ArticlePDF Available

Phylogeny of Cuscuta Subgenus Cuscuta (Convolvulaceae) Based on nrDNA ITS and Chloroplast trnL Intron Sequences

  • Real Jardín Botánico-CSIC, Consejo Superior de Investigaciones Científicas

Abstract and Figures

Within the parasitic genus Cuscuta, three subgenera have been recognized based on characters of the styles and stigmas. Cuscuta subgenus Cuscuta, with free styles and conical to elongated stigmas, is the most diversified in the Old World with about 25 species. We present the first phylogenetic study of the subgenus using nuclear ITS rDNA and chloroplast trnL intron sequences. Sequences of almost all species of the subgenus were obtained and several individuals of particular taxonomic difficulty or wide geographical distribution were sampled. Both maximum parsimony and Bayesian analyses were performed to evaluate the monophyly of the sections previously accepted in Yuncker's monograph and to investigate phylogenetic relationships between the species. The monophyly of the subgenus could not be tested with our sampling but using three species of subgenus Monogyna as outgroup, the South African section Pachystigma was sister to the remaining species of subgenus Cuscuta. Section Epistigma plus C. capitata are resolved as monophyletic in all analyses. The distinctive C. babylonica was sister to that clade on the ITS trees but it was not resolved on the trnL trees. Two monophyletic groups within section Cuscuta, first identified here, included the species of tropical African distribution in one case and C. europaea, C. approximata, and C. balansae in the other. Factors influencing the taxonomic difficulty of many species in the subgenus include lack of morphological characters, parallelism and gene flow between closely and not so closely related species. Evidence of reticulation events or within species recombination were obtained by both polyphyletic intra-individual ITS sequences and conflicting topologies of the nuclear and plastid trees.
Content may be subject to copyright.
Phylogeny of Cuscuta subgenus Cuscuta (Convolvulaceae) Based on nrDNA
ITS and Chloroplast trnL Intron Sequences
and MARI
Real Jardı
´n Bota
´nico, Consejo Superior de Investigaciones Cientı
´ficas, Plaza de Murillo 2, 28014 Madrid,
Communicating Editor: James F. Smith
ABSTRACT.Within the parasitic genus Cuscuta, three subgenera have been recognized based on characters of the
styles and stigmas. Cuscuta subgenus Cuscuta, with free styles and conical to elongated stigmas, is the most
diversified in the Old World with about 25 species. We present the first phylogenetic study of the subgenus using
nuclear ITS rDNA and chloroplast trnL intron sequences. Sequences of almost all species of the subgenus were
obtained and several individuals of particular taxonomic difficulty or wide geographical distribution were sampled.
Both maximum parsimony and Bayesian analyses were performed to evaluate the monophyly of the sections
previously accepted in Yuncker’s monograph and to investigate phylogenetic relationships between the species.
The monophyly of the subgenus could not be tested with our sampling but using three species of subgenus
Monogyna as outgroup, the South African section Pachystigma was sister to the remaining species of subgenus
Cuscuta. Section Epistigma plus C. capitata are resolved as monophyletic in all analyses. The distinctive C. babylonica
was sister to that clade on the ITS trees but it was not resolved on the trnL trees. Two monophyletic groups within
section Cuscuta, first identified here, included the species of tropical African distribution in one case and C. europaea,
C. approximata,andC. balansae in the other. Factors influencing the taxonomic difficulty of many species in the
subgenus include lack of morphological characters, parallelism and gene flow between closely and not so closely
related species. Evidence of reticulation events or within species recombination were obtained by both polyphyletic
intra-individual ITS sequences and conflicting topologies of the nuclear and plastid trees.
KEYWORDS:Convolvulaceae, Cuscuta, holocentric chromosomes, molecular phylogeny, parasitic plants.
The genus Cuscuta L. (Convolvulaceae) consists
of approximately 160 species (Yuncker 1932) of
hemiparasitic or holoparasitic herbs that attach to
the aerial parts of the host by stem haustoria. The
plants are annual or overwinter inside the host
tissue. Cuscuta (dodders) are widespread in almost
all the temperate regions of the world but most
species diversity is found in the Americas. The
genus is characterized by morphological reduction
of vegetative organs in which roots have been lost
and leaves transformed into minute scales. Cuscuta
is a taxonomically difficult genus because of these
morphological reductions, morphological parallel-
ism among species and, in some cases, infraspecific
variability. Many of the vegetative characters used
in flowering plant systematics at the species level
have been lost, thus most of the useful taxonomical
characters involve flowers and fruits.
The systematic position of Cuscuta has been
traditionally within or closely related to Convol-
vulaceae. An introduction to the history of the
systematic placement of the genus was reviewed
by Hunziker (1949–50). Many authors, e.g. Cron-
quist (1981), Takhtajan (1969), Goldberg (1986) or
Chrtek and Osbornova (1991) have accepted the
family Cuscutaceae, first proposed by Dumortier
(1829) based on its parasitic habit and the
morphological, embryological and anatomical
modifications that stem from this habit. However,
molecular data indicate that Cuscuta is a derived
member of Convolvulaceae, although the non-
parasitic members of Convolvulaceae most closely
related to Cuscuta could not be determined
´et al. 2002; Stefanovic
´and Olmstead
Cuscuta is generally divided into three subgenera
based on morphological features of the styles and
stigmas. Engelmann (1859) was the first to consider
the importance of these features for the general
classification within the genus and arranged the
species of Cuscuta into three groups. In the last
monograph of the entire genus (Yuncker 1932) and
in earlier revisions of the American species
(Yuncker 1921, 1922, 1923), Engelmann’s groups
were treated at the rank of subgenus. Studies have
shown that the three subgenera also are supported
by anatomical (Mirande 1901) and karyological
features (Pazy and Plitmann 1995; Garcı
´a and
Castroviejo 2003). The three subgenera can be
separated as follows: Subgenus Monogyna, charac-
terized by a gynoecium with styles partially or
completely united along its entire length and
stigmas globose to elongate; it is represented by
approximately nine species of the Old World, most
of them from Asia, and one species from SE North
America (C. exaltata). Subgenus Grammica has two
styles different in length, free from the base and
short, frequently globose stigmas; it is the most
species-rich and morphologically diverse group
comprising about 135 species (Yuncker 1932),
Systematic Botany sbot-32-04-19.3d 21/9/07 09:44:11 899 Cust #06-59
Author for correspondence (
Systematic Botany (2007), 32(4): pp. 899–916
#Copyright 2007 by the American Society of Plant Taxonomists
mostly from the Americas. Subgenus Cuscuta is
characterized by a gynoecium with two styles of
the same length, free from the base and conical to
elongate stigmas; this group consists of approxi-
mately 25 species originally from the Old World:
Europe, northern, eastern, and southern Africa,
Asia (except in the southeast), and the Indian
subcontinent. Some species, such as C. approximata
or C. epithymum, have been introduced and
naturalized in North and South America, Australia
and New Zealand.
Subgenus Cuscuta is the largest group in the Old
World. Many of its species can be distinguished by
their compact globose inflorescences of nearly
sessile flowers. Others, however, have cymose
branched inflorescences, clearly pedicellate flow-
ers, or both. Engelmann (1859) and Yuncker (1932)
recognized four sections based on the dehiscence
of the fruit and the morphology of styles and
stigmas. The most distinctive of them is sect.
Pachystigma, endemic to South Africa. According
to Yuncker, it consists of five species with cymose,
branched inflorescences and conical to cylindrical
stigmas that are broader than the styles (Engel-
mann based the name of the section in this
character). The species of sect. Epistigma have
subulate stigmas that are sessile or on very short
conical styles and fruits dehiscent by an irregular
line; according to Yuncker it includes four species
from central and western Asia, one of which (C.
pedicellata) is also present in Egypt, northern
Sudan, and the European parts of Turkey and
Russia. The monotypic sect. Clistococca includes C.
capitata with differentiated styles and indehiscent
fruit. This latter character is the basis for separating
the species in its own section. The largest group
within the subgenus is sect. Cuscuta, in which
Yuncker (1932) recognized 17 species widely
distributed in the Old World. It includes species
with differentiated styles, stigmas as broad as the
styles, and fruit dehiscent by a regular line.
Some of the characters used in the sectional
classification of subgenus Cuscuta are of doubtful
value for infrageneric classification. For example,
the indehiscent fruit character was used to segre-
gate the Asian species C. capitata in its own section.
However, other species in the subgenus, different
morphologically and not geographically proximal
such as C. triumvirati (Morocco and Spain, sect.
Cuscuta)andC. gerrardii (South Africa, sect.
Pachystigma), also have indehiscent fruits. The
dehiscence by an irregular line characteristic of
sect. Epistigma is, however, also present in C.
babylonica, a species that has differentiated styles
and is traditionally included in sect. Cuscuta. More
problematic are the characters used by Yuncker to
separate subsections in sect. Cuscuta. These include
the fleshiness of calyx and corolla lobes and the
number of flower parts, features that sometimes
are variable in the same species or even in the same
An important feature of all species of subg.
Cuscuta investigated to date is that they are
characterized cytologically by inverted meiosis
and holocentric chromosomes (Pazy and Plitmann
1991, 1994, 1995; Garcı
´a 2001; Garcı
´a and Castro-
viejo 2003; Guerra and Garcı
´a 2004). Species of the
other two subgenera have shown regular meiosis
and monocentric chromosomes with a wide range
of size variation. The largest have been found in
species of subg. Monogyna with chromosomes up
to 23 mm long and the shortest in species of subg.
Grammica with chromosomes as short as 0.4 mm
long, whereas in the type subgenus chromosome
size ranges between 2 and 8 mm (Pazy and
Plitmann 1995). Holocentric chromosomes lack
centromeres, and microtubules are attached along
the entire length of the chromosomes in both
mitosis and meiosis (for a review of kinetochore
function in holocentric chromosomes, see Dern-
burg 2001). This type of chromosome has also been
found in the dicot genus Drosera (Sheikh et al. 1995)
and Myristica fragans (Flach 1966), whereas in
monocots they have been documented in all genera
of Cyperaceae, Juncaceae, and Chionographis of
Melanthiaceae (Greilhuber 1995). In groups where
both types of chromosomes are present, this
remarkable feature may support the monophyly
of those species. The species of subg. Cuscuta that
have holocentric chromosomes belong to sect.
Cuscuta and one of them, C. pedicellata, to sect.
Epistigma. However, none of the species of sect.
Pachystigma have been analyzed cytologically and
the meiosis and chromosome type is unknown.
Vasudevan (1975) indicated that species of subg.
Cuscuta have monocentric chromosomes. Howev-
er, the only photograph in this publication is a plate
corresponding to C. planiflora that showed a quad-
ripartite figure of bivalents in meiotic metaphase-I,
typical of inverted meiosis and holocentric chro-
In general, species delimitation in subgenus
Cuscuta, especially species belonging to the type
section, is difficult. Morphological reduction, small
size, and lack of taxonomically useful features,
together with adaptive parallelism of the remain-
ing few characters, contribute to this problem.
However, some species such as C. europaea,C.
castroviejoi,C. triumvirati, or the species of sect.
Pachystigma are well characterized morphological-
ly and can be readily distinguished. The opposite
situation is the problematic taxonomic delimitation
Systematic Botany sbot-32-04-19.3d 21/9/07 09:44:11 900 Cust #06-59
of most of the species of sect. Cuscuta, especially
those with inflorescences in compact glomerules.
Although it is possible to distinguish clear taxo-
nomic units at the extreme ends of the variation
spectrum, there is a range of forms that merge from
one species into the other thus making it difficult to
establish specific limits. This difficulty is well
exemplified by the complex of three species that
display an extensive degree of morphological
variation and a wide geographical distribution: C.
epithymum,C. planiflora, and C. approximata. Not
only do intermediate morphological forms exist
between these, but also between other species of
more limited geographical distribution such as C.
balansae or C. palaestina. This taxonomic complexity
has led to the publication of numerous names at
specific and infraspecific ranks. The biological
factors underlying this taxonomic complexity are
unknown. In contrast to the large number of
publications concerning physiology, ultrastructure,
and molecular evolution of the plastids, and host-
parasite relationships in the genus (Costea and
Tardif 2006), little work has focused on its re-
productive biology. Plitmann (1991, 2002) is the
only author who has focused in these aspects and
he indicates that autogamy and agamospermy exist
in species of subg. Cuscuta. An important source of
morphological variation and taxonomic difficulties
in plants is derived from reticulate evolution.
´et al. (2007) have documented species
of hybrid origin in subg. Grammica from conflicting
nuclear and plastid gene trees, but these processes
have never been described in subg. Cuscuta.
