289; p ¼ 0.6836 (all comparison
groups included); and r p ¼ 0.0336, n ¼ 106;
p ¼ 0.7320 (only genera included). For the Cyt-b
gene, all groups consist almost exclusively of
vertebrates, which may on average have differed in
p-distances compared with the invertebrates that were
mostly tested on Co-1 (Kartavtsev 2009a, 2011).
Genetic variability, divergence, and introduction of an
operational criterion for delimiting a speciation mode
in genetic terms
Below, I briefly compare molecular genetic and
biochemical genetics data, and then draw some
conclusions from this and above evidence relevant to
speciation genetics.
An outcome from biochemical genetics data and their
comparison with nucleotide divergence. Evidence on
the variability of structural protein-coding genes as
assessed by electrophoretic allozyme analysis is also
useful for measuring genetic divergence. Although
they are now not very popular, they give quite a
representative view of variability at nuclear genes.
Mean heterozygosity per individual (locus) has been
recognized as the best measure of variability
(Lewontin 1978; Zhivotovsky 1983; Nei 1987).
Most popular as a measure of taxa divergence during
evolution at this level is the standard Nei’s distance,
Dn, and the inverse measure, normalized similarity,
I (Nei 1972, 1975; Pasekov 1983; Zhivotovsky 1983).
Examination of genetic diversity in wild nature species
requires analysis of both the heterozygosity (diversity)
and distances (differences), assessing different aspects
of variability, which is not always taken into account.
Note, however, that the p-distance and p can be
used as both the measures of variability and measures
of distance.
A number of surveys give similar data on hetero-
zygosity per individual H, with total mean H ¼ 0.076;
 60 Y. Ph. Kartavtsev
in vertebrates, H ¼ 0.054; for invertebrates,
H ¼ 0.100 (Aronshtam et al. 1977; Hedgecock and
Nelson 1981; Nei and Koehn 1983; Nevo et al. 1984;
Hedgecock 1986; Ward et al. 1992; Kartavtsev 2009b;
and so forth). The H value underestimates the actual
genetic diversity by approximately one-third, owing to
technical restrictions of protein electrophoresis, which
is commonly used to estimate the variability at that
level (Lewontin 1978; Nei 1975, 1987; Altukhov
1989, 1999; and others). Coefficients of intra-species
genetic similarity or divergence (distance) were
estimated in many groups of animals. The mean
genetic similarity at this level is I ¼ 0.95 (Lewontin
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
1978; Nei 1987; Altukhov 1983, 1989; Kartavtsev
2009b). According to our database, which comprises
more than 300 populations of 80 animal species,
I ¼ 0.94 ^ 0.01 (Kartavtsev 2005, 2009b; Kartavtsev
and Lee 2006). In the hierarchy of animal taxa,
subspecies have coefficients of similarity I ranging
from 0.6 to 1.0, with a mode of approximately 0.9; the
variation range is I ¼ 0.5 – 1.0 (mode about 0.8) for
semi-species and sibling species; the variation range is
0.5 – 1.0 (mode about 0.7) for species within a genus;
and 0.0 – 1.0 (mode 0.4) for genera within a family
(Avise and Aquadro 1982; Thorpe 1983; Nei 1987;
Kartavtsev 2005, 2009b; Kartavtsev and Lee 2006).
For personal use only.
This means that genetic similarity significantly
decreases with an increasing rank of the taxon and,
conversely, distance increases with an increasing taxon
rank (Figure 4; Kartavtsev 2005, 2009a,b; Kartavtsev
and Lee 2006). In other words, the allozyme data are
supportive of long-time accumulation of genetic
differences through typical allopatric speciation and
in this BSC.
Thus, in general, the current molecular genetic
evidence (see section “Genetic divergence within
species . . . ”) and the results of analysis of protein
marker genes support, firstly, the BSC basic idea that
taxon formation necessarily requires the isolation
of gene pools followed by their gradual genetic
divergence. Within species the differentiation may be
reversed, while it is normally irreversible at upper
1 d.f. ¼ 3, n ¼ 43, p ¼ 0.007 (Kartavtsev 2009a).
0.8
Average (I)
0.6
0.4
0.2
0 basis or DNA barcoding. Because of the absence of a
1 2 3 4
Taxa
Figure 4. Genetic similarity in taxa of different ranks based on
the protein markers: mean for the groups. Group 1, subspecies;
Group 2, semi-species and sibling species; Group 3, species;
Group 4, genera.
divisions or at taxa levels. Secondly, correlated
