Annals of Botany 111: 591 –609, 2013
doi:10.1093/aob/mct019, available online at www.aob.oxfordjournals.org
Patterns of plastid and nuclear variation among apomictic polyploids
of Hieracium: evolutionary processes and taxonomic implications
Torbjörn Tyler1,* and Jane Jönsson2
1
Lund University, Department of Biology, Botanical Museum, Box 117, SE-221 00 Lund, Sweden and 2Lund University,
Department of Biology, Molecular Ecology and Evolution Laboratory, Ecology Building, Sölvegatan 37,
SE-223 62 Lund, Sweden
* For correspondence. E-mail: [email protected]
Received: 26 September 2012 Revision requested: 21 November 2012 Accepted: 21 December 2012 Published electronically: 7 February 2013
† Background and Aims Apomictic species (with asexual seed production) make up for 20– 50 % of all taxonomically recognized species in northern Europe, but the phylogenetic relationships of apomictic species and the
mode of evolution and speciation remain largely unknown and their taxonomy is consequently disputed.
† Methods In the present study, plastid psbD-trnT sequences (349 accessions) and 12 nuclear microsatellite loci
(478 accessions) were used to create an overview of the molecular variation in (mainly) northern European
members of the most species-rich of all plant genera, Hieracium s.s. The results are discussed and interpreted
in the context of morphological and cytological data on the same species.
† Key Results and Conclusions The complete psbD-trnT alignment was 1243 bp and 50 polymorphisms defined
40 haplotypes. All haplotypes found in the sections of the genus distributed in the northern European lowlands
fell into one of two main groups, group H and group V, mutually separated by seven or eight polymorphisms. All
accessions belonging to H. sects. Foliosa, Hieracioides (viz. H. umbellatum) and Tridentata and all but one
accession of triploid species of H. sects. Oradea and Vulgata showed haplotypes of group V. Haplotypes of
group H were found in all accessions of H. sects. Bifida and Hieracium and in all tetraploid representatives of
H. sects. Oreadea and Vulgata. Additional haplotypes were found in accessions of the genus Pilosella and in
southern European and Alpine sections of Hieracium. In contrast, the distribution of individual haplotypes in
the two major groups appeared uncorrelated with morphology and current taxonomy, but polymorphisms
within species were only rarely encountered. In total, 160 microsatellite alleles were identified. Levels of variation were generally high with only nine pairs of accessions being identical at all loci (in all cases representing
accessions of the same species). In the neighbor-joining analysis based on the microsatellite data, accessions of
the same species generally clustered together and some smaller groups of species congruent with morphology
and/or current taxonomy were recovered but, except for H. sect. Oreadea, most larger groups were not correlated
with morphology. Although the plastid DNA sequences show too little variation and the nuclear microsatellites
are too variable to resolve relationships successfully among species or to fully understand processes of evolution,
it is concluded that both species and sections as defined by morphology are largely congruent with the molecular
data, that gene flow between the sections is rare or non-existent and that the tetraploid species may constitute the
key to understanding evolution and speciation in this genus.
Key words: Apomixis, polyploidy, phylogeny, speciation, taxonomy, chloroplast haplotypes, microsatellites,
Hieracium, Hieracium sect. Hieracium, Hieracium sect. Vulgata, Hieracium sect. Bifida, Hieracium sect.
Oreadea.
IN T RO DU C T IO N
Apomixis, i.e. reproduction by means of asexually derived seeds
(sometimes referred to as agamospermy), may be considered
both common and rare in plants. If a broad species concept is
applied, i.e. not accepting apomictic clones as species, ,1 %
of flowering plant species are known to be capable of apomixis
(Asker and Jerling, 1992, Whitton et al., 2008). However, in particular in the boreal and arctic zones, several of the most frequent, sometimes dominant, and morphologically most
polymorphic genera partly or exclusively reproduce by apomixis
(e.g. Rubus, Crataegus, Potentilla, Alchemilla, Ranunculus,
Taraxacum, Hieracium s.l., Crepis, Limonium and many
grasses). If all morphologically distinct clones in such genera
are accepted as species, as generally agreed by northern
European taxonomists, apomictic species may make up 50 %
of all species in a country like Sweden (cf. Karlsson, 1998)
and they certainly comprise a significant proportion of all discrete plant individuals in boreal and arctic habitats. The apomictic species of Hieracium alone make up for about 15 % of all
vascular plant species native to Britain and Ireland (Sell and
Murrell, 2006) and similar proportions may be obtained for
other northern European countries. In addition, apomixis is
doubtlessly widespread in plants; Carman (1997) listed .330
genera from a wide range of families including apomictic
species.
Loss of sexual reproduction has sometimes been considered
by theoreticians as giving rise to evolutionary dead ends (e.g.
Stebbins, 1950), a notion that for decades significantly reduced
the interest in apomixis and asexual reproduction among
# The Author 2013. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved.
For Permissions, please email: [email protected]
592
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
evolutionary biologists. However, as outlined above, apomictic
plants appear in no way to be subordinate or marginal in
nature. In addition, as reviewed by Hörandl and Paun (2007),
recent molecular studies have shown that theoreticians have
greatly underestimated levels of genetic variation in apomictic
organisms. However, despite the phenomenon of apomixis
being documented as early as 1841 (Smith, 1841) and thoroughly investigated in 1898 (Juel, 1898; Nogler, 2007), our
understanding of how evolution proceeds and how new
clones/taxa are formed in predominantly or exclusively apomictic groups of plants remains rather limited. Surprisingly,
apomictic taxa are rarely considered when processes of speciation are discussed.
In principle, among apomictic plants the evolution of new,
morphologically distinct and taxonomically recognizable units
may take place in four ways, each giving rise to its own
pattern and structure of variation.
(1) Extant apomicts may have evolved from hybrids between
sexual taxa. Almost all apomicts are polyploids (Mogie
et al., 2007) and allopolyploidization is generally recognized as a major mechanism for polypoloid evolution.
However, many apomictic species complexes appear to
have no extant sexual close relatives or have much wider
geographic distributions than their closest sexual relatives
(Hörandl, 2006). In many cases apomictic taxa are narrow
endemics restricted to relatively recently deglaciated areas
(Johansson, 1923; Samuelsson, 1954; Tyler, 2000). Thus,
if they evolved from sexual hybrids, many apomictic lines
have to be of a considerable age and/or must somehow
have ended up in a geographic range far distant from
where they first evolved.
(2) New apomictic taxa may evolve by accumulation of mutations, each with a small effect on the phenotype. With
time, accumulation of mutations is inevitable, at least in selectively neutral genes, but if this is the dominant process
one would expect all kinds of intermediates to occur,
making the taxonomic recognition of morphologically discrete units impossible (Tyler, 2006). At the molecular
level, a strictly hierarchical pattern of variation ideal for cladistic analysis would be expected, provided that the rate of
evolution/mutation of the molecules concerned is optimal
for the age of the taxa under study.
(3) New, morphologically readily distinguishable units may
evolve by single mutations at genes that have strong pleiotrophic effects on the phenotype (Riska, 1989; Tyler,
2006). Morphologically distinct units may then be closely
related, almost identical at the molecular level, and may
have arisen repeatedly by homoplasious mutations. In addition, if the same mutations are responsible for concerted
changes in several phenotypic characters, strong correlations among the latter would be expected (Tyler, 2006).
(4) Residual sexuality may allow for occasional sexual reproduction and hybridization also among predominantly apomictic clones, thereby giving rise to novel genetic
combinations and phenotypes. As shown by Bengtsson
(2003), even rare sexual events may have great impact
on the outcome of evolutionary processes. This is obviously the case in some apomictic complexes (e.g. Krahulcova
et al., 2000), but in others sexuality has never been
documented and stable triploidy or aneuploidy and/or retardation of reproductive organs, in combination with
peculiar embryological processes, indicate that sexual reproduction is highly unlikely (Hörandl and Paun, 2007).
The genus Hieracium (here treated in a narrow sense excluding,
for example, Pilosella) comprises perennial herbs widespread in
arctic and temperate regions of Europe, Asia and North America,
and some members behave as invasive weeds after introduction
to, for example, New Zealand. Species of Hieracium occur in
forests, semi-natural grasslands and rocky and alpine habitats.
Some species efficiently colonize roadsides and embankments,
but in the native range of the genus its main diversity is found in
natural forests or alpine habitats and in traditionally managed
semi-natural grasslands (Tyler and Bertilsson, 2009). The
species reproduce exclusively by seeds with a conspicuous
pappus, suggesting efficient wind dispersal, but, at least in sheltered forested habitats, their dispersal capacity is probably rather
limited (Tyler, 2000). The genus comprises approx. 20 diploid
species (Fehrer et al., 2009) with obligately sexual reproduction,
and thousands of polyploids which as far as is known are obligately apomictic (Mráz, 2003). With the exception of the widespread H. umbellatum, the former are all more or less confined to
unglaciated refugia along the southern edge of the present-day
range of the genus. Thus, in northern Europe including
Scandinavia, apomictic triploids, tetraploids and a few pentaploids (Tyler and Jönsson, 2009; Thomas et al., 2011) prevail
and most of these species are more or less narrow endemics confined to relatively recently deglaciated regions (Johansson,
1923; Tyler, 2000). Notably, with the exception of section
Alpina, no diploid sexual taxa are known in any of the sections
of the genus that dominate in northern Europe (i.e. H. sects.
Bifida, Foliosa, Hieracium, Oreadea, Subalpina, Tridentata,
Vulgata). Apomictic seed production is via the parthenogenetic development of the unreduced egg cell (diplospory of the
‘Antennaria-type’), generally considered to provide very limited
possibilities for sexuality. The male function varies among apomicts: many clones produce no pollen at all (Johansson, 1923)
but some produce 5–50 % viable pollen grains, and there
appears to be no close relationship between pollen production
and ploidy (triploid vs. tetraploid) or environmental conditions
(Slade and Rich, 2007; Thomas et al., 2011). Recently, some
diploid members of the genus have been shown to be able to hybridize giving rise to polyploid offspring (Mráz et al., 2005, 2011), but
the few natural hybrids found were completely seed-sterile (Mráz
et al., 2005). However, in a later experiment, artificial diploid
hybrids with some, though reduced, fertility were produced
(Mráz and Paule, 2006). A strong mentor effect has been shown
to induce autogamy in the diploids in the presence of any foreign
pollen (Mráz, 2003), further reducing the likelihood of sexual
hybridization.
The taxonomy of European Hieracium is much confused
with two co-existing ‘schools’ circumscribing species differently (Schuhwerk, 2002). According to the ‘Scandinavian
system’ followed by most British, Scandinavian and Russian
taxonomists, all constantly morphologically recognizable biotypes are recognized at the rank of species and believed to be
apomictic clones of unique but generally unknown origin
(Johansson, 1927; Sell, 1987; Tyler, 2006). More than 2000
such species have been described from Scandinavia. In this
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
system, the European species are referred to 38 sections of the
genus (Stace, 1998), some of which comprise .1000 species.
However, since these sections are rarely regarded as anything
but units of sorting and communication, their circumscription
and delimitation have not attracted much interest (Tyler, 2006)
and their evolutionary coherence is questionable. In contrast,
in the ‘C European system’ established by Zahn (1921 – 23)
and followed by most taxonomists working in central and
southern Europe, the taxa recognized as species in the
Scandinavian system are treated at subspecific or lower rank
and their circumscription and identification is given less attention. These infraspecific taxa are part of a complex formal
hierarchy including two kinds of species: (1) ‘basic species’
having a unique set of characters and believed to be ancestral
and of non-hybrid origin, and (2) ‘intermediate species’ that
combine characters from two or more basic species and
which are believed to have a hybrid origin. However, recent
molecular studies have demonstrated that both types of
species commonly have a hybrid origin (Fehrer et al., 2009).
In an attempt to somewhat reconcile the two systems, Tyler
(2006) advocated the use of informal ‘species aggregates’
based on thorough multivariate morphometric analysis to
sort the species recognized in the Scandinavian system into
units comparable with the species of the C European system.
On occasions the same specific names are used in both
systems but with highly different meanings and in North
America and in areas where Hieracium occurs as invasive
aliens only, names and concepts from both systems are often
misunderstood and mixed up, making comparisons between
regions haphazard, and effectively hampering any biogeographic analysis. As a consequence, results from genetic, evolutionary or phylogenetic studies founded on taxonomic
information or concepts need to be interpreted with great
care and in the light of the taxonomic tradition in which the
authors are working. In the present paper, the term ‘species’
will from here be used exclusively in the Scandinavian tradition to denote the smallest morphologically recognizable
groups of biotypes or clones.