Chromosome studies in subg. Cuscuta indicate that
polyploidy is frequent, but the only indication of
the possible allopolyploid origin of a species is the
work on C. approximata by Guerra and Garcı
Here we present a phylogenetic analysis based
on molecular sequence data from two regions,
nuclear ribosomal internal transcribed spacer DNA
(ITS) and chloroplast trnL intron with sampling
from nearly all species of subg. Cuscuta. Phyloge-
netic relationships were inferred using both max-
imum parsimony and Bayesian inference. The
main aims of this study were 1) to phylogenetically
analyze the sections proposed by Engelmann and
accepted by Yuncker in the subgenus, and 2) to
determine the monophyly of problematic species
such as C. epithymum,C. planiflora, and C. approx-
Taxon Sampling. Herbarium or field collected and silica
dried material was used for this study. A total of 69
specimens from 34 taxa listed in the Appendix were selected
for DNA extraction, including 28 taxa of subg. Cuscuta,4of
subg. Monogyna and 3 of subg. Grammica. The specimens
included were selected from approximately 6,500 collections
used for a taxonomic revision of the subgenus (M. A. Garcı
in preparation). Additionally, sequences of 7 species of subg.
Grammica were obtained from GenBank. Almost all species of
subg. Cuscuta recognized by Yuncker (1932) were sampled.
Three of these species could not be sampled because the
collections available are old and in some cases only the types
are available. In other cases, no permission was obtained to
extract from more recent collections. These species are C.
gerrardii (5C. cucullata), C. obtusata,andC. stenoloba
(probably a teratological form of C. epithymum). Other species
recognized by Yuncker were considered synonyms, such as
C. brevistyla and C. letourneuxii, both under C. planiflora,and
C. madagascarensis under C. abyssinica. Other species de-
scribed after Yuncker’s monograph have been sampled, all
belonging to the type section: C. pretoriana,C. rhodesiana,C.
rausii,C. castroviejoi,andC. nivea. When possible, several
accessions of the same species were extracted and amplified
because of the taxonomic problems of the group and the wide
geographical distribution of some species. Special effort was
made to represent the variability and geographical distribu-
tion of C. planiflora and C. epithymum with nine and four
accessions, respectively.
Several species of the other two subgenera were sequenced
for our phylogenetic analyses (see Appendix): C. monogyna,
C. japonica,andC. lupuliformis (subg. Monogyna)andC.
campestris,C. umbellate,andC. cuspidata (subg. Grammica).
The ITS and trnL sequences obtained for this study and
others available in GenBank, showed that the level of
divergence is much higher in species of subg. Grammica.
The inclusion of these sequences in the alignment was
difficult, especially for the ITS-2 region, therefore those
sequences were excluded and the species of subg. Monogyna
were selected as outgroup. High sequence divergence also
occurs between species in subg. Grammica.Thisisnot
surprising since it is the most diversified group in terms of
number of species, morphological variation and chromosome
size and number.
DNA Extraction, Amplification and Sequencing. Total
genomic DNA was extracted from young flowers or
inflorescences separated from haustoria to prevent host
DNA contamination. Extractions were made using E.Z.N.A.
Plant MiniPrep Kit (Omega Biotech, Doraville, Georgia).
Primer pair 1830for and 40 rev (Nickrent et al. 2004) was used
to obtain amplifications of both ITS regions, including the
5.8S of the nuclear rDNA and small flanking parts of the SSU
(18S) and LSU (26S) genes. The PCR amplification of the trnL
intron was accomplished using the primers ‘‘c’’ and ‘‘d’’
described in Taberlet et al. (1991). PCR reactions were
prepared in 25 ml volumes using Ready-to-GoHPCR Beads
(GE Healthcare, Little Chalfont UK). The cycling parameters
were the same used by Nickrent et al. (2004). Amplification
products were cleaned using the QIAquick PCR purification
kit (Qiagen, Valencia, California) or E.Z.N.A. Clean kit
(Omega Biotech). When light PCR products, fungal DNA
contamination or polymorphisms in ITS sequences were
obtained, cloning was performed using the pGEM-T Easy-
Vector II cloning kit (Promega, Fitchburg, Wisconsin). Both
strands were sequenced separately using the forward and
reverse amplification primers or the universal primers
specific to the plasmid (T7 and SP6). An ABI Prism 310 or
377 genetic Analyzer and the ABI Prism
terminator Cycle SequencingReadyReactionkitwith
AmpliTaqHDNA Polymerase (Perkin Elmer Applied Biosys-
tems, Foster City, California) was used to obtain the
electropherograms. The program Sequence Navigator (Ap-
plied Biosystems) was used to edit the resulting electropher-
ograms and to assemble contiguous sequences. Sequences
Systematic Botany sbot-32-04-19.3d 21/9/07 09:44:12 901 Cust #06-59
2007] GARCI
were then imported into SeqApp (Gilbert 1993) and aligned
manually. Sequences obtained for this study were submitted
to GenBank (see Appendix). Data matrices are available on
TreeBASE (study number S1717).
The analysis of the electropherograms obtained from the
direct sequencing of the ITS amplifications revealed several
individuals with polymorphic sites. A site was considered
polymorphic when double peaks occurred and the weakest
signal reached at least 25% of the strength of the strongest
signal (Fuertes Aguilar et al. 1999). The number of poly-
morphic sites (PS) in the species of subg. Cuscuta ranged
between 6 and 20. Both additive and nonadditive sites were
found. Following the nomenclature proposed by Fuertes
Aguilar and Nieto Feliner (2003), we here consider additive
polymorphic sites (APS) those in which the two bases
involved are also found in other accessions of the data set.
Data Analysis. Both ITS and trnL matrixes were analyzed
with parsimony and Bayesian inference. For the maximum
parsimony (MP) analyses, minimum length Fitch trees were
constructed using heuristic searches with tree bisection-
reconnection (TBR) branch swapping, collapsing branches if
maximum length is zero and with the MULPARS option on
in PAUP* 4.0b10 (Swofford 2002). Branch robustness was
estimated by bootstrap analysis (Felsenstein 1985) using 100
full heuristic replicates for the ITS data set and 10,000
replicates with the fast heuristic search (Mort et al. 2000) for
the trnL data set, as implemented in PAUP*. For trnL, a total
of 17 indels were coded using the simple indel coding as
proposed by Simmons and Ochotorena (2000).
Bayesian analyses were conducted using MrBayes 3.1
(Ronquist and Huelsenbeck 2003). The models used for the
Bayesian analyses were determined according to the Akaike
Information Criterion (AIC) as implemented in Modeltest 3.7
(Posada and Crandall 1998). For the ITS matrix, a General
Time Reversible model with a gamma shaped distribution of
rates across sites (GTR+C) was selected. A similar model but
with a proportion of invariable sites (GTR+I+C) was selected
for the DNA sequence partition of the trnL matrix whereas
the binary model was applied to the indel partition. For each
matrix, two independent and simultaneous analyses starting
from different random trees were run for 1,000,000 genera-
tions with four parallel chains and trees and model scores
saved every 100
generation. A measure of average standard
deviation of split frequencies was used as a convergence
diagnostic. Every 1,000
generation tree from the two runs
was sampled to measure the similarities between them and to
determine the level of convergence of the two runs. The first
25% of the trees were discarded as burn-in before stationarity
was reached. Both the 50% majority-rule consensus tree and
the posterior probability of the nodes were calculated from
the 15,002 remaining trees with MrBayes. Additionally, 50%
majority-rule consensus trees from each of the simultaneous
runs were obtained with PAUP* and compared with those
obtained with MrBayes.
Several alternative topologies were tested using the one-
tailed nonparametric Shimodaira-Hasegawa (SH) test (Shi-
modaira and Hasegawa 1999) in PAUP*, with 1000 bootstrap
replicates and full parameter optimization. For each of the
topologies to be tested, a constrained tree was constructed
using MacClade 4.06 (Maddison and Madisson 2003). The
MP trees under each constraint were obtained with PAUP*.
One set of topologies was tested with the ITS data set to
determine if the position of some species, which was
different in our analyses to the taxonomic proposal by
Yuncker (1932), was significantly different (Table 1). The
likelihood score under the GTR+Cmodel of the MP trees
were obtained; one of the constrained trees with best score
and the unconstrained tree were used to perform the SH
tests. The other group of SH test were performed to
determine if the topological incongruences obtained analyz-
ing independently the two data sets were statistically
significant or as a result of lack of resolution and associated
random branching. Given that the ITS trees were more
resolved and supported, the topological incongruence was
tested with the trnL data set, constraining to obtain the ITS
topologies. In this case, the TVM+I+Cmodel was selected to
obtain the likelihood scores of the best MP trees excluding
the indel data partition. Many of the MP trees had the best
score under this model and the selection of the trees to
perform the SH test was randomly done from those that had
the best likelihood scores. As an additional measure of
congruence between the two data sets, the incongruence
length difference (ILD) test (Farris et al. 1995) was performed
in PAUP* with 100 replicates of the partition homogeneity
test using the heuristic search parameters previously in-
dicated except the MULTREES option off, keeping only the
best tree per analysis.
Systematic Botany sbot-32-04-19.3d 21/9/07 09:44:12 902 Cust #06-59
TABLE 1. Results of the Shimodaira-Hasegawa tests for comparison of alternative topologies. The first group of tests were
performed to compare four alternative hypotheses to the ITS topology. The second group of tests were performed with the trnL
data set to test the potential incongruences between the nuclear and chloroplast topologies.
Shimodaira-Hasewaga test
2lnLd2lnLP Rejected
Unconstrained MP tree (ITS) 5094.956
C. babylonica as sister to sect. Cuscuta 5100.527 5.570 0.336 No
C. capitata as sister to sect. Cuscuta 5137.720 42.764 ,0.05 Yes
C. kotschyana monophyletic and sister to C. pulchella 5104.344 9.388 0.267 No
C. kurdica as sister to C. approximata clade 5126.141 30.585 ,0.05 Yes
Unconstrained MP tree (trnL) 2056.710 –
C. epithymum 3 in a clade with C. epithymum 4 2086.229 29.519 ,0.05 Yes
C. somaliensis 4 in a clade with C. somaliensis 1 and 2 2061.639 4.929 0.155 No
C. planiflora 5 in a clade with C. planiflora 1andC. palaestina 1 and 2 2084.488 27.778 ,0.05 Yes
C. planiflora 7 in a clade with C. nivea 2079.445 22.735 ,0.05 Yes
C. planiflora 8 sister to C. planiflora 1, 2, 3, 4, 5, 6 and 9, C. palaestina 1and2,
C. epilinum and C. rausii
2089.218 32.508 ,0.05 Yes
C. planiflora 9 in a clade with C. planiflora 2, 3 and 6 2088.066 31.356 ,0.05 Yes
C. planiflora 9 in a clade with C. planiflora 1andC. palaestina 1 and 2 2093.702 36.992 ,0.05 Yes
C. epilinum in a clade with C. planiflora 1andC. palaestina 1 and 2 2094.149 37.439 ,0.05 Yes
Phylogenetic Analyses of the trnL and ITS
Regions. The trnL intron length ranged from
225 bp in C. indecora (subg. Grammica) to 344 bp
in one of the accessions of C. epithymum. Following
the nomenclature and secondary structure model
proposed by Cech et al. (1994), most of the length
variation was located in the P6 and P8 regions. The
P6 stem-loop region ranged between 5 bp in C.
indecora and 40 bp in C. japonica and C. lupuliformis
(both in subg. Monogyna). Among the species of
subg. Cuscuta, the P6 region ranged from 17 to
33 bp. More length variation was observed in the
P8 region, ranging from 25 bp in C. attenuata and C.
indecora to 133 bp in one of the specimens of C.
epithymum. In both variable regions, the shortest
sequences were found in species of subg. Gram-
mica. Estimation of homology in these two regions
was difficult between the species of Grammica and
the other two subgenera, thus they were removed
for further analyses. The aligned matrix of the trnL
sequences, excluding the species of subg. Gram-
mica, resulted in 469 characters, 133 of them
parsimony informative with 17 gaps coded as
present/absent. The parsimony analysis resulted
in 564 equally parsimonious trees, 285 steps long,
CI 50.76, HI 50.24, RI 50.93. The strict
consensus tree (Fig. 1) showed identical topology
to the 50% majority-rule tree obtained by Bayesian
Unambiguous alignment of the ITS sequences of
subg. Grammica was difficult owing to sequence
divergence, and therefore these were excluded
from further analyses. The aligned matrix of the
ITS region resulted in 635 characters, including
partial sequences of the 18S and 26S regions. Of
these characters, 263 were constant and 311
parsimony-informative. Parsimony analyses re-
sulted in 8 most parsimonious trees of 839 steps,
CI 50.70, HI 50.30, RI 50.91. The parsimony
strict consensus tree (Fig. 2) showed essentially the
same topology as the 50% majority-rule consensus
tree from the Bayesian analyses and those obtained
from the two independent and simultaneous
Bayesian runs. The only topological difference
was found in the position of C. kurdica that
appeared sister to the clade of the tropical African
species in the Bayesian consensus (not shown)
whereas in the parsimony strict consensus its
position was sister to all the species of sect. Cuscuta
except C. babylonica. In both cases BS and BPP
values were low (,50 and 0.69 respectively).