evidence suggest that the geographic, allopatric, or
divergent (D1; Templeton 1981) speciation mode
(SM) prevails in nature, implying a common rule of a
gradual accumulation of small genetic differences after
the separation of gene pools and differences increase
in a hierarchy of taxa levels. This way may be
considered the main gate of species origin in the
animal world as followed from the current analysis and
classical views (Mayr 1963). Yet, there are facts
warning against simplified conclusions on the modes
of speciation. For instance, it has long been known
that the genetic “weight” of the species, say, in the
Dn scale, may be different for different animal taxa.
For example, Dn is on average 1.1 in amphibian
genera, which is an order of magnitude higher than
the corresponding value in birds (Dn ¼ 0.1; Avise and
Aquadro 1982). Other examples of this trend can be
found (Avise 1994). The range of nucleotide diversity
also shows that some animal taxa display a high-
divergence level among the species, while others
are characterized by a low value of this measure.
As already noted above, avian taxa are substantially
less differentiated at Cyt-b than amphibians and
reptiles (Johns and Avise 1998; see also Kartavtsev
2009a, 2011 and data in the section “Divergence
of DNA nucleotide sequences . . . ”). For the three
main geographic phyletic groups of Oryzias latipes,
the nucleotide diversity of Cyt-b was found to be
comparable with the within-genus divergence:
p ¼ 11.3 – 11.8% (Takehana et al. 2003). For the
other gene, Co-1, the species within the genus Cnidaria
have p ¼ 1%, while in crustaceans p ¼ 15.4% (Hebert
et al. 2002b). The difference for flatfish genera at Co-1
sequences was demonstrated too (see Figure 3), and
in a more wide context the heterogeneity among
animal taxa at the level of genus for both Cyt-b and
Co-1 is available. One-way parametric ANOVA and
Kruskall –Wallis ANOVA proved this conclusion: for
the Cyt-b—F ¼ 265.08, d.f. ¼ 3, 10,654, p , 0.01;
Kruskall – Wallis H ¼ 10.87, d.f. ¼ 3, n ¼ 32,
p ¼ 0.01; and for Co-1—F ¼ 196.91, d.f. ¼ 3,
13,886, p , 0.01; Kruskall – Wallis H ¼ 12.11,
It is possible to conclude that in the animal world
due to the D1 SM prevalence, morphologically
different or “good” species usually are quite distinct
from intra-species categories in a genetic distance
scale, and on this property the successful genotypic
identification of a species is based on the individual
universality of a distance scale, there is a need for
a complex approach to the species definition and SM
delimiting (Brower 1999; Templeton 2001; Kartavt-
sev et al. 2002; Sites and Marshall 2004; Kartavtsev
2009a). One such genetically based approach is
exemplified below.
 Sequence divergence at CO-1 and CYT-B and genetics of speciation
A scheme and an algorithmic approach to dis-
tinguish SMs (models) on the basis of key population
genetic parameters and their estimates available in the
literature have been developed (Kartavtsev 2000,
2009a,b, 2011; Kartavtsev et al. 2002). As a basis for
the evolutionary genetic concept of speciation, verbal
descriptions of seven SMs were used (Templeton
1981). Consequently, a classification scheme for seven
modes of speciation was developed (Kartavtsev et al.
2002; Kartavtsev 2005, 2009a,b, 2011). Here, in this
paper, I present a revised scheme updated for
sequence data. In short, an illustration is made of
the main elements of this scheme from all seven
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
types represented: D1 – D3 (divergent speciation) and
T1 –T4 (transformative or transilience speciation;
Figure 5). This approach leads to a relatively simple
logical and experimental scheme, which allows us to
organize an investigation of speciation in various
groups of organisms, based on a focused approach
with defined genetic terms and obtain analytic
expressions (equations) for each of the SMs
(Figure 5). Using the proposed scheme (Figure 5),
one can determine two kinds of conditions required
for the speciation: the necessity conditions and the
sufficiency conditions, which denote requirement for
the mechanisms that control species origin in genetic
For personal use only.
terms and that are sufficient for the formation and
recognition of a species. Importantly, in addition to
the general definition of the sufficient conditions, six
