Int. J. Plant Sci. 166(1):117–133. 2005.
Ó 2005 by The University of Chicago. All rights reserved.
1058-5893/2005/16601-0011$15.00
PHYLOGENY OF TRAGOPOGON L. (ASTERACEAE) BASED ON INTERNAL AND
EXTERNAL TRANSCRIBED SPACER SEQUENCE DATA
Evgeny V. Mavrodiev,1 ,*,y Mark Tancig,* Anna M. Sherwood,z Matthew A. Gitzendanner,2 ,*,y
Jennifer Rocca,*,y Pamela S. Soltis,3 ,y and Douglas E. Soltis4 ,*
*Department of Botany, University of Florida, Gainesville, Florida 32611, U.S.A.; yFlorida Museum of Natural History, University of Florida,
Gainesville, Florida 32611, U.S.A.; and zSchool of Biological Sciences, Washington State University, Pullman, Washington 99164, U.S.A.
Tragopogon L. (Lactuceae, Cichorioideae, Scorzonerinae) is an Old World genus of ca. 150 species. To
elucidate relationships in this poorly understood genus, we analyzed internal and external transcribed spacer
sequences (ITS and ETS, respectively) from 59 species from 14 of the 17 sections recognized by Borisova.
Podospermum and Lactuca were used as outgroups. ITS and ETS sequence data provide strong support for the
monophyly of Tragopogon under parsimony optimality criteria. Sequence data also support the monophyly of
several of the recognized sections within Tragopogon, including the sections Tragopogon, Majores,
Hebecarpus, Chromopappus, and Collini. However, the circumscriptions of these groups are often altered
from the taxonomic treatment of Borisova. Most members of the recognized sections Brevirostres and
Profundisulcati appear within the Brevirostres and Profundisulcati clades, respectively. Species of section
Nikitinia are also included in the Brevirostres clade. Sequence data also provide evidence for an Angustissimi
clade of T. pusillus, T. filifolius, diploid T. latifolius (2n ¼ 12), T. segetum, T. acanthocarpus, T. serotinus, and
T. charadzae, species not previously considered closely related. Our analyses also indicate possible relationships for 17 species that had not been previously placed in any section. The diploids T. porrifolius, T. pratensis,
and T. dubius, which were introduced into North America and ultimately formed the recent alloploids T. mirus
and T. miscellus, are not closely related within Tragopogon; T. dubius and T. porrifolius appear in sister
clades, and T. pratensis is more distantly related. Characters often used to establish sectional groupings, such as
the length of the achene beak, evolved multiple times in Tragopogon.
Keywords: Tragopogon, internal transcribed spacer and external transcribed spacer sequence data, phylogeny.
Introduction
terized by a distinct ‘‘Tragopogon’’ type of pollen that has six
abpolar, three equatorial, and six interapertural lacunae
(Blackmore 1982; Nazarova 1997). Koelpinia, another genus
of Scorzonerinae, has the same type of pollen (Blackmore
1982; Nazarova 1997), but this may represent an independent
origin, because Koelpinia and Tragopogon are not sister taxa
within subtribe Scorzonerinae (Mavrodiev et al. 2004).
There are several systematic treatments of Tragopogon, although none of these is comprehensive, involving all recognized species. De Candolle (1838) split Tragopogon into two
large groups that were not formally named. In one group he
placed all species possessing peduncles thickened below the
flowering capitula. In the second group the peduncles are not
thickened below the capitula. Later, Boissier (1875) split
Tragopogon into two large groups of unspecified rank based
on flower color: (1) ‘‘species with yellow flowers’’ (Flaviflora
group) and (2) ‘‘species with purplish flowers’’ (Rubriflora
group). Following these initial treatments, other early investigators of Tragopogon (Nikitin 1930; Artemczhyk 1937;
Regel 1937, 1955) described a large number of new species
but did not provide any system of classification for the genus.
Artemczhyk (1948) split all Ukrainian species of Tragopogon
into three groups, Majores, Orientalis, and Dasyrhynchiformes, and recognized these groups as separate series, but
gave descriptions for only the Ukrainian members of these series. The brief paper of Artemczhyk (1948) is also the first
Tragopogon L. (Lactuceae, Cichorioideae, Scorzonerinae)
is an Old World genus of ca. 150 species that occurs across
Eurasia from the Atlantic to the Pacific Ocean, with a center
of distribution in the Mediterranean region, the Middle East,
and eastern Europe. Tragopogon includes biennial and perennial herbs with linear or linear-lanceolate leaves; solitary,
simple, or sparingly branched stems; one or only a few capitula; and receptacles without scales. The monophyly of the genus was strongly supported in a recent phylogenetic analysis
of Scorzonerinae based on internal transcribed spacer (ITS) sequence data (Mavrodiev et al. 2004). Species of Tragopogon
are also united by several morphological characters. The
achenes of Tragopogon are usually fusiform, with five to 10
more or less distinct ribs and a beak of varying length. The involucral bracts are always in one row, ligulate flowers are yellow or purplish, and the pappus is in one row of mostly
plumose hairs (Richardson 1976). Tragopogon is also charac1
2
3
4
Author for correspondence; e-mail evgeny@ufl.edu.
E-mail magitz@ufl.edu.
E-mail psoltis@flmnh.ufl.edu.
Author for correspondence; e-mail dsoltis@botany.ufl.edu.
Manuscript received February 2004; revised manuscript received August
2004.
117
118
INTERNATIONAL JOURNAL OF PLANT SCIENCES
evolutionary treatment of Tragopogon. Artemczhyk (1948)
proposed that the center of origin of Tragopogon is the eastern Mediterranean, and he considered Tragopogon majus
Jacq., T. orientalis L., and T. dasyrhynchus Artemcz. to be
ancestral to all of the European species of the genus.
Kuthatheladze (1957) proposed a classification of Tragopogon based on the analysis of 32 species, which she distributed among five sections: Brevirostres, Collini, Angustissimi,
Majores, and Profundisulcati. According to Kuthatheladze
(1957), Tragopogon is typically a Mediterranean genus, the
oldest section in Tragopogon is Brevirostres, and the youngest is Profundisulcati. She also proposed that ancestral character states in Tragopogon include the perennial herbaceous
habit, yellow ligules, and an achene having a short beak.
In the Flora of the USSR, Borisova (1964) proposed a new
system of classification for Tragopogon, based on the analysis
of 79 species, all of which occurred in the territory of the former Soviet Union. This treatment included sections Angustissimi Kuth. (three species), Bessera Boriss. (four species),
Brevirostres Kuth. (13 species), Chromopappus Boriss. (two
species), Dasypogon Boriss. (one species), Dybjanskya Boriss.
(one species), Hebecarpus Boriss. (one species), Kemularia
Boriss. (one species), Krascheninnikovia Boriss. (one species),
Macropogon (Kuth.) Boriss. (one species), Majores Kuth.
(five species), Nikitinia Boriss. (four species), Profundisulcati
Kuth. (11 species), Collini Kuth. (¼Rubriflori [Boiss.] Boriss.)
(11 species), Sosnovskya Boriss. (four species), Tragopogon
(14 species), and Tuberosi (Kuth.) Boriss. (two species). Rechinger (1977) followed Borisova’s (1964) system of classification for Tragopogon in his Flora Iranica. Klokov (1981)
suggested that species of sections Nikitinia, Bessera, and
Brevirostres, as well as some species of section Tragopogon,
form one broad section, but he did not formally name this
group. He also suggested that the members of this unnamed
group derived from ancestral taxa having achenes with long
beaks. Tzvelev (1985) proposed a taxonomic treatment of
Tragopogon from the European portion of Russia based on
23 species. He accepted sections Brevirostres, Tragopogon,
Majores, and Collini (¼Rubriflori) but disagreed with Borisova’s (1964) recognition of sections Hebecarpus, Nikitinia,
and Bessera. Tzvelev (1985) suggested that many diploid species in the genus were of hybrid origin, but he did not suggest
possible parents for most of the known polyploids. He proposed that the ancestral Tragopogon was a perennial herb
with yellow ligules, achenes without beaks, and a chromosome number of 2n ¼ 14, with the more common chromosome number of 2n ¼ 12 as the derived state.
Based on the study of Tragopogon in the Caucasus, Nazarova (1991) accepted sections Tragopogon, Tuberosi, Collini
(¼Rubriflori), Angustissimi, Majores, Chromopappus, and
Profundisulcati of Borisova (1964). In contrast to Tzvelev
(1985), Nazarova (1991) did not believe that hybrid speciation was important in diploid Tragopogons. She also agreed
that several species of Tragopogon were allopolyploids but
did not suggest any parents. The chromosome number reported for almost all species of Tragopogon is 2n ¼ 12.
However, for T. pratense (Ishikawa 1916 in Nazarova 1991;
Clapman 1952 in Nazarova 1991) and T. crocifolius (Darlington 1956), both 2n ¼ 12 and 14 were reported. However,
the validity of reports of 2n ¼ 14 for these two species is de-
batable. The counts of 2n ¼ 14 may simply reflect the presence of B chromosomes in several populations (Nazarova
1991). Several polyploids have been reported in Tragopogon.