The same major clades (Clades 1–8) were re-
covered with the trnL and ITS data sets, but in
general the latter resulted in better resolution and
support. In both analyses, the species of subg.
Cuscuta were resolved as monophyletic using the
three species of subg. Monogyna as outgroup. Two
major clades were resolved within subg. Cuscuta.
The South African species of sect. Pachystigma were
strongly supported as monophyletic with both
data sets (Clade 1). The ITS trees included one
sample of C. africana, from which it was not
possible to obtain a trnL sequence. Clade 2,
strongly supported by both trnL and ITS, contained
the rest of the species of the subgenus. This second
large clade was resolved in three major subclades
with trnL: Clade 3 comprised the species included
by Yuncker in sect. Epistigma plus C. capitata
(monotypic sect. Cleistococca), Clade 4 the two
accessions of C. babylonica,andClade5the
remaining species of sect. Cuscuta. In the ITS trees,
the clade of C. babylonica was sister to the species in
Clade 3 whereas this relationship was not resolved
in the trnL analyses. The species of sect. Epistigma
represented by more than one sample in Clade 3
were resolved as monophyletic except for acces-
sion 3 of C. kotschyana. Both data sets resolved
a clade with the other two accessions of C.
kotschyana and the three accessions of C. pulchella.
The ITS analyses placed C. haussknechtii in a strong-
ly supported clade with the two accessions of C.
pedicellata, but this relationship was not resolved in
the trnL trees.
Clade 5 included all remaining species of sect.
Cuscuta, strongly supported as monophyletic; three
major clades (Clades 6–8) were resolved in all
analyses. The placement of C. kurdica differed in
the ITS and trnL trees; while ITS placed it as sister
to all the species in Clade 5, it was included within
Clade 7 in the trnL trees. Clade 6 consisted of
species that have a tropical African distribution;
the monophyly of these species was better sup-
ported in the trnL analyses but the ITS trees
showed a better internal resolution. Cuscuta abyssi-
nica appeared in the ITS trees as sister to the
tropical African clade, followed by the two
accessions of C. castroviejoi that were sister to
a highly supported clade that included the acces-
sions of C. rhodesiana,C. somaliensis,andC.
pretoriana. There was one additional sample of C.
somaliensis (accession 3) from which it was not
possible to obtain a trnL sequence. Two samples
showed ITS polymorphism, C. pretoriana 1 and C.
somaliensis 4. The two ribotypes of C. pretoriana 1
differed in 7 positions, only one of them additive
and the accessions were resolved as monophyletic.
However, the two ribotypes of C. somaliensis 4 were
not monophyletic, appeared in different well-
supported clades and differed in 19 positions, all
of them additive. The clone named as C. somaliensis
4B was included in a strongly supported clade with
Systematic Botany sbot-32-04-19.3d 21/9/07 09:44:12 903 Cust #06-59
2007] GARCI
Systematic Botany sbot-32-04-19.3d 21/9/07 09:44:13 904 Cust #06-59
FIG. 1. Strict consensus of 564 equally parsimonious trees obtained from the analysis of the trnL intron sequences. Numbers
above branches are bootstrap support/Bayesian posterior probability. Major clades are labeled with numbers inside circles
and referred to in the text. Accessions showing topological incongruence with the ITS consensus trees in clades with moderate
or high support are shaded. Cuscuta approximata subsp. macranthera and C. epithymum subsp. corsicana are labeled as C.
macranthera and C. corsicana, respectively.
Systematic Botany sbot-32-04-19.3d 21/9/07 09:45:07 905 Cust #06-59
FIG. 2. Strict consensus of 8 equally parsimonious trees obtained from the analysis of the ITS sequences. Numbers above
branches are bootstrap support/Bayesian posterior probability. The position of C. kurdica in parsimony and Bayesian
consensus trees is slightly different (see text). Major clades are labeled with numbers inside circles and referred to in the text.
Accessions showing topological incongruence with the trnL consensus trees in clades with moderate or high support are
shaded. Cuscuta approximata subsp. macranthera and C. epithymum subsp. corsicana are labeled as C. macranthera and C.
corsicana, respectively.
2007] GARCI
accessions 1 and 2 of this species and had an
almost identical sequence to C. somaliensis 1 from
which it differed in only one position. However, C.
somaliensis 4A appeared in a clade with C.
somaliensis 3 (having both identical sequences),
and was related to the two accessions of C.
pretoriana. The only specimen of C. rhodesiana was
embedded in the C. somaliensis/C. pretoriana clade.
Clade 7 included the different accessions of C.
approximata,C. balansae and C. europaea. This clade
had a strong bootstrap support in the ITS trees (BS
595) but weak in the trnL trees (BS ,50); however
BPP was higher in the trnL trees than in the ITS
trees (0.87 vs. 0.66). The monophyly of the three
geographically distant accessions of C. europaea
was well supported in all analyses and the ITS
sequences of the three accessions were remarkably
similar (differing at most in only one position). The
clade of C. europaea was sister to a strongly
supported monophyletic group in the ITS trees in
which two clades, not obtained in the trnL trees,
were resolved. One included four accessions of C.
balansae and the other, with low support, all
accessions of both subspecies of C. approximata.
Accession 2 and 3 of C. approximata showed six ITS
polymorphic sites, five of them additive. This
polymorphism was neither observed in accession
1ofC. approximata nor in the two samples of C.
approximata subsp. macranthera (named on the trees
as C. macranthera). The sequence of C. approximata 1
was identical to clone A of C. approximata 3
whereas the two samples of subsp. macranthera
differed from each other at two positions.
The remaining species of sect. Cuscuta were
placed in Clade 8, including the accessions of C.
planiflora,C. epithymum and other well-defined taxa
with narrower geographical distributions. The trnL
analyses resolved a clade, formed by C. epilinum
and the Ethiopian accession 9 of C. planiflora sister
to a polytomy that included the remaining acces-
sions. Internal branches of Clade 8 in the ITS trees
were more resolved and supported than in the trnL
trees. Additionally, two ITS ribotypes were sam-
pled from C. planiflora 8, which differed from each
other in 7 positions (4 of them additive) and these
were resolved in a well-supported clade. The other
specimen from which two ITS ribotypes were
sampled was C. planiflora 9, with 11 polymorphic
sites (all of them additive) and these were resolved
as polyphyletic. One of these cloned sequences (C.
planiflora 9A) was included in a clade with
accessions 2, 3, and 6 of C. planiflora that were
resolved as monophyletic on the trnL trees. The
second ribotype (C. planiflora 9B) had a sequence
identical to accessions 1 and 5 of C. planiflora,as
well as the two samples of C. palaestina and C
epilinum, all of which came out together in
a moderately supported clade. This clade was well
supported as sister to the clade formed by
accession 4 of C. planiflora and C. rausii.
Most of the topological incongruences between
the ITS and trnL trees were found in Clade 8.
Manual inspection of the consensus trees revealed
that some accessions of C. planiflora and C.
epithymum had conflicting topological positions.
Considering only the clades with moderate or high
support these accessions were: C. planiflora 7, C.
planiflora 5, the two sequence types of C. planiflora 8
and C. planiflora 9, C. epilinum and C. epithymum 3.
These conflicting accessions are shaded in Figs. 1
and 2.
Tests of Alternative Topologies and Combined
Datasets Analyses. The result of the SH tests used
to compare alternative topologies to those obtained
in our phylogenetic analyses are summarized in
Table 1. One set of these tests were performed
using the ITS data constrained to produce the
following topologies: C. babylonica as sister to the
rest of species of sect. Cuscuta (not rejected), C.
capitata as sister to sect. Cuscuta (rejected), the
accessions of C. kotschyana monophyletic and as
sister to C. pulchella (not rejected) and C. kurdica as
sister to both subspecies of C. approximata (re-
jected). The other set of SH tests were performed in
which the trnL data set was constrained to produce
each of the conflicting topologies with the ITS trees,
shaded in Figs. 1 and 2. Except for the case of C.
somaliensis 4 constrained to be sister to the trnL
sequences of C. somaliensis 1 and 2, the SH tests
rejected the ITS compatible topologies at the P5
0.05 level. Additionally, the partition homogeneity
test (ILD test) revealed significant incongruence
between the ITS and trnL data sets (P50.01). Only
when all the sequences shaded in Figs. 1 and 2 and
C. kurdica were excluded, this test revealed
congruence (P50.23) between both data parti-
The parsimony analyses of the combined data
set, excluding the conflicting OTUs, resulted in 16
MP trees, each 1087 steps in length with 437
informative characters, CI 50.72, HI 50.28, and RI
50.91. The topology of the combined strict
consensus tree (not shown) was identical to the
ITS strict consensus tree in Fig. 2, and the 8 major
clades were recovered, in general with higher
levels of BS support. Only Clade 7 had lower
support (BS 581) than in the ITS trees.
Monophyly of Subg. Cuscuta and Sect. Pachys-
tigma.The monophyly of subgenus Cuscuta as
recognized by Yuncker (1932) could not be tested
Systematic Botany sbot-32-04-19.3d 21/9/07 09:45:58 906 Cust #06-59
in this study because further sampling of subg.
Grammica should be included in the analyses.
Among the four sections of the subgenus, sect.
Pachystigma is the most distinctive by morpholog-
ical and molecular features. The inclusion of
sequences of other members of subg. Grammica
could indicate that sect. Pachystigma is more closely
related to some species of subg. Grammica than to
the rest of subg. Cuscuta (McNeal 2005). The
presence of holocentric chromosomes has been
documented in species present in all the major
groups of Clade 2 except for the tropical African
clade from which no cytological data are available.
The chromosome type in the species of sect.
Pachystigma is unknown and the phylogenetic
reconstruction cannot predict whether this section
has holocentric or monocentric chromosomes.
Cytological studies in these species in addition to
including sequences of subg. Grammica in molec-
ular phylogenetic studies will help to test the
monophyly of subg. Cuscuta as was defined by
Yuncker (1932).
All analyses provided strong support for the
monophyly of the species of sect. Pachystigma. This
group of five South African species presents
a number of diagnostic characters, such as cymose
inflorescences (Fig. 3A), broad stigmas, well de-
veloped corolla scales and pentamerous flowers,
which taken together characterize this section. All
the species are well-defined morphologically and
clearly different from each other. The only species
not included in this study was C. gerrardii (5C.
cucullata), a rare species from which no material
was available for DNA extraction. It is clearly
different from the other four species, as shown by
its short stigmas and indehiscent fruits. The seeds
are dispersed through an interstylar aperture, the
only case of this type in the subgenus. The
stigmatic morphology approaches that of species
in subg. Grammica and whether it should be
included in this section has yet to be determined.
Section Epistigma, Sect. Cleistococca, and C.
babylonica.All analyses indicate that species of
sect. Epistigma,togetherwithC. capitata,are
monophyletic. These species occur exclusively in
central and western Asia except for C. pedicellata,
which can also be found in northeastern Africa and
southeastern Europe, mainly as contaminants of
crop plants. All species of sect. Epistigma lack styles
or the styles are short and conical, and all have
fruits that dehisce by an irregular line at the base.
However, C. capitata has indehiscent fruits and
well-developed styles, characters that were used
by Engelmann (1859) and accepted by Yuncker
(1932) to segregate this species in its own section.
The enforced topology of C. capitata as sister to
species of sect. Cuscuta, that have clearly developed
styles, was rejected as significantly worse by the
SH test (P,0.05, Table 1) suggesting that this
species is closely related to sect. Epistigma. Some
species of sect. Epistigma, such as C. pedicellata,
have the apical part of the ovary somewhat
elongated, appearing as a short knob at the base
of the styles. The styles in C. capitata are clearly
conical and what is interpreted as differentiated
styles could actually be a more developed apical
projection of the ovary. Within subg. Cuscuta,
indehiscent fruits not only occur in C. capitata,
but also in the above mentioned C. gerrardii and in
C. triumvirati, species from southeastern Spain and
Morocco traditionally included in sect. Cuscuta.
Our analyses indicate that fruit indehiscence is
a convergent character state and therefore should
not be used to delimit sections in subg. Cuscuta.
Although the sectional classification in subg.