experimentally measured descriptors are introduced
(their number including mtDNA and nDNA molecu-
lar markers can be increased up to appropriate
number) to clarify how, and in which form, these
conditions are manifested in a particular case of
speciation or in a potential model. For instance, the
divergent type of speciation D1 explains classic
geographic (or allopatric) speciation (see Figure 5).
As agreed, according to the BSC, the D1 model
implies that large populations are isolated (disruption
of the gene flow) and evolve separately, accumulating
mutations, while reproductive isolating barriers
(RIBs) are caused by pleiotropic effects. The longer
the time elapsed from the isolation event, the greater
the distances between the corresponding taxa.
Accordingly, in this notation, distance descriptors
are introduced: 1. DT . DS and 4. pT . pS (where
subscripts T and S indicate genetic distances at
structural genes/sites in the putative parental taxon
and in conspecific populations or at the higher and
lower levels of taxonomic hierarchy in statu nascendi
situation). Likewise, since upon implementation of
the D1 mode, no significant genetic diversity
differences appear at either the structural gene or the
regulatory part of the genome (because the initial and
derived taxa have large effective size, Ne, and thus a
small rate in diversity decreases), the following
parameters are introduced: 2. HD ¼ HP, 3. ED ¼ EP,
and 5. pD ¼ pP (differences in heterozygosity/
61
diversity and gene expression between the daughter
and the parental taxon are absent). Finally, upon some
types of speciation, not only variability and genetic
distances but also some quantitative trait loci (QTL,
polygenes) are of major importance, which could not
be distinguished at the molecular level, but lead to the
RIB formation. Hence, the next descriptor is
introduced: 6. TM (TMþ vs. TM – , an experimental
test to detect scores for modifications or RIB
important differences at quantitative traits in nature
taxa). This test also allows one to distinguish between
an epigenetic variation and a taxonomically significant
difference.
Do all data presented above imply that speciation
always corresponds to the D1 type? Apparently not.
One good example supporting this answer is known
for two trout (Salmo trutta; Ryman et al. 1979). There
are other examples of bursts of fish evolution,
documented by molecular markers (Rutaisire et al.
2004; Duftner et al. 2005). These, as well as other
data, for instance from our database of coefficients of
similarity, indicate that sometimes small differences in
structural genes may result in the appearance of RIBs
(and thus reproductively isolated biological entities).
In the case of the trout mentioned above, the
genetic difference between the two forms Dn is 0.02
(Ryman et al. 1979), which corresponds to the level of
intra-species genetic differentiation. There are many
other examples for salmonid fishes (Kartavtsev
2009a,b, 2011), supporting the view that in these
fishes, small changes can generate biological species
during a short period of time. These evidences also
suggest an alternative SM, such as the transforma-
tional (T1) or other T-modes (Figures 5 and 6),
although in general the D1 SM prevails in this
group too.
The above original developments are close to
similar directions in evolutionary genetics. For
instance, the method of distance scaling along phyletic
lines was suggested by Avise and Walker (1999). It was
designed for the normalization of taxa weights, and as
an outcome the unification of Systematics is expected.
The estimation of gene trees’ cohesion was suggested
by Templeton (2001) to decide on species boundaries.
The approach includes the notion of genetic exchan-
geability and/or ecological interchangeability among
lineages belonging to the same species (Templeton
2001). Both approaches are operational for species
delimitation but it seems that these techniques will
hardly solve the “rigidity” of species and species
boundaries without the formalization of a species
notion. Some authors reached similar conclusions on
the basis of independent analysis of different
characters and approaches for species delimitation
(Ferguson 2002; Wiens and Penkrot 2002; Sites
and Marshall 2004). In particular, the latter authors
emphasize the idea of diffused peculiarities of
the species concept and species boundaries and,
 62 Y. Ph. Kartavtsev

