Biosystematic study of the Cyanus triumfetti group in Central Europe
Biosystematická štúdia skupiny Cyanus triumfetti v strednej Európe
Katarína O l š a v s k á, Marián P e r n ý, Jaromír K u č e r a & Iva H o d á l o v á
Institute of Botany, Slovak Academy of Sciences, Dúbravská cesta 9, SK-845 23 Bratislava,
Slovak Republic, e-mail: firstname.lastname@example.org, email@example.com,
jaromir.kucera @savba.sk, firstname.lastname@example.org
Olšavská K., Perný M., Kučera J. & Hodálová I. (2011): Biosystematic study of the Cyanus
triumfetti group in Central Europe. – Preslia 83: 59–98.
Multivariate morphometrics and an assessment of genetic diversity obtained using amplified frag
ment length polymorphism (AFLP) were used to determine the variability of the polymorphic group
Cyanus triumfetti in Central Europe. The ploidy level of the populations studied was also deter
mined; all individuals from the C. triumfetti group were diploid (2n ~ 2x ~ 22) and all those of the
related C. montanus group were tetraploid (2n ~ 4x ~ 44). A multivariate morphometric study of 71
populations revealed that three species from the C. triumfetti group occur in Central Europe, namely
‘Cyanus axillaris’, C. strictus and C. dominii. Three subspecies are recognized within the latter spe
cies, namely C. dominii subsp. dominii, C. dominii subsp. slovenicus and C. dominii subsp.
sokolensis. Morphological characters of leaves are the best features for delimiting these taxa; a shift
in characters caused by cultivation did not affect the value of key characters and differences among
the taxa remained. AFLP analysis of 38 populations from the C. triumfetti group and two from the
C. montanus group revealed a contrasting pattern of genetic variation that was related to the geo-
graphic distribution of the populations rather than the morphological variation in the C. triumfetti
group. The AFLP data revealed the following three genetically differentiated and allopatric groups:
(i) C. triumfetti s.s. and C. montanus from the Western Alps, (ii) ‘C. axillaris’ from Austria and the
Czech Republic (except the Carpathians) and (iii) ‘C. axillaris’, C. strictus and C. dominii from the
Western Carpathians and Pannonia. The striking genetic gapbetween the Austrian-Czech and the W
Carpathian-Pannonia groups and the high genetic diversity and weak genetic differentiation within
the latter group are discussed in the light of potential glacial refuges, postglacial migration routes
and/or the probability of hybridization events occurring during the evolutionary history of this
group. An identification key for the taxa of the C. triumfetti group in Central Europe is presented.
Keywords:Asteraceae, Cyanus sect. Perennes, flow cytometry, genetic variation, morphology,
Pannonia, taxonomy, Western Carpathians
The Cyanus triumfetti group is one of the many species-rich and taxonomically compli
cated aggregates of perennial knapweeds in the subtribe Centaureinae (Cass.) Dumort.
There are two taxonomic opinions about the classification of Cyanus taxa: first, they
should be included in Centaurea s.s. (e.g. Susanna & Garcia-Jacas 2007) or secondly,
placed in a separate genus, Cyanus Mill. (e.g. Greuter 2006–2009). The genus Cyanus
comprises 25 (Hellwig 2004) to 50 (Dostál 1969) species classified into two sections:
Cyanus sect. Cyanus and C. sect. Perennes (Boiss.) (valid combination at the sectional
level in the genus Cyanus is not available), differing in pollen type, basic chromosome
number and life form. Cyanus sect. Cyanus consists of annual species with pollen type
‘Cyanus’ and basic chromosome numbers of x = 8, 9 or 12. Cyanus sect. Perennes consists
Preslia 83: 59–98, 2011 59
of perennial species with pollen type ‘Montana’ and basic chromosome number of x = 10
or 11. The molecular, karyological and morphological evidence indicate that C. sect.
Perennes is monophyletic and originated from annual species of C. sect. Cyanus (Garcia-
Jacas et al. 2001, Hellwig 2004, Löser et al. 2009).
Within C. sect. Perennes, the three following unranked groups of morphologically sim
ilar taxa are proposed in the taxonomic literature: the C. triumfetti group, the C. montanus
group and the C. napulifer group (cf. Dostál 1931b, 1976b, Stefanov & Georgiev 1931,
Borhidi 1957, Bancheva & Raimondo 2003, Greuter 2006–2009). However, these three
groups have not yet been delimited using comparative biosystematic methods and several
species of uncertain taxonomic position remain in this section.
The C. triumfetti group was first defined by Borhidi (1957), who included six species in
this group, which in terms of the nomenclature used in the present paper correspond to:
C. triumfetti (All.) Dostál ex Á. Löve & D. Löve, C. graminifolius (Lam.) Olšavská,
C. baldaccii (Bald.) Holub, C. pindicola (Griseb.) Soják, C. achtarovii (Urum.) Holub and
C. pinnatifidus (Schur) Holub. An additional 10 infraspecific taxa, which were included in
the C. triumfetti group by Borhidi (1957), have been recently recognized as separate spe
cies and in terms of the nomenclature used in the present paper are listed as follows:
C. adscendens (Bartl.) Soják; C. albofimbriatus (Stef. & T. Georgiev) Greuter; C. ange-
lescui (Grinţ.) Holub; C. axillaris auct.; C. bourgaei (Boiss.) Wagenitz & Greuter;
C. dominii (Dostál) Holub; C. fuscomarginatus (K. Koch) Greuter; C. lingulatus (Lag.)
Holub; C. matthiolifolius (Boiss.) Wagenitz & Greuter; and C. strictus (Waldst. & Kit.)
Soják (Czerepanov 1995, Greuter 2006–2009, this paper). The taxa of the C. triumfetti
group differ from the remaining perennial species of C. sect. Perennes as they lack thick-
ened roots, in which they differ from the C. napulifer group, and have pale fimbriae, which
are longer than the width of the margin of the appendages of involucral bracts, in which
they differ from the C. montanus group (cf. Dostál 1976b, Bancheva & Raimondo 2003,
Štěpánek 2004). Taxa of the C. triumfetti group occur from Morocco across Europe
(except the northern parts) to Iran (Greuter 2006–2009). Several taxa are widespread
while others are narrow endemics restricted to certain geographical areas. The ecological
differentiation among the taxa is substantial and they grow on sunny steppes or in rocky
habitats, ranging from lowland to the subalpine belt.
The taxonomic treatment of the C. triumfetti group in floras and identification keys is
often inconsistent and the delimitation of the taxa within the group is difficult because of
their weak morphological differentiation. Moreover, intermediate morphotypes among
some taxa and considerable intra-population variability are also reported (Millionová
2000, Olšavská et al. 2009). Such a pattern of morphological variation could be the result
of a short time of speciation and/or weak reproductive isolation within the C. triumfetti
group. Similar taxonomic complexity due to extreme morphological variation and a high
degree of phenotypic plasticity also exists in other polymorphic groups in the Compositae,
e.g. Senecio nemorensis agg. (Hodálová 1999), Doronicum L. (Álvarez Fernández &
Nieto Feliner 2001), Centaurea phrygia agg. (Koutecký 2007), Picris hieracioides L.
(Slovák & Marhold 2007) and Centaurea stoebe L. (Španiel et al. 2008).
Diploids and tetraploids with two basic chromosome numbers (x = 10, x = 11) are
reported for taxa of the C. triumfetti group (Dostál 1976b, Goldblatt & Johnson
1979–2010, Olšavská & Perný 2009). Ongoing karyological analyses of European popula
tions (Olšavská et al. 2009, K. Olšavská, unpublished) and the majority of previous studies
60 Preslia 83: 59–98, 2011
on chromosome numbers (ca 75 reports; Goldblatt & Johnson 1979–2010, Millionová
2000, Bancheva & Greilhuber 2006, Marhold et al. 2007) report diploids with a basic
chromosome number of x = 11 (2n = 2x = 22) for the C. triumfetti group. In addition to
diploid chromosome numbers, there are four tetraploid counts with x = 11 (2n = 4x = 44)
for the C. triumfetti group from southern Europe (Guinochet 1957, Lovrić 1982, Sharkova
1996). One of these tetraploid counts is for Cyanus graminifolius (Guinochet 1957, as for
Centaurea triumfetti var. seusana Gugler). This count was not confirmed by the current
analyses of plants from the same locality and the chromosome number corrected (x = 10;
2n = 4x = 40) indicates that C. graminifolius occupies a separate position within the
C. triumfetti group (Olšavská & Perný 2009). Diploid counts with both the basic chromo
some numbers x = 10 and x = 11 (2n = 2x = 20, 2n = 2x = 22) are reported for C. lingulatus
and C. pindicola, Mediterranean taxa from the C. triumfetti group (Morales 1974, Strid &
Franzen 1981, Montserrat Martí 1987, Baltisberger 1991). However, further studies on
chromosome numbers, with special emphasis on determining the basic chromosome num
bers, are needed to bring new insights to the relationships among perennial taxa of the
Four taxa from the C. triumfetti group, treated by various authors either as species or
subspecies or varieties, are reported from Central Europe; they are listed as follows:
C. axillaris auct., C. dominii, C. strictus and C. triumfetti s.s. (Dostál 1989, Marhold &
Hindák 1998, Štěpánek 2004, Olšavská et al. 2009). However, a recent multivariate
morphometric analysis revealed that C. triumfetti s.s. does not occur in this area (Olšavská
et al. 2009).
Cyanus axillaris auct. is the most widespread taxon of the C. triumfetti group occurring
almost throughout the whole distribution range of the group (except for the southernmost
parts of Europe and Western Asia; Hellwig 2004, Greuter 2006–2009). Cyanus axillaris
auct. is characterized by decurrent leaves, which are broadly lanceolate on the lower part
of stem and non-silver fimbriae on the appendages of the involucral bracts. Because the
nomenclature of this taxon is unresolved (see Nomenclature notes) it is hereinafter
referred to as ‘C. axillaris’.
Cyanus strictus is morphologically similar to ‘C. axillaris’ and it is difficult to distin
guish between these two species (Štepánek 2004). Cyanus strictus is characterized by long
erect stems branching mainly in the upper part and narrowly lanceolate leaves with a nar
row and long decurrent part, and silver fimbriae that are twice as long as the width of the
appendage’s margin (Waldstein & Kitaibel 1805, Dostál 1976b). This species was
described from an area on the border between Slovakia and Hungary (Zempléni-hegység
Mts; Waldstein & Kitaibel 1805) and is reported from other countries in Central, eastern
and southern Europe (Czech Republic, Croatia, Montenegro, Romania, Bulgaria and
Ukraine; Horváth et al. 1995, Marhold & Hindák 1998, Mosyakin & Fedoronchuk 1999,
Nikolič 2000, Oprea 2005, Greuter 2006–2009). Outside Central Europe, the two taxa are
considered to be closely related to Cyanus strictus: taxon from western Ukraine originally
described as Centaurea ternopoliensis Dobrocz, and Cyanus angulescui from south-east
ern Romania and Moldavia (Dobrochayeva 1949, Dostál 1976b, Mosyakin & Fedoron
chuk 1999, Ciocârlan 2000). However, their taxonomic relationships need to be studied in
Cyanus dominii was described by Dostál (1931a) as Centaurea triumfetti subsp.
dominii Dostál from the Western Carpathians. This species can be clearly distinguished
Olšavská et al.: Cyanus triumfetii group in Central Europe 61
from ‘C. axillaris’ and C. strictus by its non-decurrent or very short-decurrent stem leaves
that are all approximately the same size (Dostál 1976b). There are different opinions on
the distribution of this species depending on its treatment in current Floras and other
works. Cyanus dominii is either considered to be endemic to the Western Carpathians
(Dostál 1989, Marhold and Hindák 1998, Kliment 1999) or to have a wider distribution,
including Bulgaria and Ukraine (Dostál 1976b, Andreev et al. 1992, Mosyakin &
Fedoronchuk 1999, Greuter 2006–2009).
Five varieties are described within Centaurea triumfetti subsp. dominii by Dostál
(1931a, b). However, these taxa are not mentioned in his later publications (Dostál 1950,
1976b, 1989) and only a few authors recognized them (e.g. Borhidi 1957). These five vari
eties are described as follows: (i) The nominal variety ‘Centaurea triumfetti subsp. dominii
var. eu-dominii’ described from Branisko Mts (Slovakia) is characterized by small cylin
drical involucre and narrowly lanceolate or linear rigid leaves with a revolute margin and
a glabrous upper surface and white-tomentose lower surface. (ii) Centaurea triumfetti
subsp. dominii var. slovenica Dostál described from the Western Carpathians has oblong
or lanceolate leaves that are tomentose both above and beneath. (iii) Centaurea triumfetti
subsp. dominii var. densifolius Dostál was described from Slovenský kras Karst and is
reported also from the Západné Tatry Mts in Slovakia. This variety is characterized by
tomentose linear-lanceolate or linear leaves, a short pedunculus and a tiny capitulum. Pop-
ulations from the Západné Tatry Mts were formerly described as Centaurea axillaris f.
sokolensis Pawłowski (Pawłowski 1930). (iv) Centaurea triumfetti subsp. dominii var.
romanica Dostál was described from the Domogled Mts in Romania (Dostál 1931b). Typ-
ical characters of this taxon are oblong to ovate, subglabrous leaves with serrate margins
and stem foliose up to the capitulum. This variety is no longer reported from Romania
(Oprea 2005). (v) Centaurea triumfetti subsp. dominii var. perfoliosa Dostál was
described from the Čornohora Mts in Ukraine and is very similar to Centaurea triumfetti
subsp. domini var. romanica, but differs in having tomentose leaves according to Dostál’s
description (Dostál 1931a).
Our previous analysis revealed complex morphological variation in the Cyanus
triumfetti group in the Western Carpathians and a need for further taxonomic studies was
emphasized (Olšavská et al. 2009). The aim of this study is to determine the morphologi
cal variation and relationships of Central European populations of the C. triumfetti group
using morphometric analyses based on field-collected plants and plants grown in an
experimental garden. The ploidy level of all populations studied was determined using
flow cytometry because taxonomic conclusions based on multivariate morphometrics are
often supported by ploidy level analysis (e.g. Suda & Lysák 2001, Marhold et al. 2005,
Ekrt & Štech 2008). Morphometric and karyological analyses were supplemented by
a pilot study of the genetic variability using amplified fragment length polymorphism
(AFLP) markers. AFLP is a powerful method for assessing inter- and intra-specific
genetic variation, especially in species groups that are closely related and/or have diverged
recently (Bussell et al. 2005, Meudt & Clarke 2007). An approach combining AFLP and
morphometric data has been successfully applied to a range of polymorphic plant groups,
e.g. Cardamine L. (Lihová et al. 2004, Perný et al. 2004), Festuca ser. Psammophilae
Pawlus (Šmarda et al. 2007), Viola suavis M. Bieb. (Mereďa et al. 2008), Papaver alpinum
L. s.l. (Schönswetter et al. 2009) and Prunus L. sect. Prunus (Depypere et al. 2009).
In order to confirm that C. triumfetti s.s. does not occur in the Western Carpathians and
62 Preslia 83: 59–98, 2011
adjacent parts of Pannonia, which is indicated by a previous morphometric study
(Olšavská et al. 2009), samples of this species from the Western Alps were included in the
The present study is a part of a broader investigation that aims to fill the gap in our
knowledge about the taxonomy of the perennial taxa of the genus Cyanus. There are only
a few biosystematic studies on the morphological and karyological diversity of Cyanus
taxa (Millionová 2000, Bancheva and Raimondo 2003, Bancheva & Greilhuber 2006,
Olšavská et al. 2009, Olšavská & Perný 2009). In addition, in the two preliminary studies
of Cyanus, cpDNA and nrDNA sequences or isozymes were used (Löser et al. 2009,
Bancheva et al. 2009; respectively) and some molecular phylogenetic studies on the
subtribe Centaureinae using ITS and cpDNA were based on results for only one or two
individuals of Cyanus sect. Perennes (Sussanna et al. 1995, Garcia-Jacas et al. 2001,
The specific questions addressed in this study are: (i) Which of the taxa from the
C. triumfetti group occur in Central Europe? (ii) Which of the characters used in
morphometric analyses can be used to separate the revealed taxa and which are not
strongly affected by environmental conditions? (iii) Is the AFLP fingerprinting pattern
related to the morphological variation of populations from the C. triumfetti group? (iv) Do
molecular analyses confirm the results of previous morphological analyses (Olšavská et
al. 2009), which indicate that C. triumfetti s.s. does not occur in Central Europe?
Material and methods
Between 2005 and 2009, five to 20 plants were collected from 71 populations of the
C. triumfetti group in Central Europe. In addition, samples from three populations of
C. triumfetti s.s. and two populations of C. montanus from the Western Alps were col-
lected for AFLP analyses (Table 1, Fig. 1). The number of plants collected per locality
depended on their abundance. When plants grew in clusters (i.e. each cluster corresponds
to a single genet), only one stem from a cluster was collected. Roots/rhizomes from 5–10
flowering plants (the same plants used for morphometric analyses of field-collected popu
lation samples) per population were transferred from the field and cultivated under similar
environmental conditions in an experimental garden at Višňové (49°09'40''N, 18°47'13''E,
470 m a.s.l., Žilina district, northwestern Slovakia). Herbarium specimens from plants
cultivated in an experimental garden were used to examine the genetically determined
variation in morphological traits. Fresh leaf material from cultivated plants was used for
flow cytometric analyses and leaves from selected cultivated plants were dried and pre
served in silica gel for AFLP analyses. Consequently, all analyses were conducted on the
same individuals. The aim of the AFLP analyses was to elucidate genetic variation over
a wide distribution range of all taxa rather than to assess intrapopulation variation of a few
populations. Therefore, one or two individuals from selected populations of the
C. triumfetti group were chosen so as to cover the whole study area. Selected samples of
C. montanus were used as the outgroup in the AFLP analyses because it is closely related
to the C. triumfetti group.