Ownbey (1950) showed that two allotetraploid species of
Tragopogon, T. mirus and T. miscellus, formed in North
America in the 1900s following the introduction of three diploids from Europe. The parents of T. miscellus are T. dubius
and T. pratensis; the parents of T. mirus are T. dubius and T.
porrifolius (Ownbey 1950; Ownbey and McCollum 1953; reviewed in Soltis et al. 2004). Nazarova (1991) indicated that
at least 10 species of Tragopogon from Europe are polyploids
or include polyploid cytoptypes: T. castellanus Levier
(2n ¼ 24), T. buphtalmoides (DC.) Boiss. (2n ¼ 24, 36), T.
coloratus C. A. Mey. (2n ¼ 12, 24), T. cupani Guss. ex DC.
(2n ¼ 12, 24), T. gracilis D. Don (2n ¼ 24), T. graminifolius
DC. (2n ¼ 12, 24, 36), T. pusillus Bieb (2n ¼ 12, 24), T.
reticulatus Boiss. et Huet. (2n ¼ 12, 24, 36, 56–58), T.
latifolius Boiss. (2n ¼ 12, 24), and T. tuberosus C. Koch
(2n ¼ 24). Nazarova (1991) further suggested that some of
these polyploids may have subsequently been involved in
hybridization events, yielding a complex of additional
cytotypes: T: buphtalmoides 3 T: latifolius (2n ¼ 24), T:
buphtalmoides 3 T: reticulatus (2n ¼ 24, 27–32, 35, 36).
However, the parentage of all of these Eurasian polyploids is
unknown, and they have been little studied.
In a recent molecular phylogenetic investigation of subtribe
Scorzonerinae based on ITS sequence data, we demonstrated
that Tragopogon is monophyletic and distinct from other
genera, including Geropogon (Mavrodiev et al. 2004),
a monotypic genus sometimes included within Tragopogon
(Richardson 1976). The independence of Geropogon was
previously suggested on the basis of karyological and morphological data (D’Amato 1975 in Nazarova 1991; Wilson
1983).
Relationships within Tragopogon are poorly understood.
Many species of Tragopogon have not been placed in a section;
most of these are narrow endemics that have been recognized
and named but not treated taxonomically. Furthermore, the
genus has never been the subject of a comprehensive monograph. Regional floras have provided treatments for species
only in those geographic areas. The phylogenetic utility of ITS
and external transcribed spacer (ETS) sequence data has been
demonstrated in a number of studies of genera of Asteraceae
(Clevinger and Panero 2000; Lee et al. 2002, 2003; Markos
and Baldwin 2001; Saar et al. 2003; Urbatsch et al. 2003). In
an initial effort to resolve relationships within Tragopogon,
we conducted a phylogenetic analysis of 59 species, using ITS
and ETS sequence data. We also hoped to elucidate the relationships among T. dubius, T. pratensis, and T. porrifolius, the
diploid progenitors of the recently formed North American
polyploids T. mirus and T. miscellus (Ownbey 1950).
Material and Methods
Species and Samples
We followed the treatment of Tragopogon given in the
Flora of the USSR (Borisova 1964; Rechinger 1979), because
these are the most comprehensive treatments of the genus (table 1). We sampled broadly across the entire genus, including
Table 1
Comparison of Treatments of Tragopogon
Taxa
T. acanthocarpus Boiss.
T. afghanicus K.H. Rechinger
et Koie
T. armeniacus S. Kuthath.
T. bornmuelleri M. Ownbey
et K.H. Rechinger
T. brevirostris DC.
T. capitatus S. Nikit.
T. charadzae S. Kuthath.
T. collinus DC.
T. coloratus C. A. Mey. (2n¼12)
T. dasyrhynchus Artemcz.
T. dubius Scop.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
dubjanskyi S. Niktin
elongatus S. Nikit.
filifolius Rehm. ex Boiss.
floccosus Waldst. et Kit.
heterospermus Schweigg.
kemulariae S. Kuthath.
ketzkhovelii S. Kuthath.
kindingeri
kotschyi Boiss.
krascheninnikovii S. Nikit.
T. latifolius Boiss. (2n¼12)
T. longirostris Bischoff
ex Sch. Bip.
T. major Jacq.
T. makaschwilii S. Kuthath.
T.
T.
T.
T.
T.
marginatus Boiss. et Buhse
meskheticus S. Kuthath.
montanus S. Nikit.
orientalis L.
paradoxus S. Nikit.
T.
T.
T.
T.
T.
podolicus Bess. ex DC.
porrifolius L.
pratensis Linn.
pterocarpus DC.
pusillus Bieb. (2n¼12)
T.
T.
T.
T.
T.
T.
T.
T.
T.
reticulatus Boiss. et Huet
ruber S. T. Gmel.
ruthenicus Bess. ex Claus
segetum S. Kuthath.
serotinus Sosn.
sinuatus
sosnowskyi S. Kuthath.
trachycarpus S. Nikit.
undulatus Jacq.
Borisova 1964
Kuthatheladze 1957
Sect. Profundisulcati Kuthath.
Sect. Majores (Artemcz.)
Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Brevirostres Kuthath.
Sect. Majores (Artemcz.)
Kuthath.
Sect. Sosnovskya Boriss.
Sect. Rubriflori Boiss.
Sect. Chromopappus Boriss.
Sect. Brevirostres Kuthath.
Rechinger 1977; Tzvelev 1985
Sect. Profundisulcati Kuthath.
Sect. Majores (Artemcz.) Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Brevirostres Kuthath.
Sect. Majores (Artemcz.)
Kuthath.
Sect. Nikitinia Boriss.
Sect. Rubriflori Boiss.
Sect. Brevirostres Kuthath.
Sect. Brevirostres Kuthath.
Sect. Bessera Boriss.
Sect. Kemularia Boriss.
Sect. Profundisulcati Kuthath.
Sect. Majores (Artemcz.) Kuthath.
Sect. Brevirostres Kuthath.
Sect. Collini Kuthath.
Sect. Majores (Artemcz.)
Kuthath.
Sect. Brevirostres Kuthath.
Sect. Majores (Artemcz.)
Kuthath.
Sect. Rubriflori Boiss.
Sect. Chromopappus Boriss.
Sect. Brevirostres Kuthath.
Sect. Majores (Artemcz.) Kuthath.
Sect. Brevirostres Kuthath.
Sect. Brevirostres Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Brevirostres Kuthath.
Sect. Brevirostres Kuthath.
Sect. Kemularia Boriss.
Sect. Sosnovskya Boriss.
Sect. Krascheninnikovia
Boriss.
Sect. Profundisulcati Kuthath.
Sect. Krascheninnikovia
Boriss.
Sect. Majores (Artemcz.)
Kuthath. (as syn. T. dubius
Scop.)
Sect. Profundisulcati Kuthath.
Sect. Sosnovskya Boriss.
Sect. Profundisulcati Kuthath.
Sect. Rubriflori Boiss.
Sect. Tragopogon
Sect. Majores (Artemcz.)
Kuthath.
Sect. Brevirostres Kuthath.
Sect. Hebecarpus Boriss.
Sect. Tragopogon
Sect. Chromopappus Boriss.
Sect. Tuberosi (Kuthath.)
Boriss.
Sect. Sosnovskya Boriss.
Sect. Rubriflori Boiss.
Sect. Nikitinia Boriss.
Sect. Angustissimi Kuthath.
Sect. Brevirostres Kuthath.
Sect. Angustissimi Kuthath.
Sect. Tragopogon
Sect. Majores (Artemcz.)
Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Krascheninnikovia Boriss.
Sect. Majores (Artemcz.) Kuthath.
(as syn. T. dubius Scop.)
Sect. Majores (Artemcz.)
Kuthath.
Sect. Collini Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Sosnovskya Boriss.
Sect. Rubriflori Boiss.
Sect. Tragopogon
Sect. Majores (Artemcz.) Kuthath.
Sect. Profundisulcati Kuthath.
Sect. Collini Kuthath.
Sect. Brevirostres Kuthath.
Sect.
Sect.
Sect.
Sect.
Sect.
Brevirostres Kuthath.
Majores (Artemcz.) Kuthath
Tragopogon
Chromopappus Boriss.
Tuberosi (Kuthath.) Boriss.
Sect. Sosnovskya Boriss.
Sect. Collini Kuthath.
Sect. Brevirostres Kuthath.
Sect. Angustissimi Kuthath.
Sect. Brevirostres Kuthath.
Sect. Angustissimi Kuthath.
Sect. Angustissimi Kuthath.
Sect. Brevirostres Kuthath.
Table 2
Voucher Data for Individuals Sequenced in This Study
GenBank accession numbers
Species
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
acanthocarpus
afghanicus
albinerve
armeniacus
aureus
australis
balcanicum
bornmuelleri
brevirostris
capitatus
charadzae
coelesyriacus
collinus
coloratus (2n¼12)
crocifolius
dasyrhynchus
T. dubius
T. dubjanskyi
T. elongatus
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
fibrosum
filifolius
floccosus
hayekii
heterospermus
jesdianus
kemulariae
kindingeri
kotschyi
krascheninnikovii
lamottei
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
T.
latifolius (2n¼12)
longifolius
longirostris
major
makaschwilii
marginatus
meskheticus
minor
montanus
olympicus
orientalis
podolicus
porrifolius
T. pratensis
T. pterodes
T. pusillus
T. reticulatus (2n¼12)
T. ruber
T. ruthenicus
T. samaritani
T. segetum
Voucher
ITS region of rDNA (ITS1, 5.8S, ITS2)
ETS region of rDNA
E. Nazarova 1988, s.n. (ERE)
‘‘Asia’’: K. H. Rechinger, 2193999 (USN)
Turkey: V. A. Mattews (K)
Armenia, 1972: A. Chechurov (LE)
Turkey, 1993: A. Chochrjakov (MHA)
549760 (USN)
Turkey, 1955: A. Chochrjakov (MHA)
Iran, 1957: K. H. Rechinger, 1507 (K)
M. Ownbey 274243 (WS)
Kazachstan, 1930: S. Nikitin, 12 (MW)
Georgia, 1948: S. Kuthatheladze (LE)
M. Ownbey, 274106 (WS)
Azerbajdzhan, 1952: N. N. Tzvelev (LE)
Armenia, 1979: E. Nazarova, 856 (ERE)
M. Ownbey 274740 (WS)
Russia: Volgograd, 2002: E. Mavrodiev, field
collections (Soltis lab.), det. N. N. Tzvelev
M. Ownbey, 274197 (WS)
Volgograd, 2002: E. Mavrodiev collections
(Soltis lab.)
