ArticlePDF Available

Abstract and Figures

Cannabaceae includes ten genera that are widely distributed in tropical to temperate regions of the world. Because of limited taxon and character sampling in previous studies, intergeneric phylogenetic relationships within this family have been poorly resolved. We conducted a molecular phylogenetic study based on four plastid loci (atpB-rbcL, rbcL, rps16, trnL-trnF) from 36 ingroup taxa, representing all ten recognized Cannabaceae genera, and six related taxa as outgroups. The molecular results strongly supported this expanded family to be a monophyletic group. All genera were monophyletic except for Trema, which was paraphyletic with respect to Parasponia. The Aphananthe clade was sister to all other Cannabaceae, and the other genera formed a strongly supported clade further resolved into a Lozanella clade, a Gironniera clade, and a trichotomy formed by the remaining genera. Morphological ancestral state reconstructions indicated the complex evolution pattern of most analyzed morphological characters, and it is difficult to identify morphological synapomorphies for most clades within Cannabaceae.
Content may be subject to copyright.
473
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • June 2013: 473–485
473Version of Record (identical to print version).
INTRODUCTION
Cannabaceae was first separated from Moraceae by Rendle
(1925) and contained only Cannabis L. and Humulus L. The
circumscription of this family has been expanded significantly
to include most members of Ulmaceae subfam. Celtidoideae
sensu Engler & Prantl (1893) or Celtidaceae sensu Link (1829).
Ampelocera Klotzsch has been excluded from the group and
placed in Ulmaceae following a series of molecular phyloge-
netic studies (Ueda & al., 1997; Wiegrefe & al., 1998; Sytsma
& al., 2002). Cannabaceae contains ten genera (Sytsma & al.,
2002; Mabberley, 2008) and 109 accepted and 71 putative spe-
cies (The Plant List, 2010) (Table 1).
Cannabaceae comprises taxa var ying greatly in terms of
habit and morphology. Most are trees and shrubs, but the fam-
ily also includes herbs (Cannabis) and vines (Humulus). Fruits
are usually drupes, but samaras occur in Pteroceltis Maxim.
and achenes in Cannabis and Humulus. Leaves are usually
alternate, but opposite in Lozanella Greenm., and both op-
posite and alternate in Cannabis and Humulus. Although the
morphological synapomorphies of Cannabaceae are not clear,
some morphological characters including usually unisexual
and inconspicuous flowers, antitepalous stamens, the pres-
ence of stipules, diporate or triporate pollen, and free filaments
slightly adnate to the tepals can be used to identify this family
(Judd & al., 2008). Most genera have restricted distributions,
although Aphananthe Planch., Celtis L. and Trema Lour. are
widely distributed in tropical and temperate regions (Table 1).
The modern circumscription of Cannabaceae was first
proposed by Wiegrefe & al. (1998), who carried out parsimony
analyses of chloroplast DNA restriction site data. This conclu-
sion has been supported by a subsequent parsimony analysis of
the chloroplast gene matK (Song & al., 2001), parsimony analy-
sis of the plastid regions rbcL, trnL-trnF and ndhF (Sytsma
& al., 2002), and Bayesian and parsimony analyses of the plas-
tid regions rbcL and trnL-trnF (Van Velzen & al., 2006).
Cannabaceae is sister to Moraceae and Urticaceae (Sytsma
& al., 2002; Van Velzen & al., 2006; Wang & al., 2009; Zhang
& al., 2011). However, phylogenetic relationships within Can-
nabaceae are still largely unresolved. A phylogenetic analysis
based on 33 morphological, palynological, biochemical, and
cytogenetic characters suggested that Broussonetia L’Hér. ex
Vent . of Moraceae is nested within Celtidaceae, a surprising
phylogenetic relationship (Zavada & Kim, 1996). This study
also suggested a sister-group relationship of Chaetachme
Planch. and Gironniera Gaudich., which was surprising be-
cause the two genera show strong morphological divergence
(Zavada & Kim, 1996). Some previous molecular studies
Molecular phylogenetics and character evolution of Cannabaceae
Mei-Qing Yang,1,2,3 Robin van Velzen,4,5 Freek T. Bakker,4 Ali Sattarian,6 De-Zhu Li1,2 & Ting-Shuang Yi1,2
1 Key Laboratory of Biodiversity and Biogeography, Kunming Institute of Botany, Chinese Academy of Sciences, Kunming,
Yunnan 650201, P.R. China
2 Plant Germplasm and Genomics Center, Germplasm Bank of Wild Species, Kunming Institute of Botany, Chinese Academy
of Sciences, Kunming, Yunnan 650201, P.R. China
3 University of Chinese Academy of Sciences, Beijing 100093, P.R. China
4 Biosystematics Group, Wageningen University, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands
5 Laboratory of Molecular Biology, Wageningen University, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands
6 Department of Natural Resources, Gonbad Universit y, Gonbad Kavous 4971799151, Iran
Authors for correspondence: Ting-Shuang Yi, tingshuangyi@mail.kib.ac.cn; De-Zhu Li, dzl@mail.kib.ac.cn
Abstract
Cannabaceae includes ten genera that are widely distributed in tropical to temperate regions of the world. Because of
limited taxon and character sampling in previous studies, intergeneric phylogenetic relationships within this family have been
poorly resolved. We conducted a molecular phylogenetic study based on four plastid loci (atpB-rbcL, rbcL, rp s16, trnL-trnF )
from 36 ingroup taxa, representing all ten recognized Cannabaceae genera, and six related taxa as outgroups. The molecular
results strongly supported this expanded family to be a monophyletic g roup. All genera were monophyletic except for Tre m a,
which was paraphyletic with respect to Parasponia. The Aphananthe clade was sister to all other Cannabaceae, and the other
genera formed a strongly supported clade further resolved into a Lozanella clade, a Gironniera clade, and a trichotomy formed
by the remaining genera. Morphological ancestral state reconstructions indicated the complex evolution pattern of most analyzed
morphological characters, and it is difficult to identify morphological synapomorphies for most clades within Cannabaceae.
Keywords
Cannabaceae; character evolution; classification; molecular phylogenetics
Supplementary Material
The Electronic Supplement (Figs. S1–S2) is available in the Supplementary Data section of the
online version of this ar ticle (http://www.ingentaconnect.com/content/iapt/tax).