Cuscuta based on fruit dehiscence is not problem-
atic because it only affects a few species, it may be
more important in subg. Grammica that was
divided into two sections, based on the presence
or absence of this character (Yuncker 1932).
Molecular studies in this subgenus indicate, at
least in some cases, that the indehiscent fruit type
is the result of parallelism (Stefanovic
´et al. 2007).
Cuscuta kotschyana displays a higher degree of
morphological variation within sect. Epistigma than
other species This variability affects the shape and
size of calyx and corolla lobes, length of flower
pedicels and number of flowers per glomerule. The
most common forms, occurring throughout the
distributional range of the species and morpholog-
ically similar to Boissier’s type specimen (Fig. 3B),
are here represented by accession number 1. Some
specimens, from the Arabian Peninsula and South-
ern regions of Iran and Pakistan, here represented
by C. kotschyana 2, morphologically approach C.
pedicellata. These two accessions appear as mono-
phyletic in all analyses but this is not the case for C.
kotschyana 3. This individual has the morphological
features of the plants recognized as C. kotschyana
var. caudata, representing an extreme in morpho-
logical variation in which the bracts, calyx and
corolla lobes are markedly subulate and the
flowers are frequently dark red. However, other
characters approach the form of a typical C.
kotschyana. These populations, mostly from north-
ern Iran, could be a different species but the
morphological characters are variable and the
limits between them and the typical C. kotschyana
are not clear. Moreover, the alternative topology of
all accessions of C. kotschyana monophyletic and
sister to C. pulchella, was not rejected by the SH test
Systematic Botany sbot-32-04-19.3d 21/9/07 09:45:58 907 Cust #06-59
2007] GARCI
Systematic Botany sbot-32-04-19.3d 21/9/07 09:45:58 908 Cust #06-59
FIG. 3. Representative specimens of Cuscuta subg. Cuscuta. Scale bars: 1 mm. A. Inflorescence of C. nitida (sect. Pachystigma).
B. Flowers of a typical specimen of C. kotschyana (sect. Epistigma). C–N. Representative species from sect. Cuscuta. C. Branched
inflorescence of a young specimen of C. babylonica. D. Flower of a typical specimen of C. somaliensis. E. Flower of C. balansae.F.
Flower of C. europaea. G. Glomerule of C. approximata. H. Flowers of a diploid (2n514) specimen of C. epithymum. I. Flower of
a polyploid (2n530) specimen of C. epithymum. J. Flowers of a polyploid (2n532) specimen of C. epithymum sometimes
recognized as C. epithymum subsp. kotschyi. K. Glomerule of C. planiflora. L. Flower of a papillate specimen of C. planiflora with
intermediate characters between a typical C. planiflora and C. nivea. M. Glomerule of C. nivea. N. Trimerous flower of
C. palaestina.
Cuscuta babylonica is a distinctive species, char-
acterized by the pedicellate flowers together with
the truncate calyx in which the lobes are reduced to
short projections. The truncate calyx is a character
shared with C. hausknechtii but in C. babylonica the
styles are clearly developed and do not appear to
be a mere apex of the ovary as in species of sect.
Epistigma. Cytologically, it is characterized by the
lower chromosome number found in the genus (2n
58; n54) (Pazy and Plitmann 1987) and the
largest chromosome length in subg. Cuscuta (up to
8mm). This species was included by Yuncker
(1932) in sect. Cuscuta based on the developed
styles and further placed in the monotypic subsect.
Babylonicae for its remarkable calyx features, the
pedicellate flowers and the fruit that is dehiscent
by an irregular line. Our analyses indicate that sect.
Cuscuta is monophyletic if C. babylonica is not
included. Its relationship with other sections is not
resolved in the trnL trees, whereas the ITS trees
strongly support placing this species as sister to the
clade formed by the species of sect. Epistigma plus
C. capitata. However, the alternative hypothesis of
C. babylonica sister to the other species of sect.
Cuscuta was not rejected by the SH test (P50.336).
Relative longer branches (Fig. 4) suggest that
potential long-branch attraction (LBA) could be
affecting the topology of the ITS trees, placing this
species as sister to sections Epistigma and Clesito-
cocca. Parsimony is easily affected by LBA artifacts
but Bayesian analyses may be affected as well,
even clades supported by 1.00 in posterior proba-
bility (Bergsten 2005). The more obvious morpho-
logical character shared by C. babylonica and the
species of sect. Epistigma is the fruit dehiscent by an
irregular line. Other morphological features are
shared by C. babylonica and some, but not all the
species of sect. Epistigma. For example, this species
has glomerules derived from multiple buds, as in
most species of subg. Cuscuta, but the first buds
develop flowering shoots resulting in a branched
inflorescence (Fig. 3C). This character is shared
with C. pulchella but not with the other species of
sect. Epistigma such as C. kotschyana or C. pedicellata
in which the inflorescence is not branched and is
reduced to glomerules similar to the species of sect.
The Tropical African Clade. Several species of
exclusive tropical African distribution were re-
solved as monophyletic in all analyses, although
with different support depending on the data set.
These include C. abyssinica,C. castroviejoi,C.
rhodesiana, and several accessions of C. somaliensis
and C. pretoriana. There are no obvious morpho-
logical synapomorphies for this group of species.
In C. somaliensis and C. rhodesiana, the glomerules
are normally shortly stalked, and the first or the
second flowers of the glomerule are usually
subpedicellate, with the base of the calyx swollen.
Additionally, an aborted bud may appear on the
glomerule stalk. However, these characters are not
clearly observed in C. castroviejoi,C. abyssinica,orC.
pretoriana and the glomerule structure is similar to
other species of sect. Cuscuta.ExceptforC.
castroviejoi, that was more recently described
´a 1999), the other species have been consid-
ered in African floras (e.g., Verdcourt 1963; Meeuse
and Welman 2000) as varieties of C. planiflora and
C. balansae or, in the case of C. rhodesiana,as
a synonym of C. approximata. One example is the
plant recognized as C. balansae var. mossamedensis
by Yuncker (1932) here considered as C. pretoriana
and represented in our data sets with accession
number 1. As previously discussed by Gar
(1999), there are enough morphological differences
to consider them different species and molecular
data also suggest this. However, in some cases the
differences between this and other species of sect.
Cuscuta are subtle. Examples are the populations of
C. somaliensis from northeastern Africa, Socotra and
the southeastern Arabian Peninsula, named by
Yuncker (1932) as var. socotrensis under C. balansae.
One of these plants is here represented by
accession number 1 of C. somaliensis. Morpholog-
ically, these plants resemble C. balansae but our
molecular data indicate that they are not closely
related and the external morphological similarity is
apparently a case of parallelism.
Most of the morphological variation and taxo-
nomic difficulty within the tropical African species
involves C. somaliensis and C. pretoriana. The plants
of C. somaliensis morphologically similar to the type
specimen, mostly from Kenya and Tanzania, have
densely papillate flowers, subulate calyx and
corolla lobes and well-developed corolla scales
(Fig. 3D). In C. pretoriana, flowers are smooth with
blunt calyx lobes, acute but not subulate corolla
lobes and reduced corolla scales. Transitional
forms between both species can be found showing
a gradient between papillate to granulate or even
smooth flowers. In some cases, specimens of C.
somaliensis may have scarcely papillate flowers,
and not subulate but triangular calyx and corolla
lobes, especially in populations of northeastern
Africa. Two ITS sequences have been obtained
from C. somaliensis 4, differing in 19 positions, all of
them additive. These ribotypes are placed in
different clades that contain accessions of C.
somaliensis and C. pretoriana. This additive pattern
indicates a possible hybrid origin of this specimen
in which concerted evolution has not completely
homogenized the ITS repeats towards either of the
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:05 909 Cust #06-59
2007] GARCI
progenitors. The trnL analysis resolves a clade
including the two accessions of C. pretoriana,C.
rhodesiana, and also C. somaliensis 4. This clade is
not strongly supported in the trnL trees, but
suggests a possible maternal progenitor for C.
somaliensis 4asamemberofthiscladeand
a paternal progenitor similar to C. somaliensis 3.
However, the SH test does not reject the placement
of C. somaliensis 4 in a clade with accessions 1 and 2
(Table 1) and therefore its position in the trnL trees
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:05 910 Cust #06-59
FIG. 4. One of the 8 most parsimonious trees obtained from the analysis of the ITS sequences, chosen to illustrate branch
lengths and possible effect of Long Branch Attraction (LBA) in the placement of the two accessions of C. babylonica (see text).
Branch lengths are proportional to the number of changes.
could also be explained by lack of resolution and
associated random branching.
Our sampling also recovered two ITS sequences
from C. pretoriana 1, a specimen from Burundi. The
origin of this polymorphism seems to be different
from that observed in C. somaliensis 4. A possible
origin of these two ribotypes in C. pretoriana 1is
what Bailey et al. (2003) called ‘‘shallow paralogy’’
in which a rDNA locus is duplicated, and without
substantial concerted evolution evolves indepen-
dently on different chromosomes or at different
positions on the same chromosome. These se-
quences are then seen as monophyletic on phylo-
genetic trees. In the case of C. pretoriana this
paralogy could be intra-individual or within
populations. At least one of the sequences sampled
in C. pretoriana 1 could be an early divergent
pseudogene as indicated by a six bp deletion in
clone B that affects the ITS forward priming site,
a conserved region at the 39end of 18S rDNA.
Additionally, three of the seven positions in which
both sequences differ are located in the 5.8S region,
representing the only case in our sampling of intra-
individual variation in this region.
Our data indicate that gene flow has taken or is
taking place among the tropical African species,
complicating the taxonomy of this group and the
morphological species delimitation. The taxonomic
problems not only involve C. somaliensis and C.
pretoriana but also C. abyssinica. Intermediate forms
can be found within the distributional range of
these species and additional sampling of C.
abyssinica would help to understand their patterns
of morphological variation. Moreover, karyological
data are required to know their ploidy levels and
to confirm the holocentric nature of their chromo-
somes as predicted by our phylogenetic results.
C. europaea and the C. approximata-C. bal-
ansae Clade. With different support depending
on the data set, C. europaea is sister to a clade that
contains accessions of C. balansae and both sub-
species of C. approximata. Molecular analyses
indicate that these taxa are monophyletic but
apparently there are no shared morphological
characters that could differentiate them from the
other species of sect. Cuscuta (Figs. 3E–G). One
common character is the presence of high amounts
of heterochromatin that with conventional staining
can be observed as chromocenters in both in-
terphase cell nuclei and prophase chromosomes
(areticulate nuclei). The remaining species of sect.
Cuscuta, placed in Clade 8, showed reticulate
nuclei lacking marked chromocenters (Garcı
´a and
Castroviejo 2003). These chromocenters were first
found by Pazy and Plitmann (1987) in C. babylonica
and its presence also has been documented in C.
approximata (Guerra and Garcı
´a 2004) and in the
diploid C. europaea (2n514; Garcı
´a and Castroviejo
2003). In C. europaea the number of chromocenters
is not as high as in the tetraploid C. approximata (2n
528) and they are smaller. Recent observations of
interphase nuclei of specimens of C. balansae from
Armenia, morphologically similar to accession 4,
indicate that these specimens also have areticulate
nuclei with marked chromocenters (data not
published). This character, however, is yet to be
confirmed in the typical specimens of C. balansae.
The accessions of C. approximata and C. balansae
are resolved as monophyletic, with high support
especially in the ITS trees. Morphologically, these
species are clearly different. However intermediate
specimens exist, more frequently in populations
from eastern Turkey, northern Iran, Armenia and
Azerbaijan and are represented in this study by
accession number 4 of C. balansae. General mor-
phology of these plants is similar to C. approximata,
with broad and golden-reticulate when dry calyx
lobes and larger flowers than in the typical
Anatolian specimens of C. balansae. However, sepal
tips are densely papillate and dark red as in C.
balansae, thus connecting the morphotypes of both
The direct sequencing of ITS from accession 3 of
C. approximata showed polymorphic sites as re-
vealed by the two sequences obtained by cloning.