Kazachstan, 1981: V. Bochkin, I. Rusanovich
(MHA)
Turkey, 1996: A. Khochrjakov (MHA)
Azerbajdzhan, 1929: A. Grosshejm (LE)
Hungary, 1990: 242630 (WS)
Europe, 175685 (WS)
Russia, Kaliningrad, 1973: A. Skvortzov (MHA)
Afghanistan, 1974: I. Gubanov, V. Pavlov (MW)
Armenia, 1951: E. Nazarova, 34564E (ERE)
M. Ownbey, 251956 (WS)
‘‘Persia’’: K. H. Rechinger (K)
Armenia, 1979: S. Kuthatheladze (LE)
Spain, 1982: F. Valle, G. Blanca, field collections
(Soltis lab.)
Armenia, 1979: E. Nazarova, 132916 (ERE)
M. Ownbey, 374431 (WS)
Europe, 274106 (WS)
274198 (WS)
Georgia, 1953: S. Kuthatheladze (LE)
Azerbajdzhan, 1972: S. Kuthatheladze (LE)
Georgia, 1945: D. Sosnovsky (LE)
Europe: 1373086 (USN)
‘‘Persia,’’ Shared-Bustan: K. H. Rechinger (K)
Bulgaria, 1853: (LE)
Europe: M. Ownbey, 274728 (WS)
Russia, Volgograd, 1989: V. Sagalaev (VOLG)
Spain, 1982: G. Blanca et al., field collections
(Soltis lab.)
M. Ownbey, 208347 (WS)
Bulgaria/Serbia, 1903: K. H. Rechinger,
M. Ownbey (K)
Azerbajdzhan, 1948: S. Lipshitz (LE)
Armenia, 1979: E. Nazarova, 907 (ERE)
Russia, Astrachan, 2002: E. Mavrodiev, field
collections (Soltis lab.), det. N. N. Tzvelev
Russia, Volgograd, 2002: E. Mavrodiev, field
collections (Soltis lab.)
M. Ownbey, 274420 (WS)
Georgia, 1970: T. Popova (LE)
AY645802
AY508175
AY508183
AY645803
AY645804
AY508166
AY645805
AY645806
AY508174
AY645807
AY645808
AY645809
AY645810
AY645811
AY508180
AY645812
AY645845
AY645846
AY645847
AY645848
AY645856
AY645857
AY645858
AY645813
AY645814
AY645859
AY645860
AY645815
AY645861
AY645816
AY645817
AY508182
AY645818
AY508168
AY645819
AY645820
AY508178
AY508181
AY645821
AY645823
AY645849
AY645850
AY645852
AY645851
AY645853
AY645854
AY645855
AY645862
AY645863
AY645864
AY645865
AY645866
AY645867
AY645868
AY645869
AY645871
AY645822
AY645824
AY508185
AY645825
AY645826
AY645827
AY645828
AY508184
AY508172
AY645829
AY508170
AY645831
AY645870
AY645872
AY645873
AY645874
AY645875
AY645878
AY645879
AY645880
AY508169
AY508167
AY508176
AY645881
AY645882
AY645883
AY645830
AY645832
AY645833
AY645884
AY645885
AY645886
AY645834
AY645887
AY645835
AY645836
AY645888
AY645889
AY645876
AY645877
MAVRODIEV ET AL.—PHYLOGENY OF TRAGOPOGON
121
Table 2
(Continued )
GenBank accession numbers
Species
T.
T.
T.
T.
serotinus
sinuatus
sosnowskyi
stenophyllus
T. tommasinii
T. trachycarpus
T. undulat us
Podospermum
jacquinianum
Lactuca sp.
Voucher
ITS region of rDNA (ITS1, 5.8S, ITS2)
ETS region of rDNA
Armenia, 1967: S. Kuthatheladze (LE)
Ownbey M., 274133 (WS)
Caucasus, 1830: S. Kuthatheladze (LE)
M. Ownbey, 274276 (WS)
France: M. Ownbey, 274740 (WS)
M. Ownbey, 274702 (WS)
Mongolia: I. A. Gubanov (MW)
Ukraina, Tauria: N. K. Schvedchikova (MW)
AY645837
AY645838
AY645839
AY645840
AY645841
AY645842
AY508177
AY508171
AY645890
AY645891
AY645892
AY645893
AY645894
AY645895
AY645896
Turkey: P. H. Davis, 51673 K
Gainesville: E. Mavrodiev, original collections
AY508195
L13957
AY645844
AY645843
accessions of 59 species, representing 14 of the 17 sections of
Borisova (1964). An important limitation in our sampling is
that many species of Tragopogon are difficult to obtain.
Most species occur in poorly collected areas and/or in countries experiencing political upheaval, including Iran, Iraq,
Afghanistan, and Chechnya (Russian Federation). Hence, we
relied largely on herbarium specimens and field collections of
the first author (table 2). Plants collected in the field were
either dried as pressed specimens or placed in silica gel.
We sequenced ITS and ETS from one or more species from
sections Angustissimi, Bessera, Brevirostres, Chromopappus,
Hebecarpus, Krascheninnikovia, Majores, Nikitinia, Tuberosi,
Profundisulcati, Collini (¼Rubriflori), Sosnovsky, and Tragopogon. We also sampled the single species of the monotypic
section Kemularia (table 2). The only sections of Borisova
(1964) not represented in this study are Macropogon (monotypic), Dasypogon (monotypic), and Dybjanskya (monotypic).
We were unable to obtain plants of these three small sections.
We also included 17 species of unknown systematic position
in Tragopogon: T. albinerve Freyn. et Sint., T. aureus Boiss., T.
balcanicum Velen. ex Nym., T. coelesyriacus Boiss., T. crocifolius L., T. fibrosum Freyn et Sint. ex Freyn, T. hayekii (Soo) I.
B. K. Richardson, T. jesdianus Boiss. et Buhse, T. kindingeri
Adamov, T. minor Miller, T. olympicus Boiss., T. samaritani
Heldr. et Sart., T. sinuatus Ave-Lall., T. longifolius Heldr. et
Sart., T. pterodes Panc, T. stenophyllus Jord., and T. tommasinii Sch. Bip. These species have never been attributed to any
of the formally recognized sections of Tragopogon, and they
are therefore not listed in table 1.
We also attempted to include only diploid taxa in our analyses, because of the potential difficulties and uncertainties involved in using ITS and ETS sequences in polyploid taxa
(Soltis and Soltis 1998; Álvarez and Wendel 2003). We therefore excluded from our analysis all known polyploids, such as
the North American allotetraploids T. mirus and T. miscellus
and the Eurasian polyploids T. buphtalmoides, T. castellanus,
T. coloratus, T. cupani, T. gracilis, T. graminifolius, T. reticulatus, T. latifolius, T. pusillus, and T. tuberosus. In some cases
(e.g., T. latifolius, T. reticulatus, T. coloratus, T. pusillus), we
used only the diploid races of species for which several ploidal
levels have been reported (Nazarova 1984, 1991).
Both ITS and ETS sequences were obtained for 52 taxa. Because of the poor quality or small amount of some DNA extracts, we were unable to obtain ETS sequences for several
species. Tragopogon krasheninnikovii, T. collinus, T. montanus,
T. marginatus, T. fibrosus, T. aureus, and T. sosnowskyi are represented only by ITS sequences. We conducted two sets of analyses. In one, we analyzed the 52 taxa represented by both ITS
and ETS sequences. The second analysis, of 59 taxa, also included the seven taxa for which only ITS was sequenced.
As outgroups, we included the ITS (Mavrodiev et al. 2004)
and ETS sequences of Podospermum jacquinianum Koch and
Lactuca sp. (ITS of Lactuca sp. was obtained from GenBank;
table 2). Lactuca represents Lactucinae, a subtribe that is the
close relative of subtribe Scorzonerinae (Bremer 1994).