The first molecular phylogenetic analyses of Hieracium s.l.
were published by Gaskin and Wilson (2007) and by Fehrer
et al. (2007) based on plastid and plastid and nuclear ITS
sequences, respectively. Both these studies included only a
few taxa of Hieracium s.s. and provided limited resolution and
low levels of variation among taxa. Chrtek et al. (2009) analysed
more taxa and, using the more variable nuclear ETS region
instead, were able to reveal a basal split of the genus into a
‘Western’ and an ‘Eastern’ clade, further supported by differences in genome size. This result was further elaborated by
Fehrer et al. (2009) by sequencing individual copies of the
nuclear EPS gene and additional plastid sequences. They were
thereby able to document intra-individual polymorphisms in
the majority of taxa, previously believed to be ‘basic’ or ancestral, indicative of an (ancient) hybrid origin. The basal split
between the ‘Western’ and ‘Eastern’ clades was clearly supported by plastid trnT-trnL sequences and by nuclear EPS
sequences when hybrid taxa were disregarded. Most notably,
however, they found that the majority of the taxa analysed
showed traces of hybridization, and representatives of several
of the sections of the genus that dominate in northern Europe
appeared to have a history involving hybridization between
593
representatives of the two major clades. Most recently, these
results have also found partial support in analysis of floral
scent components (Feulner et al., 2011). Still, however, these
phylogenetic analyses were based on a small number of representative taxa considered as ‘basic’ in the C European taxonomic
system. The results can only with difficulty be compared with the
sectional taxonomy of the Scandinavian taxonomic system and
provide little information about the phylogenetic relationships
of, and the evolutionary processes acting on, the individual
apomictic species.
At the level of individual species, Tyler (2006) discussed
the morphological homogeneity and the distribution of morphological characters among southern Scandinavian members
of sects. Hieracium (≈H. murorum s.l. in the C European
system), Vulgata (≈H. lachenalii s.l.), Bifida (≈ H. bifidum
s.l. and H. caesium s.l.), Tridentata (H. laevigatum s.l.) and
their intermediates, and concluded that these sections are
fairly well defined morphologically, albeit not separated by
any discontinuities. Many characters were found to be correlated, but no fixed patterns indicative of pleiotrophic effects
could be revealed. Based on multivariate analysis of morphological characters, he further divided the species treated into 38
informal species aggregates, a system further elaborated in
subsequent publications (e.g. Tyler, 2010) but not yet extended
to all Swedish species of these sections or to species from the
rest of Europe. A similar analysis restricted to Swedish and
Finnish representatives of H. sect. Oreadea was provided by
Tyler (2011).
Tyler and Jönsson (2009) investigated the ploidy of multiple
accessions of .200 species and found the species to be homogeneous with respect to ploidy, a conclusion also drawn by
Stace et al. (1997) and Thomas et al. (2011) based on more
restricted material. Tyler and Jönsson (2009) further revealed
that the morphologically most typical members of the sections
were almost exclusively triploid, whereas tetraploids dominated
among morphologically intermediate species, and they hypothesized that the tetraploids may have originated through recent (in
situ) hybridization between the triploids (although the mechanisms for such a process remain to be elucidated). In contrast,
Fehrer et al. (2009) concluded that, although there is ample evidence for ancient hybridization in the genus, in particular
between lines that have come into secondary contact after
periods of isolation in refugia during the glaciations, there are
no indications of recently ongoing hybridization among the
extant species. Thus, there is no consensus, or even any
well-founded theories, as to how the numerous apomictic
Hieracium species now found as narrow endemics in glaciated
areas, both geographically and phylogenetically far away from
their closest sexual relatives, may have come into being.
Shi et al. (1996), Stace et al. (1997) and Mráz et al. (2001)
used allozymes and Štorchova et al. (2002) used RAPD
markers to study the structure of variation in apomictic
members of H. sect. Alpina. All these studies showed that
some morphologically defined species consist of single genotypes, whereas others are somewhat variable. However, no
similar studies of the other sections of the genus have been
undertaken, and both the number of species included and the
variability of the marker systems applied were insufficient to
draw any general conclusions about evolutionary patterns or
the genetic distinctness of taxa. Similarly, Rich et al. (2008)
594
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
used AFLP markers to study genetic variation, mainly within a
single apomictic Hieracium species, and concluded that the
species was distinct at the molecular level, although every individual showed a unique molecular pattern with this hypervariable set of markers. One study with molecular markers
indicates extensive intra-specific variability and that a taxon
of Hieracium introduced to New Zealand (‘H. lepidulum’, apparently a name applied collectively to several species of H.
sect. Vulgata there) has acquired sexual reproduction as part
of the process of becoming invasive (Chapman et al., 2004),
but this conclusion is questionable since the taxonomic identity and the origin of the material studied remains obscure
and no comparison was made with patterns of morphological
variation. However, Rich et al. (2008) reported 1.5 % polymorphic loci among the seed-progeny from a single apomictic
plant using hypervariable AFLP markers, a pattern that may
indicate frequent somatic recombination or elevated mutation
rates in this apomictic line.
The aim of the present study was to obtain an overview of
the distribution and pattern of variation in both maternally
inherited and biparentally inherited markers among the northern European members of the genus Hieracium and to
compare patterns of molecular variation with morphological
variation and morphology-based taxonomic classifications.
Such an overview will be essential in understanding the applicability of these markers and enabling the formulation of stringent questions for future research. It should also provide clues
to the main processes of diversification and evolution, and indicate the validity of the current taxonomy. To fulfil these
aims, we decided to study a broad sample of taxa occurring
in the northern European lowlands, supplemented with some
alpine and more southern representatives, utilizing the
nuclear microsatellites developed by Jönsson et al. (2010) as
biparentally inherited markers, and sequencing the most variable part of the maternally inherited (Mráz et al., 2005)
plastid genome that could be identified. Since this is the first
study of molecular markers at this taxonomic level in this
species-rich group, and the primary aim was to create an overview of the variation, more effort was put into achieving a
large and representative sample than into obtaining complete
data for each individual sampled.
M AT E R IA L S A ND M E T HO DS
Plant material
The plant material for this study was selected from the approx.
800 unique accessions kept in the experimental garden of Lund
University by the first author. All accessions are of known
European origin, mainly from Scandinavia (collection data
available on request). Although all sections of the genus
present in northern Europe are represented, the vast majority
of the accessions belong to the dominant sections in the
Scandinavian lowlands [i.e. sections Bifida, Hieracium,
Oradea, Vulgata and Tridentata; sectional taxonomy according to Stace (1998) and Tyler (2006)]. Except for H. sect.
Oreadea, these sections are all widely distributed in the
Scandinavian lowlands and the material analysed includes
representatives of all sections from almost all non-alpine
Swedish provinces and some accessions from Denmark. For
almost all of the accessions, the ploidy was known from a previous study with flow cytometry (Tyler and Jönsson, 2009),
and all samples, except for those of the diploid sexual
H. umbellatum, represent polyploids (tri-, tetra- and pentaploids) believed to have an exclusively apomictic mode of reproduction (Mráz, 2003). About half of the species sampled
were represented by more than one (usually two to five, but
in one case up to 16) accessions. As far as possible, accessions
from different parts of the distribution range of the mostwidespread species were included. An accession of the
closely related genus Pilosella (P. floribunda × officinarum)
(Fehrer et al., 2007) was included as outgroup/reference. The
species-level taxonomy and nomenclature follow the most
recent treatments of these taxa (e.g. Tyler, 2006, 2010,
2011). Vouchers have been deposited in herbarium LD.
Plastid DNA sequence variation
DNA was extracted as described by Jönsson et al. (2010),
using a protocol modified from Štorchova et al. (2002) and
Doyle (1990). To find the most suitable and variable region,
ten accessions, including representatives from sections
Bifida, Hieracioides, Hieracium, Oreadea, Tridentata and
Vulgata were initially screened for variation at the following
sequences: matK-psbA, psbA-trnH and partial matK [ primers
and protocols according to Kress et al. (2005) and Ford
et al. (2009)], petL-psbE, psbJ-petA, trnV-ndhC, psbD-trnT,
atpI-atpH, trnQ-rps16, rps16-trnK, ndhF-rpl32, rpl32-trnL
[ primers and protocols according to Shaw et al. (2007)],
trnH-trnK and atpB-rbcL [ primers and protocols according
to Chiang et al. (1998)]. Based on the results of this initial
screening, the psbD-trnT spacer was identified as the most
variable and consistently amplifying region. For subsequent
analysis, psbD-trnT from 345 accessions representing 190
named species was successfully amplified using primers and
PCR protocol adopted from Shaw et al. (2007). The PCR products were sequenced in either direction by Sanger sequencing
using the BigDye Terminator v1.1 Cycle Sequencing Kit
(Applied Biosystems) on an ABI PRISM 3130 xl GENETIC
ANALYZER (Applied Biosystems/Hitachi). The sequences
were edited manually using the BioEdit software, and all ambiguous sites were discarded. The edited sequences were
manually aligned and substitutions and indels were coded as
single sites/events. A haplotype network was constructed,
based on all unambiguously polymorphic sites. To visualize
the correlation between the major haplotype groups and the
morphological variation within the four best represented sections of the genus, the former were plotted on a PCA scatterplot based on 46 independent morphological characters [the
same plot as used by Tyler and Jönsson (2009), modified
and supplemented from Tyler (2006)].
Microsatellite analysis
Variation at the ten microsatellite loci developed by Jönsson
et al. (2010) was screened in 530 accessions. The PCR products were run on an ALF express II (Amersham Pharmacia
Biotech) and the digital representation of the gels was interpreted manually using the ALFwin Fragment Analyzer 1.00.
About 25 % of the samples were run twice to check for
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
TA B L E 1. Summary of basic results from analysis of variation at
12 microsatellite loci (adopted from Jönsson et al., 2010) in 477
accessions of Hieracium (and one Pilosella)
Locus
1
2
3
4
5
6a
6b
7a
7b
8
9
10
Size of fragments (bp)
No. of fragments
Diversity
Dosage
105– 116
139– 206
111– 125
158– 179
182– 191
153– 158
120– 151
188– 238
161– 266
216– 230
114– 168
58– 101†
7
27
6
8
4
5
10
11
42
7
24
9†
Medium
High
Medium
Low
Low
Medium
High
Low
High
Low
High
Low
+
–
–
+
+
+
–
+
–
+
+*
+
In the diversity column, ‘Low’ indicates that a single allele/fragment was
completely dominating (.90 %) in the accessions and ‘High’ that five or
more alleles/fragments were found in more or less even frequencies. In the
dosage column, ‘ + ’ indicates that it was possible to interpret the banding
patterns obtained in terms of numbers of alleles at a single genetic locus.
* In some sections of the genus (see text).
†
The Pilosella accession showed additional fragments of size
164– 210 bp.
consistency of results and/or to overcome problems with the
amplification or concentrations. Accessions that did not
amplify or which produced uninterpretable results for three
or more of the loci were disregarded in subsequent analyses,
leaving us with data from 478 accessions. The level of variation revealed and the ease by which the banding patterns
could be interpreted differed among the loci (Table 1).
Primer pairs 6 and 7 each amplified two sets of fragments
that consistently differed in both length and strength. These
were treated as belonging to two different loci, henceforth
denoted ‘a’ and ‘b’. Based on the banding patterns revealed
and the ploidy of the accessions, the number of doses of
each allele at microsatellite loci 1, 4, 5, 6a, 7a, 8 and 10 was
readily inferable in most accessions, but some accessions
showed patterns indicative of duplications and/or multiple
loci. At loci 2, 3, 6b, 7b and 9, the majority of the accessions
showed banding patterns that could not be interpreted in terms
of alleles at a single locus. At locus 9, problems with interpretation was associated with members of H. sect. Bifida, whereas
all accessions from sections Hieracium, Vulgata and
Tridentata were readily interpretable; at locus 3 a very amplifying fragment that was present in almost all accessions interfered with the less strongly amplifying and polymorphic
fragments, making their relative doses impossible to infer.
Loci 2, 6b and 7b were hypervariable and commonly and consistently (as revealed by multiple analyses of the same accessions) showed more fragments in a single accession than
expected from its ploidy.
Due to these problems and problems associated with comparing results among samples with different ploidies, four different data matrices were constructed. Matrix 1a was a simple
presence/absence matrix utilizing information from all 12 loci.