Received: 31 May 2012; revision received: 12 Sept. 2012; accepted 16 Apr. 2013
474
TAX ON
62 (3) • June 2013: 473–485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
474 Version of Record (identical to print version).
suggested that Cannabis and Humulus are sister taxa (Song
& al., 2001; Sytsma & al., 2002), that Tre m a is paraphyletic with
respect to Parasponia (Sytsma & al., 2002; Yesson & al., 2004),
and that Aphahanthe is sister to the remainder of the family
(Song & al., 2001; Sytsma & al., 2002), but other intergeneric
relationships in Cannabaceae have remained unresolved (or
were weakly supported because of limited taxon and character
sampling). Van Velzen & al. (2006) analysed the phylogeny of
Cannabaceae based on two plastid markers (rbcL, trnL-trnF ).
They obtained stronger support for intergeneric relationships
and concluded that Chaetachme and Pteroceltis were distinct
genera, and recovered a large, strongly supported clade con-
taining Cannabis, Celtis, Chaetachme, Humulus, Parasponia,
Pteroceltis and Tre m a.
The current study included the three monotypic genera
(Cannabis, Chaetachme, Pteroceltis) of the family and sampled
multiple species of the other seven genera. Each sample was
sequenced for four plastid loci to estimate phylogenetic rela-
tionships in Cannabaceae. The major goals of this study were
to reconstruct intergeneric relationships in the family and to
optimise selected morphological characters on the inferred
phylogeny to identify potential synapomorphies for clades.
MATERIALS AND METHODS
Sampling. —
Thirty-six individuals representing twenty-
nine species of all ten recognized genera of Cannabaceae were
included, which is 26.6% of the accepted species (Table 1). We
sampled six species of Celtis (representing 8.2% of the accepted
species) across its distribution range (Eurasia, Africa, North
America, South America). Eight species of Trema (representing
66.7% of the accepted species) were included. For the mono-
typic genera Cannabis, Pteroceltis and Chaetachme, two to
three samples each were included. Two to three species were
sampled for the remaining genera (Aphananthe, Gironniera,
Humulus, Lozanella, Parasponia). Six species from the closely
related Moraceae, Urticaceae and Ulmaceae (all Rosales) were
chosen as outgroups based on the recent phylogenetic study of
Zhang & al. (2011). Genera currently included in Cannabaceae
with their species numbers and geographic ranges are shown
in Table 1.
DNA extraction, PCR and sequencing. —
Total genomic
DNA was extracted from silica gel-dried leaves or herbarium
specimens following the CTAB protocol of Doyle & Doyle
(1987).
Nucleotide sequences for the four chloroplast loci, i.e.,
atpB-rbcL spacer, rbcL, rp s16 and trnL-trnF spacer were gen-
erated using the following primers for both PCR amplifica-
tion and sequencing: atpB-F and rbcL-R for the atpB-rbcL
region (Chiang & al., 1998); rbcL-30F and rbcL-1400R for the
rbcL region (Zhang & al., 2011); f and 2r for the rps16 region
(Oxelman & al., 1997); and c and f for the trnL-trnF region
(Taberlet & al., 1991).
Polymerase chain reaction (PCR) amplifications were per-
formed in a volume of 25 μl containing 10–50 ng of genomic
DNA, 0.2 μmol of each primer, 10 mM Tris-HCl (pH 8.3),
50 mM KCl, 1.5 mM MgCl
2
,
20 0 μmol of ea ch dNT P and 0.5 U
Taq polymerase (Takara, Shanghai, China). The PCR cycling
parameters for all regions were as follows: a 95°C initial hot
start for 5 min, followed by 32 cycles of 94°C for 30 s, 50°C
for 40 s and 72°C for 60 s, and a final extension of 72°C for
10 min. PCR products were isolated and purified using a com-
mercial DNA purification kit (Sangon Inc., Shanghai, China)
following the manufacturer’s protocols. Cycle sequencing was
carried out using the ABI Prism BigDye Terminator Cycle
Sequencing Ready Reaction Kit with 5 ng of primer, 1.5 μl of
sequencing dilution buffer, and 1 μl of cycle sequencing mix
in a 10 μl reaction volume. Cycle sequencing conditions were
as follows: 30 cycles of 30 s denaturation (96°C), 30 s anneal-
ing (50°C), and 4 min elongation (60°C). The samples were
sequenced on an ABI 3700xl DNA analyzer. We sequenced
Table 1.
Genera currently included in Cannabaceae with their species numbers and geographic range.
Genera
Recognized
species (putative
species)
Species sampled
(individuals
sampled) Distribution
Aphananthe
Cannabis
Celtis
Chaetachme
Gironniera
Humulus
Lozanella
Parasponia
Pteroceltis
Trema
Total
5
1
73 (36)
1
6
3
2
5 (5)
1
12 (30)
109 (71)
3
1 (2)
6(7)
1 (3)
2
3
2
2
1 (2)
8 (10)
29 (36)
Madagascar, southwestern China to Japan, Malaysia, Indonesia and east Australia, Mexico
Asia, cultivated worldwide
Tropical (most), temperate (Europe 4)
Tropical Africa, southern Africa and Madagascar
Malaysia, Indonesia to Pacific
North temperate
Tropical areas of America
Malaysia, west Pacific
North and central China
Worldwide in tropical and warm climates
475
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • June 2013: 473–485
475Version of Record (identical to print version).
both strands of DNA with overlapping regions to ensure un-
ambiguous base calls.
Sequences were initially assembled using Sequencher v.4.2
(GeneCodes Corporation, Ann Arbor, Michigan, U.S.A.) and
aligned using Clustal X (Thompson & al., 1997), followed by
manual adjustments using Se-Al v.2.0 (Rambaut, 2002). Poten-
tially informative indels in regions of unambiguous alignment
were scored following the simple indel coding method (Simmons
& Ochoterena, 2000), which treats each gap as a single presence/
absence character independent of its length. A total of 82 gap
characters were coded (trnL-trnF: 24; atpB-rbcL: 27; rp s16: 31)
for inclusion in the parsimony analysis, 75 of which were phy-
logenetically informative. The matrix is available in TreeBASE
(study accession number 13867) and all new sequences generated
in this study have been deposited in GenBank under accession
numbers JN040281 to JN040432 (Appendix 1).
Phylogenetic analyses. —
Phylogenetic analyses were con-
ducted on each separate gene region as well as the combined
plastid datasets, with gaps treated as missing data and indels
coded as binary characters (simple indel coding). Phylogenetic
relationships were inferred using maximum parsimony (MP)
as implemented in PAUP* v.4.0b10 (Swofford, 2003), Bayesian
inference (BI) as implemented in MrBayes v.3.1.2 (Ronquist
& Huelsenbeck, 2003), and maximum-likelihood (ML) as
implemented in Garli v.0.96 (Zwickl, 2006).