Olšavská et al.: Cyanus triumfetii group in Central Europe 63
Table 1. – Samples from the Cyanus triumfetti and C. montanus groups used for morphometric analyses (Mor
F/G: number of field-collected/cultivated plants), karyological analyses (Kar 2x/2n: number of plants used for the
measurement of the DNA ploidy level/for chromosome counting; * published in a previous study, Olšavská et al.
2009) and AFLP analyses (AFLP). Collection data are as follows: population code, country (CZ – Czech Repub
lic, F – France, HU – Hungary, IT – Italy, SK – Slovakia); description of locality; geographic coordinates
(WGS84); altitude; date of collection; name of collector (DD – D. Dítě, IH – I. Hodálová, PE – P. Eliáš Jr., KO –
K. Olšavská, MO – M. Olšavský, MP – M. Perný, MV – M. Valachovič, RŠ – R. Šuvada, VK – V. Kolarčik).
Collection data Mor
TRI 1 SK; Devínska Kobyla Mts, Mt. Sandberg near Devínska Nová Ves;
48°11'55" N, 16°59'00" E; 340 m; 2 June 2005; KO
20/– –/– –
TRI 2 SK; Malé Karpaty Mts, S slope of Mt. Pohánska; 48°29'00" N, 17°15'55"
E; 320 m; 3 June 2005; KO, IH & MP
20/– –/– –
TRI 4 SK; Považský Inovec Mts, S slope below Tematín Castle; 48°40'15" N,
17°57'45" E; 580 m; 17 June 2005; KO & MO
20/– 6*/1* –
TRI 7 SK; Podunajská nížina Lowlands, SW slope of Mt. Zobor near Nitra;
48°20'49" N, 18°05'40" E; 362 m; 9 June 2006; KO & MP
20/4 3*/– 2
TRI 8 SK; Podunajská nížina Lowlands, Mt. Šipka near Plášťovce; 48°09'58" N,
18°59'47" E; 315 m; 10 June 2006; KO & MP
20/3 3*/– 1
TRI 16 SK; Považský Inovec Mts, Mt. Kamienka near Modrová; 48°38'37" N,
17°54'00" E; 245 m; 27 June 2006; MP
8/4 5*/4* 1
TRI 23 SK; Východoslovenská rovina Lowland, Mt. Veľký vrch near Brehov;
48°29'36" N, 21°48'35" E; 240 m; 5 July 2006; KO & MO
20/7 5*/– –
TRI 34 SK; Belianske kopce Mts, Mt. Veľký vrch near Štúrovo; 47°49'45" N,
18°37'35” E; 240 m; 18 May 2007; KO & MO
15/3 5*/– –
TRI 41 AU; Hainburger berge Mts, SW slope of Mt. Braunsberg; 48°09'15" N,
16°57'07” E; 291 m; 24 May 2007; KO & MO
17/6 6*/– –
TRI 42 AU; Wienerwald Mts, Rauhenstein Castle near Baden; 48°00'46" N,
16°12'27” E; 324 m; 24 May 2007; KO & MO
14/6 8*/– 2
TRI 43 AU; Drosendorf Stadt, view point 1.5 km N of the town; 48°52'27" N,
15°37'52” E; 434 m; 25 May 2007; KO & MO
20/8 6*/– 2
TRI 45 CZ; Moravská vrchovina Mts, rocks NE of Moravský Krumlov;
49°03'39" N, 16°19'40” E; 283 m; 26 May 2007; KO & MO
18/3 5*/– –
TRI 46 CZ; Moravská vrchovina Mts, rocks near Ivančice; 49°05'20" N,
16°20'25” E; 258 m; 26 May 2007; KO & MO
20/7 7*/– –
TRI 47 CZ; Pavlovské vrchy Mts, Mt. Děvín near Horní Věstonice; 48°51'53" N,
16°38'37” E; 424 m; 26 May 2007; KO & MO
20/6 7*/– 2
TRI 48 CZ; Pavlovské vrchy Mts, Svatý kopeček hill near Mikulov; 48°48'24" N,
16°38'44” E; 356 m; 27 May 2007; KO & MO
16/5 6*/– –
TRI 49 CZ; Biele Karpaty Mts, Žerotín hill near Radějov; 48°51'47" N,
17°19'38” E; 321 m; 28 May 2007; KO & MO
14/5 6*/– 1
TRI 56 SK; Volovské vrchy Mts, meadow near Trebejov; 48°50'06" N, 21°13'21"
E, 253 m; 7 July 2006; KO & MO
15/– 4/– 1
TRI 70 SK; Záhorská nížina Lowlands, loc. Široká near Malacky, 48°24'44" N,
17°04'14" E; 193 m; 22 June 2007; MP & MV
9/4 4/– 1
TRI 75 HU; Gerecse Mts, meadows above Tatabánya; 47°34'51" N, 18°24 59" E;
290 m; 12 May 2008; KO, MP & IH
13/– 4/– –
TRI 76 HU; Visegrádi-hegység Mts, Mt. Vaskapu near Esztergom; 47°47'10" N,
18°46'20" E; 292 m, 13 May 2008, KO, MP & IH
13/– 4/– –
TRI 77 HU; Mecsek Mts, road to television tower (Misina) near Pécs, 46°05'59"
N, 18°13 03" E; 522 m; 14 May 2008; KO & MP
13/– 4/– –
TRI 78 HU; Mecsek Mts, viewpoint Dömörkapu near Pécs, 46°05'59" N,
18°14'01" E; 404 m; 14 May 2008; KO & MP
15/– 5/– 1
TRI 79 HU; Mecsek Mts, along the road to Tubes near Pécs, 46°06'38" N,
018°11'53" E; 407 m; 14 May 2008; KO & MP
17/– 5/– –
64 Preslia 83: 59–98, 2011
Collection data Mor
TRI 80 HU; Vértes Mts, meadow by the road to Köhányás; 47°26'33" N,
18°23'34" E; 15 May 2008, KO & MP
20/– 2/– –
TRI 81 HU; Tési-fennsík Mts, Mt. Vár-Berek near Várpalota, 47°14'15" N,
18°06'04" E, 455 m; 15 May 2008; KO & MP
18/– 3/– 1
TRI 82 HU; Veszprémfajsz, Mt. Király-hegy, 47°02'02" N, 17°53'20" E; 370 m;
15 May 2008; KO & MP
12/– 5/– –
TRI 83 HU; Budai-hegység Mts, Mt. Széchenyi-hegy near Budapest; 47°29'21"
N, 18°58'52" E; 441 m; 16 May 2008; KO & MP
14/– 4/– 2
TRI 84 HU; Budai-hegység Mts, SW slope of Mt. Kis-hárs-hegy; 47°31'46" N,
18°57'55" E; 357 m; 16 May 2008; KO & MP
20/– 5/– –
TRI 87 AU; Wiener Wald Mts, Mt. Leopoldsberg near Klosterneuburg; 48°16'41"
N, 16°20'38" E; 397 m; 22 May 2008; KO & MO
20/– 2/– –
TRI 88 AU; Wiener Wald Mts, slope above Bisamberg; 48°19'09" N, 16°21'39"
E; 360 m; 22 May 2008; KO & MO
13/– 3/– –
TRI 89 AU; Leiser Berge Mts, Dörfles, above ZOO garden; 48°32'40" N,
16°21'06" E; 371 m; 22 May 2008; KO & MO
20/– 4/– 1
TRI 90 AU; Wachau, meadow E of Dürnstein; 48°23'36" N, 15°31'55" E; 275 m;
22 May 2008; KO & MO
20/– 4/– 2
TRI 92 CZ; Český Kras Karst, Radotínske údolí valley near Radotín; 49°26'13"
N, 14°19'47" E; 240m; 24 May 2008; KO & MO
20/– 4/– 1
TRI 93 CZ; České Středohoří Mts, E margin of Hlinná; 50°34'23" N, 14°06'44"
E; 464 m; 24 May 2008; KO & MO
20/– 3/– 1
TRI 94 CZ; Moravský Kras Karst, Mt. Hády near Brno; 49°13'14" N, 16°40'29"
E; 422 m; 25 May 2008; KO & MO
20/– 1/– 1
TRI 95 SK, Biele Karpaty Mts, Mt. Veterník near Skalica; 48°48'48" N,
17°43'58" E; 312 m; 25 May 2008; KO & MO
13/– 5/– –
TRI 96 CZ; Český Kras Karst, above abandoned quarry in Velká Chuchle;
50°00'56" N, 14°22'24" E; 279m; 24 May 2008; KO & MO
5/– 2/– –
TRI 22 HU; Zempléni-hegység Mts, hill with a cross above Tállya; 48°16'09" N,
21°11'35" E; 322 m; 5 July 2006; KO & MO
16/7 5*/– 1
TRI 35 HU; Bükk Mts, SE part of Mt. Bel-kö near Bélapátfalva; 48°02'30" N,
20°22'10” E; 630 m; 19 May 2007; KO & MO
14/4 5*/– 1
TRI 36 HU; Zempléni-hegység Mts, hill NW of Mád; 48°11'51" N, 21°16'00” E;
174 m; 20 May 2007; KO & MO
(locality very close to the type locality of Centaurea stricta Waldst. et Kit)
20/3 4*/– 1
TRI 37 HU; Zempléni-hegység Mts, hill with a cross above Erdöbénye;
48°15'40" N, 21°21'40” E; 227 m; 20 May 2007; KO & MO
20/6 5*/– 2
TRI 38 HU; Zempléni-hegység Mts, in vineyards by Tolcsva; 48°16'13" N,
21°23'18” E; 159 m; 20 May 2007; KO & MO
20/7 6*/– –
TRI 39 SK; Zemplínske vrchy Mts, E slope of Mt. Piliš near Veľká Bára;
48°25'52" N, 21°42'45” E; 224 m; 21 May 2007; KO & MO
15/6 6*/– –
TRI 40 SK; Zemplínske vrchy Mts, Mt. Šimonov vrch near Malá Tŕňa; 48°26'28"
N, 21°41'23” E; 173 m; 21 May 2007; KO & MO
15/1 5*/– –
TRI 112 HU; Bükk Mts, rocks on the top of Három-kó hill; 48°03'34" N,
20°28'26" E; 905 m; 9 July 2008; KO, MO & MP
14/– 4/– –
TRI 113 HU; Bükk Mts, meadow on the top of Tár-kó hill; 48°03'22" N, 20°27'42"
E; 944 m; 9 July 2008; KO, MO & MP
16/– 5/– 2
Cyanus dominii subsp. dominii
TRI 25 SK; Volovské vrchy Mts, Mt. Humenec near Veľká Lodina; 48°51'34" N,
21°09'33" E; 280 m; 6 July 2006; KO & MP
15/5 4*/2* 1
TRI 57 SK; Volovské vrchy Mts, rocks near Malá Lodina; 48°51'51" N,
21°05'54" E; 950 m, 5 July 2008, KO & MO
14/8 7/– –
Olšavská et al.: Cyanus triumfetii group in Central Europe 65
Collection data Mor
TRI 110 SK; Branisko Mts, pine forest on the W slope of Mt. Rajtopíky;
49°00'00" N, 20°51'37" E; 950 m; 5 July 2008; KO & MO
(locality cited in the protologue of ‘Centaurea triumfetti subsp. dominii
var. eu-dominii’ Dostál)
15/– 5/– 2
TRI 111 SK; Volovské vrchy Mts, Mt. Folkmarské skaly near Veľký Folkmar;
48°49'37" N, 21°0'46" E; 836 m; 5 July 2008, KO & MO
20/– 7/– 2
Cyanus dominii subsp. slovenicus
TRI 3 SK; Podtatranská kotlina Basin, Mt. Mních near Ružomberok; 49°05'15"
N, 19°19'55" E; 500 m; 11 June 2005; KO & MO
20/6 6*/– –
TRI 6 SK; Turčianska kotlina Basin, rocks by the road to Šútovo; 49°08'50" N,
19°05'30" E; 500 m; 28 May 2006; KO & MO
20/5 5*/– –
TRI 13 SK; Veľká Fatra Mts, central part of Gaderská dolina valley; 48°56'40" N,
19°00'35" E ; 580 m; 17 June 2006; KO & MO
10/4 5*/– –
TRI 14 SK; Horehronské podolie Basin, Mt. Horné lazy near Valaská; 48°49'40"
N, 19°36'00" E; 620 m; 20 June 2006; KO & MO
20/2 4*/2* 1
TRI 18 SK; Spišsko-Gemerský kras Karst, ravine by Vernár; 48°56'00" N,
20°17'15" E; 760 m; 28 June 2006; KO, IH & MP
11/1 3*/1* –
TRI 19 SK; Chočské vrchy Mts, Prosiecká dolina valley near Prosiek; 49°09'00"
N, 19°29'55" E; 640 m; 1 July 2006; KO & MO
13/2 3*/– –
TRI 27 SK; Nízke Tatry Mts, Mt. Vachtárová near Kráľová Lehota; 49°01'25" N,
19°48'05" E; 540 m; 14 July 2006; KO & MO
11/4 5*/1* –
TRI 53 SK; Súľovské vrchy Mts, cottage Súľovčanka near Súľov; 49°10'05" N,
18°34'35” E; 410 m; 2 June 2007; KO & MO
17/2 5*/– 1
Cyanus dominii subsp. sokolensis
TRI 5 SK; Slovenský kras Karst, hill above Turňa nad Bodvou; 48°36'43" N,
20°52'20" E; 349 m; 29 May 2007; KO, DD, RŠ & PE
14/5 5*/– 1
TRI 21 SK; Slovenský kras Karst, E slope above Slavec; 48°35'59" N, 20°27'59"
E; 312 m; 22 June 2008, KO & MO
20/1 3/– –
TRI 33 SK; Západné Tatry Mts, Mt. Mních near Bobrovec; 49°10'40" N,
19°38'45" E; 1380 m; 3 August 2006; KO & MO
(locality cited in the protologue of Centaurea axillaris var. sokolensis Pawł.)
9/2 6*/– –
TRI 44 SK; Západné Tatry Mts, Mt. Sokol near Bobrovec; 49°10'29" N,
19°38'17” E; 1325 m; 21 July 2007; KO & DD
(locality cited in the protologue of Centaurea axillaris var. sokolensis Pawł.)
15/1 8*/– 2
TRI 50 SK; Slovenský kras Karst, Zádielská dolina valley near Zádiel; 48°37'37"
N, 20°50'01” E; 567 m; 29 May 2007; KO, DD, RŠ & PE (locality cited
in the protologue of Centaurea triumfetti subsp. dominii var. densifolia
20/6 6*/– 2
TRI 51 SK; Slovenský kras Karst, Mt. Zajačia strana hill; 48°35'15" N, 20°43'40”
E; 250 m; 30 May 2007; KO, DD, RŠ & PE
12/4 4*/– –
TRI 109 SK; Slovenský kras Karst, Mt. Stráň near Jelšavská Teplica; 48°36'20" N,
20°16'36" E; 340 m, 22 June 2008, KO & MO
20/– 3/– 1
Intermediate morphotype between ‘Cyanus axillaris’ and Cyanus dominii
TRI 9 SK; Slovenský kras Karst, limestone rocks above Domica Cave;
48°28'40" N, 18°28'15" E; 359 m; 12 June 2006; KO & MP
11/6 6*/– –
TRI 10 SK; Slovenský kras Karst, SW slope below Krásna Hôrka Castle;
48°39'31" N, 20°35'55" E; 427 m; 12 June 2006; KO & MP
20/5 5*/– –
TRI 20 SK; Slovenský kras Karst, meadow W of Gemerská Hôrka; 48°32'25" N,
20°21'01" E; 282 m; 22 June 2008, KO & MO
20/2 3/– 1
TRI 31 SK; Pieniny Mts, SW part of Mt. Holica near Lesnica; 49°24'35" N,
20°26'20" E; 540 m; 1 August 2006; KO & DD
(locality cited in protologue of Centaurea axillaris
var. pieninica Pawł.)
12/4 1*/1* 1
66 Preslia 83: 59–98, 2011
Collection data Mor
TRI 52 SK; Slovenský kras Karst, Mt. Veľký Paklán; 48°35'18" N, 20°43'30" E;
253 m; 30 May 2007, KO, DD, RŠ & PE
20/5 2/– 1
TRI 74 SK; Pieniny Mts, Mt. Haligovské skaly near Paluby; 49°22'55" N,
20°27'35” E; 680 m; 20 July 2007; KO & DD
7/3 3*/– 1
Cyanus triumfetti s.s.