DNA Amplification and Sequencing
We isolated DNA following the general method of Doyle
and Doyle (1987), as modified by Soltis et al. (1991). For ITS
amplification and sequencing, we used forward primer
N-nc18S10 (59-AGG AGA AGT CGT AAC AAG-39) (Wen
and Zimmer 1996) and reverse primer ITS-4 (59-GGT AGT
CCC GCC TGA CCT GG-39) (White at al. 1990) or, in some
samples, reverse primer C26A (5-AGT TTC TTT TCC TCC
GCT-39) (Wen and Zimmer 1996). PCR reactions (25 mL)
contained 13 PCR Buffer (Promega, Madison, WI), 2.5 mM
MgCl2, 200 mM each dNTP, 1 M betaine, 1 mM each of forward and reverse primers, 1.25 U Taq DNA polymerase
(Promega), and ca. 10 ng of genomic DNA. The amplification profile consisted of an initial denaturation of 2 min at
95°C followed by a touch-down segment of 93°C for 1 min,
then 53°C (decreasing 1°C per cycle) for 1 min and 72°C for
2 min for five cycles, followed by 35 cycles using a 48°C annealing temperature. A final extension of 5 min was added at
the end of the cycle. The PCR products were purified using
the QIAquick PCR purification kit (Qiagen, Valencia, CA)
and the manufacturer’s protocol. We sequenced the 39 end of
the ETS. For ETS amplification and sequencing, we used forward primer L-ETS (59-GGA TCG TTC GGT GCA TTC
T-39) (Lee et al. 2002) and reverse primer 18S-ETS (59-AGA
CAA GCA TAT GAC TAC TGG CAG G-39) (Baldwin and
122
INTERNATIONAL JOURNAL OF PLANT SCIENCES
Markos 1998). The PCR conditions used for ETS were the
same as those given above for ITS.
We sequenced ITS and ETS PCR products in both directions
with the PCR primers as sequencing primers, using BeckmanCoulter Dye Terminator Cycle Sequencing Quick-Start kits
and a CEQ 8000 automated sequencer (Beckman Coulter,
Fullerton, CA) following the manufacturer’s protocols, except
that we conducted quarter-volume sequencing reactions.
Alignment and Phylogenetic Analyses
Sequences were initially aligned using the program Clustal
X (Thompson et al. 1997). Following the initial alignment,
the sequences were adjusted manually using the program SeAl, version 2.0 a11 (Rambaut 2003). Alignment of ITS and
ETS was straightforward in these taxa.
The ITS and ETS data sets for 59 and 52 taxa, respectively,
were first analyzed separately and then combined. All phylogenetic analyses were conducted using PAUP*, version 4.0 b10
(Swofford 2002). Maximum parsimony (MP) analyses were
conducted using heuristic searches with 1000 random addition
replicates with no more than 100 trees saved per replicate and
tree bisection reconnection (TBR) branch swapping, with the
MulTrees option in effect; gaps were treated as ‘‘missing.’’
Internal support for clades was measured by bootstrap
(BST; Felsenstein 1985) and jackknife (JK) analyses. BST estimates were obtained using 2500 replicates, each with 100
random addition replicates, saving no more than 1500 trees
per BST replicate, TBR branch swapping, and the MulTrees
option in effect. JK estimates were obtained using 2500 JK
replicates with 36.0% of the characters deleted (Farris et al.
1996) in each replicate and 100 random addition sequences,
saving no more than 1500 trees per replicate, TBR branch
swapping, and the MulTrees option in effect.
Character-State Reconstruction
We reconstructed the evolution of two characters (table 3)
that have been important in sectional diagnoses in Tragopogon:
life history and the length of the beak of the achene. Morphological data were adopted from Kuthatheladze (1957),
Borisova (1964), Matthews (1975), Richardson (1976), Rechinger (1977), Tzvelev (1985), and Nazarova (1991, 1995);
in some cases, data were obtained from herbarium (LE, MW,
MHA, ERE) and field collections (southeastern Russia, Armenia). We used MacClade, version 4.05 (Maddison and
Maddison 2002), using one of the most parsimonious trees
chosen at random and ‘‘all most parsimonious states’’ as the
optimization method. To assess the impact of the topology
on character-state reconstruction, we randomly selected 1000
of the 10,000 shortest trees and reconstructed character-state
changes as described.
In our reconstructions we also used Geropogon hybridus
(L.) Sch. Bip., Scorzonera austriaca Willd., Scorzonera aristata Ram. ex DC., and Takhtajaniantha pusilla (Pall.) Nazarova as outgroups. These taxa were found to be close relatives
of Tragopogon in our phylogenetic analysis of subtribe Scorzonerinae (Mavrodiev et al. 2004). The use of these alternative outgroups had no impact, however, on our characterstate reconstructions.
Results
The aligned data matrix of the ITS region of rDNA (ITS1,
5.8S, ITS2) consists of 701 characters (288 in ITS1, 168
in 5.8S, and 245 in ITS2), of which 459 are constant and
80 are parsimony-informative sites. We saved 10,000 most
parsimonious trees of length ¼ 343, Consistency Index
ðCIÞ ¼ 0:8426, Retention Index ðRIÞ ¼ 0:8402, and Rescaled
Consistency Index ðRCÞ ¼ 0:7880 (fig. 1). The aligned data
matrix of the ETS region of rDNA is 574 characters, of
which 336 are constant and 82 are parsimony-informative.
We saved 10,000 most parsimonious trees of length ¼ 344,
CI ¼ 0:8256, RI ¼ 0:8729, and RC ¼ 0:7206 (fig. 2). The
aligned data matrix of the ITS þ ETS regions for 52 taxa
is 1274 characters, of which 882 are constant and 162 are
parsimony-informative. In parsimony analyses we saved
10,000 most parsimonious trees of length ¼ 684, CI ¼
0:8173, RI ¼ 0:8381, and RC ¼ 0:6849 (fig. 3). The aligned
ITS þ ETS data set for 59 species (seven represented only by
ITS sequences) is also 1274 characters in length, of which
896 are constant and 141 are parsimony informative. We
saved 10,000 most parsimonious trees of length ¼ 616,
CI ¼ 0:8101, RI ¼ 0:8484, and RC ¼ 0:6873 (fig. 4). Parsimony analysis of the ITS data set provided strong support
(JK ¼ 100%, BST ¼ 100%) for the monophyly of the genus
Tragopogon relative to the outgroups (Lactuca sp. and Podospermum jacquinianum), but phylogenetic relationships
within Tragopogon were only partially resolved. ITS sequence data recovered a clade (JK ¼ 86%, BST ¼ 65%) that
encompasses taxa from sections Majores, Chromopappus,
and Hebecarpus and roughly corresponds to the Majores,
Chromopappus, and Hebecarpus clades detected with ETS
and ETS þ ITS sequences (see below). However, the Chromopappus and Hebecarpus clades were not obtained with ITS
alone. With the Majores s. l. clade in the ITS tree is a clade
of T. pterpodes, T. charadzae, T. dubius, T. capitatus, and T.
major, but with weak support (JK ¼ 68%, BST ¼ 52%).
However, the position of T. charadzae in this clade in the ITS
tree is problematic. This species of section Sosnovskya or
Brevirostres is morphologically very distinct from T. dubius
and the other members of this small clade. In the ETS and
ITS þ ETS trees, T. charadzae belongs to the Angustissimi
clade. Similarly, T. floccosus, placed in section Brevirostres,
is morphologically very distinct from all members of the
Majores s. l. clade, where it appears on the basis of ITS
sequence data. In contrast, in the ETS and ITS þ ETS trees,
T. floccosus is a member of the Brevirostres clade.
With ITS alone, a Collini clade with T. ruber, T. bronmuelleri, T. collinus, T. elongatus, T. jezdianus, T. montanus, and
T. marginatus received moderate (JK ¼ 83%, BST ¼ 71%)
support. The Profundisulcati clade of T. albinerve, T. meskheticus, T. makaschwilii, T. armeniacus, T. fibrosus, T. aureus, and T. kotschyi received weak support (JK ¼ 68%,
BST ¼ 57%). In addition, with ITS alone there is strong support (JK ¼ 91%, BST ¼ 87%) for a sister-group relationship
between clades Collini and Profundisulcati (fig. 1). ITS sequence data also reveal a weakly supported (JK ¼ 66%,
BST ¼ 52%) Tragopogon clade of T. hayeki, T. tommasinii,
T. longifolius, T. pratensis, T. orientalis, T. minor, and T.
trachycarpus (fig. 1). However, the Brevirostres and
MAVRODIEV ET AL.—PHYLOGENY OF TRAGOPOGON
Table 3
Morphological Characteristics of Tragopogon Data Matrix for
Character-State Reconstructions Using MacClade
Species
Geropogon hybridus
Podospermum
jacquinianum
Scorzonera aristata
Scorzonera austriaca
Takhtajaniantha pusilla
Tragopogon acanthocarpus
T. afghanicus
T. albinerve
T. armeniacus
T. australis
T. balcanicum
T. bornmuelleri
T. brevirostris
T. capitatus
T. charadzae
T. coelesyriacus
T. coloratus
T. crocifolius
T. dasyrhynchus
T. dubius
T. dubjanskyi
T. elongatus
T. filifolius
T. floccosus
T. hayekii
T. heterospermus
T. jesdianus
T. kemulariae
T. kindingeri
T. kotschyi
T. lamottei
T. latifolius
T. longifolius
T. longirostris
T. major
T. makaschwilii
T. meskheticus
T. minor
T. olympicus
T. orientalis
T. podolicus
T. porrifolius
T. pratensis
T. pterodes
T. pusillus
T. reticulatus
T. ruber
T. ruthenicus
T. samaritani
T. segetum
T. serotinus
T. sinuatus
T. stenphyllus 1
T. stenphyllus 2
T. tommasinii
T. trachycarpus
T. undulatus
Achene morphology
Life history
1
2
0
0
0
0
1
1
1
1
1
1
1
0
1
1
1
1
1
0
1
0
1
1
0
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
1
1
1
1
1
0
1
1
1
1
1
1
1
0
0
1
1
1
1
1
1
0
1
1
1
0
0
1
0
1
1
1
1
1
1
0
0
0, 1
0
0, 1
3
1
0
0
1
1
1
1
1
1
0
1
0
1
1
0, 1, 2
1
1
0
0
0
1
1
1
0
?