The MP analyses used heuristic searches with 1000 ran-
dom sequence addition replicates, tree bisection-reconnection
(TBR) branch swapping, and MULTREES on. All character
states were treated as unordered and equally weighted. To eval-
uate the relative robustness of clades in the MP trees, bootstrap
analysis (Felsenstein, 1985) was performed with 1000 repli-
cates using the same options as above except that a maximum
of 100 trees were saved per random sequence addition replicate.
For BI and ML analyses, Modeltest v.3.7 (Posada & Cran-
dall, 1998) was run for each dataset to select the best model
of sequence evolution for each gene. The models were chosen
by the Akaike information criterion (AIC) and determined by
AIC scores. For coded indels, we modeled indels as evolving
according to a stochastic binary model. Considering the im-
portance of data partitioning (Brown & Lemmon, 2007), the
combined plastid dataset was analyzed by applying separate
models to each data partition, with all parameters unlinked
across data partitions except for topology and branch length.
For the Bayesian inference, one cold and three incrementally
heated Markov chain Monte Carlo (MCMC) chains were run
for 2,000,000 generations. Trees were sampled every 100 gen-
erations. MCMC runs were repeated twice to avoid spurious
results. Stationarity of the Markov chain was ascertained by
plotting and interpreting likelihood values against number of
generations in Tracer v.1.3 (Rambaut & Drummond, 2004). The
first 5000 trees were discarded as burn-in, and the remaining
trees were used to construct majority-rule consensus trees.
The average standard deviation of split frequencies between
the two runs was 0.004, and ESS values as computed by Tracer
v.1.3 (Rambaut & Drummond, 2004) were above 600 for all
individual MCMC runs. Following Alfaro & al. (2003), we
considered posterior probabilities (PP) greater or equal to 0.95
as significant probability for a clade. For maximum likelihood
analyses, default parameters were used for the Garli searches
except that “significant topochange” was set to 0.01 and a total
of 100 ML bootstrap replicates (MLBS) were performed. The
trees obtained from Garli were used to construct 50% majority-
rule consensus trees using PAUP* v.4.0b10 (Swofford, 2003).
In order to better understand phylogenetic relationships in
the Parasponia-Trema complex and Celtis, we also expanded
our sampling by including sequences of more taxa and two more
plastid loci (matK, ndhF) from GenBank (Appendix 2). These
sequences were combined with our own sequences of four plas-
tid gene loci, and unavailable sequences were treated as missing
data. The phylogeny was reconstructed from this combined data
matrix applying the same methods as described above.
Approximately unbiased test. —
The approximately unbi-
ased (AU) test (Shimodaira, 2002), as implemented in CONSEL
v.0.1i (Shimodaira & Hasegawa, 2001) with default scaling
and replicate values (1000 bootstrap replicates), was used to
estimate relative support for all possible topologies among Lo-
zanella, Gironniera and clade E.
Ancestral character state reconstructions. —
The aim of
reconstr ucting ancestral character states was to evaluate the
evolution of morphological characters and identify potential
synapomorphies for clades. We chose eight morphological
characters that have been widely used in the classification of
genera in this family (Table 2). These morphological data were
obtained from specimen observations and the literature (Killip
& Morton, 1931; Zavada & Dilcher, 1986; Takahashi, 1989;
Takaso & Tobe, 1990; Todzia, 1993; Zavada & Kim, 1996;
Table 2.
Morphological characters and character states (for data matrix see Appendix 3).
Morphological characters States
Sexual system 0, monoecious; 1, dioecious; 2, andromonoecious; 3, monoecious or dioecious; 4, polygamous
Leaf arrangement 0, opposite; 1, alternate; 2, alternate and opposite
Pollen aperture number 0, triporate; 1, diporate; 2, pentaporate
Aestivation 0, valvate; 1, imbricate
Fruit type 0, drupe; 1, achene; 2, samara
Seed coat morphology 0, with holes; 1, without holes
Perianth at fruiting time 0, deciduous; 1, persistent
Stipule arrangement 0, intrapetiolar; 1, extrapetiolar; 2, interpetiolar
476
TAX ON
62 (3) • June 2013: 473–485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
476 Version of Record (identical to print version).
Sytsma & al., 2002; Fu & al., 2003; Zhou & Bartholomew,
2003; Sattarian & Maesen, 2006; Sattarian & al., 2006), and
all characters were discrete and coded as binary or multistate
(Table 2; Appendix 3). Ancestral state reconstructions were
performed in Mesquite v.2.74 (Maddison & Maddison, 2010)
using parsimony for tracing character evolution.
The Bayesian 50% majority-rule consensus tree of the
combined chloroplast regions was used for ancestral state
analysis, including outgroup species. In order to account for
uncertainty in phylogeny inference, ancestral state reconstruc-
tions were also performed using the Bayesian stochastic map-
ping (Huelsenbeck & al., 2003) approach as implemented in
SIMMAP v.1.5 (Bollback, 2006). Posterior probabilities for
ancestral character states were calculated using 15,000 trees
sampled from the trees (excluding burn-in) obtained in the
Bayesian analysis (see above). Morphology priors were cal-
culated using a two-step approach (http://www.simmap.com/
pgs/priors.html). In the first step, we performed an MCMC
analysis to sample the overall rate values (Gamma prior) and
bias values (Beta prior for two-state characters). In the second
step, we used the samples from the posterior distribution of
these parameters from the first step and selected the best fitting
Gamma and Beta distribution in R v.2.14.1 (R Development
Core Team, 2011).
RESU LT S
Phylogenetic analyses of individual chloroplast genes.
Sequences of rbcL and trnL-trnF were obtained for all sam-
pled accessions. However, sequences of rp s16 for Chaetachme
aristata 3 and Gironniera subaequalis were unavailable. We
failed to amplify atpB-rbcL of Aphananthe monoica, A. philip-
pinensis, Chaetachme aristata 1, C. aristata 2, C. aristata 3,
Humulus scandens, Parasponia melastomatifolia, and Tr e m a
tomentosa 1, although multiple attempts were made.
The four chloroplast datasets showed different levels of
sequence variation. Characteristics of each gene and the com-
bined dataset are presented in Table 3. The MP analysis for
rbcL resulted in five equally parsimonious trees of 425 steps:
this gene provided the lowest percentage of informative sites of
the four markers examined (12.0%). The MP analysis for atpB-
rbcL resulted in two equally parsimonious trees of 416 steps
(15.1% of characters potentially parsimony-informative). The
MP analysis of the trnL-trnF dataset resulted in three equally
parsimonious trees of 516 steps (17.5% of characters potentially
parsimony-informative). The MP analysis for rp s16 resulted
in eighteen equally parsimonious trees of 645 steps: this gene
provided the highest percentage of informative sites of all four
markers examined (20.2%). The ML and MP analyses of each
locus yielded topologies similar to the BI phylogeny and dif-
fered only in weakly supported clades (Electr. Suppl.: Fig. S1).