TRI 60 IT; Valle del Chisone valley, Forte Fenestrellum Fort; 45°01'58" N,
07°03'35” E; 1303 m; 13 June 2007; KO, MP & DD
–/– 5*/– 1
TRI 63 IT; Valle del Chisone valley, meadow close to Sestriere; 45°05'15" N,
06°26'09” E; 1980 m; 14 June 2007; KO, MP & DD
–/– 10*/– 2
TRI 67 F; Alpes Maritimes Mts, road to Colle di Tende saddle; 44°06'03" N,
07°31'24” E; 2013 m; 17 June 2007; KO, MP & DD
–/– 6*/– 2
MON 3 F; Massif du Mţ. Cenis Mts, Lac Du Mont Cenis lake; 45°14"38' N,
6°56"49' E; 2108 m; 18 June 2007 ; KO, MP & DD
–/– 4/– 1
MON 6 F; Queyras Mts; meadow near Ceillac; 44°40"23' N, 6°46"10' E; 1706 m;
4 June 2008; KO, MO & VK
–/– 3/– 2
Plant material also included samples from populations at localities (or wider areas)
from which the taxa studied were described. TRI 110 is referred to in the description of
‘Centaurea triumfetti subsp. dominii var. eu-dominii’ (Dostál, 1931a) as “Slovakia
centralis, in rupibus trachyticis montis Bránisko, altitudine 800 m s. m., leg Jos. Dostál,
1928”; TRI 50 is referred to in the description of Centaurea triumfetti subsp. dominii var.
densifolius (Dostál 1931a) as “Slovakia meridionalis, in rupibus calcareis vallis Zadielská
rokle apud opp. Košice, leg Dostál, 1928”; TRI 31 and TRI 74 are referred to in the
description of Centaurea axillaris var. pieninica Pawł. (Szafer et al. 1924) as “Pien., Tatry
Bielskie”; and TRI 33 and TRI 44 are referred to in the description of Centaurea axillaris
f. sokolensis Pawł. (Pawłowski 1931) as “Skały wapienne Sokol 1235–1320 m; Mnich
1462 m”. No typical population could be chosen for Centaurea triumfetti subsp. dominii
var. slovenica because Dostál (1931a) did not mention any specific locality in the descrip
tion of this variety (just “Slovakia, in Carpatis occidentalibus”) and this name was not
typified. Material from the type locality of C. strictus [indicated as “auf Tockayer Berg
war fast ausgeblüht auf Anfang Septembris”; lectotypified by Chrtek & Skočdopolová
(1982)] is not included in this study for technical reasons. Nevertheless, population TRI 36
is located very close to the type locality of C. strictus and localities of populations TRI 36
and TRI 37 are mentioned on labels of the original herbarium material of C. strictus col
lected by Waldstein and Kitaibel and deposited in Herbarium Kitaibelianum in BP [indi
cated as “in collibus et montibus vitiferis, Hegyallya dictis, velut ad Máda, Erdöbénye”;
“von Erdöbénye”; “Von Máda … Gebürg” (Jávorka 1926)]. Furthermore no original mate
rial of ‘C. axillaris’ was included in this study for two reasons. First, the illegitimate name
coined by Willdenow (Centaurea axillaris Willd.) and all combinations based on this
name are according to the Code of Botanical Nomenclature (McNeill et al. 2006) con
nected to the type of C. graminifolius. However, a recent karyological study revealed that
C. graminifolius differs from Central European populations of the C. triumfetti group tra
ditionally treated as ‘C. axillaris’ (Olšavská & Perný 2009, see the Introduction and
Olšavská et al.: Cyanus triumfetii group in Central Europe 67
Nomenclature notes). The second reason is that Willdenow’s (1803) description of
Centaurea axillaris lacks a specific geographical location [“Habitat in montibus Austriae,
Hungariae, Helvetiae, et Galliae australis. (v.v.)” is in the protologue; “Habitat in collinis
Austriae” is on the label of Willdenow’s original herbarium material deposited in B].
Thus, it was not possible to identify the origin of plants that fit Willdenow’s description of
Plant material from the field and experimental garden was treated similarly for the
morphometric analyses. The selected morphological characters of the terminal capitulum
(length and width of involucre, number of interior and exterior florets) were measured or
scored on fresh plants. Then herbarium specimens of the plants were prepared, with florets
68 Preslia 83: 59–98, 2011
Fig. 1. – Map showing the distribution of the populations of the Cyanus triumfetti and C. montanus groups ana-
lysed in this study:
ª ‘C. axillaris’, 䉬 C. strictus (䉬 area close to the type locality of Centaurea stricta), 䊉 Cya-
nus dominii subsp. dominii (䊉 area close to the locality cited in the protologue of 'Centaurea triumfetti subsp.
dominii var. eu-dominii'),
䊏 Cyanus dominii subsp. slovenicus, 䉴 C. dominii subsp. sokolensis (䉴 the locality
cited in the protologue of Centaurea triumfetti subsp. dominii var. densifolia; 䉴 the localities cited in the
protologue of Centaurea axillaris var. sokolensis) and
: intermediate morphotype between ‘Cyanus axillaris’
and C. dominii (
: localities close to the area cited in the protologue of Centaurea axillaris var. pieninica). For
sample site details see Table 1.
and involucral bracts attached to the paper by adhesive tape; the remaining morphological
characters were measured or scored on these specimens (for more details see Olšavská et
al. 2009). Voucher specimens are deposited in the Herbarium SAV.
Altogether, 52 morphological characters were measured or scored on fresh or herbarium
material from field-collected and cultivated plants for the multivariate morphometric
study. The characters included one semiquantitative, four binary and 35 quantitative char
acters and another 12 computed ratios (Table 2). The semiquantitative character, colour of
fimbriae of appendages of involucral bracts (AFCO) was initially coded as three binary
characters (Table 2); then the first two binary characters (AFCO1, AFCO2) were included
in the matrix. The characters include those used in a previous study (Olšavská et al. 2009)
as well as additional characters obtained by a preliminary examination that were deemed
useful for distinguishing the Central European taxa of the C. triumfetti group.
Two matrices were prepared for the morphometric analyses. The first was composed of
a total of 1148 field-collected individuals from 71 populations. The second matrix con
sisted of plants cultivated in an experimental garden and comprised of a total of 195 indi
viduals from 43 populations. Depending on the analyses listed below, the values of the
characters of the individual plants or the average values of the characters for the popula-
tions were used as operational taxonomic units (OTUs). Shapiro-Wilk test (W; Shapiro &
Wilk 1965) with associated probability (Prob < W), measures of skewness and kurtosis
were computed to detect departures from normality of characters within the groups of
individuals or groups of populations (groups were based on results of PCA and PCoA
analyses). These tests showed that all variables more or less deviated from a normal distri-
bution. Spearman correlation coefficients (Sneath & Sokal 1973, Krzanowski 1990) were
computed for each matrix of field-collected and cultivated individuals to detect pairs of
highly correlated characters and to exclude one character from a pair of highly correlated
characters from further analyses.
Principal component analysis (PCA; Krzanowski 1990) of all field-collected popula
tions (average values for populations were used as OTUs) was used to provide insights
into the overall pattern of morphological variation and to create hypotheses about popula
tion groupings. Principal coordinate analysis (PCoA1; Krzanowski 1990) was computed
on one of the groups of field-collected populations revealed by PCA. PCoA was used
because the number of populations within the revealed group was lower than the number
of characters. A separate PCoA of field-collected populations of ‘C. axillaris’ (PCoA4)
was computed to compare the grouping of populations resulting from the AFLP analyses
(see below) according to morphological pattern and geographical origin of the popula
tions. In both PCoA1 and PCoA2, the average values for the populations were used as
OTUs. The characters in the above mentioned PCA and PCoA analyses were standardized
to have zero mean and unit standard deviation and the Euclidean distance was used to
compute the secondary distance matrix.
Three canonical discriminant analyses (CDA; Klecka 1980) using field-collected indi
viduals as OTUs were computed to determine the characters that mostly contribute to the
separation of the following groups (these characters were used for the identification key):
CDA1 – C. dominii vs the group of ‘C. axillaris’ and C. strictus;CDA2–‘C. axillaris’vs
Olšavská et al.: Cyanus triumfetii group in Central Europe 69
70 Preslia 83: 59–98, 2011
Table 2. – List of characters scored or measured for morphometric analyses. Characters marked with an asterisk
were measured three times from three different floral parts or involucre bracts of the same inflorescence.
Plant part Code Character Character explanation/
Stem SI indument of stem 0 – hairy, 1 – glabrous
BRN number of branches –
LN number of leaves –
SL stem length mm
SBRL stem length up to branching mm
PEL pedunculus length mm
Ratios: SBRL/SL, (SL–PEL)/LN –
Stem leaves LAI indument on upper surface of leaves 0 – hairy, 1 – glabrous
LBI indument on lower surface of leaves 0 – hairy, 1 – glabrous
LLN maximum number of leaf lobes or indentations –
LLD depth of maximum leaf lobe or indentation mm
LUL length of the uppermost stem leaf mm
LUW width of the uppermost stem leaf mm
LUD distance of the widest part of the uppermost stem leaf from
LUBW width of base of the uppermost stem leaf mm
LUDL length of decurrent part of the uppermost stem leaf mm
LML length of the middle stem leaf mm
LMW width of the middle stem leaf mm
LMD distance of the widest part of the middle stem leaf from leaf base mm
LMBW width of base of the middle stem leaf mm
LMDL length of decurrent part of the middle stem leaf mm
Ratios: LUW/LUL, LUD/LUL, LMW/LML, LMD/LML,
Involucre BN number of involucral bracts –
AFN number of fimbries or teeth on appendage of involucre bract –
AMCO colour of margin of appendages 0 – black to dark brown,
1 – pale brown
AFCO colour of appendage fimbriae or teeth 1 – black to dark brown
2 – pale brown (AFCO 2),
3 – white
IL involucre length mm
IW involucre width mm
ID distance of the widest part of involucre from base mm
*BL average lenght of involucral bract mm
*BW average width of involucral bract mm
*BD average distance of the widest part of involucral bract mm
*AMAW average width of margin of appendage of involucral bract at
*AMMW average width of margin of appendage of involucral bract in
*AFAL average length of fimbria or teeth at apex of appendage of
AFML average length of fimbria or teeth in the middle of appendage
of involucral bract
Ratios: IW/IL, ID/IL, AFAL/AMAW, AFML/AMMW
Floret FEN number of exterior florets –
FIN number of interior florets –
*FEL average length of exterior floret mm
*FIL average length of interior floret mm
*PL average length of petal mm
*PW average width of petal mm
C. strictus; and CDA3 – separation among three subspecies of C. dominii. To determine
which characters were not strongly influenced by the environment and would therefore be
useful for identification, CDA4 was computed using individuals as OTUs and two groups:
field-collected plants vs cultivated plants. Finally, to determine differences in ecological
plasticity among revealed taxa, four CDA analyses using individuals as OTUs were per
formed: CDA5 – field-collected vs cultivated plants of ‘C. axillaris’; CDA6 – field-col
lected vs cultivated plants of C. strictus; CDA7 – field-collected vs cultivated plants of
C. dominii subsp. slovenicus and CDA8 – field collected vs cultivated plants of C. dominii
subsp. sokolensis. Only the populations for which both field-collected and cultivated
plants (Table 1) were available were used in the CDA4 to CDA8 analyses. Field-collected
vs cultivated plants of Cyanus dominii subsp. dominii were not analysed by separate CDA
because of insufficient number of plants. Classification discriminant analysis (DA; Klecka
1980), using field-collected individuals as OTUs, was performed to calculate the degree of
separation; the five groups were designated according the results of PCA and PCoA.
Given that the distribution of some characters within the groups was not normal, a non-
parametric k-nearest-neighbour method (with k = 2) DA was used.
Univariate statistics (mean, 5 and 95 percentiles) of quantitative characters and fre
quencies of qualitative characters of field-collected plants were calculated for groups
revealed by PCA, PCoA, CDA and DA. Box-and-whisker plots of the six most
discriminant characters of field collected and cultivated plants were also plotted.
The SYNTAX package (Podani 2001) was used to calculate PCoA and the remaining
morphological analyses were conducted using the SAS 8.2 statistical package (SAS Insti-
The ploidy level was newly estimated for 123 individuals from 32 populations using the PI
flow cytometry and Beckton Dickinson FACSCalibur flow cytometer (Becton Dickinson,
San Jose, CA, USA) or DAPI flow cytometry and Partec Cyflow ML instrument (Partec
GmbH, Münster, Germany) equipment with an HBO-100 mercury arc lamp. When PI
staining was used, samples were prepared using the method described by Olšavská et al.
(2009). The following simplified two-step protocol (Doležel et al. 2007) was used for
DAPI flow cytometry: young intact leaf tissue of the analyzed plant was chopped together
with an internal standard in 1 ml of ice-cold Otto I buffer (0.1 M citric acid, 0.5% Tween
20), then the sample was filtered throught 42-μm nylon mesh and 1 ml of a solution con
taining Otto II buffer (0.4 M Na
O), 2-mercaptoethanol (2 μl/ml) and DAPI
(4 μg/ml) was added to the flow-through fraction and stained for 1–2 min; flow cytometric
histograms were evaluated using Partec FloMax software (v. 2.7d; Partec GmbH Münster,
Germany). Lycopersicum esculentum Mill. 'Stupnické polní tyčkové rané' was used as an
internal standard in all flow cytometric analyses. Relationship between chromosome num
ber and DNA content was verified by using published chromosome counts (Table 1;
Olšavská et al. 2009).
A total of 55 individuals from 40 populations were chosen for AFLP analyses. Genomic
DNA was extracted from leaf fragments dried in silica gel using the NucleoSpin Plant II
Olšavská et al.: Cyanus triumfetii group in Central Europe 71
kit (Macherey-Nagel GmBH & Co KG, Düren, Germany). The quality of the extracted
DNA was checked on 1% agarose gel and extracts were quantified with Spectrophoto
meter ND-1000 (NaNoDrop Technologies, Inc.). The extracts were diluted with UHP
water (Rotisolv ® HPLC Gradient Grade; Roth, Karlsruhe, Germany) and the DNA con
centrations of the aliquots were adjusted to 15 ng/μl. The AFLP procedure (Vos et al.
1995) followed the general protocol provided by Applied Biosystems (Applied
Biosystems, 2005) with a few modifications. Restriction of the genomic DNA and ligation
were performed separately. Restriction with Tru1I (equivalent with MseI) and EcoRI
endonucleases was performed at 37 °C for 3 h and then at 65 °C for 10 min; the reaction
mixture (total volume 15 μl) contained 2.5 U of EcoRI, 1.5 U of Tru1I, 2.5 μl of 10x Tango
buffer (all reaction components were from Fermentas Inc., Ontario, Canada), 2.1 μl of
UHP water (Roth) and 20 μl of DNA (ca 300 ng). Then, 5 μl of the ligation mixture con
taining 0.5 μl 1U T4 DNA ligase (Fermentas, Inc.), 0.5 μl of T4 DNA ligase buffer, 0.6 μl
of ATP (10 mM), 2.4 μl of UHP water and 0.5 μl of each MseI and EcoRI adaptors (both
MWG Biotech) was added to each sample; the incubation continued overnight (for 16 h) at
room temperature (20 °C) in a thermal cycler (Master® Gradient, Eppendorf, Hamburg,
Germany). Both preselective and selective amplification followed the protocol of Vos et al.
(1995). The amplifications were performed in two steps in a thermocycler (Master® Gra-
dient, Eppendorf, Hamburg, Germany). For preselective amplification 1.5 μl of restric-
tion-ligation product was used as the template; the pre-amplification mixture (total vol-
ume of 3 μl) further contained 0.7 μl of 10x PCR buffer (including 15 mM MgCl
Hilden, Germany), 0.14 μl of dNTPs (10 mM of each), 0.035 μl of AmpliTaq polymerase
(5 U μl
; Qiagen), 4.225 μl of UHP water (Roth) and 0.2 μl of each primer [EcoRI-A,
MseI-C (both 50 ng/μL); MWG Biotech, Ebersberg, Germany]. Preselective amplification
was performed using the following cycle profile: initial hold at 94 °C for 2 min; 30 cycles
at 94 °C for 20 s, 56 °C for 30 s, and 72 °C for 2 min; the last step was performed at 64 °C
for 30 min and then cooled to 4 °C. The products of preselective amplification were
checked on 1% agarose gel and were quantified spectrophotometrically; the DNA concen-
trations of products were adjusted to 30 ng/μl prior to selective amplification by dilution
with UHP water (Roth). Altogether, 12 primer combinations were tested on a small num
ber of samples. The three combinations that gave the best results in terms of the clarity of
the traces, the number of bands and the number of polymorphic bands were the following
primer pairs: EcoRI-ACT/MseI-CAC, EcoRI-ACC/MseI-CAC and EcoRI-ACA/MseI-
CTG (MWB Biotech); these were then used with all samples. Selective amplification was
performed for each primer combination separately; the EcoRI primer was fluorescently
labelled with WellRED D4-PA (Sigma-Aldrich Inc., Germany). The PCR reaction mix
ture (total volume of 4.21 μl) consisted of 3.14 μl of AFLP CoreMix (Applied Biosystems,
Foster City, CA, USA), 0.65 μl of diluted preamplification product and 0.21 μl of each
selective primer (EcoRI primers at 1 μM and MseI primers at 5 μM). The PCR conditions
for selective amplification were as follows: initial cycle at 94 °C for 10 min; 10 cycles at
94 °C for 20 s, 66–56 °C for 30 s (the temperature was reduced by 1 °C after each cycle)
and 72 °C for 2 min; 25 cycles at 94 °C for 20 s, 56 °C for 30 s and 72 °C for 2 min; and the
last step was performed at 60 °C for 30 min and then cooled to 4 °C. The PCR products
were added by mixing 10 μl of a premix containing 1.4 μl of EDTA (100 mM), 1.4 μl of
sodium acetate (3 M), 0.67 μl of Glycogen (20 mg/ml) and 6.7 of μl UHP water. The PCR
products were purified by ethanol precipitation. The pellets was resuspended in 25 μl of
72 Preslia 83: 59–98, 2011
Sample Loading Solution (Beckman Coulter, Fullerton, CA, USA) followed by 0.15 μl of
CEQ DNA Size Standard – 400 (PN 608098, Beckman Coulter). The samples were sub
jected to capillary electrophoresis on a CEQ
8000 DNA Analysis System and raw AFLP
data were collected and sized using the CEQ
8000 Fragment Analysis System V.9 (both
Beckman-Coulter). The AFLP profiles were scored using the GenoGrapher 1.6.0 (avail
able at http://hordeum.msu.montana.edu/genographer). Amplified fragments between 50
and 500 base pairs (bp) were scored by visual inspection for the presence (1) or absence
(0) of peaks in the output traces. Only distinct peaks were scored as present; the resulting
binary matrices of AFLP bands were used to carry out genetic data analysis. Eight samples
were analysed on two independent occasions to estimate the reproducibility of the AFLP
data. The AFLP profiles of the replicates were scored and compared with each other to cal
culate the error rate (Bonin et al. 2004).