Matrix 1b was the same as matrix 1a excluding the hypervariable loci 2, 6b, 7b and 9. For matrix 2, in an attempt to give
strongly amplifying fragments at the less-variable loci more
595
weight without necessarily interpreting the actual dosage
effects, the presence of a single fragment was assigned the
value of 2 at loci 1, 3, 4, 5, 6a, 7, 8 and 10, a strongly amplifying fragment supplemented with one or more much weaker
allele(s) the value 1.5 and weakly amplifying alleles the
value 1, whereas at the hypervariable loci 2, 6b, 7b and 9
all fragments were assigned the value 0.25. In matrix 3 the
number of copies of each allele at loci 1, 4, 5, 6a, 7, 8 and
10 was inferred (based on strength of amplification and the
known ploidy) and the numbers were standardized such that
the sum of all alleles at each locus became 1, the alleles
detected at locus 3 were all assigned the same value such
that their sum was 1, whereas at the hypervariable loci 2, 6b,
7b and 9 all fragments were assigned the value 0.1. All
three data matrices were subjected to neighbor-joining (NJ)
analysis based on the Euclidean distances between accessions.
In addition, matrices 1a and 1b were analysed using the
Sörensen/Dice similarity coefficient (giving more weight to
joint occurrences than absences). These analyses were performed on the complete set of data, on subsets consisting of
species referred to each of the four best-represented sections
of the genus and on subsets representing the two main
groups of taxa indicated by the result of the plastid haplotype
sequence analysis (see below). All NJ trees were rooted using
the Pilosella sample as the outgroup. To get an indication of
the robustness of the clusters obtained by NJ, boostrapping
with 1000 replicates was performed. In a further attempt to
visualize the pattern of variation, matrix 1a was subjected to
principal component analysis (PCA). All statistical calculations were performed in the program PAST ver. 2.13
(Hammer et al., 2001).
R E S ULT S
Plastid DNA sequence variation
The complete psbD-trnT alignment (disregarding ambiguities)
was 1243 bp. Fifty unambiguous polymorphisms were revealed (33 substitutions and 17 insertions/deletions), defining
a total of 40 unique haplotypes in the 349 analysed accessions
(Table 2). All haplotypes found in the accessions belonging to
the sections distributed in the northern European lowlands fell
into one of two main groups, henceforth denoted group H and
group V, separated by seven or eight polymorphisms (Fig. 1 and
Table 3). Additional distinct haplotypes were found in single
accessions of Pilosella and in southern European and alpine
Hieracium sects. Alpina, Barbata, Mixta and Dremanoidea.
Since only partial sequences were available for many accessions
not all accessions could be referred to a particular haplotype, but
the wide distribution of the seven or eight polymorphisms differing between groups H and V allowed for the identification
of all sequences to this level.
Group V comprised one common haplotype (V1, found in
43 accessions) and seven rare haplotypes found in one to
three accessions and separated by one to three polymorphisms.
Among accessions with a northern European origin, all accessions belonging to H. sects. Foliosa, Hieracioides
(H. umbellatum) and Tridentata and all but one accession of
triploid species currently referred to H. sects. Oradea and
Vulgata, showed haplotypes belonging to this group. The
596
defining
the
30
haplotypes detected by analysis of the psbD –trnT
northern European Hieracium species (see Fig. 1)
spacer
(1243 bp)
in
349
accesions
of
mainly
Haplotype/
Alignment
position
75
88
109
142
202
209
210– 15
216 –17
218 –33
234– 239
266
269?
274
277
282
297
308
315
318
321
326
328
338
351
352
360
V1
V2
V3-1
V3-2
V4
V5
V6
V7
H0
H1
H2
H3
H4
H5
H6
H7
H8:1
H8:2
H9
H10:1
H10:2
H10:3
H10:4
H10:5
H11
H12
H14
H15:1
H15:2
H15:3
H15:4
H15:5
H15:6
H15:7
C
Pilosella
Alpina
Barbata
Mixta
Dremanoidea
C
C
C
C
C
C
C
C
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
?
T
T
T
T
T
T
T
T
C
?
?
?
T
A
A
A
A
A
A
A
A
A
A
A
A
A
A
G
A
A
A
A
A
A
A
A
A
A
A
?
A
A
A
A
A
A
A
A
A
?
?
?
A
G
G
G
G
G
G
G
G
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
?
A
A
A
A
A
A
A
A
A
?
?
?
A
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
G
T
T
T
T
T
T
T
T
T
T
?
T
T
T
T
T
T
T
T
T
?
?
?
T
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
A
A
C
C
C
C
C
C
C
C
?
C
C
C
C
C
C
C
C
C
?
?
?
C
T
T
T
C
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
?
T
T
T
T
T
T
T
T
T
?
?
?
T
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
?
0
0
0
0
0
0
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
0
0
0
0
0
0
0
0
?
?
?
0
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
?
0
0
0
0
0
0
0
0
0
?
?
?
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
?
1
1
1
1
1
0
1
1
0
?
?
?
1
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
G
T
T
T
T
T
?
T
T
T
T
T
T
T
T
T
?
?
?
T
T
T
T
T
T
T
C
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
?
T
T
T
T
T
T
T
T
T
?
?
?
T
T
T
T
T
T
T
A
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
?
T
T
T
T
T
T
T
T
T
?
?
?
T
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
?
C
C
T
T
T
C
C
C
C
?
?
?
C
C
C
C
C
C
C
T
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
?
C
C
C
C
C
C
C
C
C
?
?
?
C
G
G
G
G
A
G
?
G
G
G
G
G
G
G
G
G
G
G
G
A
A
A
A
A
G
G
?
G
G
G
G
A
G
G
G
G
?
?
?
G
G
G
G
G
T
G
?
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
?
G
G
G
G
G
G
G
G
G
?
?
?
G
T
T
T
T
T
T
?
T
T
T
T
T
T
T
T
T
T
T
T
T
?
G
G
T
T
T
?
T
T
T
T
T
T
T
T
T
?
?
?
T
A
A
A
A
A
A
?
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
?
A
A
A
G
A
A
A
A
A
?
?
?
A
A
A
A
A
A
A
?
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
G
?
A
A
A
A
A
A
A
A
A
?
?
?
A
G
G
G
G
G
C
?
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
?
G
G
G
G
G
G
G
G
G
?
?
?
G
A
A
A
A
A
A
?
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
C
?
A
A
A
A
A
A
A
A
A
?
?
?
A
C
C
C
C
C
C
?
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
?
C
C
C
C
C
C
G
C
C
?
?
?
C
T
T
T
T
T
G
?
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
?
G
G
G
G
G
G
G
T
T
?
?
?
T
G
G
G
G
G
A
?
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
?
G
G
G
G
G
G
G
G
G
?
?
?
G
G
G
G
G
G
G
?
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
T
G
?
G
G
G
G
G
G
G
G
G
?
?
?
G
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
TA B L E 2. Polymorphisms
Haplotype/
Alignment
position
392
411
419
456
630
655
660
675 –77
680
689
694 –699
706
712
715– 720
791
840
1003
1038
1069
1218– 36
1237
1238
1245
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
?
0
0
0
0
0
0
0
1
0
?
?
?
1
T
T
T
T
T
?
?
T
T
T
T
T
T
T
T
T
T
T
T
T
T
C
T
T
T
?
?
T
T
T
T
T
?
T
T
?
?
?
?
T
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
?
0
0
0
0
0
?
?
0
0
?
?
?
0
1
0
1
1
1
?
?
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
?
?
1
1
1
1
1
?
?
1
?
?
?
?
1
T
T
T
T
T
?
?
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
?
?
T
T
T
T
T
?
?
G
?
?
?
?
G
T
T
T
T
?
?
?
T
T
G
G
G
G
G
G
G
G
G
G
G
?
?
?
G
G
?
?
G
G
G
G
?
?
?
G
?
?
?
?
G
0
1
0
0
?
?
?
0
0
0
0
0
0
0
0
0
0
0
1
0
?
?
?
0
0
?
?
0
0
0
0
?
?
?
0
?
?
?
?
0
0
0
0
0
?
?
?
0
0
0
0
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
1
0
0
0
0
0
0
0
0
?
?
?
0
0
?
?
0
0
0
0
?
?
?
0
?
?
?
?
0
G
G
G
G
?
?
?
G
T
T
T
T
T
T
T
T
T
T
T
T
?
?
?
T
T
?
T
T
T
T
T
?
?
?
G
?
G
G
G
G
0
0
0
0
?
?
?
0
0
0
0
1
0
0
0
0
0
0
0
0
?
?
?
1
0
?
0
0
0
0
0
?
?
?
0
?
0
0
0
0
0
0
0
0
?
?
?
0
0
0
0
0
1
0
0
0
0
0
0
0
?
?
?
0
0
?
0
0
0
0
0
?
?
?
0
?
0
0
0
0
C
C
C
C
?
?
?
C
C
C
C
C
C
A
C
C
C
C
C
C
?
?
?
C
C
?
C
C
A
C
C
?
?
?
C
?
C
C
C
C
A
A
A
A
?
?
?
A
A
A
C
C
A
A
A
A
A
A
A
A
?
?
?
A
A
?
A
A
A
A
A
?
?
?
A
?
A
A
A
A
0
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
C
C
C
C
C
C
C
C
?
C
?
?
C
C
?
?
?
?
?
?
?
?
?
?
?
?
C
C
C
C
C
?
?
?
?
?
A
A
C
?
T
T
T
T
T
T
T
T
?
G
?
?
G
G
?
?
?
?
?
?
?
?
?
?
?
?
G
G
G
G
G
?
?
?
?
?
G
G
G
?
C
C
C
C
C
C
C
C
?
T
?
?
T
T
?
?
?
?
?
?
?
?
?
?
?
?
T
T
T
T
T
?
?
?
?
?
T
T
T
?
T
T
T
T
T
T
T
T
?
G
?
?
T
T
?
?
?
?
?
?
?
?
?
?
?
?
T
T
T
T
T
?
?
?
?
?
T
T
T
?
T
T
T
T
T
T
T
T
?
T
?
?
T
T
?
?
?
?
?
?
?
?
?
?
?
?
T
T
T
T
T
?
?
?
?
?
T
T
G
?
1
1
1
1
1
1
1
1
?
0
?
?
0
0
?
?
?
?
?
?
?
?
?
?
?
?
1
0
0
0
0
?
?
?
?
?
0
0
1
?
1
1
1
1
1
1
1
1
?
0
?
?
0
1
?
?
?
?
?
?
?
?
?
?
?
?
1
1
1
1
1
?
?
?
?
?
0
1
1
?
1
1
1
1
1
1
1
1
?
1
?
?
1
1
?
?
?
?
?
?
?
?
?
?
?
?
1
1
1
1
1
?
?
?
?
?
0
1
1
?
1
1
1
1
1
1
1
1
?
0
?
?
1
1
?
?
?
?
?
?
?
?
?
?
?
?
1
1
1
1
1
?
?
?
?
?
1
1
1
?
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
V1
V2
V3-1
V3-2
V4
V5
V6
V7
H0
H1
H2
H3
H4
H5
H6
H7
H8:1
H8:2
H9
H10:1
H10:2
H10:3
H10:4
H10:5
H11
H12
H14
H15:1
H15:2
H15:3
H15:4
H15:5
H15:6
H15:7
C
Pilosella
Alpina
Barbata
Mixta
Dremanoidea
369 –70
0/1 denote the absence/presence of indels.
597
598
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
H15:4 H15:5
V3:2
Dremanoidea
H15:3
H15:2
V5
H15:6
H15:7
V3:1
H6 H14
H5
H10:2
H10:5
H10:3
Pilosella
Mixta
H15:1
V7
H7
H10:1
H1
V1
H0
V2
H10:4
H12
H4
H8:1
Barbata
H11
H8:2
H2
H9
V4
Alpina
V6
H3
Group V:
Alpestria
Andryaloidea
Foliosa
Hieracioides
Italica
Oreadea* (2n = 3x)
Sabauda
Tridentata
Vulgata* (2n = 3x)
Group H:
Amplexicaulina
Bifida
Hieracium
Oreadea* (2n = 4x)
Stelligera*
Thapsoidea
Villosa
Vulgata* (2n = 4x)
F I G . 1. Most-parsimonious network of the plastid ( psbB-trnT spacer, 1243 bp, 50 polymorphic sites) haplotypes (with some ambiguity within group H due to
incomplete sequences) revealed in Swedish (black) and non-Swedish (grey) species and sections of Hieracium. The size of the circles is approximately proportional to the number of species documented. One step along the branches in the network corresponds to one evolutionary event (substitution or insertion/deletion).
Missing haplotypes are indicated as ticks on the branches. Asterisks indicate that only some members of the section belong here (as discussed in text).
only exception was the tetraploid H. constringens which is a
morphologically strongly deviating species currently referred
to H. sect. Vulgata. In addition, single accessions of
the more southernly distributed H. sects. Alpestria,
Andryaloidea, Italica and Sabauda all shared haplotype V1.