DIVERGENCE SM

D1. ADAPTIVE D2. CLINAL D3. HABITAT

Necessary Conditions for Speciation

DESCRIPTORS:
D or p, Genetic distance at structural
D1. a) Erection of extrinsic D3. a) Selection over multiple
D2. a) Selection on a cline with genes:
Reproductive Isolating habitats with no isolation by
isolation by distance; DT, pT, in suggested parent taxa,
Barriers (RIB) followed by distance; b) Origin of RIB by
b) Pleiotropic origin of RIB DS, pS, among conspecific demes,
gene flow break; b) Pleiotropic disruptive selection at genes,
DD, pD, among subspecies or sibling
origin of RIB in long time determining behavior
species;
HD, πD, Mean heterozygosity/diversity
in suggested daughter population;
Sufficient Conditions for Speciation
HP, πP, Mean heterozygosity/diversity
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11

in suggested parent population;

EP, Divergence in regulatory genes
Lack of efficient Lack of efficient hybridization
Lack of efficient hybridization among suggested parent taxa;
hybridization outside the zone inside and outside the
in the zone of contact ED, Divergence in regulatory
of contact zone of contact
genes among suggested
daughter taxa;
1. DT > DS 5. πD = πP 1. DT > DS 5. πD = πP 1. DT = DS 5. πD <= πP +
TM , Test for modification
2. ED = EP 6. TM– 2. ED ≠ EP 6. TM– 2. ED ≠ EP 6. TM– (positive);
3. HD = HP 3. HD = HP 3. HD <= HP TM–, Test for modification
4. pT > pS Φ1 (S) 4. pT > pS Φ2 (S) 4. pT = pS Φ3 (S) (negative).
RIB, Reproductive Isolation
Experimentally measurable features Barriers.
and possible descriptors for the
model (theory), Φ (S)

TRANSILIENCE SM
For personal use only.

T1. GENETIC T2. CHROMOSOMAL T3. HYBRIDOGENIC 1 T4. HYBRIDOGENIC 2

Necessary Conditions for Speciation

T1. a) Founder event causing T3. a) Hybridization of in- T4. a) Hybridization of

T2. a) Inbreeding and drift
a rapid shift in previously compartible parental species incompartible parental species
causing fixation of strongly
stable genetic system; b) RIB followed by selection for followed by inbreeding and
underdominant chromosomal
origin as by product of one or maintenance for hybrid state; selection for stabilized
mutatins; b) RIB origin as a
a small number gene b) RIB origin as a cause of recombinant; b) RIB origin as
cause of hybrid disgenesis
substitutions hybrid disgenesis a cause of hybrid disgenesis

Sufficient Conditions for Speciation

Lack of efficient Lack of efficient Lack of efficient Lack of efficient

hybridization inside and hybridization inside and hybridization in the zoneof hybridization inside and
outside the zone of contact outside the zone of contact contact outside the zone of contact

1. DT = DD 5. πD <= πP 1. DT = DD 5. πD <= πP 1. DT > DD 5. πD > πP 1. D T > DS 5. πD < πP

2. ED = EP 6. TM– 2. ED = EP 6. TM– 2. ED ≠ EP 6. TM– 2. E D ≠ EP 6. TM–
3. HD <= HP 3. HD <= HP 3. HD > HP 3. HD < HP
4. pT = pD Φ4 (S) 4. pT = pD Φ5 (S) 4.pT>pD Φ6 (S) 4. p T ≠ pS Φ7 (S)

Experimentally measurable features and possible

descriptors for the model (theory), Φ (S)

Figure 5. Schematic representation of the divergent SM, based on the population genetic principles (from Kartavtsev et al. 2002; Kartavtsev
2009a, with modifications). D1–D3, divergent SMs; T1–T4, transformative (transilience) SMs.

consequently, the necessity and applicability of several
sets of operational criteria in a multiple approach for
species identification (Sites and Marshall 2004).
This is also emphasized in the approach suggested
here (Figures 5 and 6 and relevant text). The scheme
presented in this paper was designed originally to
define a SM. However, it also contains the logical
criteria of whether species have or have not yet
originated. Thus, this approach is quite suitable for
species delimiting as having both a theoretic and
empirically operational approach. It has weakness,
which all current methods, both the non-tree-based
and tree-based, have (Sites and Marshall 2004); that
is, in some cases, the approach will require researches
 Sequence divergence at CO-1 and CYT-B and genetics of speciation 63