Genetic data analysis
Three methods were used on the entire dataset of 55 samples to obtain a general view of
the variation in the AFLP pattern:
(i) Neighbour-joining tree analysis (NJ tree) was carried out using the software
TREECON (version 1.3b; Van de Peer & De Wachter 1994) and Nei and Li’s (1979)
genetic distance. Three accessions of C. montanus were used to root the tree. Group sup-
port was assessed using the same software with repeated bootstrap analyses with 2000 rep-
(ii) The principal coordinate analysis using Jaccard’s similarity coefficient was per-
formed using SYN-TAX 2000 (Podani 2001). To achieve a better resolution two PCoA
were performed; PCoA2 was based on the complete dataset and PCoA3 on one of the
groups revealed by PCoA2.
(iii) In addition to phenetic clustering Bayesian clustering in the software BAPS 3.2
(Corander et al. 2006) was applied using the module clustering of individuals. For the
complete dataset and the main groups revealed by PCoA, NJ and Bayesian cluster analy
ses the following parameters were calculated: the total number of fragments per
taxon/group, the average, minimum and maximum number of fragments per individual
and the number of polymorphic, private and private fixed fragments per taxon/group. The
patterns in fragment sharing and the number of phenotypes were also computed.
One highly correlated pair of characters was found in both the matrix of the information on
field-collected and cultivated plants, namely the number of leaf lobes or indentations and
the depth of maximum leaf lobe or indentation (Spearman correlation coefficient = 0.965).
Character depth of maximum leaf lobe or indentation was excluded from further
multivariate analyses, which were thus based on 51 characters.
Principal component analysis of all the populations of the C. triumfetti group from
Central Europe studied resulted in five main groups separated along the first and the sec
ond axis (Fig. 2, Table 3). (i) Populations corresponding to ‘C. axillaris’ and C. strictus
Olšavská et al.: Cyanus triumfetii group in Central Europe 73
formed a group on the right side of the PCA diagram. Because only a partial separation of
‘C. axillaris’ and C. strictus was visible on the PCA diagram, this group was analysed sep
arately using PCoA1. Three distinguishable groups on the left side of PCA diagram corre
sponded to three varieties described within Cyanus dominii by Dostál (1931a). These
three groups were herein treated as (ii) Cyanus dominii subsp. dominii, (iii) Cyanus
dominii subsp. slovenicus and (iv) Cyanus dominii subsp. sokolensis (= Centaurea
triumfetti subsp. dominii var. densifolia Dostál, see also the Discussion). The inclusion of
population TRI 57 (marked on the PCA diagram by an arrow) in the group of Cyanus
dominii subsp. dominii was not clear from the PCA diagram, but we considered also geo
graphic origin of this population and the fact that these plants correspond well to Dostál’s
description of C. dominii subsp. dominii. (v) On the PCA diagram, six populations
remained in an intermediate position between the group of ‘C. axillaris’+C. strictus on
the right and three groups of subspecies of C. dominii on the left; these populations,
marked as “intermediate morphotype” in the analyses, were collected from two different
areas. Four intermediate populations were collected in Slovenský kras Karst (southern
Slovakia) where they occur in close vicinity to populations of C. dominii subsp.
sokolensis. However, they differ in their ecological demands; the intermediate populations
grow on sunny slopes in valleys, whereas C. dominii subsp. sokolensis grows on rocky
74 Preslia 83: 59–98, 2011
Fig. 2. – Principal component analysis based on 51 morphological characters of 71 field-collected populations from
the Cyanus triumfetti group in Central Europe:
ª ‘C. axillaris’, 䉬 C. strictus (䉬 area close to the type locality of
䊉 Cyanus dominii subsp. dominii (䊉 area close to the locality cited in the protologue of
‘Centaurea triumfetti subsp. dominii var. eu-dominii’),
䊏 Cyanus dominii subsp. slovenicus, 䉴 C. dominii subsp.
䉴 the locality cited in the protologue of Centaurea triumfetti subsp. dominii var. densifolia; 䉴 the local-
ities cited in the protologue of Centaurea axillaris var. sokolensis)and
: intermediate morphotype between ‘Cyanus
axillaris’andC. dominii (
: localities close to the area cited in the protologue of Centaurea axillaris var. pieninica).
The black arrow denotes population TRI 57, discussed in text.
margins of the karsts’ plateaus. The two remaining intermediate populations were col-
lected in the Pieniny Mts (northern Slovakia) from where Centaurea axillaris var.
pieninica Pawł. was described (Szafer et al. 1924). It was not possible to resolve the taxo-
nomic position of the populations from the Pieniny Mts because of the low number of
plants included in the analyses as a consequence of their rarity.
The following characters were strongly correlated with the first ordination axis in PCA
and thus had the highest influence on the differentiation within the group of ‘C. axillaris’
and C. strictus and the delimitation of the three subspecies of C. dominii: pedunculus
length, length of uppermost leaf on stem, involucre width, average width of margin of
appendage of involucral bract at apex and in the middle. The separation of the groups
along the second axis was influenced mainly by several leaf characters (distance of the
widest part of uppermost and the middle leaves from their base, length of the decurrent
part of the uppermost and the middle leaf on the stem, width of the base and length of the
middle leaf on the stem) and the AFAL/AMAW and AFML/AMMW ratios. Samples of
C. strictus were partially separated from ‘C. axillaris’ also along the third axis with which
the following characters were the most correlated: stem length, stem length up to
branching and ratio LMW/LML (Table 3).
The results of PCoA1 presented in Fig. 3 confirmed that it is possible to distinguish
between ‘C. axillaris’ and C. strictus when the PCoA diagram reveals a clear separation of
the populations into two groups. In the first group, the populations correspond to
‘C. axillaris’, whereas in the second, the populations correspond to C. strictus.
Olšavská et al.: Cyanus triumfetii group in Central Europe 75
Fig. 3. – Principal coordinate analysis based on 51 morphological characters of 46 field-collected populations
from the Cyanus triumfetti group in Central Europe:
ª ‘C. axillaris’, 䉬 C. strictus (䉬 area close to the type local-
ity of Centaurea stricta).
Table 3. – Eigenvectors showing the correlations of characters with the PCA axes (values that exceed the level of
0.2 are in bold type) and the total canonical structure expressing correlations of characters with canonical axes
(CDA1 – CDA8; values that exceed the level of 0.4 are in bold type). The values were retrieved from PCA based
on 51 morphological characters of 71 populations from the Cyanus triumfetti group (Fig. 2). CDA1 – CDA8 were
based on 51 morphological characters of individuals expressed as OTUs. CDA1 with two groups pre-defined:
C. dominii and ‘C. axillaris’+C. strictus; CDA2 with two groups pre-defined: ‘C. axillaris’ and C. strictus;
CDA3 with three groups pre-defined: C. dominii subsp. dominii, C. dominii subsp. slovenicus and C. dominii
subsp. sokolensis; CDA4 with two groups pre-defined: field-collected and cultivated plants from the same popu
lations; CDA5 with two groups pre-defined: field-collected and cultivated plants of ‘C. axillaris’ from the same
populations; CDA6 with two groups pre-defined: field-collected and cultivated plants C. strictus from the same
populations; CDA7 with two groups predefined: field-collected and cultivated plants of C. dominii subsp.
slovenicus from the same populations; CDA8 with two groups predefined: field-collected and cultivated plants of
C. dominii subsp. sokolensis from the same populations. For character explanations see Table 1.
Character PCA1 CDA1 CDA2 CDA3 CDA4 CDA5 CDA6 CDA7 CDA8
Axis1 Axis2 Axis3 Axis Axis Axis1 Axis2 Axis Axis Axis Axis Axis
SI –0.105 0.100 0.135 0.314 –0.382 –0.274 0.231 –0.227 0.008 –0.530 0.082 0.000
BRN –0.055 0.131 0.143 0.403 –0.323 –0.567 0.282 0.119 0.419 0.180 –0.158 0.024
LN –0.138 0.102 0.260 0.344 –0.794 –0.818 –0.004 –0.217 –0.461 –0.216 0.113 0.123
SL 0.074 –0.068 0.345 –0.096 –0.501 –0.095 0.043 –0.161 –0.201 0.104 0.281 0.281
SBRL –0.001 –0.111 0.337 –0.211 –0.607 –0.181 –0.228 –0.373 –0.457 –0.186 0.362 0.362
PEL 0.212 0.036 0.019 0.111 0.284 0.580 0.076 0.487 0.485 0.727 –0.173 –0.170
SBRL/SL –0.163 –0.131 0.073 –0.308 –0.329 –0.240 –0.525 –0.452 –0.729 –0.138 0.351 0.415
(SL–PEL)/LN 0.173 –0.133 0.074 –0.309 0.078 0.621 0.069 –0.089 –0.086 0.214 0.178 0.282
LAI 0.112 0.015 –0.124 0.008 0.283 0.781 –0.419 0.230 0.245 0.440 0.001 0.000
LBI –0.106 0.111 0.107 0.321 –0.346 –0.194 –0.016 –0.295 –0.071 –0.710 0.115 0.000
LLN 0.097 –0.104 –0.182 –0.256 0.362 0.190 0.107 0.220 0.250 0.233 0.038 –0.107
LUL 0.202 0.158 –0.001 0.382 0.393 0.641 0.366 0.167 0.230 0.594 0.259 –0.141
LUW 0.198 0.104 –0.136 0.233 0.535 0.716 0.314 0.096 0.220
0.435 0.342 –0.257
LUD 0.174 0.203 –0.032 0.449 0.431 0.495 0.401 0.146 0.192 0.491 0.087 –0.128
LUBW 0.147 –0.186 –0.011 –0.394 0.145 0.294 0.314 0.114 0.085 0.423 0.235 0.188
LUDL 0.141 –0.221 0.058 –0.425 0.042 0.262 0.283 0.255 0.253 0.448 0.151 0.179
LML 0.117 0.211 0.139 0.552 –0.120 0.483 0.097 –0.156 –0.102 0.198 0.356 0.420
LMW 0.138 0.141 –0.149 0.308 0.440 0.677 0.069 –0.075 –0.015 0.157 0.340 0.410
LMD 0.062 0.282 0.060 0.606 0.134 0.197 0.109 –0.209 –0.190 0.110 0.187 0.447
LMBW 0.088 –0.231 0.051 –0.549 –0.036 0.092 0.212 0.141 0.202 0.102 0.146 0.137
LMDL 0.100 –0.256 0.086 –0.562 –0.105 0.047 0.203 0.234 0.272 0.227 0.064 0.145
LUW/LUL 0.078 –0.051 –0.269 –0.104 0.408 0.556 0.100 0.017 0.063 0.001 0.255 –0.191
LUD/LUL –0.016 0.177 –0.123 0.249 0.262 –0.218 0.173 0.023 0.009 0.068 –0.136 0.005
LMW/LML 0.061 –0.025 –0.326 –0.089 0.574 0.560 –0.005 0.053 0.075 0.076 0.188 0.274
LMD/LML –0.094 0.194 –0.163 0.255 0.320 –0.396 0.064 –0.159 –0.159 –0.033 –0.098 0.212
LUL/LML 0.197 0.003 –0.149 –0.057 0.577 0.457 0.420 0.463 0.412 0.606 –0.085 –0.533
PEL/LUL 0.158 –0.067 0.036 –0.057 0.185 0.349 –0.284 0.540 0.488 0.608 –0.334 –0.124
BN 0.086 –0.065 0.097 –0.050 –0.194 0.218 0.051 0.117 0.216 –0.021 0.072 –0.149
AFN 0.176 0.009 –0.077 –0.003 0.242 0.262 0.211 0.309 0.449 0.408 –0.360 0.099
AMCO 0.173 –0.064 –0.043 –0.154 0.374 0.476 0.155 0.265 0.275 0.395 –0.006 0.101
AFCO 1 0.069 0.007 –0.102 –0.049 0.236 –0.061 0.062 0.170 0.387 0.000 0.022 0.116
AFCO 2 0.019 0.177 –0.156 0.330 0.350 0.220 –0.204 –0.054 –0.130 0.167 –0.010 0.226
IL 0.106 0.195 0.234 0.479 –0.207 0.251 0.394 0.088 0.127 0.419 0.281 –0.037
IW 0.216 0.005 0.049 0.039 0.057 0.587 0.490 0.181 0.440 –0.042 0.159 0.097
ID 0.126 0.083 0.190 0.209 –0.055 0.148 0.257 0.080 0.112 0.393 0.266 0.044
BL 0.156 0.125 0.123 0.337 0.105 0.405 0.109 0.433 0.512 0.548 –0.165 –0.420
BW 0.145 0.168 0.103 0.370 –0.076 0.507 0.505 0.100 0.286 0.072 0.024 0.013
BD 0.028 0.024 0.137 0.086 0.030 –0.018 –0.182 0.103 0.078 0.280 –0.018 –0.365
AMAW 0.214 –0.034 0.022 –0.104 0.150 0.630 0.416 0.192 0.344 0.057 –0.061 –0.074
AFAL 0.160 –0.193 0.060 –0.444 0.032 0.546 0.313 0.257 0.414 0.158 0.027 0.059
76 Preslia 83: 59–98, 2011
Character PCA1 CDA1 CDA2 CDA3 CDA4 CDA5 CDA6 CDA7 CDA8
Axis1 Axis2 Axis3 Axis Axis Axis1 Axis2 Axis Axis Axis Axis Axis
AMMAW 0.211 –0.002 0.057 0.024 0.048 0.632 0.452 0.332 0.519 0.267 –0.171 –0.037
AFML 0.153 –0.189 0.116 –0.439 –0.144 0.535 0.186 0.383 0.577 0.196 –0.221 –0.026
IW/IL 0.165 –0.127 –0.096 –0.277 0.189 0.519 0.347 0.167 0.396 –0.261 –0.034 0.127
ID/IL 0.049 –0.113 –0.016 –0.154 0.086 –0.020 –0.002 0.027 0.044 0.202 0.112 0.066
AFAL/AMAW –0.052 –0.227 0.038 –0.359 –0.090 –0.202 –0.183 0.049 0.019 0.154 0.074 0.168
AFML/AMMW –0.074 –0.223 0.059 –0.472 –0.211 –0.225 –0.460 –0.012 –0.046 –0.009 –0.065 0.014
FEN 0.162 –0.085 –0.038 –0.145 0.007 0.306 0.434 0.205 0.262 –0.152 –0.109 0.019
FIN 0.127 –0.058 –0.112 –0.102 0.135 0.582 0.400 0.135 0.431 –0.110 0.054 0.130
FEL 0.181 0.121 0.071 0.282 0.120 0.556 0.429 0.306 0.368 0.448 –0.158 –0.097
FIL 0.164 0.056 0.036 0.140 0.193 0.132 0.453 0.390 0.520 0.365 –0.140 –0.150
PL 0.173 0.154 0.136 0.390 –0.027 0.646 0.509 0.068 0.255 0.329 0.242 –0.047
PW 0.190 0.028 0.076 0.048 0.054 0.459 0.596 0.272 0.356 0.288 –0.059 –0.301
CDA1 based on field-collected plants was calculated to reveal characters differentiat
ing the two groups: (i) C. dominii and (ii) ‘C. axillaris’+C. strictus. Several characters of
the leaf on the middle of the stem, e.g. its length, distance of the widest part from leaf base,
width of base and length of the decurrent part were the most important for separating taxa
along the axis (diagram not shown; Table 3). When the two taxa ‘C. axillaris’ and
C. strictus were defined as groups, CDA2 based on individuals as OTUs revealed striking
separation by the number of leaves, stem length up to branching and the ratios LMW/LML
and LUL/LML (diagram not shown; Table 3). Finally, CDA3 of individuals of three sub-
species of C. dominii confirmed their clear separation with minor overlap. The number of
leaves, average distance between leaves, indumentum on upper surface of leaves, width of
the middle leaf on the stem and length and width of the uppermost leaf on the stem were
the most important characters for the separation of C. dominii subsp. sokolensis from the
other two subspecies along the first axis. Involucre width, average length and width of
petal in addition to the ratios SBRL/SL and AFML/AMMW contributed to the separation
of C. dominii subsp. dominii along the second axis (diagram not show; Table 3).
CDA4 of field-collected plants and cultivated plants defined as two groups revealed
only a slight shift between these groups. Cultivated plants differed from field-collected
plants mainly in pedunculus length and average length of involucral bract, and the ratios
SBRL/SL, LUL/LML and PEL/LUL (diagram not shown, Table 3). Of the most useful
characters for taxa determination identified by analyses CDA1, CDA2 and CDA3 only the
ratio LUL/LML was affected by cultivation (Fig. 4). CDA5, CDA6, CDA7 and CDA8
showed differences in ecological plasticity of studied taxa: C strictus and ‘C. axillaris’ are
more influenced by cultivation than C. dominii subsp. sokolensis and C. dominii subsp.
slovenicus (diagrams not shown, Table 3). Cultivated plants differed from field-collected
plants in various characters, depending on the particular taxon. In comparison to field-col
lected plants of ‘C. axillaris’, cultivated plants of this taxon (CDA5) had shorter stems,
fewer leaves, more branches, occurrence of branching in lower part of the stem, longer
interior florets, wider margins (AFML) and longer fimbria or teeth (AMMW) on the
appendage of the involucral bract in the middle (but the ratio of the last two characters,
AFML/AMMW, did not change). Cultivated plants of C. strictus differ from field-col
lected plants (CDA6) in the type of indumentum on stem and leaves, and characters of the
uppermost leaves (ratios that define the shape of uppermost leaves, LUW/LUL and
Olšavská et al.: Cyanus triumfetii group in Central Europe 77
78 Preslia 83: 59–98, 2011
LUD/LUL, did not change). Cultivated plants of C. strictus are less glabrous and have
larger uppermost leaves than field-collected plants. CDA7 of C. dominii subsp. slovenicus
showed that no character is strongly affected by cultivation. Cultivated plants of
C. dominii subsp. sokolensis had smaller leaves on the middle of the stem (LML, LMW,
LMD) than field-collected plants (CDA8), but the shape of these leaves (ratios
LMW/LML, LMD/LML) did not change. The difference between field-collected and cul
tivated plants may be partially caused by the low number of plants cultivated and the short
period of time for which they were cultivated.