Group H comprised four relatively common haplotypes and
23 relatively rare ones. The most distant haplotypes in this
group were separated by seven polymorphisms, but most transitional haplotypes were found, and the most common haplotypes differed by only one or two polymorphisms (Fig. 1).
All northern European accessions belonging to H. sects.
Bifida and Hieracium (both triploids and tetraploids) showed
haplotypes belonging to this group, as well as all tetraploid
representatives of H. sects. Oreadea and Vulgata (in most
cases morphologically intermediate between these sections
and H. sect. Hieracium, cf. Tyler and Jönsson, 2009).
Hieracium crinellum, a triploid Scandinavian species variously
referred to H. sects. Oreadea or Stelligera (cf. Tyler, 2011),
also exhibited a haplotype of this group (the unique haplotype
H10:5). In addition, single accessions of the more southernly
distributed H. sects. Amplexicaulina, Thapsoidea and Villosa
contained haplotypes of group H.
The various deviating haplotypes found in the single representative of Pilosella, in three accessions of the arctic – alpine
H. sect. Alpina and in the southern and central European
H. sects. Barbata, Dremanoidea and Mixta are all connected
to the path uniting haplotypes of groups H and V, but they
all have their own (one to three) synapomorphies. Thus, no
sequences truly transitional between the two major groups
were found (Fig. 1).
Microsatellite analysis
The number of different fragments amplified at each of the
microsatellite loci varied between the four (one of which completely dominating) found at locus 5 to the 42 identified at
locus 7b (Table 1). In total, 160 fragments were identified in
the 478 accessions that were successfully analysed. Levels of
variation were generally high with only nine pairs of accessions being identical at all loci (in all cases representing accessions of the same species) and many pairs of samples had only
a few common fragments in common. NJ analysis based on the
different data matrices all showed essentially the same kind of
pattern. Accessions of the same species generally cluster together (with some exceptions) and, in some of the combinations of data matrix and distance measure, some smaller
groups of species congruent with morphology and/or current
taxonomy were revealed, but most or all larger groupings
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
599
TA B L E 3. Names of species, sectional taxonomy, ploidy, haplotype(s) (cf. Fig. 1) and the number of accessions analysed
Species
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
acidolepis
ageneium
albidulum
albidulum auct. Danici
almquistianum
amplexicaule
amplificatum
anfractifolium
anfractum
anodon
austrinum
barbareifolium
basifolium
basilimbatum
bombycinum
bupleuroides
caesiiflorum
caesiomurorum
caesionigrescens
caesitium
caliginosum
calliglaucum
canipes
canovittatum
ceramotum
chlorodes
ciliatiflorum
cinerellum
coadunatum
compitale
coniops s.l.
conspectum
constringens
contaminatum s.l.
crassiceps
crebriserratum
crinellum
cruentiferum
cruentifolium
dalicum
dentatum ssp. dentatiforme
diaphanoides
dissimile
dolichorhachis
dolobratum
elegantiforme
eulasium
exaltans
expallidiforme
extensiforme
extensum
flagriferum
fulvescens
gelertii
glandulosissimum
gracilifrons
grandidens
grandiserratum
haegerstroemii
hepaticolor
hjeltii
informe
integratum
intermarginatum
involutum
isoptortum
Section
Bifida
Hieracium
Bifida
Bifida
Vulgata
Amplexicaulina
Bifida?
Hieracium
Vulgata
Oreadea
Vulgata
Vulgata
Bifida
Hieracium
Mixta
Dremanoidea
Bifida
Bifida
Bifida
Bifida
Hieracium
Bifida
Hieracium
Vulgata
Vulgata
Vulgata
Hieracium
Bifida
Bifida
Bifida
Bifida
Vulgata
Vulgata
Hieracium
Hieracium
Hieracium
Oreadea
Tridentata
Vulgata
Vulgata
Villosa
Vulgata
Bifida?
Vulgata
Tridentata
Alpestria
Oreadea
Bifida
Hieracium
Oreadea
Oreadea
Hieracium
Bifida
Bifida
Hieracium
Bifida
Hieracium
Hieracium
Hieracium
Hieracium
Hieracium
Hieracium
Hieracium
Hieracium
Bifida?
Vulgata
Ploidy level
3x
3x
3x
3x
3x
4x
4x
?
3x
3x
3x
3x
4x
3x
3x
3x
3x
4x
3x and 4x
3x
4x
4x
4x
3x
4x
4x
3x
4x
4x
3x
4x
3x
4x
3x
?
3x
3x
3x
3x
3x
3x
4x
4x
3x
3x
4x
4x
4x
3x
3x
4x
4x
4x
3x
4x
4x
3x
3x
4x
4x
4x
4x
3x
5x
4x
3x
cpHaplotype
N
H15:3
H15:2
H7
H15:2
V1
H15
H(0 –5)
H15:2
V7
V1
H(1 –5)
V1
H15:3
H15:1
Mixta
Dremanoidea
H15:2
H1
H10:1
H15:3
H2
H1
H
V1
H11
H(1 –5)
H4
H(0 –5)
H1
H
H4 and H10:2
V
V1
H15:6
H(0 –5)
H15:2
H10:1 and H10:5
V1
V
V1
H
H4 and H12
H15:2
V1
V3:2
V1
H(0 –5)
H15:1–2
V1
V1 and V5
H1
H(0 –5)
H(0 –)
H15:2
H4
H(1 –5)
H15:1–2
H5 or H15
H15:2
H2 and H4
H8:2
H1
H4
H(0 –5)
H9 and H11
V1
3
1
1
1
1
1
2
1
1
10
1
2
4
1
1
1
2
2
2
2
2
1
1
1
2
1
1
2
1
1
6
1
2
1
1
1
3
1
4
1
1
3
1
1
3
1
4
1
2
15
1
1
1
1
1
1
2
1
1
2
1
1
2
1
2
1
(Continued)
600
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
TA B L E 3. Continued
Species
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
H.
itharophyton
jaeredense
johanssonii
koehleri
kueckenthalianum ssp. tephrosoma
lacerifolium
laeticolor
lanuginosum
latifrons
lecanodes
lepidotum
lepidulum
lepistoides
lindebergii
lythrodes
maciatum
macrocentrum
macrocomum
madarodes
mallopodum
marginellum
megalodon
megavulgatum
microdon
molybdinum
myrtillinum
naevosum
nemorivagum
neopinnatifidum
neoserratifrons
niveolimbatum
obatrescens
obtusius
oestmanii
ohlsenii
oistophyllum
opeatodontum
ornatiforme
ornatum
oxylepium
pachyodon
pamphilii
panaeolum
paralium
paramaurum
patale
pectinatum
pellucidum
pendulum
persimile
phaeodermum
philanthrax
piliferum
plicatum
plumbeum
porphyrostictum
porrigens
porrigentiforme
praetenerum
pristophyllum
prolixum
psepharum
pseudodiaphanum
pseudogelertii
psilurum
pubicuspis
Section
Bifida
Vulgata
Hieracium
Hieracium
Subalpina
Bifida
Bifida
Hieracium
Oreadea
Oreadea
Vulgata
Vulgata
Hieracium
Oreadea
Oreadea
Vulgata
Vulgata
Alpestria
Bifida?
Hieracium
Hieracium
Vulgata
Vulgata
Oreadea
Bifida
Hieracium
Vulgata
Sabauda
Vulgata
Hieracium
Bifida
Tridentata
Hieracium
Vulgata
Hieracium
Hieracium
Bifida
Vulgata
Vulgata
Bifida
Hieracium
Villosa
Hieracium
Vulgata
Hieracium
Hieracium
Hieracium
Hieracium
Bifida
Hieracium
Vulgata
Hieracium
Barbata
Vulgata
Bifida
Hieracium
Bifida
Vulgata
Bifida
Vulgata
Bifida
Bifida
Vulgata
Bifida
Hieracium
Vulgata
Ploidy level
4x
3x
3x
3x
4x
4x
3x
4x
3x
4x
3x
3x
3x
3x
3x
3x
3x
4x
4x
4x
4x
4x
4x
?
4x
3x
3x
3x
3x
4x
4x
3x
3x
4x
3x
3x
3x
4x
4x
4x
3x
4x
3x
3x
4x
4x
4x
4x
3x
3x
3x
3x
4x
3x
4x
3x
4x
4x
3x
3x
3x
3x
3x
4x
3x
?
cpHaplotype
N
H8:1
V4
H2
H15:2 and H15:7
V
H7
H
H3
V
H0
V1
V1
H(1 –5)
V1
V1
V1
V1
V1
H2
H
H1
H(1 –5)
V1
V1
H10:4
H15:2
V1
V
V1
H5 or H15
H15:1– 2
V
H15:1– 2
H15:3
H
H4
H7
H15:1– 2
H(1 –5)
H5
H15:2
H
H15:2
V1
H1
H2
H6
H1 and H2
H4
H2
V
H1 and H4
Barbata
V1 and V4
H1
H15:2
H15:2
H15:–2
H4
V(1, 2, 4, 5)
H15:5
H15:1– 2
V4
H15:2
H
V
2
1
1
2
1
1
1
1
4
2
3
1
1
5
4
1
1
1
1
1
1
1
1
1
2
1
2
1
1
1
1
1
1
1
1
3
1
1
1
1
1
1
2
3
1
1
1
4
1
1
2
2
1
2
8
1
3
1
1
2
1
1
1
1
1
1
(Continued)
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
601
TA B L E 3. Continued
Species
H. punctillaticeps
H. punctillatum
H. radiiflorum
H. ravidum
H. remanens
H. resupinatum
H. sabaudum agg.
H. sarcophyllum
H. sarissatum
H. saxifragum
H. scandinaviorum
H. schlyteri
H. scioides
H. scotostictum
H. scytophyllum
H. siljense forma
H. smolandicum
H. sparsidens
H. sparsiguttatum
H. spodoleucum
H. stenocranoides
H. stenolepis
H. stenstroemii
H. stipatum
H. stiptadenium
H. striatisquameum
H. subarctoum
H. subatronitens
H. subcrassum
H. subinquilinum
H. sublaeticeps
H. subpellucidum?
H. subpunctilatum
H. subsimile
H. subterscissum
H. subulatidens
H. sudermannicum
H. svanlundii
H. symphytaceum s.l.
H. taeniifolium
H. torticeps
H. triangulare s.s.
H. triangulare ‘sublacerifolium’
H. trichelliceps
H. triviale
H. umbellatum
H. valentius
H. variicolor
H. villatingense
H. xestocarenum
Pilosella floribunda × officinarum
H. sect. Alpina sp.
H. sect. Andryaloides sp.
H. sect. Foliosa sp.
H. sect. Hieracium sp. nov.
H. sect. Thapsoidea sp.
H. sect. Tridentata sp.
H. sect. Tridentata sp.
H. sect. Tridentata sp.
Section
Vulgata
Vulgata
Hieracium
Bifida
Hieracium
Vulgata
Sabauda
Bifida
Hieracium
Oreadea
Bifida
Vulgata
Hieracium
Hieracium
Vulgata
Bifida
Vulgata
Hieracium
Hieracium
Vulgata
Hieracium
Bifida
Hieracium
Vulgata
Hieracium
Vulgata
Vulgata
Vulgata
Hieracium
Vulgata
Bifida
Vulgata
Vulgata
Vulgata
Hieracium
Hieracium
Hieracium
Oreadea
Italica
Vulgata
Hieracium
Hieracium
Hieracium
Vulgata
Vulgata
Hieracioides
Hieracium
Hieracium
Vulgata
Tridentata
Pilosella
Alpina
Andryaloidea
Foliosa
Hieracium
Thapsoidea
Tridentata
Tridentata
Tridentata
Ploidy level
3x
3x
3x
4x
4x
3x
3x
4x
4x
3x
3x
3x
4x
3x
3x
4x
3x
4x
3x
3x
3x?
3x
4x
4x
4x
3x
4x
3x
4x
3x
4x
4x
3x
3x
4x
3x?
4x
3x
3x
4x
3x
3x
4x
3x
3x
2x
4x
4x
3x
3x
3x?