Phylogenetic analyses of combined chloroplast genes.
The chloroplast genome typically behaves as a single, non-
recombining region, and there was no significant conflict in
well-supported clades among the different gene trees. We thus
combined all four chloroplast loci in our analysis. The com-
bined dataset included 4538 unambiguously aligned positions
with 731 parsimony-informative characters and gaps treated
as missing. Parsimony analysis of the combined data yielded
one most parsimonious tree of 2048 steps (consistency index,
CI = 0.746; retention index, RI = 0.842). The coded indels of
the combined data provided another 82 characters, of which
77 (93.9%) were parsimony-informative (PI indels). When the
PI indels were excluded from the analyses, the same relation-
ships were observed, but some clades received lower support.
ML and BI analyses of the combined dataset yielded to-
pologies similar to the MP phylogeny (Fig. 1) and differed only
in weakly supported clades (P > 0.05). The Bayesian combined
chloroplast tree was selected to represent our results.
In the plastid combined analysis (Fig. 1), Cannabaceae
was strongly supported as monophyletic (MPBS = 91%; MLBS
= 97%; PP = 1.0). The monophyly of each genus of Cannab-
aceae was also strongly supported except for Parasponia
nested within Trema . Aphananthe (MPBS = 100%; MLBS =
100%; PP = 1.0) was sister to the rest of the family. Three
strongly supported clades were recovered: the Lozanella clade,
the Gironniera clade and the clade formed by the remaining
genera (MPBS = 98%; MLBS = 100%; PP = 1.0). Within the
last clade, three clades were resolved, and Chaetachme and
Pteroceltis were strongly supported as sister genera (MPBS
= 100%; MLBS = 100%; PP = 1.0). Likewise, Cannabis and
Humulus were also strongly supported as sister genera (MPBS
= 100%; MLBS = 100%; PP = 1.0).
The phylogeny from the expanded data matrix using ad-
ditonal GenBank sequences (Electr. Suppl.: Fig. S2) is largely
Table 3.
Characteristics of individual and combined datasets.
Dataset Taxa
Aligned
length
Variable sites Parsimony-informative sites
Consistency
index
Retention
index
Model
selected by
AIC
Within
ingroup
Entire
dataset
Within
ingroup
Entire
dataset
rbcL 42 1227 221 170 147 112 0.616 0.763 TIM+T+G
atpB-rbcL 34 1074 300 161 162 119 0.858 0.906 TVM+G
trnL-trnF 42 1117 329 174 196 158 0.789 0.887 GTR+G
rps16 40 1120 392 241 226 182 0.777 0.86 GTR+G
Combined 42 4538 1242 746 731 571 0.746 0.842 GTR+I+G
477
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • June 2013: 473–485
477Version of Record (identical to print version).
0.01
Celtis sinensis
Ulmus macrocarpa
Pilea cadierei
Debregeasia saeneb
Ficus tikoua
Broussonetia papyrifera
Cudrania tricuspidata
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus scandens
Humulus yunnanensis
Humulus lupulus
Celtis madagascariensis
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Chaetachme aristata 1
Chaetachme aristata 2
Chaetachme aristata 3
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema tomentosa 2
Trema cannabina 2
Trema angustifolia
1.00/92/91
1.00/100/100
1.00/100/100
1.00/83/89
1.00/100/100
1.00/100/100
0.83/-/-
1.00/91/97
1.00/100/100
0.56/-/-
1.00/98/100
0.54/-/-
0.70/-/-
1.00/100/100
1.00/100/100
1.00/98/98
1.00/92/90
1.00/97/94
1.00/100/100
1.00/100/100
1.00/100/100
1.00/83/77
1.00/100/100
1.00/100/100
1.00/87/94
1.00/100/100
1.00/100/100
1.00/100/100
1.00/100/100
0.95/-/56
0.99/58/73
1.00/98/95
1.00/87/86
0.99/58/73
A
B
C
Moraceae
Ulmaceae
Urticaceae
E
D
Fig. 1.
Bayesian tree based on the combined data of four chloroplast genes. A, B and C represent three unresolved nodes. Clades D and E are
discussed in the text. PP values from the BI analysis, and bootstrap values (%) of the MP and ML analyses are shown (PP/MPBS/MLBS). Dashes
indicate that branches are supported by less than 50% PP, MPBS, or MLBS.
478
TAX ON
62 (3) • June 2013: 473–485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
478 Version of Record (identical to print version).
congruent with that based on the four loci sequenced by us.
Exceptions were: (1) Celtis was not monophyletic in the BI
and ML analyses, which suggested that C. schippii is sister
to a clade formed by all genera except Aphananthe, Gironn-
niera and Lozanella. However, this relationship only obtained
weak support. Celtis was supported as monophyletic in the
MP analysis without BS support; (2) the relationships among
Celtis, the Cannabis-Humulus clade, the Chaetachme-Ptero-
celtis clade, and the Parasponia-Tre m a clade are slightly dif-
ferent between the two phylogenies, but relationships among
these clades were not resolved in both analyses (Fig. 1; Electr.
Suppl.: Fig. S2). The expanded data did not improve the reso-
lution of the phylogeny of Cannabaceae, and some apparently
spurious results may have been caused by missing data. We
thus only included this phylogeny as a supporting figure and
use it only to address relationships in the Parasponia-Tre m a
clade.
Ancestral character state reconstructions. —
The an-
cestral condition of the sexual system of the family is equiv-
ocal in the parsimony analysis, and the Bayesian analysis
indicated monoecy being the ancestral state with a slightly
higher probability (69.8%; Fig. 2A). Four shifts to monoecy
and dioecy occurred in Gironniera, the Cannabis-Humulus
clade, Chaetachme and Tre m a angustifolia, two shifts to an-
dromonoecy in Celtis, and one shift to dioecy in Lozanella.
Polygamy independently evolved at least five times in Tre m a
and Celtis.
Alternate leaves represent the ancestral state for Can-
nabaceae in both the parsimony and Bayesian analyses with
high probobility (99.6%; Fig. 2B). One shift to opposite leaves
occurred in Lozanella, and one shift to opposite and alternate
leaves in the Cannabis and Humulus clade.