Non-parametric classificatory discriminant analysis of the five groups identified by the
ordination analyses – (i) ‘C. axillaris’, (ii) C. strictus, (iii) C. dominii subsp. dominii,(iv)
C. dominii subsp. slovenicus and (v) C. dominii subsp. sokolensis – showed that more than
90% of the individuals were correctly classified. Some morphological overlap among the
taxa was exhibited by 25 plants of ‘C. axillaris’ (4.1%) that were misclassified as
C. strictus and by nine of C. dominii subsp. slovenicus (7.4%) that were misclassified as
C. dominii subsp. dominii (Table 4).
Table 4. – Results of non–parametric discriminant analysis of individuals from the Cyanus triumfetti group with
the following five pre–defined groups: ‘C. axillaris’ (axi). C. strictus (str). C. dominii subsp. dominii (dom).
C. dominii subsp. slovenicus (slo) and C. dominii subsp. sokolensis (sok).
Actual group Group membership predicted
(number of observations and percentage classified into groups)
axi str dom slo sok
axi 579 (94.61%) 25 (4.08%) 3 (0.49%) 1 (0.16%) 4 (0.65%)
str 0 (0%) 149 (99.33%) 0 (0%) 0 (0%) 1 (0.67%)
dom 0 (0%) 0 (0%) 64 (100%) 0 (0%) 0 (0%)
slo 0 (0%) 0 (0%) 9 (7.38%) 111 (90.98%) 2 (1.64%)
sok 0 (0%) 0 (0%) 1 (0.91%) 0 (0%) 109 (99.09%)
Box-and-whisker plots of the six groups of field-collected and cultivated plants
revealed by multivariate morphometric analyses showed only a slight overlap in the
depicted discriminant characters (Fig. 4). Univariate statistics of all quantitative characters
and frequencies of all qualitative characters also showed that the recognized taxa differed
in several characters. The univariate analyses showed that C. dominii differ from
‘C. axillaris’ and C. strictus in having very short decurrent leaves (there is an even smaller
overlap in this character between cultivated plants of these taxa); ‘C. axillaris’ has fewer
stem leaves than C. strictus and also the width and the ratio expressing the shape of the
middle leaf on the stem differ in these two species (LMW/LML; Electronic Appendix 1,
Fig. 4). Examination of qualitative characters showed that most plants of C. dominii subsp.
sokolensis have leaves with a tomentose upper surface (99.1%) while most plants of
Olšavská et al.: Cyanus triumfetii group in Central Europe 79
Fig. 4. – Box-and-whisker plots displaying the variation in selected morphological characters among five taxa
from the Cyanus triumfetti group in Central Europe and comparison of field-collected (F) and cultivated plants
(C); ‘C. axillaris’ [axi; n(F) = 612, n(C) = 71], C. strictus [str; n(F) = 150, n(C) = 34], C. dominii subsp. dominii
[dom; n(F) = 64, n(C) = 13], C. dominii subsp. slovenicus [slo; n(F) = 122, n(C) = 26], C. dominii subsp.
sokolensis [sok; n(F) = 110, n(C) = 19] and intermediate morphotype between ‘C. axillaris’ and C. dominii [int;
n(F) = 90, n(C) = 25]. Boxes define the 25th and 75th percentiles, horizontal lines show the median, whiskers are
from the 10th to 90th percentiles and asterisks indicate outliers.
C. dominii subsp. dominii and subsp. slovenicus have leaves with a glabrous upper surface
(96.8% and 76.9% respectively). Most C. strictus plants have white fimbrie on the append
ages of the involucral bracts (99.9%), whereas no more than 42% of the plants of other
taxa have the same colour of appendages (Electronic Appendix 1). The differentiation
80 Preslia 83: 59–98, 2011
Fig. 5. – Shape of the (1) uppermost leaves on stem, (2) middle leaves on stem, (3) involucral bracts and (4) invo
lucres of the Cyanus triumfetti group in Central Europe (a) ‘C. axillaris’, (b) C. strictus, (c), C. dominii subsp.
dominii, (d) C. dominii subsp. sokolensis, (e) C. dominii subsp. slovenicus. Drawings are based on the mean val
ues of the measurements of the characters; the decurrent parts of the leaves are marked by arrows. Figures were
drawn by Zlata Komárová and Katarína Olšavská.
among ‘C. axillaris’, C. strictus and the three subspecies of C. dominii are also apparent
when involucral bract, involucres, the uppermost and the middle leaf on the stem of the
taxa are illustrated based on the mean values of the characters measured (Fig. 5).
The 123 individuals from 32 populations of ‘C. axillaris’, C. strictus and C. dominii were
diploid (2n ~ 2x ~ 22) (Table 1). Data for the populations of the C. triumfetti group
reported here confirm and extend previous records (Olšavská et al. 2009). The seven
plants from two populations of C. montanus were tetraploid (2n ~ 4x ~ 44) (Table 1),
which accords with the chromosome number 2n = 44 already cited for this taxon in France
(Guinochet 1957), Switzerland (Krahenbühl & Küpfer 1992) and Austria (Lipper &
Neighbour-joining tree (Fig. 6) of the C. triumfetti and C. montanus groups identified four
main clusters. Three of them, (i) C. montanus from the Western Alps, (ii) C. triumfetti s.s.
from Western Alps and (iii) cluster containing samples of ‘C. axillaris’ from Austria and
the Czech Republic (except the Carpathians; hereinafter referred to as Austria-Czech
group) had high bootstrap support (100%, 100% and 92%, respectively). (iv) The fourth
cluster containing the remaining samples of ‘C. axillaris’, C. strictus and C. dominii from
the Western Carpathians and Pannonia (hereinafter referred to as W Carpathians-Pannonia
group) had only very low bootstrap support (< 40%). Within the latter cluster, individuals
of C. strictus from the Zempléni-hegység Mts and the Bükk Mts formed two distinct clus-
ters with intermediate bootstrap support (82% and 77%, respectively).
On the ordination diagram of the PCoA analysis of the entire dataset based on AFLP
data (55 individuals; PCoA2; Fig. 7a), there are four groups identical to the clusters on the
NJ tree: (i) C. montanus from the Western Alps, (ii) C. triumfetti s.s. from the Western
Alps (iii) the Austria-Czech group of ‘C. axillaris’ and (iv) the W Carpathians-Pannonia
group of ‘C. axillaris’, C. strictus and C. dominii. Individuals of the C. triumfetti group
were clearly separated along the first and second axis while those of the C. montanus
group were divided from the remaining individuals mainly along the third axis. The extra
PCoA based on 35 individuals from the W Carpathians-Pannonia group (PCoA3; Fig. 7b)
showed only a tendency to form groupings of individuals corresponding to the groups
identified by the morphometric analyses (‘C. axillaris’, C. strictus, C. dominii and inter
mediate morphotype). The samples of C. dominii together with those of the intermediate
morphotype were situated mostly on the left side of PCoA3 diagram while most samples
of ‘C. axillaris’ and C. strictus were on the right side. The exception is one individual from
population TRI 56 (marked on the diagram by an arrow), which, according to its morphol
ogy, belongs to ‘C. axillaris’, but on the PCoA diagram of AFLP data appears on the left
side in the C. dominii group. The samples of C. strictus formed two groups as on the NJ
tree, reflecting their geographical origin (Zempléni-hegység Mts and Bükk Mts). Interest
ingly, three individuals of C. strictus from the Bükk Mts were clearly separated from all
remaining samples of the W Carpathians-Pannonia group along the second axis.
The optimal partition with the highest log marginal likelihood (–4419.3013) produced
by BAPS consisted of three clusters that corresponded to the following groups: (i)
Olšavská et al.: Cyanus triumfetii group in Central Europe 81
82 Preslia 83: 59–98, 2011
Fig.6. – Neighbour-joining tree of AFLP data for 55 individuals from the Cyanus triumfetti and C. montanus groups:
♠ C. montanus, ¬ C. triumfetti s.s., ♥ ‘C. axillaris’, 䉬 C. strictus, 䊉 C. dominii subsp. dominii, 䊏 C. dominii
䉴 C. dominii subsp. sokolensis and : intermediate morphotype between ‘C. axillaris’ and
C. dominii (cross). Bootstrap values above 40% based on 2000 replicates are shown.
Olšavská et al.: Cyanus triumfetii group in Central Europe 83
Fig. 7. – Ordination diagrams of principal coordinate analyses based on AFLP data; (a) PCoA2 calculated for 55
individuals from the Cyanus triumfetti and C. montanus groups; (b) PCoA3 calculated for 35 individuals from the
C. triumfetti group from the Western Carpathians and Panonia:
♠ C. montanus, ¬ C. triumfetti s.s., ♥ ‘C. axillaris’,
䉬 C. strictus, 䊉 C. dominii subsp. dominii, 䊏 C. dominii subsp. slovenicus, 䉴 C. dominii subsp. sokolensis and
: intermediate morphotype between ‘C. axillaris’ and C. dominii (cross). The black arrow denotes population
TRI 56, discussed in text.
C. triumfetti s.s. + C. montanus from the Western Alps, (ii) ‘C. axillaris’ from Austria and
the Czech Republic (= Austria-Czech group of ‘C. axillaris’) and (iii) ‘C. axillaris’,
C. strictus and C. dominii from the Western Carpathians and Pannonia (= W Carpathians-
Pannonia group of ‘C. axillaris’, C. strictus and C. dominii). According to the results of
Bayesian clustering, the species C. triumfetti and C. montanus from the Western Alps
formed one cluster but the NJ tree and the PCoA analyses indicated two groups.
The AFLP analysis resulted in 264 scorable bands for 55 individuals with 24 (9%)
monomorphic bands. Control replicates proved that the AFLP data were highly reliable
(repeatability 99.79%). Across the whole dataset, 55 unique AFLP multilocus phenotypes
were detected. No pair or group of individuals had an identical AFLP phenotype. The
average total number of fragments per individual was highest for C. montanus (74) and
lowest for C. triumfetti s.s. (67). Table 5 shows the statistics for the AFLP bands. The W
Carpathians-Pannonia group of ‘C. axillaris’, C. strictus and C. dominii had the most
polymorphic bands (186 out of 210; 89%) and C. triumfetti s.s. the lowest (65 out of 89;
73%). The number of private bands varied considerably from 74 in the W Carpathians-
Pannonia group of ‘C. axillaris’, C. strictus and C. dominii to seven in C. montanus.No
fragments were private fixed in either the Austria-Czech group of ‘C. axillaris’orW
Carpathians-Pannonia group of ‘C. axillaris’, C. strictus and C. dominii, while three and
nine were fixed in C. triumfetti s.s. and C. montanus, respectively. Presence of private
fixed AFLP fragments in C. triumfetti s.s. and C. montanus may be a consequence of the
low number of samples. Examination of AFLP bands that were shared in the whole AFLP
data set revealed that there were no fragments that were shared between any taxon or
group of taxa.
Table 5. – Distribution of AFLP fragments across the investigated Cyanus samples; N ind – number of individu-
als. N phen – number of AFLP multilocus phenotypes. N fragm – (min) average (max) number of fragments per
individual. N pr – number of private fragments per taxa. N prf – number of private fixed fragments per taxa.
Taxon N ind N phen (min) N fragm (max) N pr N prf
C. montanus 3 3 (71) 74 (77) 7 9
C. triumfetti s.s. 5 5 (65) 67 (71) 8 3
Austria-Czech group 12 12 (60) 70 (80) 20 0
W Carpathians-Panonian group 35 35 (63) 71 (77) 74 0
Morphometric and genetic variation of ‘Cyanus axillaris’
As the AFLP analyses revealed that populations of ‘C. axillaris’ differentiate into two
groups a detailed analysis of the variation in the infraspecific morphology of this species
was undertaken. PCoA4 computed for 37 field-collected populations of ‘C. axillaris’
revealed a certain amount of morphological differentiation within the species (Fig. 8). The
populations were divided into two groups along the first axis. Within the group on the right
side of the diagram another division of populations along the second axis was visible.
However, this morphological variation of ‘C. axillaris’ was incongruent with the geo
graphic origin of the populations and AFLP data (Austria-Czech group of ‘C. axillaris’vs
‘C. axillaris’ from the W Carpathians-Pannonia group; Fig. 8).
84 Preslia 83: 59–98, 2011
Morphological differentiation of the Cyanus triumfetti group in Central Europe
The current morphological analyses suggest that the following three species of the
C. triumfetti group in Central Europe can be distinguished: ‘C. axillaris’, C. strictus and
C. dominii. These species can be recognized by a combination of morphological characters
of which the leaf characters are the best for delimiting these species. Cyanus dominii can be
clearly distinguished from both ‘C. axillaris’andC. strictus as C. dominii has non-decurrent
stem leaves, whereas ‘C. axillaris’ differs from C. strictus mainly in the number of leaves,
stem length up to branching and in the ratio of width to length of the leaf on the middle of the
stem (LMW/LML). Interestingly, the ratio of length of fimbria/width of margin of append
age of involucral bract that is considered as a crucial character in some identification keys to
Central-European knapweeds (Dostál 1976b, 1989, Simon 1992) was not confirmed as suit
able for classification of these species. The analyses indicate there is a large overlap in this
character (measured in the current analyses as two ratios AFAL/AMAW, AFML/AMMW,
see Table 2) and in other similar characters such as the length of the fimbria (AFAL, AFML)
and width of the margin of the involucral bract (AMAW, AMMW); this result questioned the
importance of these characters for delimitating these species (Fig. 8). However, the morpho
logical analyses showed that differences among the taxa were mainly in the colour of the
Olšavská et al.: Cyanus triumfetii group in Central Europe 85
Fig. 8. – Ordination diagram of principal coordinate analysis based on 51 morphological characters calculated
for 37 field-collected populations of ‘Cyanus axillaris’ in Austria and the Czech Republic except the Carpathians
♥, grey symbols indicate population assigned to the Austrian-Czech group by AFLP analyses) and from the
Western Carpathians and Pannonia (
♣, black symbols indicate population assigned to the W Carpathians-
Pannonian group by AFLP analyses). The grey symbols indicate a population from which selected individuals
were analysed by AFLP.
margin and the fimbriae of the appendages of the involucral bracts (see identification key
below and Electronic Appendix 1).
As depicted by the multivariate morphometric analyses, infraspecific morphological
variation within each of the three species was very conspicuous. This intraspecific varia
tion has taxonomic implications for C. dominii, which includes three groups of popula
tions recognized as subspecies based on their morphology and distribution: C. dominii
subsp. dominii, C. dominii subsp. sokolensis and C. dominii subsp. slovenicus (see below).
Cultivation of plants of ‘C. axillaris’, C. strictus and C. dominii showed that mainly leaf
and stem characters were influenced by environmental conditions. However, the shift in
characters caused by cultivation did not affect the determination value of key characters and
differences among taxa remained. For example, the number of leaves was lower in cultivated
plants of all taxa, but this character was still useful for identifying the taxa (Fig. 4).
Pattern of genetic variation
Principal coordinate analyses of the AFLP data revealed three genetically differentiated
groups within the C. triumfetti aggregate corresponding to their geographic origin
(Fig. 7a). The group of C. triumfetti s.s. from the Western Alps and the Austria-Czech
group of ‘C. axillaris’ also had high bootstrap support in the neighbour-joining tree, but
there was no support for the internal node of the W Carpathians-Pannonia group of
‘C. axillaris’, C. strictus and C. dominii (Fig. 6). Thus, it was difficult to infer the genetic
relationships of the latter group. In the case of C. triumfetti s.s., the genetic analyses con-
firmed the results of a previous study, which revealed the morphological separation of this
Alpine taxon from Central European taxa (Olšavská et al. 2009). The pattern of AFLP
variation of the C. triumfetti group in Central Europe is more remarkable considering the
morphological differentiation of the taxa assessed in this study. Populations of the
C. triumfetti group from Central Europe are divided into two genetically distant and
allopatric groups, which do not correspond to the three species recognized by the
morphometric analyses. All samples from the Austria-Czech genetic group should be
treated as ‘C. axillaris’ based on their morphology and those from the W Carpathian-
Pannonia group are of the three morphologically well-differentiated species. This result
was unexpected because the populations of three different species from the one area are
genetically closer to each other than are populations of one species – ‘C. axillaris’ – from
two different areas. Such a pattern of genetic variation raises some evolutionary hypothe
ses. For example, the genetic variation recorded for ‘C. axillaris’ might indicate that popu
lations of this species from the Austria-Czech and W Carpathian-Pannonia groups may
differ in their biogeographic history. The fact that ‘C. axillaris’ is widely distributed
throughout Europe could imply that this species survived periods of glaciation in several
glacial refugia and/or it colonized Europe by more postglacial migration routes. A similar
scenario of postglacial colonization of Europe by other groups from the family
Compositae was revealed by phylogeographic analyses based on genetic markers (e.g.