3x
3x
3x
3x
4x
3x
4x
3x
cpHaplotype
V1 and V6
V1
H15:1
H15:2
H15:3
V1
V
H1 and H10:3
H15:1–2
V
H
V4
H2
H15:2
H(1 –5)
H(1 –5)
V3:2
H(0 –5)
H (ej 2 –5)
V1
H12
H1
H5
V1
V1
V
H4
V1
H15:2
V1
H15:2
H2
V1
V1
H15:2
H15:1–2
H4
V1
V
H(1 –5)
H15:2
H1
H1
V1
V1
V1
H1
H2
V(1, 2, 5)
V
Pilosella
A2
V1
V1
H15:2
H14
V3:1
V(1, 2, 5)
V1
N
2
1
1
6
1
3
1
3
1
1
1
1
1
3
2
2
1
4
1
1
1
2
2
1
3
1
2
1
1
1
1
1
1
2
2
3
1
1
1
1
2
1
1
3
3
3
3
2
2
1
1
2
1
1
1
1
1
1
5
Due to incomplete sequences, some haplotypes could only be referred to one of the major haplotype groups or to a subgroup of two or more similar
haplotypes.
appear largely random. This impression is further strengthened
by the bootstrap analysis giving weak or no support for most or
all basal branches in the topologies (not shown). The best fit
between the results of the NJ analysis and morphology/
taxonomy was obtained by the analysis of data matrix 1a
( presence/absence data utilizing information from all 12
loci) with the Sörensen/Dice similarity coefficient; this is the
only analysis that will be discussed further. In the complete
602
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
NJ dendrogram including all species analysed (not shown)
most (51 out of 56) accessions of members of section
Oreadea (including H. crinellum) cluster together with only
few (five) admixtures of species from other sections. The
H. plicatum aggregate of H. sect. Vulgata [sensu Tyler
(2006), with the addition of the morphologically somewhat
similar H. constringens and H. expallidiforme] and the
H. stenolepis aggregate (H. coadunatum, H. plumbeum and
H. stenolepis) also form separate clusters, but most larger clusters appear to be uncorrelated with both taxonomy/morphology or to the groupings revealed by the analysis of plastid
haplotypes. When the same analysis was conducted separately
for accessions harbouring plastid haplotypes of groups H and
V, respectively, largely the same pattern appeared (not
shown). Except for H. sect. Oreadea, there is hardly any correlation between these dendrograms and current sectional taxonomy. Further, accessions belonging to morphologically
distant members of sections from southern Europe appear
interspersed among the accessions of the dominant sections
in the northern European lowlands.
If the morphologically/taxonomically most distant accessions
are omitted and the analysis is performed separately for each of
the well-represented sections Bifida, Hieracium, Oreadea and
Vulgata + Tridentata (Figs 2 – 5) some more logical groupings
show up (to be discussed in more detail below), but the
general pattern remains that accessions of the same species
cluster together, whereas most larger clusters appear to be
formed more or less randomly, and in general only the smallest
clusters receive reasonable bootstrap support (not shown).
In the PCA analysis based on matrix 1, including only accessions of H. sects. Bifida, Hieracium, Oreadea, Tridentata and
Vulgata, only approx. 25 % of the variation in the original data
was explained by the first two axes, and these axes heavily
rely on the presence/absence of only few (approximately five
each) of the 160 fragments in the original data as revealed by
the character loadings. In the resulting scatterplot (not shown),
sections Hieracium and Vulgata are fairly well separated along
the first axis but members of the other sections appear interspersed and the second axis appears unrelated to other existing
data. Apparently, presence/absence of individual markers is
largely uncorrelated, making it impossible to reduce the variation to a few components as attempted by the PCA analysis.
DISCUSSION
The integrity and potential of the species
One of the structural controversies in Hieracium taxonomy is
whether or not it is at all possible to identify the basic apomictic clones/taxa based on morphology, and the other way round,
whether or not the species identified this way constitute the
real monotopic apomictic clones. Previous studies using
various molecular markers and focusing on individual
species have revealed significant variation within most morphologically defined species (Shi et al., 1996; Stace et al.,
1997; Mráz et al., 2001, Štorchova et al., 2002; Rich et al.,
2008), but studies have been restricted to the alpine sections
of the genus and the sampling has not been suitable for investigating to what extent that variation overlaps with that of other
species and whether or not species are monotopic. However,
H. ravidum (1275)
H. ravidum (1277)
H. ravidum (1276)
H. ravidum (1274)
H. ravidum (1278)
H. caesiiflorum (1253)
H. helsingicum (1259)
H. ravidum (722)
H. caesiiflorum (1091)
H. caesiiflorum (1254)
H. compitale (1432)
H. informe (1286)
H. longimanum (913)
H. prolixum (981)
H. canitiosum (1388)
H. acidolepis (1332)
H. acidolepis (1396)
H. acidolepis (1405)
H. albidulum auct. Dan, (1213)
H. molybdinum (685)
H. molybdinum (928)
H. sublividum (1014)
H. dissimile (1293)
H. longimanum (910)
H. prolixum (1105)
H. fulvescens (1418)
H. laeticolor (896)
H. longimanum (851)
H. longimanum (911)
H. caesiomurorum (630)
H. basifolium (1359)
H. basifolium (X2)
H. basifolium (843)
H. sp. nov. (1169)
H. basifolium (563)
H. maculosum (920)
H. laeticolor (897)
H. exaltans (1425)
H. scandinaviorum (917)
H. scandinaviorum (994)
H. scandinaviorum (611)
H. gelertii (1216)
H. pseudogelertii (1219)
H. sublaeticeps (1284)
H. amplificatum (1417)
H. amplificatum (1444)
H. amplificatum (1290)
H. prolixum (714)
H. acudentulum (828)
H. caesitium (1433)
H. pendulum (421)
H. porrigens (267)
H. porrigens (84)
H. porrigens (974)
H. porrigens (975)
H. caesitium (848)
H. coniops (1064)
H. coniops (250)
H. coniops (1386)
H. coniops (1443)
H. coniops (373)
H. pendulum (1157)
H. praetenerum (1270)
H. mediiforme (927)
H. coniops (248)
H. cinerellum (1437)
H. cinerellum (1455)
H. sijiense (1002)
H. sijiense (1429)
H. albinotum (1346)
H. caesionigrescens (1092)
H. caesionigrescens (633)
H. caesionigrescens (306)
H. calatharium (1096)
H. lacerifolium (895)
H. opeatodontum (1195)
H. sarcophyllum (442)
H. sarcophyllum (990)
H. sarcophyllum (726)
H. itharophyton (1125)
H. itharophyton (889)
H. gracilifrons (1354)
H. albidulum (1081)
H. calliglaucum (1097)
H. stenolepis (1179)
H. stenolepis (35)
H. stenolepis (753)
H. coadunatum (1110)
H. plumbeum (1267)
H. plumbeum (710)
H. plumbeum (266)
H. plumbeum (1212)
H. plumbeum (1266)
H. plumbeum (1342)
H. plumbeum (310)
H. cf. plumbeum (22)
H. stenolepis (757)
H. porrigentiforme (1376)
H. oxylepium (1362)
Pilosella floribunda × officinarum (1325)
F I G . 2. Neighbor-joining dendrogram based on a partial analysis of presence/
absence data and the Sörensen/Dice coefficient for 12 nuclear microsatellite
loci, including only members of Hieracium sect. Bifida and the outgroup.
Samples are referred to by the name of the species and the accession
number (in brackets).
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
H. triviale (205)
H. triviale (211)
H. triviale (1415)
H. triviale (1032)
H. triviale (1029)
H. sp. (1409)*
H. sp. (10)*
H. dolobratum (499)*
H. sp. (1422)*
H. paralium (1371)
H. paralium (1392)
H. paralium (1307)
H. pseudodiaphanum (867)
H. megavulgatum (537)
H. neopinnatifidum (103)
H. neopinnatifidum (108)
H. neopinnatifidum (1320)
H. sp. (544)*
H. sp. (556)*
H. cf. neopinnatifidum (163)
H. maciatum (380)
H. sp. (1412)*
H. cf. obatrescens (557)*
H. sp. (1395)*
H. sp. (821)*
H. sp. (347)*
H. sp. (9)*
H. resupinatum (1173)
H. resupinatum (1390)
H. subsimile (1188)
H. subsimile (1189)
H. resupinatum (1174)
H. phaeodermum (1158)
H. phaeodermum (1159)
H. eustictum (1115)
H. elongatifrons (791)
H. dalicum (640)
H. naevosum (101)
H. naevosum (385)
H. lepidotum (1357)
H. sp. (789)*
H. scytophyllum (1177)
H. scytophyllum (1424)
H. pristophyllum (1167)
H. trichelliceps (1318)
H. trichelliceps (1362)
H. trichelliceps (1193)
H. constringens (1106)
H. plicatum (663)
H. constringens (861)
H. constringens (865)
H. constringens (621)
H. plicatum (970)
H. plicatum (971)
H. plicatum (969)
H. schlyteri (968)
H. schlyteri (999)
H. xanthostylum (1033)
H. lepidotum (312)
H. lepidotum (372)
H. lepidotum (903)
H. lepidotum (610)
H. xestocarenum (235)*
H. sp. (498)*
H. cf. oestmanii (1350)
H. austrinum (1328)
H. diaphanoides (339)
H. subinquilinum (477)
H. abfractum (839)
H. abfractum (849)
H. isoptortum (1368)
H. isoptortum (1391)
H. sp. (1382)*
H. pristophyllum (1387)
H. jaedrense (1328)
H. punctillaticeps (1399)
H. punctillaticeps (142)
H. punctillaticeps (141)
H. almquistianum (1410)
H. almquistianum (832)
H. adunans (1288)
H. adunans (1289)
H. linguiforme (1301)
H. lepidiceps (1376)
H. punctillatum (1378)
H. subirriguum (1185)
H. subirriguum (1186)
H. pubicucpis (1171)
H. subatronitens (1315)
H. barbareifolium (1379)
H. barbareifolium (1384)
H. villatingense (841)
H. luebeckii (1137)
H. subpunctillatum (1356)
H. cf. almquistianum (986)
H. stipatum (480)
H. subpunctillatum (1353)
H. smolandicum (1360)
H. smolandicum (1178)
H. lepidulum (1132)
H. lepidulum (1133)
H. villatingense (1401)
H. lepidulum (374)
H. subatronitens (1184)
H. ceramotum (1072)
H. ceramotum (1073)
H. ceramotum (1419)
H. involutum (1348)
H. cuneolatum (1111)
H. subpellucidum (1430)
H. chlorodes (1333)
H. diaphanoides (869)
H. taeniifolium (1347)
H. diaphanoides (49)
H. ceramotum (1103)
H. subarctoum (465)
H. subarctoum (466)
H. ornatum (1306)
H. megalodon (1303)
H. madarodes (1349)
H. oblaqueatum (1148)
H. causiatum (1436)
H. cruentifolium (332)
H. cruentifolium (333)
H. cruentifolium (804)
H. cruentifolium (1411)
H. dolobratum (796)*
H. dolobratum (797)*
H. involutum (1296)
H. dolichorachis (1448)
H. impressiforme (1120)
H. impressiforme (1121)
H. ornatum (944)
H. lepidotum (651)
H. lepidotum (872)
H. striatisquameum (576)
H. acroleucum (826)
H. acroleucum (827)
H. spodoleucum (1361)
H. macrocentrum (1446)
H. schlegelii (1311)
H. minuriens (1059)
H. lugubre (914)
H. conspectum (320)
H. canovittatum (565)
H. progrediens (1309)
H. ornatiforme (1305)
Pilosella floribunda × officinarum (1325)
A
B
C
D
F I G . 3. Neighbor-joining dendrogram based on a partial analysis of presence/
absence data and the Sörensen/Dice coefficient for 12 nuclear microsatellite
603
ploidy analyses including multiple samples of the same
species failed to detect any variation in ploidy in taxonomically undisputed species (Tyler and Jönsson, 2009). In the present
study, a few cases in which a species contained more than one
plastid haplotype (within the same major haplotype group)
were revealed (Table 3), but for most species too few accessions were analysed to draw any general conclusions.
However, for most of the well-known and relatively widespread species, for which multiple accessions were analysed,
no polymorphisms were detected. This holds for, for
example, H. anodon (ten accessions), H. basifolium (four),
H. cruentifolium (five), H. lindebergii (five), H. plumbeum
(eight) and H. ravidum (eight).
All nine pairs of accessions that were identical across all
microsatellite markers consisted of pairs of accessions belonging to the same species and, in the majority of cases where
two or more accessions of the same species were included,
these clustered together in the NJ analysis (Figs 2 – 5).
However, there were also several exceptions to this pattern and
the outcome heavily depended on how the alleles were coded
(see Materials and methods) and which subset of accessions
was included in the analysis. It also has to be kept in mind that
the bootstrap support for most branches in the NJ was low.