Φ1 (S) ∈ {(DT > DS) ⊂ (ED = EP) ⊂ (HD = HP) ⊂ (pT > pS) ⊂ ( πD = πP) ⊂ TM–} (D1)
Φ2 (S) ∈ {(DT > DS) ⊂ (ED ≠ EP) ⊂ (HD = HP) ⊂ (pT > pS) ⊂ ( πD = πP) ⊂ TM–} (D2)
Φ3 (S) ∈ {(DT = DS) ⊂ (ED ≠ EP) ⊂ (HD <= HP) ⊂ (pT = pS) ⊂ ( πD <= πP) ⊂ TM+} (D3)
Φ4 (S) ∈ {(DT = DD) ⊂ (ED ≠ EP) ⊂ (HD <= HP) ⊂ (pT > pD) ⊂ ( πD <= πP) ⊂ TM–} (T1)
Φ5 (S) ∈ {(DT = DD) ⊂ (ED = EP) ⊂ (HD <= HP) ⊂ (pT > pD) ⊂ ( πD <= πP) ⊂ TM–} (T2)
Φ6 (S) ∈ {(DT > DD) ⊂ (ED ≠ EP) ⊂ (HD > HP) ⊂ (pT > pD) ⊂ ( πD > πP) ⊂ TM–} (T3)
Φ7 (S) ∈ {(DT > DS) ⊂ (ED ≠ EP) ⊂ (HD > HP) ⊂ (pT > pS) ⊂ ( πD < πP) ⊂ TM–} (T4)

Figure 6. Analytic representation of seven SMs (from Kartavtsev 2009a, with modifications). D1–D3, divergent SMs; T1–T4,
transformative (transilience) SMs. Descriptors: D, genetic distances for structural gene; DT, pT, in putative parental taxon; DS, pS, among
conspecific demes; DD, pD, among subspecies or sibling species; HD, pD, mean heterozygosity/diversity in putative daughter population;
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11

HP, pP, mean heterozygosity/diversity in putative parental population; EP, divergence at regulatory genes in putative parental taxon;
ED, divergence at regulatory genes in putative daughter taxon; TMþ, test for modification (positive); TM – , test for modification (negative).

to make qualitative judgments because of the variety of
ways for species to originate. Potentially, the approach
developing is close to the Population Aggregation
Analysis (PAA) in Davis and Nixon’s (1992) version
because it is based on the population-based
parameters such as DT, HD, and so on (see Figures 5
and 6). However, this PAA1 approach could easily be
converted to the mode PAA2 as defined by Brower
(1999). Developed notations (see Figures 5 and 6;
see also Sites and Marshall 2004) may even have
For personal use only.
less fit. The other limitation is that generalizations
(deductions) are only possible in a framework of the
genetic terms defined. But individuals comprising
species are phenotypes. Thus, genotype/phenotype
correspondence should be defined in an appropriate
form and genotype-and-environment interaction, or
ecological interchangeability in Templeton’s (2001)
sense should also be introduced somehow. Partly,
it is supposed, when QTL or other complimentary
descriptors will be introduced. An advance is that this