However, the specimen here designed as accession
1 of the same species showed no polymorphic sites
and a unique ITS sequence. This latter specimen
was collected from the same population as the
specimen with a single 45S locus located on one of
the longer pairs of chromosomes (Guerra and
´a 2004); this locus was associated with the
nucleolus after prolonged hematoxylin staining. In
other populations, one additional pair of smaller
chromosomes is associated with the nucleolus,
indicating the possible presence of a second
nucleolar organizer region (NOR) on these chro-
mosomes and therefore a second active rDNA
array. Considering that populations of accessions 3
and 1 of C. approximata are close geographically
(both from the Iberian Peninsula), the number of
rDNA sites and/or homogenization through con-
certed evolution may occur at the population level
in this species. In our sampling, molecular data
suggest that the second locus in accession 1 may
have been lost, whereas in accession 3 both loci
have been retained. The other specimen sampled
(accession 2) is from a distant population (Paki-
stan) in which both sequence types are present,
differing in a few positions from the sequences of
the Spanish populations. The monophyly of both
sequence types is weakly supported (BS 552, BPP
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:56 911 Cust #06-59
2007] GARCI
50.80). Similar results are obtained if the ITS
sequences of C. approximata subsp. macranthera are
removed from the analyses. These data indicate
that both sequence types are not shallow paralogs
derived from a recent locus duplication. Guerra
and Garcı
´a (2004) suggested a possible allopoly-
ploid origin of C. approximata based on the banding
pattern and distribution of the 5S and 45S sites of
the two chromosome complements. Further sam-
pling is necessary to elucidate the origin of the
sequence types and to confirm the possible
allopolyploid origin of this species.
Turanian distribution that resemble C. europaea, but
with acute calyx and corolla lobes and in general,
smaller flowers (Plitmann 1978) have been recog-
nized as C. kurdica. One of these specimens was
included in our phylogenetic study and its position
on the ITS trees is the only case in which the
topology obtained with parsimony and Bayesian
inference was slightly different. All ITS clones
obtained gave the same sequence and no poly-
morphic sites were observed. Nevertheless, these
two topological positions are weakly supported
and therefore the position of this taxon is not
resolved but arises from the polytomy of Clade 5.
Based solely on morphology, it might be expected
that this plant is related to C. europaea. In the trnL
trees, although weakly supported, it appears in
Clade 6 with C. europaea,C. approximata, and C.
balansae. However, the enforced topology of the ITS
sequence of C. kurdica to be sister to the clade of C.
approximata was rejected by the SH test. A possible
explanation for its position on the ITS trees is that
this sequence is the result of recombination after
a hypothetical hybridization event between par-
entals from two different major clades (A
and Wendel 2003). Additional sampling is required
to explain its position within sect. Cuscuta and also
to discard a possible origin of this sequence by
PCR-recombination (Cronn et al. 2002).
The C. epithymum and C. planiflora Groups.
The ITS data recover two strongly supported
clades that include the four accessions of C.
epithymum, but these were not resolved as mono-
phyletic in Clade 8. In the trnL parsimony strict
consensus tree, the accessions of C. epithymum are
resolved as monophyletic but with BS,50 and not
resolved in the Bayesian consensus tree. Three
distinctive taxa, well defined morphologically and
with a narrower area of distribution, are included
within the clades of C. epithymum. These taxa are C.
triumvirati,C. nivea,andC. epithymum subsp.
corsicana. Both C. triumvirati and C. epithymum
subsp. corsicana have a number of morphological
and ecological similarities, such as the pedicellate
flowers, thread-like stems originating in the bracts
of the inflorescences, and parasitism of shrubby
hosts of the western Mediterranean mountains.
One possible explanation is that taxa considered as
C. epithymum are actually different species with
convergent morphological characters. In our sam-
pling, in addition to subsp. corsicana, we have
included specimens of the two subspecies of C.
epithymum generally recognized in European floras
and that were characterized morphologically by
Feinbrun (1970, 1972). Accessions 1 and 2 are two
distant representatives of the type subspecies, with
membranous flowers, glomerules more than 7 mm
in diameter and style and stigma length together
longer than the ovary (Fig. 3H). These plants are
diploids (2n514, Garcı
´a and Castroviejo 2003) and
are the most common form of the species in central
and northern Europe, growing on a wide variety of
hosts, but especially common on shrubs of the
genera Ulex,Genista and Erica. The plant labeled C.
epithymum 4 is a polyploid (2n532) that has the
morphological features of C. epithymum subsp.
kotschyi (Fig. 3J), with fleshier calyx lobes, smaller
glomerules, combined style and stigma length
similar or slightly longer than the ovary, and
mostly parasitic on chamaephytes. These charac-
ters, however, are variable and a continuous range
of morphological variation can be observed be-
tween both subspecies of C. epithymum, especially
in the western Mediterranean. The specimen
named C. epithymum 3 is representative of these
morphologically intermediate populations (Fig. 3I)
that typically grow on basophilous shrubs and
chamaephytes, mostly of the genera Thymus,
Genista,orArtemisia. Chromosome counts have
revealed that they are polyploids with 2n530
´a and Castroviejo 2003) and the general
flower and glomerule sizes display great variation
even in the same population. Another variable
morphological feature is the presence of more or
less developed conical projections on the tips of
calyx and corolla lobes linking these forms with C.
nivea and with the plants recognized by Yuncker
(1932) as C. planiflora var. godronii. The topological
position of C. epithymum 3 in the ITS and trnL trees
is incongruent; with ITS it appears in a well
supported clade with C. epithymum 4 whereas the
analyses of the trnL data set place it together with
C. epithymum 2 in a moderately supported clade
that also contains accession 8 of C. planiflora
(discussed later). The enforced topology of C.
epithymum 3 sister to C. epithymum 4 in the trnL
trees was rejected by the SH test. These results
indicate that reticulation may be occurring in the
species complex, making taxonomic delimitation
very difficult. The absence of polymorphic sites in
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:56 912 Cust #06-59
the polyploid specimens of C. epithymum, indicates
a possible homogenization towards one of the
rDNA arrays by concerted evolution.
Plants with whitish and mostly pentamerous
flowers in compact glomerules, well developed
corolla scales, and fleshy and turgid calyx lobes
have been included under the name C. planiflora.
Other characters such as corolla and especially
calyx lobe shape or flower and glomerule sizes are
variable and, as in the case of C. epithymum, there is
a continuous range of variation between well-
defined morphological forms. Except for C. plani-
flora 7, all accessions of this species are placed
(with weak support) in one clade on the ITS trees,
whereas this relationship is not resolved with trnL.
Accession 1, from Sicily, C. planiflora 4 from Saudi
Arabia, and C. planiflora 5 from isolated popula-
tions in the Tibesti mountains in Chad are geo-
graphically distant individuals of the most typical
forms of the species, with smaller flowers and
glomerules (Fig. 3K). Accessions 2 and 3 are from
Macronesia (Madeira and Canary Islands, respec-
tively), whereas specimen 6 is from the southern
part of the Iberian Peninsula. These plants were
recognized by Yuncker (1932) as var. episonchum of
C. approximata but our analyses indicate that they
are not closely related to this species. Morpholog-
ically, the glomerules and the flowers are in
general larger, although these characters are vari-
able even in the same population. Similar plants
are found in other regions of the western Mediter-
ranean, especially in Morocco, but no samples from
this country have been included in our sampling.
These three accessions appear as monophyletic in
all analyses and share a 12 bp deletion in the P8
region of the trnL intron.
Plants from the eastern Mediterranean and
northeastern Africa, here represented by accession
9 and recognized by many authors as C. brevistyla,
are difficult to separate using morphological
characters. Our accession 9 from Ethiopia showed
polymorphic ITS and the two sequences differed in
11 additive positions. One of the ITS ribotypes
comes out in a clade with accessions 1 and 5 of C.
planiflora,C. palaestina, and C. epilinum, all with
identical or almost identical sequences. The second
ITS sequence type, however, is placed with the
Macronesian and Spanish accessions, although
their sequences are not identical (differing in 4 or
5 positions). The presence of these two sequence
types in the same individual may be relevant from
a biogeographic perspective because this provides
evidence of the connection between the Macro-
nesian and western Mediterranean populations
with those from northeastern Africa. Putative
phytogeographical links between Macronesia and
eastern Africa have been proposed based on
disjunct distributions of several plant groups
(Andrus et al. 2004). However, not many conclu-
sions can be made with the data obtained in this
study and further sampling including additional
populations from north Africa is needed to un-
derstand these relationships.
Accession 7 of C. planiflora taxonomically corre-
sponds with C. planiflora var. papillosa, distinguish-
able from a typical specimen of C. planiflora by its
papillate flowers (Fig. 3L). As discussed by Garcı
(2001) it resembles C. nivea (Fig. 3M) in that its
papillae densely cover the calyx and corolla, but
a number of other characters are more similar to C.
planiflora. In the ITS analysis, it is placed with C.
nivea in a well supported clade sister to accessions
3 and 4 of C. epithymum. However, on the trnL trees
it is placed in a moderately supported clade with
the two accessions of C. palaestina and accession 1
of C. planiflora. Although these plants are poly-
ploids with 2n534 (Garcı
´a 2001), no ITS
polymorphisms were found. The topological in-
congruence, supported by the SH tests, and the
morphological features suggest a likely allopoly-
ploid origin, in which the ITS arrays have been
homogenized by concerted evolution towards the
paternal repeat type. The ITS sequence is almost
identical to that of C. nivea, differing in 2
autapomorphic positions, suggesting a paternal
progenitor in this species or in an ancestral plant
close to this extant species. The maternal pro-
genitor would be a plant similar to a typical form
of C. planiflora. These papillate forms of C. nivea
and C. planiflora have a sympatric distribution in
the western Mediterranean (central and eastern
Iberian Peninsula and Morocco to Algeria). Where-
as C. nivea is restricted to higher elevations (500–
2100 m), the papillate and nonpapillate forms of C.
planiflora share similar habitats and hosts at lower
elevations (0–750 m) where their morphological
variation ranges overlap, especially in papillae
Accession 8 of C. planiflora is another example of
topological incongruence between nuclear and
plastid gene trees. This Afghan specimen is
a representative of the most common form of C.
planiflora in Central Asia, with more membranous
calyx lobes and thinner and frequently truncated
corolla scales. The analysis of the trnL data set
places it as related to C. epithymum (accessions 2
and 3) with moderate support, whereas ITS places
it in the clade of the C. planiflora group. The range
of morphological variation in these Asian plants in
some cases approaches more the typical specimens
of C. planiflora and in other cases small flowered
forms of C. epithymum. As suggested by these
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:56 913 Cust #06-59
2007] GARCI
results, hybridization and gene flow may be
involved in its morphological complexity. The
two ITS sequences sampled are highly supported
as monophyletic, suggesting they are shallow
paralogs resulting from duplication-divergence
events and lack of concerted evolution.
A number of species that can be readily
distinguished morphologically are placed among
the clades of C. planiflora. One of them is C.
palaestina, a central and eastern Mediterranean
tetraploid species (2n528) with a number of
distinctive morphological features, such as the
cucullate corolla lobes, smaller flower size, mem-
branous calyx lobes and mostly trimerous or
tetramerous flowers (Fig. 3N). Transitional forms
towards C. planiflora, with larger flowers and
fleshier calyx lobes also exist within its geograph-
ical range, and frequently the distinction between
these species is not clear. However, C. rausii,an
endemic to the Greek island of Karpathos (Garcı
1998), is morphologically distinctive. Although
molecular data indicate a close relationship with
C. planiflora and C. palaestina, morphologically it is
different, especially owing to its pedicellate flowers
and angled calyx tube. However, other features,
such as the cucullate corolla lobes and the
tetramerous flowers ally it with C. palaestina and
C. planiflora. Although C. rausii may well have
originated after the long isolation of Karpathos and
the rest of the Cardaegean area during the
Pleistocene (Greuter 1971), additional sampling is
required to rule out the idea that C. rausii is
actually a teratological form of either C. planiflora
or C. palaestina.
The other distinctive species that appears within
the C. planiflora clades is C. epilinum.Itis
a hexaploid (2n542) that exclusively parasitizes
cultivated flax (Linum usitatissimum)orother
linicolous plants such as Camelina sativa.This
species, as other flax weeds, have been subject to
strong selection by the traditional methods of
cultivation and harvesting this crop. It has a num-
ber of characters that have favored its expansion as
a contaminant of flax seeds. These include sparsely
branched stems, seeds joined in pairs approaching
the diameter of the host seeds and rapid fruit and
seed development with a high germination per-
centage. Hjelmqvist (1950), based on some mor-
phological similarities, proposed the origin of C.
epilinum from C. europaea in mountainous regions
of central Asia; however, our data do not support
this origin. The analyses of the trnL data set place
this species with the Ethiopian sample of C.
planiflora (accession 9), whereas the ITS sequence
is identical to one of the ribotypes of that
individual and to other accessions of C. palaestina
and C. planiflora. These data suggest that C.
epilinum may well have differentiated from a wild
form of C. planiflora in the Nile Valley or the Near
East, regions in which flax was cultivated since
prehistoric times (Zohary and Hopf 2000).