Among the 21 species that were represented by four or more
accessions, no species were found to be monomorphic, and the
percentage of polymorphic bands ranged between 19 % in
H. lepistoides and 86 % in H. anodon, with a mean of 54 %
(to be compared with .95 % polymorphic bands in each of
the better represented sections of the genus). However, since
multiple accessions were only included for species known to
be morphologically variable and/or exceptionally geographically widespread, levels of polymorphism in an average species
may be assumed to be at the lower end of this range. Thus, the
conservative conclusion is that samples of the same morphologically defined species on average were more genetically
similar to each other and tended to cluster together at a higher
frequency than any random samples, but that the present data
are insufficient to convincingly demonstrate whether species
are monotopic or not.
However, ample data now exist that show that considerable
genetic variation is harboured in morphologically almost invariant and presumably obligately apomictic polyploid
species of Hieracium (Shi et al., 1996; Stace et al., 1997;
Štorchova et al., 2002; Rich et al., 2008; this study). Rich
et al. (2008) even found variation among progeny arrays.
This pattern is certainly a product of the deliberate choice of
highly variable and rapidly evolving markers for studies of
variation among closely related species, but how this variation
arises is still unresolved. Punctual mutations are the easy
answer, but levels of variation within Hieracium species
appear unexpectedly high for a completely asexual system
loci, including only members of Hieracium sects. Vulgata and Tridentata
and the outgroup. Samples are referred to by the name of the species and
the accession number (in brackets) and species belonging to H. sect.
Tridentata are indicated by an asterisk. The branches leading to subclusters
comprising species of (A) H. subsimile aggregate, (B) H. plicatum aggregate,
(C) H. anfractum/atronitens aggregate and (D) most of the tetraploid species
(in H. diaphanoides aggregate and H. austrinum aggregate) are indicated by
capital letters.
604
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
H. lepistoides (905)
H. lepistoides (906)
H. lepistoides (907)
H. johanssonii (891)
H. lepistoides (1135)
H. cf. intermarginatum (1441)
H. morulum (930)
H. scioides (1247)
H. integratum (658)
H. integratum (660)
H. philanthrax (968)
H. philanthrax (1160)
H. glandulosissimun (874)
H. variicolor (775)
H. variicolor (858)
H. variicolor (1027)
H. variicolor (1028)
H. variicolor (1252)
H. lanuginosum (901)
H. neoserratifrons (1146)
H. neoserratifrons (689)
H. subterscissum (1250)
H. subterscissum (1440)
H. subterscissum (1453)
H. sarissatum (1364)
H. ciliatiflorum (635)
H. oistophyllum (591)
H. oistophyllum (695)
H. oistophyllum (696)
H. caliginosum (1223)
H. caliginosum (1224)
H. chloromaurum (1225)
H. sparsidens (1400)
H. sparsidens (1406)
H. sp. nov (1355)
H. persimile (960)
H. marginellum (683)
H. marginellum (925)
H. stiptadenium (1439)
H. stiptadenium (1451)
H. hepaticolor (1366)
H. hepaticolor (792)
H. pectinatum (1245)
H. limitianeum (1358)
H. sudermannicum (1191)
H. mallopodum (922)
H. patale (1069)
H. sp. nov (1375)
H. crassiceps (1393)
H. flagriferum (1402)
H. scotostictum (1326)
H. scotostictum ‘Leopard ’(921)
H. scotostictum (1001)
H. jaedrense (1234)
H. haegerstroemii (963)
H. subcrassum (1013)
H. canipes (855)
H. canipes (1098)
H. canipes (1099)
H. cf. stenstroemii (1180)
H. valentius (1024)
H. valentius (1404)
H. stenstroemii (1012)
H. triangulare (1192)
H. sp. nov (1431)
H. koehleri (1398)
H. marginellum (1143)
H. mimeticum (684)
H. integratifrons (1124)
H. baliophyllum (1089)
H. hystrix (377)
H. obtextum (405)
H. philanthrax (1162)
H. tubaticeps (1198)
H. transtrandense (1434)
H. cuprimontanum (1226)
H. dentifolium (1112)
H. pellucidum (1156)
H. grandidens (1036)
H. anfractifolium (599)
H. grandidens (575)
H. ciliatiflorum (318)
H. grandidens (353)
H. albovittatum (1034)
H. pellucidum (881)
H. pellucidum (956)
H. pellucidum (955)
H. grandidens (355)
H. incrassans (1236)
H. incrassans (1237)
H. incrassans (1238)
H. remanens (987)
H. triangulare (1022)
H. triangulare (1249)
H. varicolor (1201)
H. paramaurum (1263)
H. duderhultense (343)
H. cf. koehleri (1241)
H. aethalodes (829)
H. pachyodon (947)
H. hjeltii (882)
H. hjeltii (884)
H. ohlsenii (407)
H. stenocranoides (458)
H. tenebricosum (1145)
H. contaminatum (321)
H. grandifoliatum (1117)
H. strictipes (764)
H. sp. nov (798)
H. albovittatum (1222)
H. stenocranoides (574)
H. basilimbatum (602)
H. sp. nov (1337)
H. grandifoliatum (1232)
H. grandifoliatum (875)
H. myrtillinum (1263)
H. panaeolum (948)
H. panaeolum (1403)
H. ageneium (622)
H. subulatidens (489)
H. subulatidens (492)
H. subulatidens (486)
H. cf. porphyrostictum (1037)
H. crebriserratum (1202)
H. crebriserratum (329)
H. crebriserratum (1339)
H. tenebricosum (1018)
H. tenebricosum (1019)
H. expallidiforme (648)
H. expallidiforme (871)
H. marginelliceps (1140)
H. psilurum (718)
H. torticeps (1196)
H. torticeps (1167)
H. torticeps (1021)
H. obtusius (1243)
H. cf. onychodontum (1369)
H. grandiserratum (582)
H. radiiflorum (606)
H. sparsiguttatum (524)
H. isodontum (1338)
H. orbicans (942)
H. orbicans (943)
H. orbicans (941)
H. sp. nov (1039)
Pilosella floribunda × officinarum (1325)
F I G . 4. Neighbor-joining dendrogram based on a partial analysis of presence/
absence data and the Sörensen/Dice coefficient for 12 nuclear microsatellite loci, including only members of Hieracium sect. Hieracium and the outgroup. Samples are
referred to by the name of the species and the accession number (in brackets).
*
H. crinellum (1204)
H. crinellum (1205)
H. crinellum (1207)
H. crinellum (808)
H. extensiforme (562)
H. extensiforme (732)
H. extensiforme (1321)
H. extensiforme (802)
H. extensiforme (1394)
H. extensiforme (1397)
H. extensiforme (1383)
H. extensiforme (1373)
H. extensiforme (1380)
H. cf. anodon (737)
H. anodon (738)
H. extensiforme (1045)
H. anodon (805)
H. anodon (1044)
H. cf. extensiforme (1329)
H. extensiforme (1042)
H. extensiforme (560)
H. svanlundii (561)
H. lindebergii (1211)
H. lindebergii (733)
H. lindebergii (1367)
*
H. lindebergii (743)
H. lindebergii (1047)
H. anodon (735)
H. saxifragum (532)
H. saxifragum (742)
H. latifrons (409)
H. latifrons (410)
*
H. latifrons (1046)
H. latifrons (739)
H. latifrons (741)
H. anodon (734)
H. anodon (736)
H. extensiforme (1206)
H. anodon (1049)
H. anodon (448)
H. lecanodon (664)
*
H. lecanodon (669)
H. lecanodes (667)
H. cf. anodon (803)
H. extensiforme (449)
H. microdon (97)
H. eulasium (1372)
H. eulasium (1377)
*
H. eulasium (1374)
H. eulasium (1381)
H. extensum (58)
H. lythrodes (1138)
H. lythrodes (1139)
H. lythrodes (521)
*
H. lythrodes (679)
H. lythrodes (678)
H. anodon (740)
H. extensiforme (807)
Pilosella floribunda × officinarum (1325)
F I G . 5. Neighbor-joining dendrogram based on a partial analysis of presence/
absence data and the Sörensen/Dice coefficient for 12 nuclear microsatellite
loci, including only members of Hieracium sect. Oreadea and the outgroup.
Samples are referred to by the name of the species and the accession
number (in brackets). The branches leading to each of the seven subclusters
comprising accessions of a single distinct species, as discussed in the text,
are indicated with asterisks.
(cf. Bengtsson, 2003) and asexual recombination and various
processes that may speed up the mutation rate may be hypothesized (cf. Rich et al., 2008). The adaptive value of such processes in systems incapable of generating variation by means
of sexual recombination may be imagined, but without
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
605
7
6
Haplotypes of group H
Haplotypes of group V
Bifida
5
4
3
Hieracium
PCA 1
2
1
0
–1
–2
–3
Tridentata
–4
Vulgata
–5
–6
–5
–4
–3
–2
–1
0
1
2
3
4
5
6
7
PCA 2
F I G . 6. The distribution of the two main groups of plastid haplotypes among Swedish members of Hieracium sects. Bifida, Hieracium, Vulgata and Tridentata
plotted onto a PCA scatterplot based on 47 independent morphological characters. The current limits of the sections (thick lines) as well as the position of all
species of the southern half of Sweden that have yet not been analysed with respect to chloroplast haplotypes (small black dots) are roughly indicated.
readily comparable data from the same loci in asexually reproduced cell-lines of ‘normal’ sexual species it is difficult to tell
whether or not any such processes are operational in
Hieracium. Nevertheless, the assumption that apomictic
species, even if narrowly defined, are devoid of genetic variation, and therefore incapable of adapting and evolving, obviously does not hold, and should be dismissed.
Patterns of variation among sections, inter-sectional gene flow
and sectional taxonomy
If only triploid species are considered, the basic division of the
plastid haplotypes into two well-separated groups (Fig. 1) corresponds almost perfectly with the current sectional taxonomy. All
typical triploid representatives of H. sects. Vulgata and Oreadea,
and all representatives (irrespective of ploidy) of H. sect.
Tridentata and the diploid H. umbellatum, share haplotypes of
group V, whereas all representatives of H. sects. Bifida and
Hieracium are characterized by haplotypes of group H. In addition, the two representatives of H. sect. Alpina also share the
same, but well-differentiated, haplotype. A basic division of
the plastid haplotypes found in species of Hieracium into two
major groups was previously revealed by Chrtek et al. (2009)
and Fehrer et al. (2009) based on different marker sequences.
Even if the results are difficult to compare due to taxonomic
incongruencies and only marginally overlapping taxon sampling, their ‘Western’ and ‘Eastern’ clades may correspond to
the haplotype groups H and V, respectively, revealed in the
present study. Included in their ‘Western’ clade are the taxa
H. bifidum and H. murorum, roughly corresponding to H.
sects. Bifida and Hieracium in the taxonomy adopted here, and
H. schmidtii and H. stelligerum which plausibly correspond to
the tetrapoid members of H. sect. Oreadea here represented by
H. eulasium, and to H. crinellum, respectively, which in the
present study were all found to have haplotypes of group
H. Similarly, H. laevigatum, roughly corresponding to H. sect.
Tridentata, and H. umbellatum, which are characterized by haplotypes of group V as revealed by the present study, belong to the
‘Eastern’ clade of Fehrer et al. (2009). However, H. lachenalii,
roughly corresponding to H. sect. Vulgata is found in the
‘Western’ clade of Fehrer et al. (2009), but was interpreted as
an ‘interclade hybrid’ by Chrtek et al. (2009), and is dominated
by haplotypes of group V as revealed by the present study. The
present result that tetraploid species, as opposed to the triploid
ones, currently referred to H. sect. Vulgata, contain haplotypes
of group H provides an easy explanation for the discrepancy
between these studies. However, notwithstanding this congruence with previous studies based on a few representative
species, the present result that the haplotype group, almost
without exception, corresponds to the current sectional classification is both new, promising and surprising given the complexity revealed by most other morphological and molecular markers
applied so far.
The obvious correlation between the two major haplotype
groups and the morphology and sectional classification of
members of H. sects. Hieracium, Bifida, Vulgata and
Tridentata is also apparent when plotted onto a PCA scatterplot
based on morphology (Fig. 6). Based on these data, it is tempting
606
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
to suggest a taxonomic recircumscription of H. sect. Vulgata,
such that all tetraploid members are excluded and transferred
to H. sect. Hieracium or H. sect. Bifida. As previously discussed
by Tyler and Jönsson (2009), most or all of these tetraploid
species show some morphological affinities with members of
either H. sect. Hieracium or H. sect. Bifida (cf. also Fig. 6),
making a sectional transfer possible to justify from the morphological point as well. Members of the widespread
H. diaphanoides and H. ornatum species aggregates are tetraploid and show a close resemblance to members of H. sect.
Hieracium with respect to, for example, indumentum characters,
and they all share haplotypes with the latter section. However, as
discussed by Tyler (2006), at least most of these morphologically intermediate species have the ecology and habitat requirements typical of H. sect. Vulgata. Further, the fact that the
present haplotype data allows for no discrimination whatsoever
among the sections characterized by each of the two haplotype
groups, and the lack of support for the sections from the
nuclear markers to be discussed below, indicate that taxonomic
rearrangements based on the single plastid ‘character’ may be
better avoided. However, the sectional placement of these
species ought to be analysed with particular care in future
studies based on different sets of markers.