properties of the tree-based method (see below). As
in PAA2, it is suggested to use not only genotypic
scores (character states) but other suitable values of
descriptors (qualitative, QT and quantitative traits,
QTL, etc.); they could be represented as per-
individual sets of the records or as vector scores for
implementing a multi-dimensional analysis (canoni-
cal, principal component analysis, etc.) with the aims
of testing a null hypothesis (H1) of the absence of
vectors’ gatherings and, if rejected, the alternative
hypothesis (H2) will be tested for discrimination
among them and taking solution in the frame of logic
suggested (Figure 5); and obtaining a solution
whether vectors’ genetic ( ¼ phylogenetic) unity is
available, both as a distance value and a coalescent
signature obtained from a tree; again solving H1 and
H2. To obtain phyletic signal it will be necessary to
develop new descriptors in Figure 5 scheme and
introduce them in the set of equations D1 – T4 (and
others when developed) in Figure 6. These special
descriptors, such as the branch length or the
parsimony outcomes to the current operational
taxonomic unit of a consensus tree built for several
gene sequences, could be operational criteria among
others. The approach is basically empirical but
different from such others reviewed (Sites and
Marshall 2004), having a general genetics and
population genetics theory basis and formalization as
equations of the set theory. Such an approach has its
own limitations and advances. One limitation is that
it is restricted to sexually reproducing species, for
which basic population genetic principles are more or
approach is wider than many other suggested for
species delimiting (see Sites and Marshall 2004) in its
ability to define different SMs (or take into account
the differences in species types). Also, by weighting
the members of equations in a specific way, it is
possible to further develop the approach as a frame-
work for future theory, the genetic theory of
speciation.
Acknowledgements
This work was supported in part by Far Eastern
Branch of Russian Academy of Science grant number
09-I-P23-07 and the Russian Foundation for Basic
Research grants numbers 07-04-00186, 08-04-91200.
Declaration of interest: The authors report no
conflicts of interest. The authors alone are responsible
for the content and writing of the paper.
References
Altukhov YuP. 1983. Geneticheskie protsessy v populyatsiyakh
(Genetic Processes in Populations). Moscow: Nauka Publ.
Altukhov YuP. 1989. Genetic processes in populations. 2nd ed.,
Moscow: Nauka.
Altukhov YuP. 1999. Genetic processes in populations. 3rd ed.,
Moscow: Nauka.
Aronshtam AA, Borkin LYa, Pudovkin AI. 1977. Isozymes in
population and evolutionary genetics Genetika izofermentov
[Genetics of Isozymes]. Moscow: Nauka. p 199–249.
Avise JC. 1994. Molecular markers, natural history and evolution.
New York: Chapman & Hall.
 64 Y. Ph. Kartavtsev
Avise JC. 2000. Phylogeography: The history and formation of
species. Cambridge: Harvard University Press.
Avise JC, Aquadro CF. 1982. A comparative summary of genetic
distances in the vertebrates: Pattern and correlations. Evol Biol
15:151–185.
Avise JC, Walker D. 1999. Species realities and numbers in sexual
vertebrates: Perspectives from an asexually transmitted genome.
Evolution 9(3):992–995.
Brower AVZ. 1999. Delimitation of phylogenetic species with
DNA sequences: A critique of Davis and Nixon’s population
aggregation analysis. Syst Biol 48:199–213.
Creer S, Malhotra A, Thorpe RS. 2003. Assessing the phylogenetic
utility of four mitochondrial genes and a nuclear intron in the
Asian pit viper genus, Trimeresurus: Separate, simultaneous, and
conditional data combination analyses. Mol Biol Evol 20(8):
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
1240–1251.
Davis JI, Nixon KC. 1992. Populations, genetic-variation, and the
delimitation of phylogenetic species. Syst Biol 41:421–435.
Dobzhansky Th. 1955. Evolution, genetics and man. New York/
London: John Wiley & Sons/Chapman & Hall. p 398.
Duftner N, Koblmuller S, Sturmbauer C. 2005. Evolutionary
relationships of the Limnochromini, a tribe of Benthic deepwater
cichlid fish endemic to Lake Tanganyika. East Africa. J Mol Evol
60(3):277–289.
Felsenstein J. 2004. Inferring phylogenies. Sunderland, MA:
Sinauer Associates.