1
1
1
1
1
Note. Achene morphology: 0 ¼ beak of achene is very short,
practically absent (usually #0.5 mm); 1 ¼ beak is clearly present
(usually >2 mm). Life history: 0 ¼ perennial; 1 ¼ biennial; 2 ¼
annual; 3 ¼ other.
123
Angustissimi clades, which are both present in the ETS
and ETS þ ITS trees, are not recovered in the ITS trees.
With ETS sequence data alone a Majores clade of T. stenophyllus, T. coelesyriacus, T. capitatus, T. dubius, T. major,
T. pterodes, T. afgh anicus, and T. kemulariae received moderate support (JK ¼ 83%, BST ¼ 2%) (fig. 2). Sister to the
Majores clade is the Chromopappus clade, consisting of T.
longirostris and T. coloratus (JK ¼ 71%, BST ¼ 66%), although support for the Majores þ Chromopappus sister
group is less than 50%. A Hebecarpus clade (JK ¼ 89%,
BST ¼ 82%) of T. crocifolius, T. australis, T. balcanicum, T.
samaritani, T. sinuatus, and T. porrifolius is sister to
Majores þ Chromopappus (JK ¼ 63%, BST ¼ 60%) (fig. 2).
Analysis of ETS data also recovered a Profundisulcati clade
(JK ¼ 73%, BST ¼ 61%) of T. albinerve, T. meskheticus, T.
makaschwilii, T. armeniacus, and T. kotschyi. The Profundisulcati clade is sister (JK ¼ 65%, BST ¼ 50%) to the Collini
clade (JK ¼ 75%, BST ¼ 68%) of T. ruber, T. bronmuelleri,
T. elongatus, and T. jezdianus.
Also receiving support with ETS sequences alone is an Angustissimi clade (JK ¼ 93%, BST ¼ 79%) of T. pusillus, T.
filifolius, T. latifolius (2n ¼ 12), T. segetum, T. acanthocarpus, T. serotinus, and T. charadzae. This clade does not correspond to any formally named section but contains both of
the species of section Angustissimi (T. segetum and T. sosnowskyi) (fig. 4; see below).
A Brevirostres clade of T. kindingeri, T. brevirostris, T.
dasyrhynchus, T. reticulatus, T. dubjanskyi, T. ruthenicus, T.
heterospermus, T. undulatus, T. podolicus, and T. floccosus
also received support (JK ¼ 90%, BST ¼ 76%) in analyses
based only on ETS. The Brevirostres clade is part of a polytomy with T. lamottei, T. hayeki, T. tommasinii, T. longifolius, and species from section Tragopogon: T. pratensis, T.
orientalis, T. minor, and T. trachycarpus. Tragopogon hayeki,
T. tommasinii, T. longifolius, T. orientalis, T. pratensis, and
T. minor form a moderately supported Tragopogon clade
(JK ¼ 88%, BST ¼ 74%) in the ITS þ ETS tree, and T. lamottei is sister to this clade, but with weak support
(JK ¼ 65%) (fig. 3).
The ITS and ETS trees do not exhibit any major conflicts.
In fact, there is close agreement between the ITS and ETS
trees, although the ETS tree is better resolved and supported
than the ITS tree (cf. figs. 1, 2). All major differences involve
clades that are resolved and supported with ETS but unresolved with ITS. The primary differences are (1) the ETS tree
detected the Chromopappus and Hebecarpus clades, whereas
ITS did not; (2) species of section Brevirostres formed a moderately supported Brevirostres clade with ETS (JK ¼ 90%,
BST ¼ 76%), whereas only a portion of this clade is recovered with ITS (and without support >50%); and (3) a moderately supported Angustissimi clade (JK ¼ 93%, BST ¼ 79%)
with several species from the Caucasus is present in the ETS
tree but not in the ITS tree. Given the lack of major conflict,
we therefore combined the ITS and ETS data sets. Our analyses of the combined ITS þ ETS data set for 52 taxa produced
a generally well-resolved tree for Tragopogon (fig. 3) that is
similar to that produced when the seven species represented
only by ITS sequences were added to the analysis (fig. 4).
However, support for several clades is lower in the latter
trees than in the 52-taxon ITS þ ETS tree, most likely as
Fig. 1 One of 10,000 shortest trees saved of length 339 resulting from maximum parsimony analyses of the Tragopogon internal transcribed
spacer data matrix. Arrows indicate branches that collapse in the strict consensus. Numbers above branches indicate branch lengths. Jackknife/
bootstrap values for nodes receiving >50% support are indicated in italics below the branches. Clade names roughly correspond to sections of
Borisova (1964).
124
Fig. 2 One of 10,000 shortest trees saved of length 243 resulting from maximum parsimony analyses of the Tragopogon external transcribed
spacer data matrix. Arrows indicate branches that collapse in the strict consensus. Numbers above branches indicate branch lengths. Jackknife/
bootstrap values for nodes receiving >50% support are indicated in italics below the branches. Clade names roughly correspond to sections of
Borisova (1964).
125
126
INTERNATIONAL JOURNAL OF PLANT SCIENCES
the result of more missing ETS data. We will first discuss the
52-taxon ITS þ ETS topology (fig. 3) and then return to the
placement of the seven taxa for which only ITS data are present (fig. 4).
Discussion
Phylogeny of Tragopogon
There are no well-supported differences between the ITS
and ETS trees. Thus, there are no instances of hard incongruence (see Seelanan et al. 1997). Because of the lack of major
conflict, the ITS and ETS data sets were combined.
The ITS þ ETS tree provided strong support (JK ¼ 100%,
BST ¼ 100%) for the monophyly of the genus Tragopogon
relative to the outgroups (fig. 3). This result agrees with the
findings of a recent molecular phylogenetic investigation of
subtribe Scorzonerinae using ITS sequences alone (Mavrodiev
et al. 2004). Sequence data also provide support for the
monophyly of several sections of Tragopogon, although relationships among sections are not always clear or well supported. We provide our best inference of phylogenetic
relationships within Tragopogon below, based on the combined ITS þ ETS tree.
The Majores clade (JK ¼ 77%, BST ¼ 72%) contains T.
stenophyllus, T. coelesyriacus, T. capitatus, T. dubius, T. pterodes, and T. afghanicus. Several of these species (i.e., T. dubius, T. capitatus, and T. afg hanicus) have been placed in
section Majores by some authors (Borisova 1964; Rechinger
1977). Tragopogon stenophyllus was included in T. crocifolius by Richardson (1976), but our trees place T. crocifolius
in the Hebecarpus clade, separate from the Majores clade
(see below). Thus, molecular phylogenetic data indicate that
T. stenophyllus is a species separate from T. crocifolius. This
result is based on two samples of T. stenophyllus from different collections (fig. 3). The other species in the Majores
clade, T. pterodes and T. coelesyriacus, have not been previously placed in any section of Tragopogon (fig. 3).
All species of the Majores clade have achenes with a long
beak and long peduncles, which are more or less plump below the flowering capitula. However, the color of the ligules
varies within the clade. For example, T. stenophyllus, T. pterodes, T. coelesyriacus, and T. afghanicus have purple or violet ligules; in contrast, T. dubius and T. capitatus have yellow
ligules. We agree with Tzvelev (1985) that the color of ligules
is not an important character in determining sections within
Tragopogon. Based on sequence data and morphology, section Majores could be expanded to include T. pterodes, T.
stenophyllus, and T. coelesyriacus. However, Tragopogon remains so poorly understood that any formal changes should
await a thorough molecular and morphological analysis of
the entire genus.
The Chromopappus clade is moderately supported
(JK ¼ 80%, BST ¼ 70%) and contains T. coloratus and T.
longirostris. Tragopogon coloratus is the type of section
Chromopappus (Borisova 1964). However, T. longirostris is
usually included in section Krascheninnikovia (Borisova
1964) and is morphologically similar to the type species of
that section, T. krascheninnikovii (Borisova 1964). Also, T.
krascheninnikovii is similar in appearance to T. coloratus
(Nazarova 1995). However, T. krascheninnikovii (which is
represented only by an ITS sequence) is placed in the Majores
clade in the ITS þ ETS tree (fig. 4), as a part of a wellsupported (JK ¼ 90%, BST ¼ 87%) clade with T. coelesyriacus and T. stenophyllus. Our data indicate that the small section Krascheninnikovia (two species) is not monophyletic.
Both species of the Chromopappus clade are morphologically
very similar to T. dubius and other members of the Majores
clade and in some treatments are actually included in section
Majores (Kuthatheladze 1957). Hence, morphology and
DNA data agree in suggesting a close relationship between
Majores and Chromopappus. In our analyses, the Majores
and Chromopappus clades are sisters, but with weak support
(JK ¼ 68%, BST ¼ 51%; fig. 3).