All the enforced topologies in which the trnL
data set was constrained to produce the compatible
ITS topologies in Clade 8 were rejected by the SH
test, suggesting that lack of resolution and random
branching is not the origin of the topological
incongruences. In our discussion we suggest that
hybridization between species may be involved,
but also within species processes as recombination
may occur (Linder and Rieseberg 2004), especially
in the morphologically and geographically diverse
C. planiflora complex.
ACKNOWLEDGEMENTS. The authors thank to Dr. Freitag
and the curators of the following herbaria for the use of their
specimens: B, BM, BR, E, G, M, MA, MO, NY, SIU, UPS and
W; and Dan Nickrent and Sasˇa Stefanovic
´for their critical
comments and help to improve the manuscript. Partial
funding for this study came from the Anthos Project
´LVAREZ, I. and J. F. WENDEL. 2003. Ribosomal ITS sequences
and plant phylogenetic inference. Molecular Phylogenetics
and Evolution 29: 417–434.
J. FRANCISCO-ORTEGA. 2004. Using molecular phyloge-
nies to test phytogeographical links between East/South
Africa-Southern Arabia and the Macaronesian islands –
a review, and the case of Vierea and Pulicaria section
Vieraeopsis (Asteraceae). Taxon 53: 333–346.
2003. Characterization of angiosperm rDNA polymor-
phism, paralogy and pseudogenes. Molecular Phylo-
genetics and Evolution 29: 435–455.
Cladistics 21: 163–193.
CECH, T. R., S. H. DAMBERGER, and R. R. GUTELL. 1994.
Representation of the secondary and tertiary structure of
group-I introns. Nature Structural Biology 1: 273–278.
COSTEA, M. and F. J. TARDIF. 2006. The biology of Canadian
weeds. 133. Cuscuta campestris Yuncker, C. gronovii
Willd. ex Schult. C. umbrosa Beyr. ex Hook. C. epithymum
(L.) L. and C. epilinum Weihe. Canadian Journal of Plant
Science 86: 293–316.
WENDEL. 2002. PCR-mediated recombination in amplifi-
cation products derived from polyploid cotton. Theoret-
ical and Applied Genetics 104: 482–489.
CRONQUIST, A. 1981. An integrated system of classification of
flowering plants. New York: Columbia University Press.
CHRTEK, J. and J. OSBORNOVA. 1991. Notes on the synantropic
plants of Egypt 3. Grammica campestris and other species
of family Cuscutaceae. Folia Geobotanica et Phytotaxono-
mica 26: 287–314.
DERNBURG, A. F. 2001. Here, there and everywhere: kineto-
chore function on holocentric chromosomes. The Journal
of Cell Biology 153: F33–F38.
DUMORTIER, B. C. J. 1829. Analyse des familles des plantes, avec
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:56 914 Cust #06-59
l’indication des principaux genres qui s’y rattachent.
Tournay: J. Casterman.
ENGELMANN, G. 1859. Systematic arrangement of the species
of the genus Cuscuta, with critical remarks on old species
and descriptions of new ones. Transactions of the Academy
of Sciences of St. Louis 1: 453–523.
Constructing a significance test for incongruence.
Systematic Biology 44: 570–572.
FEINBRUN, N. 1970. A taxonomic review of European
Cuscutae. Israel Journal of Botany 19: 16–29.
———. 1972. Cuscuta. Pp. 74–77 in Flora Europaea, vol. 3, eds.
T. G. Tutin, V. H. Heywood, N. A. Burges, D. M. Moore,
D. H. Valentine, S. M. Walters, and D. A. Webb.
Cambridge: Cambridge University Press.
FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an
approach using the bootstrap. Evolution 39: 783–791.
FLACH, M. 1966. Diffuse centromeres in a dicotyledoneus
plant. Nature 209: 1369–1370.
1999. Nuclear ribosomal DNA (nrDNA) concerted
evolution in natural and artificial hybrids of Armeria
(Plumbaginaceae). Molecular Ecology 8: 1341–1346.
——— and G. NIETO FELINER. 2003. Additive polymorphisms
and reticulation in an ITS phylogeny of thrifts (Armeria,
Plumbaginaceae). Molecular Phylogenetics and Evolution
28: 430–437.
´A, M. A. 1998. Cuscuta rausii (Convolvulaceae), a new
species from Greece. Annales Botanici Fennici 35: 171–174.
———1999. Cuscuta subgenus Cuscuta in Ethiopia with the
description of a new species. Annales Botanici Fennici 36:
———. 2001. A new western Mediterranean species of
Cuscuta (Convolvulaceae) confirms the presence of
holocentric chromosomes in subgenus Cuscuta.Botanical
Journal of the Linnean Society 135: 169–178.
——— and S. CASTROVIEJO. 2003. Estudios citotaxono
de las especies ibe
´ricas del ge
´nero Cuscuta (Convolvul-
aceae). Anales del Jardı
´n Bota
´nico de Madrid 60: 33–44.
GILBERT, D. G. 1993. SeqApp, version 1.9a157. Bloomington,
Indiana, USA: Biocomputing Office, Biology Depart-
ment, Indiana University.
GOLDBERG, A. 1986. Classification, evolution and phylogeny
of the families of dicotyledons. Smithsonian Contributions
to Botany 58: 1–314.
GREILHUBER, J. 1995. Chromosomes of the Monocotyledons
(General Aspects). Pp. 379–414 in Monocotyledons:
systematics and evolution, eds. P. J. Rudall, P. J. Cribb,
D. F. Cutler, and C. J. Humphries. Kew: Royal Botanic
GREUTER, W. 1971. Betrachtungen zur Pflanzengeographie
der Su
¨s. Opera Botanica 30: 49–64.
´A. 2004. Heterochromatin and
rDNA sites distribution in the holocentric chromosomes
of Cuscuta approximata (Convolvulaceae). Genome 47:
HJELMQVIST, H. 1950. The flax weeds and the origin of
cultivated flax. Botaniska Notiser 2: 257–298.
HUNZIKER, A. T. 1949–50. Las especies de Cuscuta en
Argentina y Uruguay. Trabajos del Museo Bota
´nico de la
Universidad Nacional de Co
´rdoba 1: 1–333.
LINDER, C. R. and L. H. RIESEBERG. 2004. Reconstructing
patterns of reticulate evolution in plants. American
Journal of Botany 91: 1700–1708.
MADDISON, W. P. and D. R. MADISSON. 2003. MacClade, v.
4.06. Sunderland: Sinauer Associates.
MCNEAL, J. R. 2005. Systematics and plastid genome evolution in
the parasitic plant genus Cuscuta (dodder). Ph.D. Disserta-
tion. Pennsylvania: Pennsylvania State University.
MEEUSE, A. D. J. and W. G. WELMAN. 2000. Cuscuta. Pp. 3–15
in Flora of Southern Africa, Convolvulaceae, vol. 28(1), ed.
G. Germishizen. Pretoria: National Botanical Institute.
MIRANDE, M. 1901. Recherches physiologiques et anatomi-
ques sur les Cuscutacee
´s. Bulletin Scientifique de la France
et de la Belgique 34: 1–284.
MORT, M. E., P. S. SOLTIS,D.E.SOLTIS, and M. L. MABRY. 2000.
Comparison of three methods for estimating internal
support on phylogenetic trees. Systematic Biology 49:
´N, and R. L.
MATHIASEN. 2004. A phylogeny of all species of
Arceuthobium (Viscaceae) using nuclear and chloroplast
DNA sequences. American Journal of Botany 91: 125–138.
PAZY, B. and U. PLITMANN. 1987. Persisting demibivalents:
a unique meiotic behavior in Cuscuta babylonica Choisy.
Genome 34: 533–536.
——— and ———. 1991. Unusual chromosome separation in
meiosis of Cuscuta.Genome 34: 533–536.
——— and ———. 1994. Holocentric chromosome behavior
in Cuscuta (Cuscutaceae). Plant Systematics and Evolution
191: 105–109.
——— and ———. 1995. Chromosome divergence in the
genus Cuscuta and its systematic implications. Caryologia
48: 173–180.
PLITMANN, U. 1978. Cuscutaceae. Pp. 222–237 in Flora of Turkey
and the East Aegean islands, vol. 6, ed. P. H. Davies.
Edinburgh: Edinburgh University Press.
———. 1991. Reproductive adaptations of parasitic higher
plants. The case of Cuscuta (Cuscutaceae). Pp. 133–144 in
Flora et Vegetatio Mundi, vol. 9, Contributiones Selectae ad
Floram et Vegetationem Orientis, eds. T. Engel, W. Frey,
and H. Kuerschner. Berlin, Sttutgart: J. Cramer.
———. 2002. Agamospermy is much more common than
conceived: a hypothesis. Israel Journal of Plant Sciences,
supplement 50: S111–S117.
POSADA, D. and K. A. CRANDALL. 1998. Modeltest: testing the
model of DNA substitution. Bioinformatics 14: 917–818.
Bayesian phylogenetic inference under mixed models.
Bioinformatics 19: 1572–1574.
SHEIKH, S. A., K. KONDO, and Y. HOSHI. 1995. Study of
diffused centromeric nature of Drosera chromosomes.
Cytologia 60: 43–47.
SHIMODAIRA, H. and M. HASEGAWA. 1999. Multiple compar-
isons of log-likelihoods with applications to phylo-
genetic inference. Molecular Biology and Evolution 16:
SIMMONS, M. P. and H. OCHOTORENA. 2000. Gaps as character
in sequenced-based phylogenetic analysis. Systematic
Biology 49: 369–381.
´, S., L. KRUEGER, and R. G. OLMSTEAD. 2002.
Monophyly of the Convolvulaceae and circumscription
of their major lineages based on DNA sequences of
multiple chloroplast loci. American Journal of Botany 89:
——— and R. G. OLMSTEAD. 2004. Testing the phylogenetic
position of a parasitic plant (Cuscuta, Convolvulaceae,
Asteridae): Bayesian inference and the parametric
bootstrap on data drawn from three genomes. Systematic
Biology 53: 384–399.
———, M. KUZMINA, and M. COSTEA. 2007. Delimitation of
major lineages within Cuscuta subgenus Grammica
(Convolvulaceae) using plastid and nuclear DNA
sequences. American Journal of Botany 94: 568–589.
SWOFFORD, D. L. 2002. PAUP*: phylogenetic analysis using
parsimony (*and other methods). Version 4.0b10. Sun-
derland: Sinauer Associates.
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:57 915 Cust #06-59
2007] GARCI
Universal primers for amplification of three non-coding
regions of chloroplast DNA. Plant Molecular Biology 17:
TAKHTAJAN, A. 1969. Flowering Plants: Origin and Dispersal.
Edinburgh: Oliver and Boyd.
VASUDEVAN, K. N. 1975. Contribution to the cytotaxonomy
and cytogeography of the flora of the Western Hima-
layas (with an attempt to compare it with the flora of the
Alps). Part I. Bericht der Schweizerischen Botanischen
Gesellschaft 85: 57–84.
VERDCOURT, B. 1963. Cuscuta. Pp. 3–12 in Flora of Tropical East
Africa, Convolvulaceae, eds. C. E. Hubbard and E. Milne-
Redhead. London: Crown Agents for Oversea Govern-
ments and Administrations.
YUNCKER, T. G. 1921. Revision of the North American and
West Indian species of Cuscuta.Illinois Biological Mono-
graphs 6: 91–231.
———. 1922. Revision of the South American species of
Cuscuta I. American Journal of Botany 9: 557–575.
———. 1923. Revision of the South American species of
Cuscuta II. American Journal of Botany 10: 1–17.
———. 1932. The genus Cuscuta.Memoirs of the Torrey
Botanical Club 18: 109–331.
ZOHARY, D. and M. HOPF. 2000. Domestication of plants in the
Old World,3
edition Oxford University Press.
APPENDIX 1. Voucher specimens and GenBank accession
numbers used in the molecular phylogenetic analyses.
Specimens are arranged by subgenera and sections as
recognized by Yuncker (1932). GenBank numbers are in the
order trnL,ITS.
Sect. Pachystigma Engelm., Cuscuta africana Willd., South
Africa, Cape Province, Carlquist 5082 (MO), _, DQ924574;
Cuscuta angulata Engelm., South Africa, Cape Province,
Orchard 469 (MO), EF152065, DQ924575; Cuscuta natalensis
Baker, (1) South Africa, Natal, Hilliard and Burtt 14528 (M),
AJ457113, DQ924576; (2) South Africa, Natal, Stewart 2113 (E),
AJ457114, DQ924577; Cuscuta nitida E. Mey. ex Choisy, (1)
South Africa, Cape Province, Pillans 10905 (MO), AJ457115,
DQ924573; (2) South Africa, Cape Province, Salter 9196 (BM),
AJ457117, DQ924572.