Ferguson JWH. 2002. On the use of genetic divergence for
identifying species. Biol J Linn Soc 75(4):509–516.
Garcia-Machado E, Chevalier-Monteagudo PP, Solignac M. 2004.
Lack of mtDNA differentiation among Hamlets (Hypoplectrus,
Serranidae). Mar Biol 144:147–152.
For personal use only.
Graur D, Li WH. 1999. Fundamentals of molecular evolution.
Sunderland, MA: Sinauer Associates.
Hall B. 2001. Phylogenetic trees made easy: A how-to manual for
molecular biologists. Sunderland, MA: Sinauer Associates.
Hebert PDN, Givinska A, Ball SL, deWaard JA. 2002a. Biological
identification through DNA barcodes. Proc R Soc Lond B
270(1512):02PB0653.1–02PB0653.9.
Hebert PDN, Ratnasingham S, deWaard JA. 2002b. Barcoding
animal life: Cytochrome c oxidase subunit 1 divergences among
closely related species. Proc R Soc Lond B 270(1512):
03BL0066.S1– 03BL0066.S4.
Hedgecock D. 1986. Population genetic bases for improving
cultured crustaceans, FIFAC/FAO Symposium on Selection,
Hybridization and Genetic Engineering in Aquaculture of
Fish and Shellfish for Consumption and Restocking. Bordeaux.
Hedgecock D, Nelson C. 1981. Genetic variation of enzymes and
adaptive strategies in crustaceans Genetika i razmnozhenie
morskikh zhivotnykh [Genetics and Reproduction of Marine
Animals]. Vladivostok: Dal’nevost. Nauchn. Tsentr Akad. Nauk
SSSR. p 105–129.
Johns GC, Avise JC. 1998. A comparative summary of genetic
distances in the vertebrates from the mitochondrial cytochrome
b gene. Mol Biol Evol 15(11):1481–1490.
Kartavtsev Y. 2000. Genetic aspects of speciation, species
differentiation and biodiversity. In: Proceedings of the Inter-
national Meeting of Biodiversity in Asia. September. Tokyo p 27.
Kartavtsev YuP. 2005. Molekulyarnaya evolyutsiya i populyatsion-
naya genetika (Molecular Evolution and Population Genetics).
Vladivostok: Dal’nevost. Gos. Univ. (Far Eastern State Univ.
Publ.).
Kartavtsev YPh. 2009a. Analysis of sequence diversity at
mitochondrial genes on different taxonomic levels. Applicability
of DNA based distance data in genetics of speciation and
phylogenetics. In: Mahoney CL, Springer DA, editors. Genetic
diversity. New York: Nova Science Publishers. p 1– 50.
Kartavtsev YPh. 2009b. Molekulyarnaya evolyutsiya i populyat-
sionnaya genetika [Molecular Evolution and Population
Genetics]. 2nd ed., Vladivostok: Dal’nevost. Gos. Univ. [Far
Eastern State University Publ.]. (Content, chapter abstracts, and
headings in English.) p 280.
Kartavtsev YPh. 2011. Sequence divergence at mitochondrial genes
in animals. Applicability of DNA data in genetics of speciation,
phylogenetics and molecular ecology. Mar Genomics 49:71–81.
Kartavtsev YP, Lee J-S. 2006. Analysis of nucleotide diversity at
genes Cyt-b and Co-1 on population, species, and genera levels
Russian. J. Genetics 42(4):341–362. (In Russian, Translated
in English).
Kartavtsev YP, Hanzawa N. 2007. Inferences in Leuciscinae
(Pisces Cyprinidae) phylogeny and taxonomy based on
cytochrome b sequence distances and on enzyme loci diversity.
Korean J Genet 29(4):427–435.
Kartavtsev YuF, Sviridov VV, Hanzawa N, Sasaki T. 2002. Genetic
divergence of Far-Eastern dace species. Russ J Genet 38(11):
1285–1297.
Kartavtsev YP, Lee Y-M, Jung S-O, Byeon H-K, Son Y, Lee J-S.
2007a. The complete mitochondrial genome of the bullhead
torrent catfish, Liobagrus obesus (Siluriformes, Amblycipididae):
Genome description and phylogenetic considerations inferred
from the Cyt b gene. Gene 396:13– 27.
Kartavtsev YP, Park T-J, Vinnikov KA, Ivankov VN, Sharina SN,
Lee J-S. 2007b. Cytochrome b (Cyt-b) gene sequences analysis in
six flatfish species (Pisces, Pleuronectidae), with phylogenetic
and taxonomic insights. Mar Biol 152(4):757 –773.
Kartavtsev YPh, Sharina SN, Goto T, Chichvarkhin AY, Balanov
AA, Vinnikov KA, Ivankov VN, Hanzawa N. 2008. Cytochrome
oxidase 1 (Co-1) gene sequence analysis in six flatfish species
(Teleostei Pleuronectidae) of Russia Far East with inferences in
phylogeny and taxonomy. Mitochondrial DNA 19(6):479 –489.
Kartavtsev YPh, Sharina SN, Goto T, Balanov AA, Hanzawa N.
2009a. Sequence diversity at cytochrome oxidase 1 (Co-1) gene
among sculpins (Scorpaeniformes, Cottidae) and some other
scorpionfish of Russia Far East with phylogenetic and taxonomic
insights. Genes Genomics 31(2):191–205.