Triporate pollen is inferred to be the ancestral state for
Cannabaceae in both the parsimony and Bayesian analyses
Fig. 2A–H.
Reconstruction of ancestral states of morphological traits using parsimony and Bayesian approaches. Posterior probabilities for each
character state are indicated as pie char ts.
monoecious
dioecious
andromonoecious
monoecious and dioecious
Sexual system
opposite
alternate
opposite and alternate
Leaf arrangement
polygamous
Ulmus macrocarpa
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Chaetachme aristata 2
Trema levigata
Trema tomentosa 1
Parasponia andersoni
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema tomentosa 2
Trema cannabina 2
Trema angustifolia
Chaetachme aristata 1
Chaetachme aristata 3
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Trema levigata
Trema tomentosa 1
Parasponia andersoni
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
Ulmus macrocarpa
Chaetachme aristata 2
Chaetachme aristata 1
Chaetachme aristata 3
Trema tomentosa 2
Trema cannabina 2
AB
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Humulus scandens
Humulus yunnanensis
Humulus scandens
Humulus yunnanensis
479
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • June 2013: 473–485
479Version of Record (identical to print version).
with a relatively high probability (85.8%; Fig. 2C). One shift
to diporate pollen occurred in the Tr ema-Parasponia clade.
The ancestral state of aestivation for Cannabaceae is
equivocal in the parsimony analysis (Fig. 2D). The Bayesian
analysis inferred imbricate aestivation as the ancestral condi-
tion for the family with a slightly higher probability (66.2%;
Fig. 2D). Shifts to valvate aestivation occurred at least two
times in Aphananthe and the Chaetachme-Pteroceltis-Tre m a -
Parasponia clade. Within the latter clade, two reversals to im-
bricate aestivation took place in Pteroceltis and Parasponia.
Drupes are the ancestral state for Cannabaceae in both the
parsimony and Bayesian analyses with high probability (96.2%;
Fig. 2E). One shift to samaras occurred in Pteroceltis, and one
shift to achenes in the Cannabis-Humulus clade.
Seed coat without holes is reconstructed as the ancestral
state by both the parsimony and Bayesian analyses (90.1%;
Fig. 2F). One shift to seed coat with holes occurred in clade D
(as defined in Fig. 1), and two reversals took place in Humulus
and the Trema-Parasponia clade.
Persistent perianth is the ancestral condition for the fam-
ily in both the parsimony and Bayesian analyses with slightly
higher probability (63.1%) (Fig. 2G). Four shifts to deciduous
perianth occurred in Celtis, Chaetachme, Lozanella and Trema
levigata.
Extrapetiolar stipules is the ancestral state for Cannabaceae
supported by both the parsimony and Bayesian analyses with
high probability (97.2%; Fig. 2H). Three shifts to intrapetiolar
stipules occurred in Chaetachme, Lozanella and Parasponia,
and one shift to interpetiolar stipules occurred in Humulus.
DISCUSSION
The monophyly of Cannabaceae.
Including all recog-
nized genera and using four plastid loci, our molecular data
strongly supported the monophyly of Cannabaceae (MPBS =
91%; MLBS = 97%; PP = 1.0). Members of this family share a
9-bp deletion at position 247 in the trnL-trnF intergenic spacer.
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Chaetachme aristata 2
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
triporate
diporate
pentaporate
Aperture number
Ulmus macrocarpa
Trema tomentosa 2
Trema cannabina 2
Chaetachme aristata 1
Chaetachme aristata 3
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
valvate
imbricate
Aestivation
Ulmus macrocarpa
Chaetachme aristata 2
Chaetachme aristata 1
Chaetachme aristata 3
Trema tomentosa 2
Trema cannabina 2
C D
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Humulus scandens
Humulus yunnanensis
Humulus scandens
Humulus yunnanensis
480
TAX ON
62 (3) • June 2013: 473–485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
480 Version of Record (identical to print version).
The recently expanded Cannabaceae can be identified by the
following morphological characters: presence of cystoliths,
absence of laticifers, antitepalous stamens, triporate or diporate
pollen grains, curved embryo, two carpels, and superior ovary
with apical placentation. However, most of these characters are
shared with Moraceae and/or Urticaceae. More detailed studies
should be carried out to identify synapomorphies of the family.
Intergeneric relationships among Cannabaceae. —
Phylo-
genetic relationships within Cannabaceae were not completely
resolved in previous studies (Song & al., 2001; Sytsma & al.,
2002; Van Velzen & al., 2006). However, including all recog-
nized genera, we were able to greatly improve phylogenetic
resolution. Most nodes except nodes A, B and C obtained strong
support (Fig. 1). Whether these nodes reflect rapid evolution
remains to be tested. Below we discuss phylogenetic relation-
ships in the family based on our results.
The Aphananthe clade. —
Our results strongly supported
Aphananthe (MPBS = 100%, MLBS = 100%, PP = 1.0) as sister
to all other genera of Cannabaceae, in agreement with most pre-
vious molecular studies (Song & al., 2001; Sytsma & al., 2002;
Van Velzen & al., 2006). Aphananthe differs from the other
genera by its base chromosome number of x = 13 (Oginuma
& al., 1990), asymmetrical ovules (Takaso, 1987), the presence
of flavonols (Giannasi, 1978), and a unique seed coat morphol-
ogy (Takaso & Tobe, 1990), supporting the finding that this
genus has an isolated position in the family. Our molecular
study strongly suggested the genus to be a member of Can-
nabaceae, with which it shares pollen structure (Kuprianova,
1962; Takahashi, 1989), leaf vernation pattern (Terabayashi,
1991), and gynoecial vasculature (Omori & Terabayashi, 1993).
Phylogenetic relationships in clade E. —
The Giron-
niera clade, the Lozanella clade and clade D together formed a
strongly supported monophyletic group (MPBS = 100%, MLBS
= 100%, PP = 1.0), marked as clade E in Fig. 1. Most members
of this group contain glycoflavones (Giannasi, 1978). Excep-
tions to this are Cannabis, which produces f lavonols as well
as glycoflavones (Clark, 1978), and Humulus which contains
flavonols (Alaniya & al., 2010). In addition, all members of
clade E share two deletions in trnL-trnF (one 100-bp deletion
at position 125–224; one 8-bp deletion at position 270–277).