Sect. Cleistococca Engelm., Cuscuta capitata Roxb., (1)
Afghanistan, Rasoul 3571 (W), AJ457112, DQ924583; (2)
Afghanistan, Hedge & Wendelbo 516 (E), AJ457111, DQ924584
Sect. Epistigma Engelm., Cuscuta haussknechtii Yunck.,
Iran, Rechinger 47689 (W), EF152066, DQ924580; Cuscuta
kotschyana Boiss., (1) Iran, Assadi 27520 (E), AJ457102,
DQ924585; (2) Iran, Le
´onard 5674 (E), AJ457103, DQ924586;
(3) Iran, Rechinger 40710 (G), AJ457104, DQ924587; Cuscuta
pedicellata Ledeb., (1) Afghanistan, Freitag 945 (herb.
Freitag), AJ457105, DQ924582; (2) Iran, Rechinger 46349 (W),
AJ457107, DQ924581; Cuscuta pulchella Engelm., (1) Afgha-
nistan, Podlech 18495 (M), AJ457108, DQ924588; (2) Afghani-
stan, Volk 756 (B), AJ457109, DQ924589; (3) Afghanistan, Volk
71/358 (M), AJ457110, DQ924590.
Sect. Cuscuta,Cuscuta abyssinica A. Rich., Kenya, Fries
2216 (UPS), EF152067, DQ924631; Cuscuta approximata Bab.,
(1) Spain, Garcı
´a 833 (MA), EF152068, DQ924599; (2) Pakistan,
Ghafour 5091 (E), AJ428062, DQ924596 (ITS clone A),
DQ924595 (ITS clone B); Spain, Garcı
´a 1286 (MA), AJ428061,
DQ924597 (ITS clone A), DQ924598 (ITS clone B); Cuscuta
approximata subsp. macranthera (Boiss.) Feinbrun, (1)
Greece, Raus et al. 21952 (B), AJ457098, DQ924600; (2) Greece,
Raus et al. 21950 (B), AJ457099, DQ924601; Cuscuta babylo-
nica Aucher ex Choisy, (1) Afghanistan, Andersen and
Cornelius 412 (W), AJ457100, DQ924579; (2) Turkey, Davis
and Coode D36629 (E), AJ457101, DQ924578; Cuscuta balansae
Boiss. & Reut., (1) Crete, Bergmeier and Mattha
¨s 3457 (B),
AJ428059, DQ924591; (2) Crete, Bergmeier and Mattha
¨s 3441
(B), AJ457119, DQ924592; (3) Turkey, Wahak s.n. (B),
AJ457120, DQ924593; (4) Iran, Lamond 3726 (E); AJ428060,
DQ924594; Cuscuta castroviejoi M. A. Garcı
´a, (1) Ethiopia,
De Wilde 6669 (MO), EF152069, DQ924638; (2) Ethiopia, De
Wilde 5114 (MO), AJ457135, DQ924639; Cuscuta epilinum
Weihe, Cultivated, Garcı
´a 3919 (MA), AJ430075, DQ924610;
Cuscuta epithymum (L.) L., (1) Germany, Gottschlich 19077
(MA), AJ430072, DQ924609; (2) Spain, Garcı
´a 998 (MA),
AJ430069, DQ924606; (3) Spain, Garcı
´a 725 (MA), AJ430070,
DQ924607; (4) Spain, Garcı
´a 1270 (MA), AJ430071, DQ924608;
Cuscuta epithymum subp. corsicana (Yunck.) Lambinon,
Corse, Martı
´nez Ortega 512 (MA), AJ457121, DQ924605;
Cuscuta europaea L., (1) India, Stainton 8822 (E), AJ457122,
DQ924602; (2) Spain, Garcı
´a 995 (MA), AJ430076, DQ924603;
(3) Germany, Larsen et al. KL37596 (MA), AJ430077,
DQ924604; Cuscuta kurdica Engelm., Turkey, McNeill 569
(W), EF152070, DQ924613; Cuscuta nivea M. A. Garcı
Spain, Garcı
´a 1018 (MA), AJ457097, DQ924614; Cuscuta
palaestina Boiss., (1) Greece, Garcı
´a 981 (MA); AJ457123,
DQ924616; (2) Greece, Raus and Schiers 15642 (B), AJ457124,
DQ924617; Cuscuta planiflora Ten., (1) Sicily, Castroviejo et al.
15378SC (MA), AJ457096, DQ924618; (2) Madeira, Press and
Short 1232 (BM), AJ457125, DQ924619; (3) Canary Islands,
Jarvis and Murphy 79 (BM), AJ457126, DQ924620; (4) Saudi
Arabia, Mandeville 2436 (BM), AJ457127, DQ924622; (5) Chad,
Scholz 259 (B), AJ457128, DQ924623; (6) Spain, Garcı
(MA), AJ457129, DQ924621; (7) Spain, Valde
´s-Bermejo 5148
(MA), AJ457130, DQ924624; (8) Afghanistan, Ko
¨ie 2343 (NY);
EF152071, DQ924636 (ITS clone A), DQ924637 (ITS clone B);
(9) Ethiopia, Thulin 1610 (UPS), EF152072, DQ924632 (ITS
clone A), DQ924633 (ITS clone B); Cuscuta pretoriana Yunck.,
(1) Burundi, Reekmans 4242 (MO), EF152073, DQ924627 (ITS
clone A), DQ924628 (ITS clone B); (2) Namibia, Giess et al.
5837 (MO); AJ457132, DQ924626; Cuscuta rausii M. A.
´a, Greece, Raus 8471 (B), AJ457133, DQ924615; Cuscuta
rhodesiana Yunck., Zambia, Richards 16851 (BR), EF152074,
DQ924640; Cuscuta somaliensis Yunck., (1) Yemen, Wood 74/
337 (BM), AJ457134, DQ924629; (2) Socotra, Thulin and Gifri
8619 (UPS), AJ457131, DQ924625; (3) Ethiopia, De Wilde 6653
(MO), _, DQ924630; (4) Tanzania, Abdallah et al. 378 (UPS),
EF152075, DQ924634 (ITS clone A), DQ924635 (ITS clone B);
Cuscuta triumvirati Lange, (1) Spain, Garcı
´a 1077 (MA),
AJ430073, DQ924611; (2) Morocco, Jury et al. 18299 (MA),
AJ430074, DQ924612.
Cuscuta japonica Choisy, USA, Texas, Ketchersid and
Heintze s.n. (SIU), EF152064, DQ924571; Cuscuta lehmanni-
ana Bunge, Afghanistan, Freitag 3921 (Herb. Freitag);
AJ428055, _ ; Cuscuta lupuliformis Krock, Germany, Schu-
macher s.n. (MA); AJ428058, DQ924570; Cuscuta monogyna
Vahl., Spain, Garcı
´a 1290 (MA), AJ428054, DQ924569.
Cuscuta campestris Yunck., Spain, Garcı
´a 910 (MA);
AJ428057, _ ; Cuscuta cuspidata Engelm., USA, Texas, Garcı
et al. 3918 (MA), AJ428056, _ ; Cuscuta umbellata Kunth,
Cape Verde Islands, Medina s.n. (MA), AJ428053, EF192271.
GenBank accessions of sequences used but not obtained for
this study: Cuscuta attenuata Waterf., AF348404, AF348405;
Cuscuta australis R. Br., AY558828, AY554398; Cuscuta
cephalanthi Engelm., AY558829, AY554399; Cuscuta chinen-
sis Lam., AY558830, AY558824; Cuscuta compacta Choisy,
AY558831, AY558825; Cuscuta gronovii Willd., AY558836,
AY554402; Cuscuta indecora Choisy, AY558837, AY554403.
Systematic Botany sbot-32-04-19.3d 21/9/07 09:47:59 916 Cust #06-59
... Overall, members of this genus occur on all continents, except Antarctica, with most species reported in the Americas and Mexico, which are also considered their centre of diversity (Yuncker, 1932;Stefanović et al., 2007). In Africa, only a handful of studies have reported dodder occurrence (Zerman and Saghir, 1995;Garcia, 1999;Garcia and Martin, 2007;Garcia et al., 2014). However, their distribution patterns remain unknown. ...
... These markers have been extensively used to infer phylogenetic relationships, character evolution, and biogeography across the genus (Garcia and Martin, 2007;McNeal., et al., 2007;Garcia et al., 2014). In our case, all our sequences were resolved alongside respective C. campestris, C. kilimanjari, and C. reflexa taxa from GenBank, indicating that they indeed belonged to these species. ...
... In our case, all our sequences were resolved alongside respective C. campestris, C. kilimanjari, and C. reflexa taxa from GenBank, indicating that they indeed belonged to these species. Additionally, these phylogenetic reconstructions confirmed the monophyly of subgenus Monogynella, with all members basal to subgenera Cuscuta and Grammica, consistent with earlier reports (Garcia and Martin, 2007;McNeal et al., 2007;Stefanović et al., 2007;Garcia et al., 2014). Apart from these three, other Cuscuta species have also been reported in Africa, although most of them belong to subgenus Cuscuta (Zerman and Saghi, 1995;Garcia, 1999;Garcia and Martin, 2007). ...
... Overall, members of this genus occur on all continents, except Antarctica, with most species reported in the Americas and Mexico, which are also considered their center of diversity (Yuncker, 1932;Stefanovi c et al., 2007). In Africa, only a handful of studies have reported dodder occurrence (Zerman and Saghir, 1995;García, 1999;García and Martin, 2007;García et al., 2014). However, their distribution patterns remain unknown. ...
... We validated this identification by sequencing rbcL, trnL, and ITS regions from representative individuals, and performed phylogenetic reconstruction alongside other species from GenBank. These markers have been extensively used to infer phylogenetic relationships, character evolution, and biogeography across the genus (García and Martin, 2007;McNeal et al., 2007;García et al., 2014). In our case, all our sequences were resolved alongside respective C. campestris, C. kilimanjari, and C. reflexa taxa from GenBank, indicating that they indeed belonged to these species. ...
... In our case, all our sequences were resolved alongside respective C. campestris, C. kilimanjari, and C. reflexa taxa from GenBank, indicating that they indeed belonged to these species. Additionally, these phylogenetic reconstructions confirmed the monophyly of subgenus Monogynella, with all members basal to subgenera Cuscuta and Grammica, consistent with earlier reports (García and Martin, 2007;McNeal et al., 2007;Stefanovi c et al., 2007;García et al., 2014). Apart from these three, other Cuscuta species have also been reported in Africa, although most of them belong to subgenus Cuscuta (Zerman and Saghi, 1995;García, 1999;García and Martin, 2007). ...
Full-text available
Invasive holoparasitic plants of the genus Cuscuta (dodder) threaten African ecosystems due to their rapid spread and attack on various host plant species. Most Cuscuta species cannot photosynthesize and hence rely on host plants for nourishment. After attachment through a peg-like organ called a haustorium, the parasites deprive hosts of water and nutrients, which negatively affects host growth and development. Despite their rapid spread in Africa, dodders have attracted limited research attention, although data on their taxonomy, host range, and epidemiology are critical for their management. Here, we combine taxonomy and phylogenetics to reveal the presence of field dodder (Cuscuta campestris) and Cuscuta kilimanjari (both either naturalized or endemic to East Africa), in addition to the introduction of the giant dodder (Cuscuta reflexa), a south Asian species, in continental Africa. These parasites have a wide host range, parasitizing species across 13 angiosperm orders. We evaluated the possibility of C. reflexa to expand this host range to tea (Camelia sinensis), coffee (Coffea arabica), and mango (Mangifera indica), crops of economic importance to Africa, for which haustorial formation and vascular-bundle connections in all three crops revealed successful parasitism. However, only mango mounted a successful post-attachment resistance response. Furthermore, species distribution models predicted high habitat suitability for Cuscuta spp. across major tea- and coffee-growing regions of Eastern Africa, suggesting an imminent risk to these crops. Our findings provide relevant insights into a poorly understood threat to biodiversity and economic wellbeing in Eastern Africa, and provide critical information to guide development of management strategies to avert Cuscuta spp. spread.
... Cuscutaceae. Based on recent molecular studies, Cuscuta should remain in Convolvulaceae (Garcia & Martin 2007). The study of micro-morphological characters is an important step in the establishment of relationships between the comprising taxa. ...
... In the present study, 12 populations from three species of Cuscuta (C. australis Hook.f., C. campestris, and C. chinensis) were obtained from northern parts of Iran during field work from March to July 2016 ( (Garcia & Martin 2007). ...
... Although, important factors that could be responsible for this incongruence are: differences in number of species and choice of molecular markers. In a study carried out by Garcia & Martin (2007), the position of some taxon is not resolved and arises from the polytomy of a clade. They believed that, factors influencing the taxonomic difficulty of many species in the subgenus Cuscuta, include lack of morphological characters, parallelism and gene flow between closelyand not-so-closely related species. ...