The Hebecarpus clade is strongly supported (JK ¼ 91%,
BST ¼ 86%) and is sister to the Majores þ Chromopappus
clade (JK ¼ 80%, BST ¼ 69%). This clade contains T. crocifolius, T. australis, T. balcanicum, T. samaritani, T. sinuatus,
and T. porrifolius. In Borisova’s (1964) system, section Hebecarpus includes only T. porrifolius. The other species of our
Hebecarpus clade have not been previously placed in any section of Tragopogon. However, T. australis is often included
in T. porrifolius at the rank of subspecies (Richardson 1976).
Similarly, Boissier (1875) indicated a very close relationship
between T. samaritani and T. crocifolius, and Richardson
(1976) considered T. samaritani to be a subspecies of T. crocifolius. However, our data indicate that both T. australis
and T. samaritani are distinct species: T. australis is well supported, as distinct from T. porrifolius, and T. samaritani is
distinct from T. crocifolius (fig. 3). The morphological description of section Hebecarpus (Borisova 1964) is very brief,
and it is difficult to understand the differences between this
section and sections Chromopappus or Krascheninnikovia.
This may be the reason why Tzvelev (1985) included T. porrifolius (sect. Hebecaprus) in section Majores.
The clade ðMajores þ ChromopappusÞ þ Hebecarpus received moderate (or low) support (JK ¼ 80%, BST ¼ 69%)
and is consistent with a broad treatment of section Majores
(i.e., including sections Majores, Chromopappus, and Hebecarpus), as proposed by Kuthatheladze (1957) and Tzvelev
(1985). The morphological characters differentiating the Majores, Chromopappus, and Hebecarpus clades require additional study.
The placement of T. kemulariae is problematic. In the ETS
tree, T. kemulariae is a member of the Majores clade
(JK ¼ 83%, BST ¼ 79%) (fig. 2), whereas in the ITS tree it
occurs in the clade corresponding largely to section Angustissimi and Brevirostres. In the ITS þ ETS tree, the relationship
of T. kemulariae is unclear (fig. 3); morphologically, T. kemulariae is very similar to T. dubius (Nazarova 1995) of the
Majores clade. However, Kuthatheladze (1957) placed T.
kemulariae in section Profundisulcati, and Borisova (1964)
placed this taxon in a monotypic section Kemularia.
The Profundisulcati clade is well supported (JK ¼ 99%,
BST ¼ 98%) (fig. 3) and contains T. albinerve, T. meskheticus, T. makaschwilii, T. armeniacus, and T. kotschyi. Three
of the taxa in the Profundisulcati clade (T. meskheticus, T.
makaschwilii, T. armeniacus) were placed in section Profundisulcati (Borisova 1964). Kuthatheladze (1957) placed section Profundisulcati as a series within section Majores,
Fig. 3 One of 10,000 shortest trees saved of length 591 resulting from maximum parsimony analyses of the Tragopogon internal þ external
transcribed spacer data matrix for 52 taxa. Arrows indicate branches that collapse in the strict consensus. Numbers above branches indicate
branch lengths. Jackknife/bootstrap values for nodes receiving >50% support are indicated in italics below the branches. Clade names roughly
correspond to sections of Borisova (1964).
Fig. 4 One of 10,000 shortest trees saved of length 616 resulting from maximum parsimony analyses of the Tragopogon internal þ external
transcribed spacer data matrix with seven additional species represented only by ITS sequences (boldface). Arrows indicate branches that collapse
in the strict consensus. Numbers above branches indicate branch lengths. Jackknife/bootstrap values for nodes receiving >50% support are
indicated in italics below the branches.
MAVRODIEV ET AL.—PHYLOGENY OF TRAGOPOGON
129
Fig. 5 Reconstruction of the diversification of achene morphology in Tragopogon. Character states were mapped onto one randomly selected
maximum parsimony tree using MacClade and the all most parsimonious state optimization method.
but based on our phylogenetic analysis the Profundisulcati
clade is very distinct from the Majores clade. Tragopogon albinerve has not been previously placed in any section of
Tragopogon, but morphologically it is very similar to another
member of section Profundisulcati, the polyploid species T.
buphtalmoides (Matthews 1975).
Tragopogon kotschyi of the Profundisulcati clade was
placed in section Sosnovskya (Rechinger 1977), a small section whose species are morphologically intermediate between
species of sections Brevirostres and Profundisulcati (Borisova
1964). In our phylogenetic analyses, section Sosnovskya is
not monophyletic. The four species of section Sosnovskya
used in our analysis, T. reticulatus, T. charadzae, T. kotschyi,
and T. marginatus, all belong to distinct clades (Brevirostres,
Angustissimi, Profundisulcati, and Collini, respectively), each
of which receives support >50% (figs. 1–4). Nazarova
(1991) disagreed with the recognition of section Sosnovskya
in her treatment of Tragopogon. Morphological investigations of section Profundisulcati are needed.
Our data also indicate that the Profundisulcati clade includes T. aureus and T. fibrosus. Both of these taxa are
placed in the Profundisulcati clade based only on their ITS
sequences (fig. 4). These two species have not been placed in
any section and are poorly understood morphologically. For
example, the achenes of T. aureus are unknown (Matthews
1975). Tragopogon fibrosus is a rare, narrow endemic of
North Anatolia known only from three specimens, two of
which are very old (1895) and inaccessible (LD). Matthews
(1975) remarked that this species is known only from the
type collections, but in our treatment we used the unpublished collections of T. fibrosus obtained by A. Khokhriakov
(MHA) from Turkey (table 1).
The Collini clade is well supported (JK ¼ 98%, BST ¼
97%) and contains T. ruber, T. bronmuelleri, T. elongatus,
and T. jezdianus. Three additional species, T. montanus, T.
collinus, and T. marginatus, for which we have only ITS sequences, are also placed in the Collini clade (fig. 4). Most
of these taxa (T. ruber, T. montanus, T. elongatus, and T.
collinus) are members of section Collini (¼Rubriflori)
(Borisova 1964). Members of the Collini clade are perennial,
with purplish ligules and with achene beaks that are equal to
or shorter than the body of the achene. Tragopogon marginatus, with yellow ligules, has been placed in either section
Sosnovskya (Borisova 1964) or section Collini (Kuthatheladze 1957). The rare endemic species T. bronmuelleri was
placed in section Profundisulcati (Rechinger 1977). In our
analysis, however, this species is sister to T. ruber, a species
of section Collini (figs. 3, 4). The diagnostic attributes of the
Collini clade require investigation.
The sister-group relationship of the Profundisulcati and
Collini clades is well supported (JK ¼ 99%, BST ¼ 96%).
This result provides some support for a broader treatment of
section Collini that combines species of sections Collini and
Profundisulcati. However, our data also support the reciprocal monophyly of the Collini and Profundisulcati clades, so
either a broad section Collini or smaller sections Collini and
130
INTERNATIONAL JOURNAL OF PLANT SCIENCES
Fig. 6 Reconstruction of the diversification of life history in Tragopogon. Character states were mapped onto one randomly selected maximum
parsimony tree using MacClade and the all most parsimonious state optimization method.
Profundisulcati would be consistent with the phylogenetic
tree.
The Brevirostres clade is well supported (JK ¼ 95%,
BST ¼ 86%) and consists of T. kindingeri, T. brevirostris, T.
dasyrhynchus, T. reticulatus, T. dubjanskyi, T. ruthenicus, T.
heterospermus, T. undulatus, T. podolicus, and T. floccosus
(fig. 3). Many of the species of the Brevirostres clade (T. reticulatus, T. brevirostris, T. dasyrhynchus, T. undulatus, and
T. podolicus) are from section Brevirostres (Kuthatheladze
1957; Tzvelev 1985; Nazarova 1991). The clade also contains two species (T. ruthenicus and T. dubjanskyi) placed in
section Nikitinia by Borisova (1964). Tzvelev (1985) also included T. ruthenicus and T. dubjanskyi in section Brevirostres
(see also Klokov 1981).
The Brevirostres clade also contains T. heterospermus,
a species of section Bessera (Borisova 1964). However, Tzvelev (1985) included all Bessera species in section Brevirostres,
and the diagnostic morphological attributes of section Bessera are unclear (Tzvelev 1985). However, not all species of
section Brevirostres appear in the Brevirostres clade. Tragopogon serotinus and T. filifolius, species that are usually included in section Brevirostres (Kuthatheladze 1957; Borisova
1964), both belong to a distinct, well-supported clade that
does not correspond to a section, which we are calling the
Angustissimi clade (JK ¼ 91%, BST ¼ 72%).
The Tragopogon and Brevirostres clades are sisters in the
ITS þ ETS tree, a relationship that receives conflicting levels
of support (JK ¼ 90%, BST ¼ 67%) (fig. 3). The Tragopogon clade is fairly well supported (JK ¼ 88%, BST ¼ 74%)
and is composed of T. hayeki, T. tommasinii, T. longifolius,
T. pratensis, T. orientalis, T. minor, and T. trachycarpus
(fig. 3). This clade contains the type species of the genus
Tragopogon (T. pratensis). This species and three others (T.
orientalis, T. trachycarpus, and T. minor) have been placed in
section Tragopogon. The other taxa of the Tragopogon clade
(T. hayeki, T. tommasinii, and T. longifolius) have not been
placed in any section of Tragopogon. However, two of these
taxa (T. hayeki and T. tommasinii) are morphologically very
similar to T. pratensis s. l. (Richardson 1976; see also
Boissier 1875). Some workers have included T. orientalis
within T. pratensis (Richardson 1976). However, our results
clearly indicate that T. pratensis is phylogenetically distinct
from T. orientalis. A well-supported (JK ¼ 92%, BST ¼ 84%)
subclade of T: hayeki þ T: tommasinii is sister to T:
longifolius þ T: orientalis (JK ¼ 91%, BST ¼ 88%). These
four species form a well-supported clade (JK ¼ 98%, BST ¼
95%) that is in turn sister (JK ¼ 74%, BST ¼ 75%) to a subclade (JK ¼ 73%, BST ¼ 84%) of T: pratensis þ T: minor.