Chauvet et al. 2004, Mráz et al. 2007, Font et al. 2009). Based on the results of AFLP anal
yses, it is assumed that the ‘C. axillaris’ populations from the Austria-Czech group are not
connected by gene flow with populations of ‘C. axillaris’ in the Western Carpathians and
Pannonia. Such a genetic gap between the populations from different parts of Central
Europe is recorded for other plant groups. For example, the Dryas octopetala L. popula
86 Preslia 83: 59–98, 2011
tions in the Tatras are connected to northeastern Europe and Siberia, whereas the alpine
populations are connected to northwestern Europe (Skrede et al. 2006). Rosa pendulina L.
expanded into Central Europe from two refugia (roughly the Alps vs Southern
Carpathians) and now its populations (circumscribed by their cpDNA haplotypes) are
divided by a contact zone in the Danube valley (Fér et al. 2007). However, the results of
AFLP analysis of the C. triumfetti group recorded in this study could not confirm or reject
any phylogeographic scenario about ‘C. axillaris’ in Central Europe due to the lack of
material from other parts of its distribution, including potential refugia.
The low genetic differentiation of the W Carpathian-Pannonia group of ‘C. axillaris’,
C. strictus and C. dominii and high genetic variation within this group can be interpreted in
two different ways. The first scenario suggests that C. strictus and C. dominii originated by
the isolation of two populations of some widely distributed ancestor. A taxon that may be
considered as the putative ancestor is ‘C. axillaris’, which also currently occurs in this area.
The distribution of these species suggests long-term isolation of the Western Carpathians
populations that could be maintained over time by the specific substrate and vegetation
requirements of this species (for a similar example, see Pečinka et al. 2006). Cyanus dominii
and C. strictus grow exclusively on dolomites, limestones and tuffs, and they occur in rela
tively small spatial isolates of steppe and rocky-slope vegetation that are separated by large
areas of continuous forest at least since the end of the last glaciation, i.e. approx. 10,000
years BP (Krippel 1986). The alternative scenario suggests occasional gene flow and/or
hybridization events among previously diverged taxa occurring in this area. Our ongoing
hybridization experiments show the possibility of crossing among ‘C. axillaris’, C. strictus
and C. dominii, as well-developed seeds were obtained in 48–83% of the hybrids (K.
Olšavská et al., unpublished). These two scenarios (common ancestor and short time of
speciation vs gene flow or hybridization) on their own or in combination could result in very
low genetic differentiation of morphologically well-separated taxa.
Relative to the morphological delimitation of populations or taxa there are two patterns
of genetic differentiation recognized in the studies on genetic differentiation at the species
or below-species level published over the last decade. The first and more common pattern
is the congruence between the genetic and morphological variation, resulting often in
a definite taxonomic classification of the taxa (e.g. Lihová et al. 2004, Perný et al. 2004,
Šmarda et al. 2007). The second and rarer conclusion is that the patterns of genetic and
morphological variation are incongruent. This incongruence is attributed to various fac
tors, such as the polygenic regulation of quantitative traits (Roldán-Ruiz et al. 2001,
Archak et al. 2003), hybridization and introgression (Greimler & Jang 2007, Lihová et al.
2007), human selection and domestication (Depypere et al. 2009) or natural selection
(Kamada et al. 2007, Pérez-Barrales et al. 2009). In the present study, the incongruence
between the AFLPs and the morphology evident in the W Carpathians and Pannonia group
and in ‘C. axillaris’ may be attributed to past hybridization events and natural selection for
some polygenic quantitative morphological characters in addition to biogeographic
history as discussed above. These processes have yet to be identified.
Classification of the Cyanus triumfetti group
Analysing the morphological data and geographical distribution led to the conclusion that
the rank of species is appropriate (according to the classical phenetic species concept;
Olšavská et al.: Cyanus triumfetii group in Central Europe 87
Sneath & Sokal 1973, Stuessy 2009) for the following four groups of populations of the
C. triumfetti group studied: C. triumfetti s.s., ‘C. axillaris’, C. strictus and C. dominii.
These four species are differentiated by several morphological characters and have non-
overlapping distributions. The three subspecies within C. dominii provide an example of
allopatric differentiation where the classification at the subspecies level is the most ade
quate (cf. Perný et al. 2004, Galbany-Casals et al. 2006, Conti 2007, Kropf 2008, Peruzzi
& Passalacqua 2008). In the Western Carpathians, similar examples of infraspecific differ
entiation addressed by biosystematic studies include e. g. Aconitum firmum Rchb. (Mitka
et al. 2007), Cardamine amara L. (Marhold et al. 2002) or Scilla drunensis (Speta) Speta
(Kochjarová 2005). The three subspecies of C. dominii share a synapomorphic character –
leaves are not decurrent or very shortly – that can be used to discriminate them from all
other taxa of the C. triumfetti group.
Cyanus triumfetti s.s. Results of AFLP analyses presented in this paper accord with
previous multivariate morphometric analyses that excluded C. triumfetti s.s. from the
Western Carpathians and adjacent parts of Pannonia (Olšavská et al. 2009). The PCoA2
diagram based on AFLP data (Fig. 7a) and results of Bayesian clustering indicate that
C. montanus subsp. montanus may be genetically closer to C. triumfetti s.s. than to other
taxa of the C. triumfetti group. However, the apparent clustering of C. montanus and
C. triumfetti s.s. may be a consequence of uneven sampling of the main four groups (the
largest portion of the variation is among the larger groups). In addition, C. montanus and
C. triumfetti are distinctly separated by numerous private and private fixed fragments of
both taxa and the absence of shared fragments.
‘Cyanus axillaris’. The morphological analyses showed that ‘C. axillaris’ is a taxo-
nomically critical species with high intraspecific variation. The morphometric analyses
that are presented indicate morphological differentiation within ‘C. axillaris’, but this dif-
ferentiation is incongruent with the AFLP data and geographical distribution. Therefore,
no groups of populations were distinguished at any taxonomic level based on the results
presented (Fig. 8). However, ‘C. axillaris’ has a broad distribution, and therefore, addi-
tional studies of material from a broader range of areas are necessary to determine whether
the morphological and genetic variation outlined for Central European populations may
be applicable to neighbouring areas and if the variation of ‘C. axillaris’ deserves some tax
onomic delimitation. Until the problem of the intraspecific variation of this species is
elucidated, nomenclatural questions can not be resolved (see Nomenclature notes).
Cyanus strictus. The NJ tree (Fig. 6) and PCoA3 (Fig. 7b) of the AFLP markers indi
cate that the samples of C. strictus form two separate groups that reflect their geographical
origin (the Zempléni-hegység Mts and the Bükk Mts). This separation is also partially vis
ible in the results of morphometric analyses (Fig. 2, 3) where three populations from the
Bükk Mts appear closer to the populations of ‘C. axillaris’ than the remaining populations
of C. strictus. However, there are no clear groups in a separate morphometric analysis of
individuals of C. strictus (results not shown); thus, the degree of infraspecific morphologi
cal differentiation was not sufficient to recognize any intraspecific taxa within this species.
Cyanus dominii. The considerable intraspecific variation of this species resulted in the
description by Dostál (1931a) of five varieties in Europe. The morphometric analyses pre
sented in the current paper confirmed the broad morphological variation of the Western
Carpathians populations of C. dominii and support their classification into the three follow
ing morphologically distinct taxa: C. dominii subsp. dominii, C. dominii subsp. slovenicus
88 Preslia 83: 59–98, 2011
and C. dominii subsp. sokolensis. These morphological analyses also indicate that the pop
ulation from Slovenský kras Karst [where Centaurea triumfetti subsp. dominii var.
densifolius was described (Dostál 1931a)] and from the Zapadné Tatry Mts [where
Centaurea axillaris f. sokolensis Pawł. was described (Pawłowski 1931)] belong to the
same subspecies C. dominii subsp. sokolensis. The three subspecies of C. dominii have
distinct distributions restricted to certain mountains in Slovakia (Fig. 1). The results of
AFLP analyses provide no genetic differentiation of these three subspecies, possibly
because of the rather low number of samples. Overall, the results of this study may stimu
late further research on populations from the C. triumfetti group in Bulgaria or Ukraine,
where C. dominii is reported by several authors (Dostál 1976b, Andreev et al. 1992,
Mosyakin & Fedoronchuk 1999, Greuter 2006–2009).
Intermediate morphotypes. Based on the morphological analyses the intermediate
plants from the Pieniny Mts (northern Slovakia) are similar in most characters to
C. dominii but their leaves have a broad and long decurrent part that is typical of
‘C. axillaris’. Interestingly, Dostál (1931a) also mentions that plants from the Pieniny Mts
are very similar to C. triumfetti subsp. dominii var. slovenica but differ in their distinctly
decurrent leaves. Unfortunately, only a few plants were included in the current analyses
due to their rare occurrence in this area. Hence, there is insufficient data to support or
reject any hypotheses about their taxonomic position. Therefore, this study raises the
question of whether populations from the Pieniny Mts are of hybrid origin or are part of
the intraspecific variation of C. dominii or ‘C. axillaris’. Intermediate plants from
Slovenský kras Karst may be more obviously attributed to a recent hybridization event.
These intermediate plants have mixed morphological characters of both putative parental
taxa. They have a high number of branches and stem leaves, which are typical of
C. dominii subsp. sokolensis, but they have narrow lanceolate leaves and 60% of the plants
with decurrent leaves on the middle of the stem similar to ‘C. axillaris’. These plants grow
in Slovenský kras Karst in the vicinity of populations of C. dominii subsp. sokolensis
(5–10 km), but they prefer different ecological conditions and grow on sunny slopes in
valleys, which is more typical of ‘C. axillaris’ (whereas C. dominii subsp. sokolensis
grows on rocky margins of karst plateaus). Hybrids between ‘C. axillaris’ and C. dominii
are reported by Dostál (1950; described as Centaurea triumfetti subsp. hybr. sillingeri
Dostál) from other parts of Slovakia, namely the Malé Karpaty Mts, Súľovské skaly Mts,
Nízke Tatry Mts and Malá Fatra Mts. According to the results of the morphological analy
ses, the population TRI 2 from the Malé Karpaty Mts belongs to ‘C. axillaris’, whereas
samples collected in the Súľovské skaly Mts (TRI 53) and Nízke Tatry Mts (TRI 27)
belong to C. dominii subsp. slovenica.
Identification key of the taxa from the Cyanus triumfetti group in Central Europe
For confident identification, it is recommended that at least several plants from a locality
should be identified using this key because of the high intrapopulation variation of the
taxa. It is necessary to use a combination of characters (not just one) because of the mor
phological overlap between some taxa in some characters. The character values given in
the key represent the 10–90 percentiles; the morphological characters of inflorescences
and flowers were measured or scored for the terminal capitulum.
Olšavská et al.: Cyanus triumfetii group in Central Europe 89
1a Middle stem leaves decurrent ..........................................................................................................................2
1b Middle stem leaves not decurrent or rarely shortly decurrent (up to 7.5 mm) ........... C. dominii (Dostál) Holub
01a Stems with numerous branches (3–20); number of stem leaves 22–39; leaves tomentose on both sur
faces; middle stem leaves linear (6.2–12.8 mm wide, ratio of width to length 0.08–0.14); endemic to
Slovenský kras Karst and Západné Tatry Mts ..................C. dominii subsp. sokolensis (Pawł.) Olšavská
01b Stems sparingly branched (0–4); number of stem leaves 7–29; leaves usually glabrous or tomentose
above; middle stem leaves lanceolate to linear-lanceolate (8.9–25.8 mm wide, ratio of width to length
02a Leaves usually glabrous above; middle stem leaves linear-lanceolate (8.9–21.3 mm wide); involucre
cylindrical (6.9–10.5 × 12.3–14.9 mm); external floret 29.4–38.9 mm long with petals 8.3–10.9 mm long
and 0.7–1.2 mm wide; endemic to Volovské vrch Mts and Branisko Mts ........ C. dominii subsp. dominii
02b Leaves glabrous or tomentose above; middle stem leaves lanceolate (11.8–25.8 mm wide); involucre
ovoid (11.8–25.8 × 13.5–16.4 mm); external floret 35.0–47.3 mm long with petals 11.1–16.8 mm long
and 1.1–1.8 mm wide; endemic to mountains in Central Slovakia
.......................................................................................C. dominii subsp. slovenicus (Dostál) Olšavská
2a Fimbriae white or silver; number of stem leaves 18–29; stems tomentose; stem leaves narrowly lanceolate;
uppermost stem leaves 2.8–5.8 mm wide, middle stem leaves 4.6–11.6 mm wide; ratio of width to length of
leaves 0.07–0.15; Zempléni-hegység Mts and Bükk Mts .............................. C. strictus (Waldst. & Kit.) Soják
2b Fimbriae pale to dark brown or white; number of stem leaves 10–20; stems tomentose to subglabrous; stem
leaves lanceolate; uppermost stem leaves 4.1–11.2 mm wide, middle stem leaves 7.1–16.5 mm wide; ratio of
width to length of leaves 0.1–0.22; Central Europe .......................................................................‘C. axillaris’
[Intermediate morphological types between ‘C. axillaris’ and C. dominii occur in Slovenský kras Karst and
Pieniny Mts. Intermediate plants from the Pieniny Mts possess tomentose and lanceolate middle stem leaves
(6.7–18.2 mm wide) and an ovoid involucre (8.6–10.4 × 12.8–15.1 mm) – both characters are similar to
C. dominii; but their leaves have broad and long decurrent part – typical of ‘C. axillaris’. Plants from Slovenský
kras Karst have many branches (2–11) and stem leaves (16–29) – both characters are typical of C. dominii subsp.
sokolensis; but they have narrow lanceolate leaves and 60% of the plants have decurrent middle stem leaves –
characters similar to those of ‘C. axillaris’.]
Allioni (1774) named Cyanus triumfetti after Giovanni Battista Triumfetti (1658–1708),
professor of botany and director of the Botanical Garden in Rome. In the original descrip
tion Allioni used the epithet “triumfetti” as the appropriate Latin genitive to Latin form of
the name Johannes Baptista Triumfettus. The other often used form “triumfettii”was
formed by adding the appropriate genitive inflection (-i) to italian form of the name G. B.
Triumfetti. According to the rules of the Code of Botanical Nomenclature (hereinafter the
Code; Art. 60.; McNeill et al. 2006) both forms are correct, but the original spelling
“triumfetti” has priority.
When Willdenow (1803) published the name Centaurea axillaris Willd., he cited in the
synonymy two previously published names, Centaurea variegata Lam. and Centaurea
seusana Chaix. According to the rules of the Code (Art. 52.1., 52.2.; McNeill et al. 2006),
because Willdenow did not use the epithet of one of the cited names this name is illegiti
mate. Both names Centaurea variegata and Centaurea seusana belong to the taxon occur
ring in the south-eastern part of France; the correct name of this taxon in the genus Cyanus
is Cyanus graminifolius (Lam.) Olšavská. The current karyological study indicates that
C. graminifolius differs in its basic chromosome number and ploidy level from other taxa
of the C. triumfetti group and is recognized as a separate taxon (Olšavská & Perný 2009).
Thus, on the one hand, there are plants occurring in Central, eastern and southern Europe
90 Preslia 83: 59–98, 2011
bearing the name C. axillaris according to numerous authors (e.g. Presl & Presl 1819,
Čelakovský 1871, Velenovský 1891, Hayek 1901, Adamovič 1911, Jávorka 1924, Polívka
et al. 1928, Hegi 1929, Dostál 1931a, 1931b, 1950, 1954, 1989, Hayek & Markgraf 1931,
Stefanov & Georgiev 1931, Prodan & Nyárády 1964, Soó & Kárpáti 1968, Soó 1970,
1972, Marhold & Hindák 1998, Štěpánek 2002, 2004), whereas on the other hand, the
name C. axillaris is typified by the type of C. graminifolius and belongs to the French
plants according to the rules of the Code (Art. 7.5.; McNeill et al. 2006). Štěpánek (2004)
tried to resolve this problem by introducing a new combination Cyanus triumfetti subsp.
axillaris (J. Presl & C. Presl) Štěpánek. Štěpánek treated the name Cyanus axillaris J. Presl
& C. Presl as the legitimate name of a new species because when Presl & Presl (1819) pub
lished this name, they did not refer to Willdenow or any other author. The most valid opin
ion in this case is that of Greuter which treats Presls’ name as a combination based on an
illegitimate name (Greuter 2006–2009) according to the rules of the Code (Art. 33.3.;
McNeill et al. 2006). Article 33.3. of the Code should be applied here because Presls did
not refer to the authors of the basionyms also for some other names [e.g. Presls did not
refer to Aira glauca Schrad. (that they used in the genus Koeleria Pers.), to Tussilago alba
L. (used in the genus Petasites Mill.), to Centaurea nigrescens Willd. (used in the genus
Cyanus)]. Therefore, the name Cyanus axillaris J. Presl et C. Presl should be correctly
cited as Cyanus axillaris (Willd.) J. Presl et C. Presl. As the basionymum for new combi-
nations at the rank of subspecies, Greuter prefers to use the name Centaurea montana
subsp. axillaris (Willd.) Čelak published by Čelakovský (1871). In some of the studies
that concentrate on Central and eastern European flora, authors prefer to use the name
Centaurea triumfetti subsp. aligera (Gugler) Dostál instead of Cyanus axillaris (Dostál
1976a, Soó 1980, Wagenitz 1987, Meusel & Jäger 1992, Simon 1992, Horváth et al. 1995,
Ciocârlan 2000, Haupler & Muer 2000, Oprea 2005). This name was published by Gugler
(1907) as Centaurea variegata var. aligera Gugler. While Gugler referred to Willdenow’s
name in the protologue and did not adopt the epithet “axillaris”, the name Centaurea
variegata var. aligera is also treated as nomen illegitimum (cf. Greuter 2006–2009)
according to the rules of the Code (Art. 52.1., 52.2.; McNeill et al. 2006). As a result, all
the available names, and any combinations based on them referred to Willdenow’s name
Centaurea axillaris Willd., have to be typified by the type of Cyanus graminifolius. Thus,
the taxon that is now called Cyanus axillaris or Centaurea triumfetti subsp. aligera differs
from the taxon to which those names were applied when following the rules of the Code.