The phylogenetic relationships among Gironniera, Lozanella
and clade D within this clade were not resolved. Our Bayesian
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
drupe
achene
samara
Fruit type
Ulmus macrocarpa
Chaetachme aristata 2
Chaetachme aristata 1
Chaetachme aristata 3
Trema tomentosa 2
Trema cannabina 2
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
with holes
without holes
Seed coat morphology
Ulmus macrocarpa
Chaetachme aristata 2
Chaetachme aristata 1
Chaetachme aristata 3
Trema tomentosa 2
Trema cannabina 2
E F
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Humulus scandens
Humulus yunnanensis
Humulus scandens
Humulus yunnanensis
481
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • June 2013: 473–485
481Version of Record (identical to print version).
inference indicated that Gironniera is sister to a branch compris-
ing Lozanella and clade D, but this conclusion is inconsistent
with the result of the AU test, which suggested that Lozanella is
sister to a branch comprising Gironniera and clade D.
Phylogenetic relationships in clade D. —
The monophyly
of Clade D received strong support (MPBS = 100%, MLBS
= 98%, PP = 1.0). Most members of this clade have a base
chromosome number of x = 10, but Humulus has x = 8 and
Chaetachme x = 15 (Oginuma & al., 1990; Sytsma & al., 2002).
Four strongly supported subclades (Celtis, Cannabis-Humu-
lus, Chaetachme-Pteroceltis, Tre m a-Parasponia) were recon-
structed in accordance with results of Van Velzen & al. (2006),
but relationships among them were largely unresolved. The
phylogenetic position of Cannabis and Humulus has been con-
troversial for a long time with relationships suggested to Celtis
(Chase & al., 1993), to Pteroceltis-Celtis (Wiegrefe & al., 1998),
to Pteroceltis-Celtis-Tre m a (Song & al., 2001) or to Pteroceltis
(Sytsma & al., 2002). The limitations of these studies were that
they did not sample all genera and did not discuss morphologi-
cal evidence to support these relationships. Cannabis, Humulus
and Pteroceltis share a distinctive S-type sieve-element plastid
(Behnke, 1989), but this character is homoplastic because these
three genera did not form a clade. Including all recognized
genera, our phylogenetic analysis indicated that the Cannabis-
Humulus clade is sister to all other genera of clade D with low
support. Additionally, the sister relationship between Celtis
and the Pteroceltis-Chaetachme-Tre ma-Parasponia clade was
implied with low statistical support in our analysis.
The Chaetachme-Pteroceltis subclade. —
The monotypic
Chaetachme is endemic to tropical Africa, southern Africa
and Madagascar, and the monotypic Pteroceltis is one of the
so-called Tertiary relict trees (Chai & al., 2010) restricted to
temperate habitats in several parts of China (Li & al., 2012).
These two genera have previously been placed in Celtidaceae
(Link, 1829; Grudzinskaya, 1967), but their phylogenetic po-
sitions were not resolved in previous studies. A phylogenetic
study employing rbcL indicated that Chaetachme and Ptero-
celtis formed a clade with low support (Ueda & al., 1997), but
the two genera were not sister in the study of Sytsma & al.
(2002). Our results show strong support for a sister relationship
persistent
deciduous
Perianth on the fruit
Ulmus macrocarpa
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Celtis sinensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
stipule arrangement
Trema tomentosa 2
Trema cannabina 2
Chaetachme aristata 2
Chaetachme aristata 1
Chaetachme aristata 3
G H
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
Humulus scandens
Humulus yunnanensis
Debregeasia saeneb
Pilea cadierei
Cudrania tricuspidata
Broussonetia papyrifera
Ficus tikoua
Aphananthe aspera
Aphananthe monoica
Aphananthe philippinensis
Gironniera celtidifolia
Gironniera subaequalis
Lozanella permollis
Lozanella trematoides
Cannabis sativa 1
Cannabis sativa 2
Humulus lupulus
Celtis madagascariensis
Pteroceltis tatarinowii 1
Pteroceltis tatarinowii 2
Trema levigata
Trema tomentosa 1
Parasponia andersonii
Parasponia melastomatifolia
Trema orientalis
Trema domingense
Trema micrantha
Trema cannabina 1
Trema discolor
Trema angustifolia
Ulmus macrocarpa
Chaetachme aristata 2
Chaetachme aristata 1
Chaetachme aristata 3
Trema tomentosa 2
Trema cannabina 2
extrapetiolar
intrapetiolar
interpetiolar
Humulus scandens
Humulus yunnanensis
Celtis sinensis
Celtis iguanaea
Celtis ehrenbergiana 2
Celtis biondii
Celtis tournefortii
Celtis ehrenbergiana 1
482
TAX ON
62 (3) • June 2013: 473–485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
482 Version of Record (identical to print version).
between Chaetachme and Pteroceltis (MPBS = 100%, MLBS
= 100%, PP = 1.0), which is consistent with Van Velzen & al.
(2006). Synapomorphies for the sister-group relationship in-
clude a 5-bp insertion at position 646 650 of rps16. However,
Pteroceltis and Chaetachme show some differences in both
morphological characters and chromosome number. Pteroceltis
has serrate leaf blades, extrapetiolar stipules, leaves 3-veined
from base, unarmed twigs and samaras, whereas Chaetachme
has entire leaf blades, interpetiolar stipules, pinnate leaf vena-
tion, spiny twigs and drupes. Furthermore, the chromosome
number of Pteroceltis is x = 10, while that of Chaetachme is x
= 15. Synapomorphies for this clade are unknown. The geo-
graphical distribution of the two genera is non-overlapping, but
Pteroceltis had a wider range during the Tertiary as indicated
by fossil occurrences in the Oligocence of Germany (Weyland,
1937) and in the Middle Eocene of Tennessee (Manchester
& al., 2009). This may imply that these two genera may have
been geographically closer to each other than they are today.