Kartavtsev YPh, Sharina SN, Goto T, Rutenko OA, Pitruk DL,
Semenchenko AA, Hanzawa N. 2009b. Sequence diversity
at cytochrome oxidase 1 (Co-1) gene among pricklebacks
(Actinopterigii, Stichaeidae) and some other percoids fishes of
Russia Far East with inferences in phylogeny and taxonomy.
Aquat Biol 8(1):95–103.
Lewontin RC. 1978. Geneticheskie osnovi evolutsii. [Transl.
(1974). The Genetic Basis of Evolutionary Change. New York:
Columbia Univ.] Moscow: Mir.
Li WH. 1997. Molecular evolution. Sunderland, MA: Sinauer
Associates.
Li WH, Zarkhih A. 1995. Statistical tests of DNA phylogenies.
Syst Biol 44:49–63.
Machordom A, Macpherson E. 2004. Rapid radiation and cryptic
speciation in squat lobsters of the genus Munida (Crustacea,
Decapoda) and related genera in the South West Pacific:
Molecular and morphological evidence. Mol Phylogenet Evol
33(2):259–279.
Mayr E. 1963. Animal species and evolution. Cambridge: Harvard
University Press.
Nei M. 1972. Genetic distances between populations. Am Nat
106(949):283 –292.
Nei M. 1975. Molecular population genetics and evolution.
Amsterdam: North Holland.
Nei M. 1987. Molecular evolutionary genetics. New York:
Columbia University Press.
Nei M, Koehn R. 1983. In: Nei M, Koehn RK, editors. Evolution
of genes and proteins. Sunderland, MA: Sinauer Associates.
Nei M, Kumar S. 2000. Molecular evolution and phylogenetics.
New York: Oxford University Press.
Nevo E, Beiles A, Ben-Shlomo R. 1984. The evolutionary
significance of genetic diversity: Ecological, demographic and
life history correlates. In: Mani GS, editor. Evolutionary
 Sequence divergence at CO-1 and CYT-B and genetics of speciation
dynamics of genetic diversity. Lect Notes Biomath. Vol. 53.,
p 4 –213.
Pasekov VP. 1983. Genetic distances Geneticheskie rasstoyaniya.
Itogi Nauki Tekhn Obshch Genet 8:4–75.
Rutaisire J, Boot AJ, Masemba C, Nyakaana S, Muwanika VB.
2004. Evolution of Labeo victorianus predates the Pleistocene
desiccation of Lake Victoria: Evidence from mitochondrial DNA
sequence variation. S Afr J Sci 100(11–12):607–608.
Ryman NF, Allendorf FW, Stahl G. 1979. Reproductive isolation
with little genetic divergence in sympatric populations of brown
trout (Salmo trutta). Genetics 92:247–262.
Sanderson MJ, Shaffer HB. 2002. Troubleshooting molecular
phylogenetic analyses. Annu Rev Ecol Syst 33:49–72.
Sasaki T, Kartavtsev YP, Uematsu T, Sviridov VV, Hanzawa N.
2007. Phylogenetic independence of Far Eastern Leuciscinae
Mitochondrial DNA Downloaded from informahealthcare.com by IBI Circulation - Ashley Publications Ltd on 09/05/11
(Pisces: Cyprinidae) inferred from mitochondrial DNA analysis.
Gene Genet Syst 82:329–340.
Sites JW, Marshall JC. 2004. Operational criteria for delimiting
species. Annu Rev Ecol Evol Syst 35:199– 227.
Statsoft. 2001. Statistica for windows. Statistica 6: Users Guide.
Tulsa: StatSoft.
Swofford DL, Olsen GJ, Waddel PJ, Hillis DM. 1996. Phylogenetic
inference. In: Hillis DM, Moritz C, Mable B, editors. Molecular
systematics. Sunderland, MA: Sinauer Associates. p 407–514.
For personal use only.
65
Takehana Y, Nagai N, Matsuda M, Tsuchiya K, Sakaizumi M.
2003. Geographic variation and diversity of the Cytochrome b
gene in Japanese wild populations of medaka, Oryzias latipes.
Zool Sci 20(10):1279–1291.
Templeton AR. 1981. Mechanisms of speciation—population
genetic approach. Annu Rev Ecol Syst 12:23–48.
Templeton AR. 2001. Using phylogeographic analyses of gene trees
to test species status and processes. Mol Ecol 10(3):779– 791.
Thorpe JP. 1983. Enzyme variation, genetic distance and
evolutionary divergence in relation to levels of taxonomic
variation. In: Oxford JS, Rollinson D, editors. Protein
polymorphism: Adaptive and taxonomic significance. London:
Academic. p 131–152.
Ward RD, Skibinski DOF, Woodwark M, et al. 1992. Protein
heterozygosity, protein structure, and taxonomic differentiation.
Evolutionary Biology 26:73–159.
Ward RD, Zemlak TS, Innes BH, Last PA, Hebert PDN. 2005.
DNA barcoding Australia fish species. Philos Trans R Soc Lond
B 360:1847–1857.
Wiens JJ, Penkrot TA. 2002. Delimiting species using DNA and
morphological variation and discordant species limits in spiny
lizards (Sceloporus). Syst Biol 51(1):69–91.
Zhivotovsky LA. 1983. Statistical methods of analyzing gene
frequencies in natural populations. Itogi Nauki Tekhn Obshch
Genet 8:76–104.