To resolve this problem, a new name needs to be chosen for European populations treated
currently as Cyanus axillaris, auct.
Taxonomic treatment and typification of Cyanus dominii
Cyanus dominii (Dostál) Holub in Folia Geobot. Phytotax. 12: 308. 1977
Cyanus dominii (Dostál) Holub subsp. dominii
Basionym: Centaurea triumfetti subsp. dominii Dostál in Acta Bot. Bohem. 10: 71. 1931.
Ind. loc.: “Slovakia centralis, in rupibus trachyticis montis Bránisko, altitude 800 m s. m., leg. Jos. Dostál, 1928”.
Neotype (designated here): “Slovakia, Volovské vrchy Mts, Mt. Humenec near Veľká Lodina village; 48°51'34"
N, 21°09'33" E; 280 m; leg. & det. K. Olšavská, 26. 7. 2007” (SAV).
Olšavská et al.: Cyanus triumfetii group in Central Europe 91
≡ Centaurea dominii (Dostál) Dubovik in Bot. Zhurn. (Moscow & Leningrad)75: 1579. 1990.
≡ Cyanus strictus subsp. dominii (Dostál) Soják in Čas. Nár. Muz., Odd. Přír. 140: 131. 1972.
≡ Cyanus triumfetti subsp. dominii (Dostál) Dostál in Folia Mus. Rerum Nat. Bohemiae Occid., Bot. 21: 14. 1984.
– Centaurea triumfetti subsp. dominii var. eu-Dominii Dostál in Acta Bot. Bohem. 10: 71. 1931, nom. inval.
Note: Neotype of Cyanus dominii was designated here because the holotype cited in
protologue of the name Centaurea triumfetti subsp. dominii and other original material
belonging to this name is missing. The specimen collected at Veľká Lodina in the Volovské
vrchy Mts fits the protologue as well as our results and will serve as the nomenclatural type
of this taxon until Dostál’s original material is rediscovered.
Cyanus dominii subsp. slovenicus (Dostál) Olšavská, comb. nova
Basionym: Centaurea triumfetti subsp. dominii var. slovenica Dostál in Acta Bot. Bohem. 10: 71. 1931.
Ind. loc.: “Slovakia, in Carpatis occidentalibus”. Lectotype (designated here): “Centaurea Triumfetti All. ssp.
Dominii Dostál var. slovenica Dostál, Lubochňa – před hotelem Bratislava, Det. Dr. Jos. Dostál, 27/VII. 1919” (PRC).
Note: The herbarium sheet from the surrounding of Lubochňa village in the north part of
the Veľká Fatra Mts determined by Dostál in 1929 is selected here as a lectotype of Cyanus
dominii subsp. slovenicus. In the protologue of Centaurea triumfetti subsp. dominii var.
slovenica (Dostál 1931a) no specimens are cited, so the most representative specimen
from the area mentioned in the protologue (“Slovakia, in Carpatis occidentalibus”) was
Cyanus dominii subsp. sokolensis (Pawł.) Olšavská, comb. nova
Basionym: Centaurea axillaris var. sokolensis Pawł. in Kosmos (Lvov) 55: 70. 1930.
Ind. loc.: “Skały wapienne Sokol 1235–1320 m; Mnich 1462 m”. Lectotype (designated here): “Centaurea
axillaris W. var. sokolensis m. ad int., Tatry, pasmo Sivego Wierchu, Góra Sokol 1320 m, skały wap., B.
Pawłowski., 27. 7. 1927” (KRAM 305241).
≡ Centaurea triumfetti subsp. sokolensis (Pawł.) Dostál in Květena ČSR: 1686. 1950 (p. p.)
≡ Centaurea triumfetti subsp. dominii var. densifolia f. sokolensis (Pawł.) Dostál in Acta Bot. Bohem. 10: 72. 1931
= Centaurea triumfetti subsp. dominii var. densifolia Dostál in Acta Bot. Bohem. 10: 72. 1931
= Centaurea triumfetti subsp. dominii var. densifolia f. densifolia Dostál in Acta Bot. Bohem. 10: 72. 1931
Note: One of two syntypes cited in the protologue of the name Centaurea axillaris var.
sokolensis Pawł. (Pawłowski 1931) is here selected as the lectotype of Cyanus dominii
subsp. sokolensis. The specimen collected on Mt. Sokol is selected as lectotype because
Cyanus dominii subsp. sokolensis was named after this mountain.
See http://www.preslia.cz for Electronic Appendix 1
We thank Birgit Gemeinholzer and Kim Govers for their assistance with the AFLP analyses, Barbora Šingliarová
and Stanislav Španiel for their assistance with flow cytometric analyses, Karol Marhold and Pavol Mereďa Jr. for
valuable discussions, Zlata Komárová for drawing the plants, Dušan Senko for his help with map preparation and
the persons listed in Table 1 for their assistance in the field. Tony Dixon kindly corrected our English. This study was
supported by the Grant Agency of the Ministry of Education of the Slovak Republic and the Slovak Academy of Sci
ences (project no. VEGA 2/0053/08) and the EDIT scholarship for outstanding women scientists 2008 (to K. O.).
92 Preslia 83: 59–98, 2011
Morfometrická analýza 71 populácií (1148 jedincov) skupiny Cyanus triumfetti zo strednej Európy ukázala, že na
tomto území možno rozlíšiť tri druhy: ‘C. axillaris’, C. strictus a C. dominii. Cyanus dominii odlišujú od ‘C. axil
laris’aC. strictus nezbiehavé alebo krátko-zbiehavé byľové listy; k tomuto druhu patria populácie zo stredného
a východného Slovenska. Cyanus strictus a‘C. axillaris’ sa líšia farbou bŕv príveskov zákrovných listeňov, po
čtom a tvarom byľových listov. K druhu C. strictus boli priradené populácie zo severovýchodnej časti Maďarska
a juhovýchodného Slovenska, k druhu ‘C. axillaris’ patria všetky študované populácie z Rakúska a Českej repub
liky a niektoré populácie z Maďarska a Slovenska. Výsledky morfologických analýz potvrdili veľkú vnútrodruho
vú variabilitu stredoeurópskych populácií C. dominii a možnosť ich klasifikácie do troch poddruhov: C. dominii
subsp. dominii (Volovské vrchy, Branisko), C. dominii subsp. slovenicus (stredné Slovensko) a C. dominii subsp.
sokolensis (Slovenský kras, Západné Tatry). Kultivačné experimenty ukázali, že ‘C. axillaris’, C. strictus a C. do
minii sa líšia v miere ekologickej plasticity, pričom najvýraznejšie ekologicky ovplyvnené boli byľové a listové
znaky. Posun v analyzovaných znakoch neovplyvnil možnosť jednotlivé taxóny rozlíšiť a rozdiely medzi taxónmi
zostali zachované aj pri pestovaní v rovnakých podmienkach. Analýza AFLP markerov 38 populácií zo skupiny
C. triumfetti a dvoch populácií zo skupiny C. montanus (55 jedincov) odhalila tri geneticky odlišné a alopatrické
skupiny: (1) C. triumfetti s.s. a C. montanus zo Západných Álp, (2) ‘C. axillaris’ z Rakúska a Českej republiky
(okrem Karpát), a (3) ‘C. axillaris’, C. strictus a C. dominii zo Západných Karpát a Panónie. Genetická separácia
vzoriek C. triumfetti s.s. potvrdila výsledky predchádzajúcich morfologických analýz, ktoré vylúčili výskyt tohto
druhu na území Strednej Európy. Stredoeurópske vzorky skupiny C. triumfetti boli rozdelené do dvoch skupín,
ktoré nekorelovali s taxónmi potvrdenými na základe morfologickej diferenciácie. Analýza AFLP markerov uká
zala genetické odčlenenie vzoriek z Rakúska a Českej republiky patriacich k druhu ‘C. axillaris’, ako aj veľkú ge
netická variabilitu a slabú genetickú diferenciáciu vzoriek zo Západných Karpát a Panónie patriacich druhom
‘C. axillaris’, C. strictus a C. dominii. Získaný obraz genetickej variability je diskutovaný v súvislosti s potenciál-
nymi glaciálnymi refúgiami, cestami postglaciálnej migrácie a hybridizačnými udalosťami, ktoré sa mohli
vyskytovať v evolučnej histórii skupiny. Ploidná úroveň študovaných populácií bola stanovená pomocou
prietokovej cytometrie: diploidná ploidná úroveň (2n ~ 2x ~ 22) bola potvrdená pre jedincov zo skupiny
C. triumfetti a tetraploidná (2n ~ 4x ~ 44) pre jedincov zo skupiny C. montanus.
Adamovič L. (1911): Flora jugoistochne Srbije (Flora serbiae Austro-Orientalis). – Dionichke Tiskare, Zagreb.
Allioni C. (1774): Auctuarium ad synopsim methodicam stirpium horti regii taurinensis. – Mélanges Philos.
Math. Soc. Roy. Turin (1770–1773) 5: 53–96.
Álvarez Fernández I. J. & Nieto Feliner G. (2001): A multivariate approach to assess the taxonomic utility of
morphometric characters in Doronicum (Asteraceae, Senecioneae). – Folia Geobot. 36: 423–444.
Andreev N., Anchev M., Kozhuharov S., Markova M., Peev D. & Petrova A. (1992): Opredelitel na visshite
rastenia v Bulgaria [Identification key to the vascular plants in Bulgaria]. – Nauka i izkustvo, Sofia.
Applied Biosystems (2005): AFLP® Plant mapping. Protocol. – PE Applied Biosystems, Foster City.
Archak S., Gaikwad A. B., Gautam D., Rao E. V. V. B., Swamy K. R. M. & Karihaloo J. L. (2003): Comparative
assessment of DNA fingerprinting techniques (RAPD, ISSR and AFLP) for genetic analysis of cashew
(Anacardium occidentale L.) accessions of India. – Genome 46: 362–369.
Baltisberger M. (1991): Cytological investigations of some Greek plants. – Fl. Medit. 1: 157–173.
Bancheva S. T., Geraci A. & Raimondo F. M. (2009): Variabilità genetica in tre specie del genere Cyanus
(Centaureinae, Asteraceae) della Bulgaria [Genetic variation within three species of the genus Cyanus
(Centaureinae, Asteraceae) from Bulgaria]. – In: Anonymus (ed.), Book of abstracts, 104. Congresso della
Società Botanica Italiana Onlus, Campobasso, p. 3, Società Botanica Italiana, Campobasso.
Bancheva S. T. & Greilhuber J. (2006): Genome size in Bulgarian Centaurea s.l. (Asteraceae). – Pl. Syst. Evol.
Bancheva S. T. & Raimondo F. M. (2003): Biosystematic studies of seven Balkan species from genus Cyanus
(Compositae). – Bocconea 16: 507–527.
Bonin A., Bellemain E., Eidesen P. B., Pompanon F., Brochmann C. & Taberlet P. (2004): How to track and assess
genotyping error in population genetic studies. – Mol. Ecol. 13: 3261–3273.
Borhidi I. (1957): Centaurea achtarovii Urumoff dans les Carpathes orientaux. Annotations sur le groupe
Centaurea triumfettii All. – Ann. Hist.-Nat. Mus. Natl. Hung. 7: 219–225.
Bussell J. D., Waycott M. & Chappill J. A. (2005): Arbitrarily amplified DNA markers as characters for phylo
genetic inference. – Persp. Pl. Ecol. Evol. Syst. 7: 3–26.
Olšavská et al.: Cyanus triumfetii group in Central Europe 93
Čelakovský L. (1871): Prodomus der Flora von Böhmen, Vol. 2. – Comité für die naturwissenschaftliche
Durchforschung Böhmens, Prag.
Chauvet S., Van Der Velde M., Imbert E., Guillemin M. L., Mayol M., Riba M., Smulders M. J. M., Vosman B.,
Ericson L., Bijlsma R. & Giles B . E. (2004): Past and current gene flow in the selfing, wind-dispersed species
Mycelis muralis in western Europe. – Mol. Ecol. 13: 1391–1407.
Chrtek J. & Skočdopolová B. (1982): Waldstein’s collection in the herbarium of the National Museum in
Prague. – Sborn. Nár. Mus. v Praze, Přír. Vědy 38B: 201–238.
Ciocârlan V. (2000): Flora ilustrată a României [Illustrated flora of Romania]. – Editura Ceres, Bucureşti.
Conti F. (2007): Morphometric study and taxonomy of Genista pulchella Vis. s.l. (Fabaceae), a south European
species. – Bot. J. Linn. Soc. 153: 245–254.
Corander J., Marttinen P., Sirén J. & Tang J. (2006): BAPS: Bayesian analysis of population structure, Manual v.
4.1. – Department of Mathematics, University of Helsinki, Helsinki, URL: [http://www.rni.helsinki.fi/-
Czerepanov S. K. (1995): Vascular plants of Russia and adjacent states (the former USSR). – Cambridge Univer
sity Press. Cambridge.
Depypere L., Chaerle P., Breyne P., Vander Mijnsbrugge K. & Goetghebeur P. (2009): A combined morphometric
and AFLP based diversity study challenges the taxonomy of the European members of the complex Prunus L.
section Prunus. – Pl. Syst. Evol. 279: 219–231.
Dobrochayeva D. M. (1949): Voloshky URSR, ikh poshyrennya ta istoriya rozvytku [Knapweeds of USSR, their
distribution and evolution]. – Bot. Zhurn. 6: 63–77.
Doležel J., Greilhuber J. & Suda J. (2007): Estimation of nuclear DNA content in plants using flow cytometry. –
Nature Protoc. 2: 2233–2244.
Dostál J. (1931a): Distributio speciarum, varietatum, formarumque Čechoslovenicarum Centaureae e Sectione
Montanae. – Acta Bot. Bohem. 10: 67–74.
Dostál J. (1931b): Étude préliminaire de l’evolution et de la systématique des types du genre Centaurea sect.
Cyanus subsec. Montanae. – Preslia 10: 57–69.
Dostál J. (1950): Květena ČSR [Flora of Czechoslovakia]. – Nakladatelství ČSAV, Praha.
Dostál J. (1954): Klíč k úplné květeně ČSR [Key to the complete flora of ČSR]. – Nakladatelství ČSAV, Praha.
Dostál J. (1969): Taxonomická studie o čeledích Lycopodiaceae a Asteraceae a o rodech Lycopodium
a Centaurea [Taxonomic study about families Lycopodiaceae and Asteraceae and about genera Lycopodium
and Centaurea]. – Acta Univ. Palack. Olomuc. Fac. Rerum Nat., Ser. 2, Biol. 31: 5–35.
Dostál J. (1976a): New nomeclatural combinations and taxa of the Compositae subtribe Centaurinae in Europe. –
In: Heywood V. H. (ed.), Flora Europaea. Notulae systematicae ad Floram Europaeam spectantes No. 18, Bot.
J. Linn. Soc. 71: 191–210.
Dostál J. (1976b): Centaurea L. – In: Tutin T. G., Heywood V. H., Burges N. A., Moore D. M., Valentine D. H.,
Walters S. M. & Webb D. A. (eds), Flora Europaea 4: 254–301, Cambridge University Press, Cambridge.
Dostál J. (1989): Nová květena Československé socialistické republiky [New flora of the Czechoslovak Socialist
Republic]. – Academia, Praha.
Ekrt L. & Štech M. (2008): A morphometric study and revision of the Asplenium trichomanes group in the Czech
Republic. – Preslia 80: 325–347.
Fér T., Vašák P., Vojta J. & Marhold K. (2007): Out of the Alps or Carpathians? Origin of Central European popu
lations of Rosa pendulina. – Preslia 79: 367–376.
Font M., Garcia-Jacas N., Vilatersana R., Roquet C. & Susanna A. (2009): Evolution and biogeography of
Centaurea section Acrocentron inferred from nuclear and plastid DNA sequence analyses. – Ann. Bot. 103:
Galbany-Casals M., Sáez L. & Benedí C. (2006): A taxonomic revision of Helichrysum sect. Stoechadina
(Asteraceae, Gnaphalieae). – Can. J. Bot. 84: 1203–1232.
Garcia-Jacas N., Susanna A., Garnatje T. & Vilatersana R. (2001): Generic delimitation and phylogeny of the
subtribe Centaurinae (Asteraceae
): a combined nuclear and chloropast DNA analysis. – Ann. Bot. 87:
Garcia-Jacas N., Uysal T., Romashchenko K., Suárez-Santiago V. N., Ertugˇrul K. & Susanna A. (2006):
Centaurea revisited: a molecular survey of the Jacea group. – Ann. Bot. 98: 741–753.
Goldblatt P. & Johnson D. E. (eds) (1979–2010): Index to plant chromosome numbers. – Missouri Botanical Gar
den, St. Louis, URL: [http://www.mobot.mobot.org/w3T/search/ipcn.html; accessed 1 March 2010].
Greimler J. & Jang C.-G. (2007): Gentianella stiriaca, a case of reticulate evolution in the northeastern and east
ern Central Alps. – Taxon 56: 857–870.
94 Preslia 83: 59–98, 2011
Greuter W. (2006–2009): Compositae (pro parte majore). – In: Greuter W. & Raab-Straube E. von (eds),
Compositae, Euro+Med Plantbase: the information resource for Euro-Mediterranean plant diversity, URL:
[http://ww2.bgbm.org/EuroPlusMed; accessed 1 March 2010].