The Parasponia-Trema complex. —
Tre m a is a pantropi-
cally distributed pioneer plant (Yesson & al., 2004), and Para-
sponia is distributed in Southeast Asia and the Pacific Islands
and is the only genus of the family that is known to fix nitrogen
(B eck ing, 1983). Parasponia differs from Trema in having con-
nate intrapetiolar stipules and imbricate perianth lobes in male
flowers (Soepadmo, 1977), and the ability to fix nitrogen (Beck-
ing, 1983; Sturms & al., 2010). However, our molecular results
strongly suggested that Parasponia is nested within Tr e ma
(Fig. 1). These results are consistent with previous phylogenetic
analyses (Zavada & Kim, 1996; Sytsma & al., 2002; Yesson
& al., 2004; Van Velzen & al., 2006). Parasponia and Tre m a
shared three atpB-rbcL insertions (7-bp insertion at position
386–392, 5-bp insertion at position 524–528, 7-bp insertion at
position 942–948), one trnL-trnF insertion (6-bp insertion at
position 30–35) and one rps16 insertion (4-bp insertion at posi-
tion 44– 47). The expanded dataset also strongly supported that
Parasponia is nested within Trema (MLBS = 100%, MPBS =
98%, PP = 1.0), and five Parasponia species formed a weakly
supported clade (Electr. Suppl.: Fig S2). Furthermore, we found
that diporate pollen is a good synapomorphy for this clade
(Fig. 2C). Additionally, both genera have a basically lineate
seed coat where the exotestal cells are characteristically lon-
gitudinally elongate (Takaso & Tobe, 1990), and pollen which
lacks a granular layer in the exine (Takahashi, 1989). However,
these two characters are shared with Lozanella. Based on this
evidence, we suggested that Parasponia should be merged with
Tre m a. The taxonomy of Trema is disputed because of the diffi-
culty in finding good morphological features to delimit species
(Yesson & al., 2004). Important diagnostic characters including
leaf texture, venation, size, shape, colour and pubescence are
variable and probably subject to ecotypic and ontogenetic varia-
tion. This may be the main reason why multiple individuals
from one species did not form one clade in a previous molecular
study (Yesson & al., 2004) and our study (see Trema cannabina
and Trema tomentosa in Fig. 1).
The Cannabis-Humulus subclade.
Cannabis and Humu-
lus were strongly supported as sister genera (MLBS = 100%,
MPBS = 100%, PP = 1.0), in agreement with some previous
molecular studies that sampled only one specimen from each
genus (Song & al., 2001; Sytsma & al., 2002; Song & Li, 2002).
Many morphological characters support this clade, including
their herbaceous habit, palmately lobed or compound leaves,
achenes, and female inflorescences which are bracteate spi-
cate cymes.
Morphological character evolution in Cannabaceae. —
Most morphological characters analyzed showed a complex
evolution pattern and changed more than twice. For example,
at least twelve shifts occurred among the five states of the
sexual system. Both monoecy and dioecy occur frequently in
the same family, but rarely in the same genus (Baker, 1959).
Two or more sexual system states are present within several
genera of Cannabaceae, including Cannabis, Celtis, Chaeta-
chme, Gironniera, Humulus and Trem a . Monoecy and dioecy
even co-occur in the same species of Cannabis, Chaetachme,
Gironniera, Humulus and Trema . Further studies should be
carried out to reveal the underlying ecological and evolutionary
mechanisms of frequent mating system transitions in Canna-
baceae. Other characters such as seed coat with holes, valvate
aestivation, deciduous perianth and intrapetiolar stipules also
have changed several times in the family.
The expanded Cannabaceae include most genera of Ulm-
aceae or Celtidaceae, which makes the family morphologically
highly diverse. Our character evolution analyses showed that
some morphological characters traditionally used to delimit
generic boundaries can not be used for this purpose. For ex-
ample, unisexual or polygamous flowers, combined with other
morphological characters, have been used to separate Tre m a
and Celtis (Fu & al., 2003). In contrast to this, our analyses
indicated that both genera have complex sexual system, and
each state is homoplastic.
It is difficult to find mor phological synapomorphies for
most clades of Cannabaceae. For example, all morphologi-
cal characters shared by the Chaetachme-Pteroceltis clade are
homoplastic. However, some morphological characters such as
diporate pollen, achenes or opposite and alternate leaves have
evolved only once in the family and can be regarded as synapo-
morphies of certain clade: diporate pollen for the Parasponia-
Tre m a clade, achenes and opposite or alternate leaves for the
Cannabis-Humulus clade.
ACKNOWLEDGEMENTS
We thank Ulrike Bertram (the Ecological-Botanical Garden,
Bayreuth) and Shudong Zhang (the Kunming Institute of Botany,
Chinese Academy of Sciences, Kunming) for their help in obtain-
ing mater ial, the Botanic Garden Conservation International (BGCI)
instit ution for offering samples, and Andrew Hipp (the Morton Arbo-
retum, Chicago) for revising the manuscript. The DNA analysis was
undertaken at the Molecular Biology Experiment Center, Kunming
Instit ute of Botany. This study was supported by the National Natural
Science Foundation of China (g rant no. 40830209 and no. 31270274),
the Knowledge Innovation Program of the Chinese Academy of Sci-
ences (grant no. KSCX2-YW-R-136) and the Talent Project of Yunnan
Province (project no. 2011CI042).
483
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • June 2013: 473–485
483Version of Record (identical to print version).
LITERATURE CITED
Alaniya, M., Kavtaradze, N.S., Skhirtladze, A., Pizza, C. & Pia-
cente, S. 2010. Flavonol glycoside from Humulus lupulus. Chem.
Nat. Compounds 46: 641–642.
ht tp://dx .d oi.org /10.100 7/s1060 0-010- 9699 -x
Alfaro, M.E., Zoller, S. & Lutzoni, F. 2003. Bayes or bootstrap? A
simulation study comparing the perfor mance of Bayesian Markov
chain Monte Carlo sampling and bootstrapping in assessing phy-
logenetic confidence. Molec. Biol. Evol. 20: 255–266.
http://dx.doi.org/10.1093/molbev/msg028
Baker, H. 1959. Reproductive methods as factors in speciation in f low-
ering plants. Cold Spring Harbor Symp. Quant. Biol. 2 4: 177–191.
ht tp://dx .d oi.org /10.1101/SQB.19 59.0 24.01.019
Becking, J. 1983. The Parasponia parviflora-Rhizobium symbiosis:
Host specif icity, growth and nitrogen fixation under various condi-
tions. Pl. & Soil 75: 309–342.
ht tp://dx .d oi.org /10.100 7/BF023 69969
Behnke, H.D. 1989. Sieve-element plastids, phloem proteins, and the
evolution of flowering plants. IV. Hamamelidae. Pp. 105–128 in:
Crane, P.R. & Blackmore, S. (eds.), Evolution, systematics, and
fossil histor y of the Hamamelidae, vol. 2, Lower Hamamelidae.
Oxford: Clarendon Press.
Bollback, J. 2006. SIMM AP: Stochastic character mapping of discrete
traits on phylogenies. B. M. C. Bioinf. 7: 88.
ht tp://dx .d oi.org /10.1186/1471-2105 -7-88
Brown, J.M. & Lemmon, A.R. 2007. The importance of data partition-
ing and the utility of Bayes factors in Bayesian phylogenetics. Syst.