Molecular data indicate that T. lamottei is a distinct species. This species appears as sister to all other species of the
MAVRODIEV ET AL.—PHYLOGENY OF TRAGOPOGON
Tragopogon clade (JK ¼ 65%; fig. 3). However, the phylogenetic position of T. lamottei is not well supported. Tragopogon lamottei has been considered closely related to T.
pratensis (T. pratensis ssp. lamottei [Rouy] O. de Bolós &
J. Vigo), but our results do not support this.
Some species of section Tragopogon not sampled here (T.
karelinii S. Nikit., T. tomentosulus Boriss., T. gracilis D.
Don., T. subalpinus S. Nikit., and T. sibiricus Ganesh.) have
purplish ligules, and the systematic placement of all of these
Asian species is uncertain. For example, Nikitin (1937)
placed all of these taxa (except T. sibiricus) with species from
section Collini (¼Rubriflori). Species from section Tragopogon are mostly biennial plants, so the only clear morphological difference between sections Tragopogon and Brevirostres
is the length of the achenes and the achene beak (Tzvelev
1985). All members of section Brevirostres and the Brevirostres clade have achenes with very short beaks (excluding
T. reticulatus) and yellow ligulate flowers. Members of section Tragopogon and the Tragopogon clade have achenes
with obvious beaks (excluding T. trachycarpus, with a very
short beak) that usually are shorter than the body of the
achene. Klokov (1981) suggested that species of section
Brevirostres and some species of section Tragopogon be combined to form one broad section.
Artemczhyk (1948) suggested that T . heterospermus and
T. podolicus (sect. Brevirostres) are closely related to T. orientalis and T. pratensis from section Tragopogon. According
to our topology, however, T. heterospermus and T. podolicus
are in the Brevirostres clade, while T. orientalis and T. pratensis are in the Tragopogon clade.
Tragopogon kindingeri is also morphologically similar to
T. pratensis (Richardson 1976) but is sister to the rest of the
Brevirostres clade and has not been previously placed in any
section of Tragopogon. Although some have suggested that
T. longifolius is a subspecies of T. brevirostris (Richardson
1976; also Boissier 1875), our results place T. longifolius as
a part of a well-supported subclade (JK ¼ 98%, BST ¼ 95%),
with T. orientalis, T. tommasinii, and T. hayeki, in the Tragopogon clade. We agree with the conclusions of Artemczhyk
(1948) and Borisova (1964), who suggested that T. brevirostris
is a narrow endemic of the Caucasus and that T. longifolius
should not be included in T. brevirostris.
We identified a previously unrecognized clade (JK ¼ 91%,
BST ¼ 72%) that contains T. pusillus (sect. Tuberosi or Collini), T. filifolius (sect. Brevirostres), T. latifolius (2n ¼ 12)
(sect. Profundisulcati or Majores), T. segetum (sect. Angustissimi), T. acanthocarpus (sect. Profundisulcati or Majores),
T. serotinus (sect. Brevirostres), and T. charadzae (sect.
Sosnovskya or Brevirostres) (fig. 3). We are calling this clade
Angustissimi because it contains both species of section
Angustissimi (T. segetum and T. sosnowskyi). Tragopogon
sosnowskyi is included in the subclade based only on its ITS
sequence (fig. 4). Within this clade, T. serotinus and T. charadzae form a clade (JK ¼ 89%, BST ¼ 71%). These two species are morphologically similar and were included in section
Brevirostres by Kuthatheladze (1957).
Diagnostic attributes of the Angustissimi clade are problematic. Most species of the Angustissimi clade have similar
geographic distributions that center on the Caucasus. Tragopogon segetum, T. filifolius, T. serotinus, T. sosnowsky, and
131
T. charadzae are all restricted to the Caucasus. Tragopogon
acanthocarpus occurs in a local area in the south Caucasus
and in several provinces of northern and central Iran. Tragopogon latifolius is mostly a polyploid species, distributed in
the south Caucasus (Nazarova 1995) and Turkey (Matthews
1975), but we used only the diploid race of T. latifolius. This
rare race of T. latifolius is known only from a few places in
Armenia (Nazarova 1984). Tragopogon pusillus is reported
from Turkey (Matthews 1975), Iran (Borisova 1964), Crimea
(Tzvelev 1985), and Turkmenistan (‘‘Turkomania’’) (Borisova
1964). However, Matthews (1975, p. 667) indicated that the
Turkish record of T. pusillus ‘‘needs confirmation.’’ Rechinger
(1977) described an Iranian population of T. pusillus as
a new species (T. gongylorrhizus Rech.). The distribution of
Fig. 7 Types of achenes in Tragopogon. A, T. ruber: achene with
short beak (b). B, T. ruthenicus: achene without beak. C, T. capitatus:
achene with a long beak (b). Scale bar ¼ 10 mm. Original photographs.
132
INTERNATIONAL JOURNAL OF PLANT SCIENCES
T. pusillus in Crimea and Turkmenistan is very local. For example, in Turkmenistan, T. pusillus is known only from one
location (Borisova 1964). The close relationship of the flora
of the Crimea/Caucasus region to the Turkmenistan flora is
also seen in the moss flora (Abramov et al. 1987). In Crimea,
T. pusillus is known from just a narrow area on the southern
coast (Tzvelev 1985), and the nature of this disjunct area
of T. pusillus is the subject of discussion (Tzvelev 1985).
Tzvelev (1985) suggested that many species from Crimea, especially T. pusillus, with clear ‘‘roots’’ in the Caucasus, are the
result of the previous existence of a mountain region between
Crimea and the modern Caucasus before the origin of the
modern Black Sea. Additional morphological investigations
are needed to clarify the diagnostic attributes of this geographically well-demarcated clade of Tragopogon.
Character Evolution
The length of the beak of the achene has been an important taxonomic character in Tragopogon (Borisova 1964;
Richardson 1976; Tzvelev 1985), differing among species
and sections. It is difficult to describe all variants of the ratio
between the length of the beak and the length of the body of
the achene. Following Klokov (1981) and Tzvelev (1985), we
split all Tragopogon species into two large groups based on
achene morphology: the first group contains species without
a beak, and the second group contains species with a beak,
either long or short (fig. 7). The outgroups have achenes
without beaks.
Using the 1000 randomly selected shortest trees with the
all most parsimonious states option, the ancestral state of
Tragopogon is an achene with a beak (fig. 5; fig. 7A, 7C).
The ancestral state of each clade (Majores, Chromopappus,
Hebecarpus, Profundisulcati, Collini, Brevirostres, Tragopogon, and Angustissimi) is also an achene with a beak.
Achenes without beaks appear to have evolved two times,
once in the ancestor of the Brevirostres clade and again in T.
trachycarpus in the Tragopogon clade. Within the Brevirostres clade there was a subsequent reversal to a beaked
achene in T. reticulatus (fig. 5).
Mavrodiev et al. (2004) discussed the evolution of habit
(annual, biennial, or perennial) across the entire subtribe
Scorzonerinae, using 20 representatives of Tragopogon (biennial or perennial). That broad analysis revealed that the an-
cestral state of Tragopogon is the perennial habit, with
multiple changes to biennial. In this article, we have repeated
the reconstruction of habit, using many more species (59) of
Tragopogon. Our results confirm our earlier reconstruction
while providing new insights within Tragopogon and with
close relatives of Tragopogon as outgroups. The ancestral
habit of the genus Tragopogon is again reconstructed as perennial (fig. 6). However, the ancestral state of some clades,
such as Majores, Chromopappus, Hebecarpus, and Tragopogon, is biennial. The ancestral state of the Brevirostres and
Angustissimi clades is equivocal.
Our phylogenetic analysis has provided the first assessment
of phylogenetic relationships within the large, taxonomically
complex genus Tragopogon. Many of the sections recognized
by morphology are supported as monophyletic by analyses of
gene sequence data. In addition, possible relationships are indicated for numerous species of Tragopogon that have never
been assigned to any section within the genus. We observed
a general correlation between phylogeny and geography,
with closely related taxa occurring in the same geographic region. For example, most species of the Angustissimi clade
have identical geographic distributions that center on the
Caucasus. This phylogenetic analysis should be considered
only a first step toward resolving relationships within this
poorly understood genus. Not only are additional molecular
phylogenetic analyses required that employ additional genes
and more taxa, but in addition, the genus requires a thorough
morphological analysis.
Acknowledgments
This research was supported by National Science Foundation (NSF)-North Atlantic Treaty Organization grant DGE0209500, an REU supplement to NSF grant DEB-9707868,
the University of Florida Research Foundation, and the
School of Biological Sciences, Washington State University,
for support of undergraduate research. We thank the Royal
Botanic Gardens, Kew, for access to specimens of Tragopogon and for permission to remove samples for DNA analysis. We also thank Professor Estella A. Nazarova (Botanical
Institute of Armenia [Erevan]) for important consultations
and for the collecting of material.