Gugler W. (1907): Die Centaureen des Ungarischen Nationalmuseums. – Ann. Hist.-Nat. Mus. Natl. Hung. 6:
Guinochet M. (1957): Contribution à l’étude caryologique du genre Centaurea L. sens. lat. – Bull. Soc. Hist. Nat.
Afrique N. 48: 282–300.
Haeupler H. & Muer T. (2000): Bildatlas der Farn- und Blütenpflanzen Deutschlands. – Verlag Eugen Ulmer,
Hayek A. von (1901): Die Centaurea-Arten Österreich-Ungarns. – Denkschr. Kaiserl. Akad. Wiss., Wien. Math.-
Naturwiss. Kl. 72: 595–773.
Hayek A. von & Markgraf F. (1931): Prodromus florae peninsulae Balcanicae, Vol. 2. – Repert. Spec. Nov. Regni
Veg. Beih. 30: 577–960.
Hegi G. (1929): Illustrierte Flora von Mitteleuropa, Vol. 4(2). – J. F. Lehmanns Verlag, München.
Hellwig F. H. (2004): Centaureinae (Asteraceae) in the Mediterranean: history of ecogeographical radiation. –
Pl. Syst. Evol. 246: 137–162.
Hodálová I. (1999): Multivariate analysis of the Senecio nemorensis group (Compositae) in the Carpathians with
a new species from the East Carpathians. – Folia Geobot. 34: 321–335.
Horváth F., Dobolyi Z. K., Morschhauser T., Lőkös L., Karas L. & Szerdahelyi T. (1995): Flóra adatbázis,
taxonlista és attribútum-állomány 1.2 [Flora database 1.2. Taxon-list and store of the attributes]. – MTA
Ökológia és Botanikai Kutatóintézete, Vácrátót.
Jávorka S. (1924): Magyar Flóra [Hungarian Flora]. – A Studium Kiadása, Budapest.
Jávorka S. (1926): Kitaibel Herbáriuma [Herbarium Kitaibelianum]. – Ann. Hist.-Nat. Mus. Natl. Hung. 24:
Kamada T., Yamashiro T. & Maki M. (2007): Intraspecific morphological and genetic differentiation in
Scrophularia grayana (Scrophulariaceae). – J. Pl. Res. 120: 437–443.
Klecka W. R. (1980): Discriminant analysis. – In: Sage University Papers, Series: Quantitative Applications in
the Social Sciences no. 19, Sage Publications, Beverly Hills, London.
Kliment J. (1999): Komentovaný prehľad vyšších rastlín flóry Slovenska, uvádzaných v literatúre ako endemické
taxóny [Commented list of the vascular plants of Slovakia reported in literature as endemic taxa]. – Bull. Slov.
Bot. Spoločn. 21 (Suppl. 4): 1–434.
Kochjarová J. (2005): Scilla bifolia group in the Western Carpathians and adjacent part of the Pannonian lowland:
annotated chromosome counts. – Preslia 77: 317–326.
Koutecký P. (2007): Morphological and ploidy level variation of Centaurea phrygia agg. (Asteraceae) in the
Czech Republic, Slovakia and Ukraine. – Folia Geobot. 42: 77–102.
Krähenbühl M. & Küpfer P. (1992): [Reports no. 92–97]. – In: Kamari G., Felber F. & Garbari F. (eds), Mediterra
nean chromosome number reports 2, Fl. Medit. 2: 255–258.
Krippel E. (1986): Postglaciálny vývoj vegetácie Slovenska [Postglacial vegetation development in Slovakia]. –
Kropf M. (2008): Intraspecific patterns of European mountain plants: a morphometric analysis confirms molecu
lar results in the submediterranean oreophyte Anthyllis montana L. (Fabaceae) – Taxon 57: 511–524.
Krzanowski W. J. (1990): Principles of multivariate analysis. – Clarendon Press, Oxford.
Lihová J., Kučera J., Perný M. & Marhold K. (2007): Hybridization between two polyploid Cardamine
(Brassicaceae) species in northwestern Spain: discordance between morphological and genetic variation pat
terns. – Ann. Bot. 99: 1083–1096.
Lihová J., Marhold K., Tribsch A. & Stuessy T. F. (2004). Morphometric and AFLP Re-evaluation of tetraploid
Cardamine amara (Brassicaceae) in the Mediterranean. – Syst. Bot. 29: 134–146.
Lipper W. & Heubl G. R. (1988): Chromosomenzahlen von Pflanzen aus Bayern und angrenzenden Gebieten. –
Ber. Bayer. Bot. Ges. 59: 13–22.
Löser C. J., Akaydin G. & Hellwig F. H. (2009): Incomplete lineage sorting among annuals and perennials in the
genus Cyanus. – In: Anonymus (ed.), Program and Abstract, Systematics 2009: 1st Meeting of BioSyst EU,
Leiden, p. 63–64, National Herbarium of the Netherlands & National Museum of Natural History, Leiden.
Lovrić A.-Z. (1982): [Reports]. – In: Löve A. (ed.), IOPB chromosome number reports LXXVII, Taxon 31:
Marhold K. & Hindák F. (eds) (1998): Checklist of non vascular and vascular plants of Slovakia. – Veda,
Olšavská et al.: Cyanus triumfetii group in Central Europe 95
Marhold K., Huthmann M. & Hurka H. (2002): Evolutionary history of the polyploid complex of Cardamine
amara (Brassicaceae): isozyme evidence. – Pl. Syst. Evol. 233: 15–28.
Marhold K., Jongepierová I., Krahulcová A. & Kučera J. (2005): Morphological and karyological differentiation
of Gymnadenia densiflora and G. conopsea in the Czech Republic and Slovakia. – Preslia 77: 159–176.
Marhold K., Mártonfi P., Mereďa P. jun. & Mráz P. (eds) (2007): Chromosome number survey of the ferns and
flowering plants of Slovakia. – Veda, Bratislava.
McNeill J., Barrie F. R., Burdet H. M., Demoulin V., Hawksworth D. L., Marhold K., Nicolson D. H., Prado J.,
Silva P. C., Skog J. E., Wiersema F. H. & Turland N. J. (eds) (2006): International Code of Botanical Nomen
clature (Vienna Code). Adopted by the Seventeenth International Botanical Congress Vienna, Austria, July
2005. – Regnum Vegetabile 146: 1–568, A. R. G. Gantner Verlag, Ruggell, Liechtenstein.
Mereďa P. jun., Hodálová I., Mártonfi P., Kučera J. & Lihová J. (2008): Intraspecific variation in Viola suavis in
Europe: parallel evolution of white-flowered morphotypes. – Ann. Bot. 102: 443–462.
Meudt H. M. & Clarke A. C. (2007): Almost forgotten or latest practice? AFLP applications, analyses and
advances. – Trends Pl. Sci. 12: 106–117.
Meusel H. & Jäger E. J. (1992): Vergleichende Chorologie der zentraleuropäischen Flora 3. – Gustav Fischer
Millionová H. (2000): Taxonomická studie vytrvalých druhů rodu Cyanus Mill. v České republice [Taxonomic
study of perennial species of the genus Cyanus Mill. in Czech Republic]. – Dipl. Thesis, Department of Bot
any, Charles University, Prague.
Mitka J., Sutkowska A., Ilnicki T. & Joachimiak A. J. (2007): Reticulate evolution of high-alpine Aconitum
(Ranunculaceae) in the Eastern Sudetes and Western Carpathians (Central Europe). – Acta Biol. Cracov., Ser.
Bot. 49: 15–26.
Monteserrat Martí J. M. (1987): Números cromosomáticos de plantas occidentales, 420–421. – Anales Jard. Bot.
Madrid. 43: 425–426.
Morales F. M. J. (1974): [Reports]. – In: Löve Á. (ed.), IOPB chromosome number reports XLVI, Taxon 23: 805.
Mosyakin S. L. & Fedoronchuk M. M. (1999): Vascular plants of Ukraine: a nomenclatural checklist. – National
Academy of Sciences of Ukraine, M. G. Kholodny Institute of Botany, Kiev.
Mráz P., Gaudeul M., Rioux D., Gielly L., Choler P., Taberlet P. & the IntraBioDiv Consortium (2007): Genetic
structure of Hypochaeris uniflora (Asteraceae) suggests vicariance in the Carpathians and rapid post-glacial
colonization of the Alps from an eastern Alpine refugium. – J. Biogeogr. 34: 2100–2114.
Nei M. & Li W.-H. (1979): Mathematic model for studying genetic variation in terms of restriction
endonucleases. – Proc. Natl. Acad. Sci. U.S.A. 76: 5269–5273.
Nikolić T. (2000): Flora Croatica: Index florae Croaticae. Pars 3. – Nat. Croat. 9 (Suppl. 1): 1–324.
Olšavská K. & Perný M. (2009): Cyanus graminifolius. – In: Marhold K. (ed.), IAPT/IOPB chromosome data 7,
Taxon 58: 181–183.
Olšavská K., Perný M., Mártonfi P. & Hodálová I. (2009): Cyanus triumfettii subsp. triumfettii (Compositae) does
not occur in the western Carpathians and adjacent parts of Pannonia: karyological and morphological evi
dence. – Nord. J. Bot. 27: 21–36.
Oprea A. (2005): Lista critică a plantelor vasculare din România [Checkist of the vascular flora of Romania]. –
Editura Universitatii Alexandru Ioan Cuza, Iaşi.
Pawłowski B. (1931): Spis ważniejszych roślin, znalezionych w Tatrach słowackich w grupie Siwego Wierchu
i u jej podnóża [List of remarkable plants recorded in Slovak Tatras in Sivý vrch group and its foothills]. –
Kosmos 55: 695–711.
Pečinka A., Suchánková P., Lysak M. A., Trávníček B. & Doležel J. (2006): Nuclear DNA content variation
among Central European Koeleria taxa. – Ann. Bot. 98: 117–122.
Pérez-Barrales R., Pino R., Albaladejo R. G. & Arroyo J. (2009): Geographic variation of flower traits in Narcis
sus papyraceus (Amaryllidaceae): do pollinators matter? – J. Biogeogr. 36: 1411–1422.
Perný M., Tribsch A. & Anchev M. E. (2004): Intraspecific differentiation in the Balkan diploid Cardamine acris
(Brassicaceae): molecular and morphological evidence. – Folia Geobot. 39: 405–429.
Peruzzi L. & Passalacqua N. G. (2008): Taxonomy of the Onosma echioides (L.) L. complex (Boraginaceae)
based on morphometric analysis. – Bot. J. Linn. Soc. 157: 763–774.
Podani J. (2001): SYN-TAX 2000. Computer programs for multivariate data analysis in ecology and systematics.
User’s manual. – Scientia Publishing, Budapest.
Polívka F., Domin K. & Podpěra J. (1928): Klíč k úplné květeně republiky československé [Key to the complete
flora of the Czechoslovak republic]. – R. Promberger, Olomouc.
Presl J. S. & Presl C. B. (1819): Flora Čechica [Czech Flora]. – In comissis apud J. G. Calve, Pragae.
96 Preslia 83: 59–98, 2011
Prodan J. & Nyárády E. I. (1964): Centaurea L. – In: Săvulescu T. (ed.), Flora republicii populare Romîne [Flora
of People's Republic of Romania], Vol. 9, p. 785–951, Editura Academiei Republicii Populare Romîne,
Roldán-Ruiz I., Gilliland T. J., Calsyn E., Gillilad T. J., Coll R., Van Eijk M. J. T. & De Loose M. (2001): Estimat
ing genetic conformity between related ryegrass (Lolium) varities. II. AFLP characterization. – Molec. Breed.
SAS Institute (2000): SAS online Doc
, Version 8. – SAS Institute, Cary.
Schönswetter P., Solstad H., García P. E. & Elven R. (2009): A combined molecular and morphological approach
to the taxonomically intricate European mountain plant Papaver alpinum s.l. (Papaveraceae): taxa or infor
mal phylogeographical groups? – Taxon 58: 1326–1343.
Shapiro S. S. & Wilk M. B. (1965): An analysis of variance test for normality (complete samples). – Biometrika
Sharkova S. T. (1996): [Reports]. – In: Kamari G., Felber F. & Garbari F. (eds), Mediterranean chromosome num
ber reports 6, Fl. Medit. 6: 328–333.
Simon T. (1992): A magyarországi edényes flóra határozója [Identification key of the vascular flora of Hungary].
– Tankönyvkiadó, Budapest.
Skrede I., Eidesen P. B., Portela R. P. & Brochmann C. (2006): Refugia, differentiation and postglacial migration
in arcticalpine Eurasia, exemplified by the mountain avens (Dryas octopetala L.) – Mol. Ecol. 15:
Slovák M. & Marhold K. (2007): Morphological evaluation of Picris hieracioides L. (Compositae, Lactuceae)in
Slovakia. – Phyton (Horn) 47: 73–102.
Šmarda P., Šmerda J., Knoll A., Bureš P. & Danihelka J. (2007): Revision of Central European taxa of Festuca ser.
psammophilae Pawlus: morphometrical, karyological and AFLP analysis. – Pl. Syst. Evol. 266: 197–232.
Sneath P. H. A. & Sokal R. R. (1973): Numerical taxonomy. – W. H. Freeman, San Francisco.
Soó R. (1970): A Magyar flóra és vegetácio rendsztertani-növényveföldrajz kézikönyve [Synopsis of the flora
and vegetation of Hungary], Vol. 4.– Akadémiai Kiadó, Budapest.
Soó R. (1972): Systematisch-nomenklatorische Bemerkungen zur Flora Mitteleuropas mit Beziehungen zur
südosteuropäischen Flora. – Feddes Repert. 83: 129–212.
Soó R. (1980): A Magyar flóra és vegetácio rendsztertani-növényveföldrajz kézikönyve [Synopsis of flora and
vegetation of Hungary], Vol. 6.– Akadémiai Kiadó, Budapest.
Soó R. & Kárpáti Z. (1968): Magyar flóra. Harasztok – Virágos növények. Növényhatározó II. [Hungarian flora.
Ferns – Flowering plants. Identification key II]. – Tankönyvkiadó, Budapest.
Španiel S., Marhold K., Hodálová I. & Lihová J. (2008): Diploid and tetraploid cytotypes of Centaurea stoebe
(Asteraceae) in Central Europe: morphological differentiation and cytotype distribution patterns. – Folia
Geobot. 43: 131–158.
Stefanov B. & Georgiev T. (1931): Prinos za razgranichenie na vidovet’ na roda Centaurea L. ot sect. Cyanus D.
C. [Contribution to the distribution of species of the genus Centaurea L.] – Spis. Bulg. Akad. Nauk. 44:
Štěpánek J. (2002): Centaurea L. – In: Kubát K., Hrouda L., Chrtek jun. J., Kaplan Z., Kirschner J. & Štěpánek J.
(eds), Klíč ke květeně České republiky [Key to the flora of the Czech Republic], p. 673–676, Academia,
Štěpánek J. (2004). Cyanus Mill. – In: Štěpánek J. & Štěpánková J. (eds), Květena České republiky [Flora of the
Czech Republic] 7: 451–458, Academia, Praha.
Strid A. & Franzen I. A. (1981): [Reports]. – In: Löve Á. (ed.), IOPB chromosome number reports LXXIII, Taxon
Stuessy T. F. (2009): Plant taxonomy: the systematic evaluation of comparative data. Ed. 2. – Columbia Univer
sity Press, New York.
Suda J. & Lysák M. A. (2001): A taxonomic study of the Vaccinium sect. Oxycoccus (Hill) W. D. J. Koch
(Ericaceae) in the Czech Republic and adjacent territories. – Folia Geobot. 36: 303–319.
Susanna A. & Garcia-Jacas N. (2007): Compositae: tribe Cardueae Cass. – In: Kadereit J. W. & Jeffrey C. (eds),
Flowering plants. Eudicots. Asterales, p. 123–147, Berlin, Heidelberg.
Susanna A., Garcia-Jacas N., Soltis D. E. & Soltis P. S. (1995): Phylogenetic relationship in tribe Cardueae
(Asteraceae) based on ITS sequences. – Amer. J. Bot. 82: 1056–1068.
Szafer W., Kulczyński S. & Pawłowski B. (1924): Rośliny polskie [Plants of Poland].–Książnica Atlas, Lwów,
Olšavská et al.: Cyanus triumfetii group in Central Europe 97
Van de Peer Y. & De Wachter R. (1994): TREECON for Windows : a software package for the construction and
drawing of evolutionary trees for the Microsoft Windows environment. – Computer Applic. Biosci. 10:
Velenovský J. (1891): Flora Bulgarica – Fr. Řivnáč, Pragae.
Vos P., Hogers R., Bleeker M., Reijans M., Van de Lee T., Hornes M., Frijters A., Pot J., Poleman J., Kuiper M. &
Zabeau M. (1995): AFLP: a new technique for DNA fingerprinting. – Nucl. Acids Res. 23: 4407–4414.
Wagenitz G. (1987): Hegi, Illustrierte Flora von Mitteleuropa, Ed. 2, Vol. 4 (2): Abchträge, Berichtigungen und
Ergänzungen zum Nachdruck der 1. Auflage von Band VI/2 (1928/9). – Verlag Paul Parey, Berlin &
Waldstein F. & Kitaibel P. (1805): Descriptiones et icones plantarum rariorum Hungariae. Vol. 2. – Matthiae
Andreae Schmidt, Wien.
Willdenow C. L. (1803): Species plantarum. Ed. 3, Vol. 3 (3). – Impensis G. C. Nauk, Berolini.
Received 19 April 2010
Revision received 31 August 2010
Accepted 1 September 2010
98 Preslia 83: 59–98, 2011