The pattern revealed by the nuclear microsatellite markers is
far less easy to interpret. With the possible exception of
H. sect. Oreadea, discussed separately below, none of the analyses attempted here revealed any obvious correlation between
the current sectional taxonomy and the patterns of nuclear molecular variation (Figs 2 – 4). As is also evident from the underlying raw data, most alleles are shared among species from
different sections. Not even when the most variable loci
were disregarded (matrix 1b) could any resolution at the sectional level or differentiation between the major groups
revealed by the plastid haplotypes be obtained (not shown).
Whether this homoplasy is the result of an inherently high
rate of synonymous mutation at these loci or of gene flow
among species and sections, or both, is difficult to tell from
these data alone, but the overall high level of variation and
the fact that polymorphisms in most cases were also revealed
within species speaks in favour of the former explanation.
However, the result that not even accessions representing morphologically strongly deviating sections from geographically
remote areas of southern Europe cluster separately from the
northern European sections is unexpected.
In contrast to the congruence observed between the two
main haplotype groups and the morphology-based current sectional classification of the northern European taxa, the distribution of individual haplotypes within the two major
haplotype groups appears largely random. All of the more
common individual haplotypes were found to be shared by
species referred to different sections and to species aggregates
with no apparent morphological connections, and some were
even shared with morphologically strongly deviating
members of southern European sections (Table 3). Given the
low number of polymorphic sites separating individual haplotypes within the major haplotype groups, this may not come as
a surprise. Based on these data it is thus not possible to tell
whether this pattern is the result of homoplasy/recurrent synonymous mutations or of random reticulate evolution/sexual
transfer of plastids between distantly related lineages with
contrasting morphologies. However, since the two main haplotype groups show such close a correlation with morphology
and current sectional taxonomy, sexual transfer of plastids
may, at most, have taken place between species of the same
section. Given the apparent close relationship of the sections
concerned here and the absence of clear morphological or
phenological discontinuities between them (Tyler, 2006) it
becomes difficult to believe that sexual geneflow, if occurring
at all, should be restricted to species of the same section. We
are therefore inclined to believe that transfer of plastids
between morphologically well-differentiated species is rare
or non-existent, and flow of plastids between, for example,
H. sects. Hieracium and Bifida on the one hand and H. sects.
Tridentata and Oreadea on the other has clearly never occurred. However, it should be noted that this line of reasoning
only applies if it is assumed that plastids may be transferred
independently of the morphological characters used to
delimit the sections. If the progeny of sexual crosses in all
cases inherit not only the plastids but also most taxonomically
important characters from the maternal parent, then the resultant hybrid species will be classified taxonomically as a
member of the same section as the maternal parent and the
inter-sectional hybrid origin will remain undetected.
Species aggregates and relationships among individual species
As discussed above, due to the relatively limited number and
possible polytopic origin through homoplasious mutations, the
present plastid haplotype data do not provide much information
concerning the relationships of individual species. In contrast,
since accessions of the same species show a clear tendency to
cluster together based on the data from nuclear microsatellites,
it may be worthwhile also studying the species that appear as
neighbours in these analyses in some detail. Not disregarding
the obvious weaknesses of these data discussed in the context
of the integrity of individual species above, it should be noted
that several, though far from all, of the larger clusters shown in
Figs 2 – 5 partially or fully coincide with the informal species
aggregates suggested by Tyler (2006 and later). In H. sect.
Bifida (Fig. 2), H. sarcophyllum aggregate as represented by
H. sarcophyllum (three accessions), H. itharophyton (two) and
H. gracilifrons (one) is resolved as a separate cluster, and the
basal grade comprising H. oxylepium (one), H. stenolepis
(four), H. plumbeum (seven), H. coadunatum (one) and
H. calliglaucum (one) is made up of species belonging to
H. stenolepis aggregate. The best agreement with current
understanding of the morphological patterns is, however,
found in H. sect. Vulgata (Fig. 3). Here, two accessions of
H. subsimile cluster tightly together with three accessions of
H. resupinatum (both H. subsimile aggregate) which in turn
appear as neighbours to two accessions of the morphologically
somewhat similar H. phaeodermum. The H. plicatum/constringens aggregate as defined by their peculiar involucral indumentum is perfectly recovered in a cluster formed by H. trichelliceps
(three accessions), H. constringens (three), H. plicatum (four),
H. schlyteri (two) and H. xanthostylum (one). Notably, with
the addition of H. expallidiforme which shows the same kind
of involucral indumentum but is currently referred to H. sect.
Hieracium based on other characters, the same cluster is also
recovered in the analysis including all accessions (not shown).
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
This in spite of the fact that H. constringens was found to be the
only tetraploid member of H. sect. Vulgata that shared its haplotype with the triploid members of this section. The main group of
tetraploid species taking an intermediate position between H.
sects. Vulgata and Hieracium and sharing plastid haplotypes
with the latter section, as discussed above (H. diaphanoides/
acidodontum/ornatum aggregate), is also relatively well recovered, and a large cluster is formed by species with affinities to
the morphologically connected H. anfractum aggregate and
H. atronitens aggregate (e.g. H. anfractum, H. punctillaticeps,
H. almquistianum, H. punctatum, H. barbareifolium). In H.
section Hieracium (Fig. 4) there appear to be fewer clusters
that corroborate morphological affinities, but it has previously
also been acknowledged that the pattern of morphological variation in this section is less well structured or understood (Tyler,
2006).
All three sections discussed here have representatives distributed throughout the Scandinavian lowlands but no geographic
patterns, except those constrained by the geographic distribution
of individual species accessions which cluster tightly together,
are reflected in the NJ dendrograms based on the microsatellite
data. As an example, the subcluster comprising members
of the H. plicatum/constringens aggregate discussed above
includes accessions from throughout Sweden, e.g. material of
the widespread H. plicatum from the southern Swedish province
of Småland and the north-eastern-most Swedish province of
Norrbotten. Except for a basal cluster made up of species morphologically similar to H. torticeps, not even representatives
of the ‘park-Hieracia’, i.e. species of H. sect. Hieracium introduced to northern Europe from central Europe as admixture to
grass seed in the late 19th century, cluster together but rather
appear interspersed with native Scandinavian species in the NJ
dendrogram (Fig. 4).
To conclude, even if the present data from nuclear microsatellite markers does not allow for any firm conclusions and may
definitely not be taken as the basis for taxonomic rearrangements, it does corroborate some patterns previously observed
based on careful analysis of a multitude of morphological characters. This observation, in combination with the fact that accessions of the same species, in the majority of cases, cluster tightly
together, clearly shows that morphology and molecular data are
correlated in Hieracium. As a consequence, future studies aimed
at resolving the phylogenetic relationships and natural taxonomy
may benefit from both molecular and traditional morphological
data. Unfortunately, however, the present set of molecular data is
too weak to allow for a meaningful formal analysis of the relative
phylogenetic signal of individual morphological characters (i.e.
the taxonomic significance of the character), an approach that
might otherwise have been of great help to guide future studies
based on morphology, as the relative importance of and the
best way to code these characters constitute the main obstacle
to further advances in that field (Tyler, 2006).
Hieracium sect. Oreadea
Hieracium sect. Oreadea is the only section of the genus that
was reasonably well recovered in the analysis of the complete
dataset of nuclear microsatellites. Ninety-one per cent of the
accessions referable to this section turned up in a separate subcluster which to .90 % consisted of members of this section
607
(not shown). This is somewhat surprising in view of that Tyler
(2011), in the most recent revision of the Swedish representatives of this section, concluded that this section is far more
diverse and heterogenous than the more species-rich but wellknown H. sects. Bifida, Hieracium and Vulgata. With respect
to ploidy level, the section comprises both triploids and tetraploids (Tyler, 2011) and, as was the case for H. sect. Vulgata discussed above, the triploids (except for H. crinellum to be
discussed below) were all found to have plastid haplotypes of
group V (the subtypes V1 and V5) whereas the three teraploid
species examined all have haplotypes of group H. In the analysis
of nuclear microsatellites in H. sect. Oreadea (Fig. 5), accessions of the three tetraploid species H. extensum, H. eulasium
and H. lecanodes appear in two separate but neighbouring subclusters in the lower half of the dendrogram.
Hieracium crinellum, which was the only triploid member
of this section found to have a haplotype of group H, is the
only analysed representative of a morphologically very deviating, though still not well-defined or sufficiently investigated,
group of species that Tyler (2011) suggested might better be
referred to H. sect. Stelligera. After having studied the type
species of that latter section he no longer holds to this suggestion, but H. crinellum remains sufficiently morphologically deviating to be treated as distinct from H. sect. Oreadea. In the
analysis of nuclear microsatellites (Fig. 5), the four accessions
of H. crinellum form a separate subcluster on a relatively long
branch at the very top of the dendrogram, indicating its somewhat deviating position within the section.
The tetraploid Hieracium lecanodes is another peculiar and
morphologically strongly deviating species (Tyler, 2011) and
it has been suggested to have affinities to H. umbellatum or
even be a hybrid of that diploid species. However, the result
that H. lecanodes (two accessions) has an unique haplotype
of group H (haplotype H0, cf. Fig. 1) while both most
members of H. sect. Oreadea and H. umbellatum have haplotypes of group V does not favour that suggestion.
In general, concerning the species-level taxonomy of the
Swedish members of H. sect. Oreadea, Tyler (2011) concluded that the section was made up of a relatively low
number of morphologically distinct and narrowly distributed
species, as well as two or three highly variable, widely distributed, taxonomically enigmatic and plausibly paraphyletic/ancestral ones. This view is strongly supported by the present
analysis of nuclear microsatellites in which two to five accessions of each of H. crinellum, H. lindebergii, H. saxifragum,
H. latifrons, H. lecanodes, H. eulasium and H. lythrodes all
form separate subclusters, while multiple accessions of the
widespread species H. anodon and H. extensiforme form the
grade/matrix from which the clusters formed by the more distinct species branch off (Fig. 5).
Evolutionary processes, hypotheses and prospects for the future
Even if the basic background information needed to formulate stringent hypotheses were considered lacking, one of the
aims of the present study as formulated in the introduction
was to provide clues to the main processes of diversification
and evolution among polyploid and apomictic Hieracium
species. As discussed above, it is now evident from several independent studies that significant genetic variation resides
608
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
within individual apomictic species and they will thus be
capable of evolving further. In addition, even if further
studies are clearly needed to demonstrate this, there exist no
data indicating that the morphologically defined species are
generally polytopic or random assemblages of superficially
similar clones or ramets. In contrast, taxa defined based on
morphology appear clearly correlated to patterns of molecular
variation, and the variation found within species may possibly
be the result of accumulation of mutations alone. However, the
fundamental question of how speciation proceeds in this
system remains unanswered. Plastid haplotype data appear to
indicate that sexual gene flow is rare or absent, at least
between members of the different sections, but the large and
mainly non-hierarchical pattern of variation revealed by the
nuclear markers may be interpreted as indicative of extensive
gene flow between species and groups of species (if not
simply an artefact of the choice of marker as discussed
above). Fehrer et al. (2009) concluded that hybridization
appears to have played a very important role in the evolution
of the genus in the distant past but was unable to find any
traces of ongoing reticulation. This conclusion appears fully
concordant with the data of the present study, but is unable
to explain how such a huge number of locally endemic
species have ended up in recently deglaciated areas and does
not explain the evolution of the individual extant species.
In this context, the tetraploid members of H. sects. Vulgata
and Oreadea are of particular interest. It has previously been
suggested, based on morphology, that these tetraploids may
have evolved from sexual crosses between triploid members of
these sections and H. sect. Hieracium (Tyler and Jönsson,
2009) and the present finding that all but one of these tetraploid
and morphologically intermediate species have plastids characteristic of the latter section further strengthens the idea of a
hybrid origin, with members of H. sect. Hieracium acting as
the seed parent. Most of these species also cluster together in
the analysis of nuclear markers, indicating some sort of
common ancestry or relatedness. The latter may be taken as an
indication that they are all descendants of the same hybridization
event, or at least have evolved from crosses between the same
pair of taxa, but the fact that they contain different plastid haplotypes does not favour that hypothesis. Nevertheless, understanding the evolution of these tetraploids may present the key to a
better understanding of speciation and evolutionary processes
in this system and should thus be the focus of future studies.