Literature Cited
Abramov II, AL Abramova, IV Sirotina 1987 The Crimea-Caucasian
relationships of the moss flora of the Turkomania. Nov Sist
Nizshikh Rast 24:179–184. (In Russian.)
Álvarez I, JF Wendel 2003 Ribosomal ITS sequences and plant
phylogenetic inference. Mol Phylogenet Evol 29:417–434.
Artemczhyk IV 1937 Tragopogon in Ukrainian flora. Trans Bot Inst
Charkov State Univ 2:23–65. (In Ukrainian.)
——— 1948 On the origin of sand-land Tragopogon species of the
European part of the USSR. Sci Trans Biol Fac Chernigov State Univ
1:85–91. (In Ukrainian with Russian abstract.)
Baldwin BG, S Markos 1998 Phylogenetic utility of the external
transcribed spacers (ETS) of 18S–26S rDNA: congruence of ETS
and ITS trees of Calycadenia (Compositae). Mol Phylogenet Evol
10:449–463.
Blackmore S 1982 Palynology of subtribe Scorzonerinae (Compositae: Lactuceae) and its taxonomic significance. Grana 21:
149–160.
Boissier EP 1875 Tragopogon. Pages 744–755 in EP Boissier, ed.
Flora orientalis. Vol 3. H. Georg, Geneva.
Borisova AG 1964 Tragopogon. Pages 115–196 in BK Shishkin, ed.
Flora of the USSR. Vol 29. Nauka, Moscow. (In Russian.)
Bremer K 1994 Asteraceae, cladistics and classification. Timber,
Portland, OR.
Clevinger JA, JL Panero 2000 Phylogenetic analysis of Silphium and
subtribe Engelmanninae (Asteraceae: Heliantheae) based on ITS
and ETS sequence data. Am J Bot 37:565–572.
Darlington CD 1956 Chromosome botany and the origins of
cultivated plants. 3rd ed. Hafner, New York.
MAVRODIEV ET AL.—PHYLOGENY OF TRAGOPOGON
De Candolle AP 1838 Prodromus systematis naturalis regni vegetabilis. Vol 7. Treuttel & Wurtz, Paris.
Doyle JJ, JL Doyle 1987 A rapid DNA isolation procedure from
small quantities of fresh leaf tissue. Phytochem Bull 19:11–15.
Farris JS, VA Albert, M Kallersjo, D Lipscomb, AG Kluge 1996
Parsimony jackknifing outperforms neighbor-joining. Cladistics 12:
99–124.
Felsenstein J 1985 Confidence limits on phylogenetics: an approach
using the bootstrap. Evolution 39:783–791.
Klokov MV 1981 Ukrainian sand’s floristic complex. Pages 90–133
in MV Klokov, ed. News of higher and lower plant systematics.
Naukova Dumka, Kiev. (In Russian.)
Kuthatheladze SI 1957 Caucasian representatives of Tragopogon
genus. Monogr Bot Inst Acad Sci Georgia, ser. A, 2:3–79. (In
Russian with Georgian abstract.)
Lee J, BG Baldwin, LD Gottlieb 2002 Phylogeny of Stephanomeria
and related genera (Compositae–Lactuceae) based on analysis of
18S–26S nuclear rDNA ITS and ETS sequences. Am J Bot 89:
160–168.
——— 2003 Phylogenetic relationship among the primary North
American genera of Cichoriae (Compositae) based on analysis of
18S–26S nuclear rDNA ITS and ETS sequences. Syst Bot 28:
616–626.
Maddison WP, DR Maddison 2002 MacClade, version 4.05:
analysis of phylogeny and character evolution. Sinauer, Sunderland, MA.
Markos S, BG Baldwin 2001 Higher-level relationship and major
lineages of Lessingia (Compositae, Astereae) based on nuclear
rDNA internal and external transcribed spacer (ITS and ETS)
sequences. Syst Bot 26:168–183.
Matthews VA 1975 Tragopogon. Pages 657–668 in PH Davis, ed.
Flora of Turkey and East Aegean Islands. Vol 5. Edinburgh
University Press, Edinburgh.
Mavrodiev EV, CE Edwards, DC Albach, MA Gitzendanner, PS Soltis,
DE Soltis 2004 Phylogenetic relationships in subtribe Scorzonerinae (Asteraceae: Cichorioideae: Cichorieae) based on ITS sequence
data. Taxon 53:699–712.
Nazarova EA 1984 Chromosome numbers in the Caucasian representatives of the families Asteraceae, Brassicaceae, Fabaceae,
Linoniaceae. Bot Zh 63:972–975. (In Russian.)
——— 1991 Karyotypical evolution in genus Tragopogon L.
(Lactuceae, Asteraceae). Pages 116–134 in AL Takhtajan, ed. Flora,
vegetation, and vegetable resources of Armenia. Academy of
Science of Armenia, Erevan. (In Russian with Armenian abstract.)
——— 1995 Tragopogon. Pages 111–131 in AL Takhtajan, ed. Flora
of Armenia. Vol. 9. Koeltz, Koeningstein. (In Russian.)
——— 1997 Karyosystematic investigation of the genus Scorzonera
L. s. l. (Lactuceae, Asteraceae). Caryologia 50:250–261.
Nikitin SA 1930 On systematic of sand races of Tragopogon. Proc
Main Bot Gard USSR 29:484–491. (In Russian.)
——— 1937 New species of the genus Tragopogon. Bot Mater Herb
Bot Inst USSR Acad Sci 7:257–274. (In Russian.)
Ownbey M 1950 Natural hybridization and amphiploidy in the
genus Tragopogon Am J Bot 37:487–499.
133
Ownbey M, GB McCollum 1953 Cytoplasmic inheritance and
reciprocal amphiploidy in Tragopogon. Am J Bot 40:788–796.
Rambaut A 2003 Se-Al: a manual sequence alignment editor. http://
evolve.zoo.ox.ac.uk/software.html?id=seal.
Rechinger KH 1977 Tragopogon. Pages 83–120 in KH Rechinger,
ed. Flora Iranica. Vol 122. Akademische Druck und Verlagsanstalt,
Graz.
Regel C 1937 Beitrag zur Kentnis einiger Tragopogon. Scr Horti Bot
Univ Vytauti Magni 5:19–31.
——— 1955 Die Verbreitung einiger europaeisch-asiatischer
Tragopogon-Arten. Ber Schweiz Bot Ges 65:251–262.
Richardson IBK 1976 Tragopogon. Pages 322–326 in TG Tutin, VH
Heywood, NA Burges, CH Valentine, eds. Flora Europaea. Vol 4.
Cambridge University Press, Cambridge.
Saar DE, NO Polans, PD Sorensen 2003 A phylogenetic analysis of
the genus Dahlia (Asteraceae) based on internal and external
transcribed spacer regions of nuclear ribosomal DNA. Syst Bot 28:
627–639.
Seelanan TH, H Schnabel, JF Wendel 1997 Congruence and consensus in the cotton tribe. Syst Bot 22:259–290.
Soltis DE, TG Collier, ML Edgerton 1991 Chloroplast DNA
variation within and among genera of the Heuchera group
(Saxifragaceae): evidence for chloroplast transfer and paraphyly.
Am J Bot 78:1091–1112.
Soltis DE, PS Soltis 1998 Choosing and approach and an appropriate
gene for phylogenetic analysis. Pages 1–42 in DE Soltis, PS Soltis, JJ
Doyle, eds. Molecular systematics of plants. 2nd ed. Kluwer,
Boston.
Soltis DE, PS Soltis, JC Pires, A Kovarik, JA Tate 2004 Recent and
recurrent polyploidy in Tragopogon (Asteraceae): cytogenetic,
genomic, and genetic comparisons. Bot J Linn Soc 82:485–501.
Swofford DL 2002 PAUP*: phylogenetic analysis using parsimony
(*and other methods). Version 4.10b. Sinauer, Sunderland, MA.
Thompson JD, TJ Gibson, F Plewniak, F Jeanmougin, DG
Higgins 1997 The Clustal X Windows interface: flexible strategies
for multiple sequence alignment aided by quality analysis tools.
Nucleic Acids Res 24:4876–4882.
Tzvelev NN 1985 Genus Tragopogon L. (Asteraceae) in the European part of the USSR. Pages 238–250 in T Egorova, ed. News in
higher plant systematics. Vol 22. Nauka, Leningrad. (In Russian.)
Urbatsch LE, RP Roberts, V Karaman 2003 Phylogenetic evaluation
of Xylothamia gundlachia and related genera (Asteraceae, Astereae)
based on ETS and ITS nrDNA sequence data. Am J Bot 90:
634–649.
Wen J, EA Zimmer 1996 Phylogeny and biogeography of Panax L.
(the ginseng genus, Araliaceae): inference from its sequences of the
nuclear ribosomal DNA. Mol Phylogenet Evol 6:167–177.
White TJ, T Brunce, S Lee, J Taylor 1990 Amplification and direct
sequencing of fungal ribosomal RNA genes for phyloginetics. Pages
315–322 in M Innis, D Gelfand, J Sinsky, T White, eds. PCR
protocols: a guide to methods and applications. Academic Press,
San Diego, CA.
Wilson DF 1983 Karyotypes of Tragopogon (Compositae: Lactuceae).
Brittonia 35:341–350.