Full-text available
Cuscuta is the only parasitic genus in Convolvulaceae family. This genus is globally distributed, with most species in the tropics, subtropics, and some in the temperate regions. In this study, the micro-morphological features and molecular evidences of 12 populations from three species of Cuscuta (C. australis, C. campestris, and C. chinensis) have been considered. In total, seven quantitative and two qualitative characters of pollen were selected and measured. The most important characters include: shape, ornamentation of tectum, exine thickness and colpus length of the pollen. Based on this study, the seed shape and surface support at least for separation of C. australis from other two species. Using nuclear (nrDNA ITS) marker, we reconstructed phylogenetic relationships within three species of Cuscuta. This data set was analyzed by phylogenetic methods including Bayesian, Maximum likelihood, and Maximum parsimony. In phylogenetic analyses, all members of three species formed a well-supported clade (PP=1, ML/BS=100/100) and divided into two major clades (A and B). Clade A is composed of specimens of C. australis. Two species of C. campestris and C. chinensis are nested in clade B. Neighbor-Net diagram demonstrated separation of the studied populations. The results showed that, micro-morphological and molecular data provide reliable evidence for separation of these species.
... Different accessions of C. approximata Bab., for example, have polymorphism in the ITS, and the location of C. kurdica Engelm. differed between the ITS and trnL trees (García and Martín 2007). These contrasting topologies may indicate hybridization events similar to those reported in subgenus Grammica (e.g. ...
... Stefanović and Costea 2008;García et al. 2014). Evidence such as this has not been observed in C. africana, C angulata and C. nitida with ITS, 26S, trnL nor rbcL sequence analyses (García and Martín 2007). Chromosome number and size, as well as the number of rDNA sites in C. nitida, were within the range of variation already reported for species of the genus Cuscuta. ...
Full-text available
Cuscuta is a cytogenetically diverse genus, with karyotypes varying 18-fold in chromosome number and 127-fold in genome size. Each of its four subgenera also presents particular chromosomal features, such as bimodal karyotypes in Pachystigma. We used low coverage sequencing of the Cuscuta nitida genome (subgenus Pachystigma), as well as chromosome banding and molecular cytogenetics of three subgenus representatives, to understand the origin of bimodal karyotypes. All three species, C. nitida, C. africana (2n = 28) and C. angulata (2n = 30), showed heterochromatic bands mainly in the largest chromosome pairs. Eighteen satellite DNAs were identified in C. nitida genome, two showing similarity to mobile elements. The most abundant were present at the largest pairs, as well as the highly abundant ribosomal DNAs. The most abundant Ty1/Copia and Ty3/Gypsy elements were also highly enriched in the largest pairs, except for the Ty3/Gypsy CRM, which also labelled the pericentromeric regions of the smallest chromosomes. This accumulation of repetitive DNA in the larger pairs indicates that these sequences are largely responsible for the formation of bimodal karyotypes in the subgenus Pachystigma. The repetitive DNA fraction is directly linked to karyotype evolution in Cuscuta.
... In C. planiflora (2n = 14, 26, 28, and 34), the populations with 2n = 34 have an asymmetrical karyotype that together with morphological features indicate that it is an allopolyploid (García, 2001). Both, C. epithymum and C. planiflora are taxonomically difficult as revealed by the high number of infraspecific taxa that have been described (García and Martín, 2007), and their variation in chromosome numbers and karyotypes may indicate cryptic diversity in these species' complexes. FIGURE 6 | Reconstruction of the position of the 5S (left) and 35S (right) rDNA sites. ...
Full-text available
Karyotypes are characterized by traits such as chromosome number, which can change through whole-genome duplication and dysploidy. In the parasitic plant genus Cuscuta (Convolvulaceae), chromosome numbers vary more than 18-fold. In addition, species of this group show the highest diversity in terms of genome size among angiosperms, as well as a wide variation in the number and distribution of 5S and 35S ribosomal DNA (rDNA) sites. To understand its karyotypic evolution, ancestral character state reconstructions were performed for chromosome number, genome size, and position of 5S and 35S rDNA sites. Previous cytogenetic data were reviewed and complemented with original chromosome counts, genome size estimates, and rDNA distribution assessed via fluorescence in situ hybridization (FISH), for two, seven, and 10 species, respectively. Starting from an ancestral chromosome number of x = 15, duplications were inferred as the prevalent evolutionary process. However, in holocentric clade (subgenus Cuscuta), dysploidy was identified as the main evolutionary mechanism, typical of holocentric karyotypes. The ancestral genome size of Cuscuta was inferred as approximately 1C = 12 Gbp, with an average genome size of 1C = 2.8 Gbp. This indicates an expansion of the genome size relative to other Convolvulaceae, which may be linked to the parasitic lifestyle of Cuscuta. Finally, the position of rDNA sites varied mostly in species with multiple sites in the same karyotype. This feature may be related to the amplification of rDNA sites in association to other repeats present in the heterochromatin. The data suggest that different mechanisms acted in different subgenera, generating the exceptional diversity of karyotypes in Cuscuta.
... Nas últimas décadas, diversas análises morfológicas e evolutivas com espécies do gênero Cuscuta foram realizadas, descrevendo, por exemplo, a morfologia dos grãos de pólen (Welsh et al. 2010) e das escamas infraestaminais (Riviere et al. 2013) e as estratégias reprodutivas (Wright et al. 2011(Wright et al. , 2012. Com base em todas as evidências corroboradas e considerando as recentes análises filogenéticas com o gênero (García & Martín 2007, Stefanović et al. 2007, García et al. 2014, Costea et al. (2015) classificaram todas as 194 espécies atualmente reconhecidas de Cuscuta em quatro subgêneros: subg. ...
Full-text available
RESUMO Cuscuta é amplamente distribuído e possui aproximadamente 200 espécies de parasitas volúveis. Estima-se que ocorram 26 espécies no Brasil, porém, não há estudos atuais publicados para a flora do país. Por meio de extensa revisão bibliográfica, de materiais de herbários e expedições a campo, foram reconhecidos 15 táxons de Cuscuta na Região Sul do Brasil (Estados do Paraná, Santa Catarina e Rio Grande do Sul): Cuscuta boliviana, C. campestris, C. corniculata, C. epilinum, C. incurvata, C. indecora var. neuropetala, C. obtusiflora, C. odorata, C. orbiculata, C. platyloba, C. racemosa, C. taimensis, C. xanthochortos var. xanthochortos, C. xanthochortos var. carinata e C. xanthochortos var. lanceolata. Uma nova espécie foi descrita (C. taimensis P.P.A. Ferreira & Dettke) e novas ocorrências para os Estados e uma para o Brasil foram confirmadas durante este estudo. São fornecidas chave de identificação, descrições morfológicas, ilustrações, além de dados de distribuição geográfica e habitat dos táxons.
... A study by Nickrent [1] indicates that leading economic implication or damage is to the host plant accelerated by species from four genera, namely: Cassytha, Cuscuta, Striga, Orananche, and Arceuthobium. Different studies [2][3][4] have indicated that dodder is the most important and challenging parasitic group of weed of over 70 varieties and 200 species. ...
Full-text available
Parasitic plants proliferation globally is daunting and a threat to our ecosystems. In this study we explore holoparasites with limitation to dodders (Cuscuta spp. & Cassytha filiformis). An experiment was performed to ascertain anatomical and morphological characteristics of dodder capsule and its stem. We present dodder infestation stages, development phases and close observable internal and external microscopic features. A distinct haustorium trait of dodders is shown by micrographs. The study finds that dodder seeds possess high ecological dispersal character with vast adaptability through morphological analysis. This ramifies their ecological phenotypic plasticity. Externally, dodder stems attack the host phloem through haustoria that suck nutrients from the sap weakening it.
... Possibly the first molecular phylogenetic study to include Cuscuta, Soltis & al. (1997) showed that the genus was sister to Ipomoea L. Data from nuclear, plastid, and mitochondrial genomes were used by Stefanović & Olmstead (2004) to show that Cuscuta was nested within Convolvulaceae, but despite extensive analysis, its exact position in the family could not be determined. Subsequent molecular phylogenetic work has provided a robust framework for the species within the genus (García & Martin, 2007;McNeal & al., 2007b;Stefanović & al., 2007;García & al., 2014). Van der Kooij & al. (2000) compared six dodder species for chlorophyll content, photosynthetic capacity, and plastid ultrastructure. ...
Full-text available
Angiosperms that morphologically and physiologically attach to other flowering plants by means of a haustorium have evolved 12 times independently resulting in 292 genera and ca. 4750 species. Although hemiparasites predominate, holoparasitism has evolved in all but two clades, Cassytha (Lauraceae) and Krameria (Krameriaceae). Santalales contains the largest number of genera (179) and species (2428) among the 12 parasitic plant lineages whereas Orobanchaceae is the largest single family with 102 genera and over 2100 species. This review presents the current state of knowledge on the molecular phylogenetic relationships among all clades of parasitic angiosperms. These methods have been particularly important in revealing the closest non‐parasitic relatives of holoparasites, plants that exhibit reduced morphologies, increased substitution rates, and frequent horizontal gene transfers, all of which confound phylogenetics. Although comprehensive molecular phylogenies are still lacking for many of the large genera, nearly complete generic level sampling exists, thus allowing unprecedented understanding of the evolutionary relationships within and among these fascinating plants.
... The worst economic damage in important host crops is caused by species from only four genera: Cuscuta, Arceuthobium, Orobanche, and Striga [18]. The genus Cuscuta L. (dodders) is one the most diverse and challenging groups of parasitic plants with more than 200 species and over 70 varieties [19][20][21]. The stem of a field dodder plant is threadlike and twining, and it is either leafless or the leaves are reduced to hardly visible scales. ...
The recently-developed statistical method known as the "bootstrap" can be used to place confidence intervals on phylogenies. It involves resampling points from one's own data, with replacement, to create a series of bootstrap samples of the same size as the original data. Each of these is analyzed, and the variation among the resulting estimates taken to indicate the size of the error involved in making estimates from the original data. In the case of phylogenies, it is argued that the proper method of resampling is to keep all of the original species while sampling characters with replacement, under the assumption that the characters have been independently drawn by the systematist and have evolved independently. Majority-rule consensus trees can be used to construct a phylogeny showing all of the inferred monophyletic groups that occurred in a majority of the bootstrap samples. If a group shows up 95% of the time or more, the evidence for it is taken to be statistically significant. Existing computer programs can be used to analyze different bootstrap samples by using weights on the characters, the weight of a character being how many times it was drawn in bootstrap sampling. When all characters are perfectly compatible, as envisioned by Hennig, bootstrap sampling becomes unnecessary; the bootstrap method would show significant evidence for a group if it is defined by three or more characters.
As a result of a worldwide taxonomic revision of Cuscuta L. subgenus Cuscuta, C. castroviejoi M.A. Garcia is described as a new species endemic to Ethiopia. The taxonomic status of some little known African species of the subgenus that appear in Ethiopia is discussed. Two previously recognized species of the subgenus, C. approximata Bab, and C. pedicellata Ledeb., are excluded for the country while C. pretoriana Yunck. and C. somaliensis Yunck. are recorded for the first time. A key to the species of the subgenus in Ethiopia is provided.
Drosera dichrosepala was exposed to different doses of Gamma radiation. Fragmented and fused chromosomes were observed as a consequences. Diffused centromeres in every fragment chromosome was detected by centromeric banding (Cd-banding) and was supported by its typical disjunction and totally lack of lagging chromosomes or micronuclei from anaphase to telophase. C-banding revealed that the fragment chromosomes were highly heterochromatic and fragmentation might be occurred at the terminal regions of chromosomes. Fluorescent banding suggested that most of the fragment chromosomes were rich in GC base composition in the species. Alteration of karyotype due to Gamma irradiation also indicated that spontaneous fragmentation or fusion of chromosomes might be a possible factor for promoting the bimodal karyotype in this genus.
Cuscuta L. plants are holoparasites with high chromosome divergence. Species representing the three subgenera of the genus were checked for their chromosome number, length and behaviour. Chromosome numbers range between 2n = 8 and 2n = 60, and their size, although roughly uniform within species, varies between 1–23μm. On the whole, size or number of Cuscuta chromosomes cannot be used as diagnostic criteria, except for very few species with unique numbers or sizes. Some associations between certain systematic entities and chromosome characters can be inferred. Species of the Old World of subgenus Cuscuta Engelm. are characterized by inverted meiosis and apparently by holocentric chromosomes, whereas species of the two other subgenera, subgenus Motiogyna Engelm. and subgenus Grammica Yuncker, undergo the regular process of meiosis and have monocentric chromosomes. The presence of holocentric chromosomes does not seem to be associated with their number.