However, the presently available molecular markers may
not appear sufficient tools to gain a better understanding of
these processes. The variation in the psbD-trnT sequence,
and that in the ETS and EPS sequences applied by Chrtek
et al. (2009) and Fehrer et al. (2009), is apparently too
limited to resolve the relationships of the species beyond the
basic split of the genus into two major groups that now
appears well established. This is in spite of the fact that
these sequences have been selected as the ones showing
most variation among a relatively large set of plastid sequences
tested (J. Fehrer, Academy of Sciences of the Czech Republic,
pers. comm.; the present study). In contrast, the nuclear microsatellites applied here are clearly too polymorphic and rapidly
evolving to be suitable for studies at this taxonomic scale. The
fact that even representatives of morphologically and geographically distant sections shared most of the alleles indicates
that the mode of evolution of these markers makes them unsuitable for studies at the sectional level and above.
Microsatellite markers with longer themes that are less
rapidly evolving may be developed, but the resolving power
of the present markers may also be better utilized by focusing
on groups of taxa believed to be closely related. Such systems
may be found on relatively isolated and recently emerged or
deglaciated islands. Ideally, both local endemics likely to
have evolved in situ and more widespread species that may
be hypothesized to be their progenitors should be present in
such a system, and the number of species should be appropriately small. If such a system is found and thoroughly investigated, the present molecular markers in combination with
morphology may be sufficient to gain a full understanding of
how speciation proceeds and how variation is obtained and
structured. A better understanding of these processes in
Hieracium will be important not only for researchers in this
huge plant genus but would also be beneficial for our understanding of evolutionary processes in asexual systems in
general and for our understanding of the evolutionary value
of sex as such (cf. Bengtsson, 2003).
ACK N OW L E DG E M E N T S
This study was financed by a grant to the first author from the
Swedish Taxonomy Initiative. Mikael Hedrén is thanked for
fruitful discussions about how to interpret the microsatellite
electropherograms. Sofie Nordström, Maren Wellenreuther
and Bengt Hansson are all thanked for invaluable help in the
laboratory.
L I T E R AT U R E C I T E D
Asker SE, Jerling L. 1992. Apomixis in plants. Boca Raton, FL: CRC.
Bengtsson BO. 2003. Genetic variation in organisms with sexual and asexual
reproduction. Journal of Evolutionary Biology 16: 189–199.
Carman JG. 1997. Asynchronous expression of duplicate genes in angiosperms may cause apomixix, bispory, tetraspory, and polyemryony.
Biological Journal of the Linnean Society 61: 51–94.
Chapman H, Robson B, Pearson ML. 2004. Population genetic structure of a
colonising, triploid weed, Hieracium lepidulum. Heredity 92: 182–188.
Chiang T-Y, Schaal BA, Peng C-I. 1998. Universal primers for amplification
and sequencing a noncoding spacer between the atpB and rbcL genes of
chloroplast DNA. Botanical Bulletin of Academia Sinica 39: 245 –250.
Chrtek J Jr., Zahradniček J, Krak K, Fehrer J. 2009. Genome size in
Hieracium subgen. Hieracium (Compositae) strongly correlates with
major phylogenetic groups. Annals of Botany 104: 161– 178.
Doyle JJ, Doyle JH. 1990. Isolation of plant DNA from fresh tissue. Focus
12: 13– 15.
Fehrer J, Gemeinholzer B, Chrtek J Jr., Bräutigam S. 2007. Incongruent
plastid and nuclear DNA phylogenies reveal ancient intergeneric hybridization in Pilosella hawkweeds (Hieracium, Cichorieae, Asteraceae).
Molecular Phylogenetics and Evolution 42: 347 –361.
Fehrer J, Krak K, Chrtek J Jr. 2009. Intra-individual polymorphism in
diploid and apomictic polyploid hawkweeds (Hieracium, Lactuceae,
Asteraceae): disentangling phylogenetic signal, reticulation, and noise.
BMC Evolutionary Biology 9: 239. http://dx.doi.org/10.1186/1471-21489-239.
Feulner M, Schuhwerk F, Dötterl S. 2011. Taxonomical value of inflorescence scent in Hieracium s. str. Biochemical Systematics and Ecology
39: 732 –743.
Ford CS, Ayres KL, Toomney N, et al. 2009. Selection of candidate coding
DNA barcoding regions for use in land plants. Botanical Journal of the
Linnean Society 159: 1 –11.
Tyler & Jönsson — Plastid and nuclear variation among apomictic polyploids
Gaskin JF, Wilson LM. 2007. Phylogenetic relationships among native and
naturalized Hieracium (Asteraceae) in Canada and the United States
based on plastid DNA sequences. Systematic Botany 32: 478–485.
Hammer Ø, Harper DAT, Ryan PD. 2001. PAST: Paleontological Statistics
Software Package for Education and Data Analysis. Palaeontologia
Electronica 4: 4. http://palaeo-electronica.org/2001_1/past/issue1_01.htm.
Hörandl E. 2006. The complex causality of geographical parthenogenesis.
New Phytologist 171: 525–538.
Hörandl E, Paun O. 2007. Patterns and sources of genetic diversity in apomictic plants: implications for evolutionary potentials. In: Hörandl E,
Grossniklaus U, van Dijk PJ, Sharbel TF. eds. Apomixis: evolution,
mechanisms and perspectives. Ruggell, Liechtenstein: A.R.G. Gantner
Verlag, 170 –194.
Johansson K. 1923. Växtgeografiska spörsmål rörande den Svenska
Hieracium-floran. Svensk Botanisk Tidskrift 17: 116–213.
Johansson K. 1927. Enumerantur Hieracia vulgata Suecicae. Arkiv för
Botanik 21A n. 15.
Juel O. 1898. Parthenogenesis bei Antennaria alpina (L.) R.Br. Botanischer
Centralblatt 74: 369–372.
Jönsson J, Wellenreuther M, Tyler T. 2010. Ten polymorphic microsatellite
markers for Hieracium s.s. (Asteraceae). Conservation Genetics
Resources 2: 295– 300.
Karlsson T. 1998. Förteckning över svenska kärlväxter. Svensk Botanisk
Tidskrift 91: 241– 560.
Krahulcova A, Krahulec F, Chapman HM. 2000. Variation in Hieracium
subgen. Pilosella (Asteraceae): what do we know about its sources?
Folia Geobotanica 35: 319– 338.
Kress WJ, Wurdack KJ, Zimmer EA, Weight LA, Janzen DH. 2005. Use
of DNA barcodes to identify flowering plants. Proceedings of the
National Academy of Sciences of the USA 102: 8369–8374.
Mogie M, Britton NF, Steward-Cox JA. 2007. Asexuality, polyploidy and
the male function. In: Hörandl E, Grossniklaus U, van Dijk PJ, Sharbel
TF. eds. Apomixis: evolution, mechanisms and perspectives. Ruggell,
Liechtenstein: A.R.G. Gantner Verlag, 196–214.
Mráz P. 2003. Mentor effects in the genus Hieracium s.str. (Compositae,
Lactuceae). Folia Geobotanica 38: 345–350.
Mráz P, Paule J. 2006. Experimental hybridization in the genus Hieracium
s.str.: crosses between diploid taxa. Preslia 78: 1– 26.
Mráz P, Chrtek J, Kirschner J. 2001. Genetic variation in the Hieracium
rohacsense Group (Hieracium sect. Alpina). Phyton 41: 269–276.
Mráz P, Chrtek J Jr., Fehrer J, Plačkova I. 2005. Rare recent natural hybridization in Hieracium s.str.: evidence from morphology, allozymes and
chloroplast DNA. Plant Systematics and Evolution 255:177–192.
Mráz P, Chrtek J, Fehrer J. 2011. Interspecific hybridization in the genus
Hieracium s. str.: evidence for bidirectional gene flow and spontaneous
allopolyploidization. Plant Systematics and Evolution 293: 237–245.
Nogler GA. 2007. The discovery of parthenogenesis: a long journey to the
truth. In Hörandl E, Grossniklaus U, van Dijk PJ, Sharbel TF. eds.
Apomixis: evolution, mechanisms and perspectives. Ruggell, Liechtenstein:
A.R.G. Gantner Verlag, 25–35.
Rich TCG, McDonnell EJ, Lledó MD. 2008. Hieracium cyathis (Asteraceae)
and its relation to other apomictic taxa. Botanical Journal of the Linnean
Society 156: 669–680.
Riska B. 1989. Composite traits, selection responce, and evolution. Evolution
43: 1172–1191.
Samuelsson G. 1954. Maps of a selection of Scandinavian Hieracium species.
Kungliga Svenska Vetenskaps-Akademiens Handlingar 5 n. 3.
609
Schuhwerk F. 2002. Some thoughts on the taxonomy of Hieracium. Berichte
der Bayeriche Botaniche Gessellschaft 72: 193–198.
Sell PD. 1987. An introduction to the study of the British Hieracia. 1. History
and classification. Watsonia 16: 365– 371.
Sell PD, Murrell G. 2006. Flora of Great Britain and Ireland, Vol. 4.
Cambridge: Cambridge University Press.
Shaw J, Lickey EB, Schilling EE, Small RL. 2007. Comparison of whole
chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: the tortoise and the hare III. American
Journal of Botany 94: 275– 288.
Shi Y, Gornall RJ, Draper J, Stace CA. 1996. Intraspecific molecular variation in Hieracium sect. Alpina (Asteraceae), an apomictic group. Folia
Geobotanica et Phytotaxonomica 31: 305–313.
Slade K, Rich TCG. 2007. Pollen studies in British Hieracium sect. Alpina
(Asteraceae). Watsonia 26: 443– 450.
Smith J. 1841. Notice of a plant which produces perfect seeds without any
apparent action of pollen. Transactions of the Linnaean Society London
18: 509– 512.
Stace CA. 1998. Sectional names in the genus Hieracium (Asteraceae) sensu
stricto. Edinburgh Journal of Botany 55: 417– 441.
Stace CA, Gornall RJ, Shi Y. 1997. Cytological and molecular variation in
apomictic Hieracium sect. Alpina. Opera Botanica 132: 39–51.
Stebbins GL. 1950. Variation and evolution in plants. New York, NY:
Columbia University Press.
Štorchova H, Chrtek J, Bartish IV, Tetera M, Kirschner J, Stepanek J.
2002. Genetic variation in agamospermous taxa of Hieracium sect.
Alpina (Compositae) in the Tatry Mts. (Slovakia). Plant Systematics
and Evolution 235: 1 –17.
Thomas S, Bailey JP, Rich TCG. 2011. Pollen and chromosome studies in
Hieracium sect. Alpestria (Asteraceae). Nordic Journal of Botany 29:
244–248.
Tyler T. 2000. Detecting migration routes and barriers by examining the distribution of species in an apomictic species complex. Journal of
Biogeography 27: 979–988.
Tyler T. 2006. Patterns of morphometric variation and a new supraspecific
classification of apomictic taxa of Hieracium (Asteraceae) from
Denmark and southern Sweden. Plant Systematics and Evolution 261:
39–88.
Tyler T. 2010. Dalarnas skogs- hag- och krattfibblor. [Hieracium sect.
Hieracium, sect. Vulgata and sect. Bifida in the province of Dalarna].
Trollius 40: 1 –38 (+ 107 pp. supplementary material available at www.
dalafloran.se/fibblor_index.htm).
Tyler T. 2011. Hieracium sect. Oreadea (Asteraceae) in Sweden: from a complete mess to a preliminary taxonomic classification. Nordic Journal of
Botany 29: 538– 589.
Tyler T, Bertilsson A. 2009. Hieracum (Asteraceae) of subalpine Dalarna,
Sweden, revisited: decline in species richness attributable to both forestry
and overgrowth. Nordic Journal of Botany 27: 419–424.
Tyler T, Jönsson J. 2009. Ploidy analysis of apomicic Hieracium (Asteraceae)
reveal unexpected patterns and variation. Nordic Journal of Botany 27:
490–502.
Whitton J, Sears CJ, Baack SJ, Otto SP. 2008. The dynamic nature of apomixis in the angiosperms. International Journal of Plant Sciences 169:
169–182.
Zahn KH. 1921– 23. Compositae: Hieracium. In: Engler A. ed. Das
Pflanzenreich. Leipzig: W. Engelmann, IV/280 (1 –1705).