Biol. 56: 643655. ht tp://dx .d oi .org /10.1080/10 63515 070154 6249
Chai, X.Y., Chen, S.L. & Xu, W. 2010. Using inter-simple sequence re-
peat markers to analyze the genetic structure of natural Pteroceltis
tatarinowii populations and implications for species conservation.
Pl. Syst. Evol. 285: 65–73.
http://dx.doi.org/10.1007/s00606-009-0256-7
Chase, M.W., Soltis, D.E., Olmstead, R.G., Morgan, D., Les, D.H.,
Mishler, B.D., Duvall, M.R., Price, R.A., Hills, H.G., Qiu, Y.L.,
Kron, K.A., Rettig, J.H., Conti, E., Palmer, J.D., Manhart,
J.R., Sy tsma, K.J., Michaels, H.J., Kress, W.J., Karol, K.G.,
Clark, W.D., Hedren, M., Gaut, B.S., Jansen, R.K., Kim, K. J.,
Wimpee, C.F., Smith, J.F., Furnier, G.R., Strauss, S.H., Xiang,
Q.Y., Plunkett, G.M., Soltis, P.S., Swensen, S.M., Williams,
S.E., Gadek, P.A., Quinn, C.J., Eguiarte, L.E., Golenberg, E.,
Learn, G.H., Graham, S.W., Barrett, S.C.H., Dayanandan, S.
& Albert , V.A. 1993. Phylogenetics of seed plants: An analysis of
nucleotide sequences from the plastid gene rbcL. Ann. Missouri
Bot. Gard. 80: 528–580. http://dx.doi.org/10.2307/2399846
Chiang, T.Y., Schaal, B.A. & Peng, C.I. 1998. Universal primers for
amplification and sequencing a noncoding spacer bet ween the
atpB and rbcL genes of chloroplast DNA. Bot. Bull. Acad. Sin.
39: 245 –250.
Clark, M.N. 1978. A study of infraspecific f lavonoid variation of Can-
nabis sativa L. (Cannabaceae). Dissertation, University of British
Columbia, Va n cou ver, Canada.
Doyle, J. & Doyle, J. 1987. A rapid DNA isolation procedu re for small
quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. Amer.
19: 11 –15.
Engler, A. & Prantl, K. 189 3. Die natürlichen Pflanzenfamilien, vol. 3.
Berlin: Bor nträger.
Felsenstein, J. 1985. Confidence limits on phylogenies: An approach
using the bootstrap. Evolution 39: 783 –791.
http://dx.doi.org/10.2307/2408678
Fu, L., X in, Y. & Whittemore, A. 2003. Ulmaceae. Pp. 1–19 in: Wu,
Z.Y., Raven, P.H. & Hong, D.Y. (eds.), Flora of China, vol. 5. Bei-
jing: Science Press; St. Louis: Missouri Botanical Garden Press.
Giannasi, D.E. 1978. Generic relationships in Ulmaceae based on fla-
vonoid chemistry. Tax o n 27: 331–344.
htt p://dx.doi.org/10.2307/1220369
Grudzinskaya, I.A. 1967. Ulmaceae and reasons for distinguishing
Celtidoideae as a separate family Celtidaceae Link. Bot. Zhurn.
S.S.S.R. 52: 172–349.
Huelsenbeck, J.P., Nielsen, R. & Bollback, J.P. 2003. Stochastic map-
ping of morphological characters. Syst. Biol. 52: 131–158.
ht tp://dx .d oi.org /10.1080/10 63515 039 019278 0
Judd, W.S., Campbell, C.S., Kellogg, E.A., Stevens, P.F. & Dono-
ghue, M.J. 2008. Plant systematics: A phylogenetic approach, 3rd
ed. Sunderland, Massachusetts: Sinauer.
Killip, E.P. & Morton, C.V. 1931. The genus Lozanella. J. Washington
Acad. Sci. 21: 336 –339.
Kuprianova, A. 1962 . Palynological data and the systematics of the
Fagales and Urticales. Pp. 17–25 in: Sladkov, A.N. (ed.), Reports
of Soviet Palinologists for the First International Conference on
Palinology (Tucson, USA). Moskow: Publishing House of USSR
Academy of Sciences..
Li, X.H., Shao, J.W., Lu, C., Zhang, X.P. & Qiu, Y.X. 2012 . C hloro -
plast phylogeography of a temperate tree Pteroceltis tatarinowii
(Ulmaceae) in China. J. Syst. Evol. 50: 325–333.
http://dx.doi.org/10.1111/ j.1759-68 31. 2 01 2.0 0 2 03.x
Link, H.F. 1829. Handbuch zur Erkennung der nutzbarsten und am
häufigsten vorkommenden Gewächse, vol. 2, Berlin: in der Haude
und Spenerschen Buchhandlung.
Mabberley, D.J. 2008. Mabberley’s plant-book: A portable diction-
ary of plants, their classification and uses. New York: Cambridge
University Press.
Maddison, W.P. & Maddison, D.R. 2010. Mesquite: A modular sys-
tem for evolutionar y analysis, version 2.74. University of British
Columbia & Oregon State University. http://mesquiteproject.org
Manchester, S.R., Chen, Z.D., Lu, A.M. & Uemura, K. 2009. Eastern
Asian endemic seed plant genera and their paleogeog raphic his-
tory throughout the Northern Hemisphere. J. Syst. Evol. 47: 1– 42.
http://dx.doi.org/10.1111/j.1759-6831.2009.00001.x
Oginuma, K., Raven, P.H. & Tobe, H. 1990. Karyomor phology and
relationships of Celtidaceae and Ulmaceae (Urticales). Bot. Mag.
(To k y o) 10 3: 113 –131. ht tp://dx .d oi.org /10.100 7/BF02489 620
Omori, Y. & Terabayashi, S. 1993. Gynoecial vascular anatomy and its
systematic implications in Celtidaceae and Ulmaceae (Urticales).
J. Pl. Res. 106: 249–258. ht t p://dx.doi .o rg /10.100 7/ BF0234 4593
Oxelman, B., Liden, M. & Berglund, D. 1997. Chloroplast r ps16 intron
phylogeny of the tribe Sileneae (Caryophyllaceae). Pl. Syst. Evol.
206: 393 410. ht tp://dx .d oi.org /10.100 7/BF00 987959
Posada, D. & Crandall, K.A. 1998. MODELTEST: Testing the model
of DNA substitution. B. M. C. Bioinf. 14: 817– 818.
http://dx.doi.org/10.1093/bioinformatics/14.9.817
R Development Core Team 2011. R: A language and envi ronment for
statistical computing. R Foundation for Statistical Computing, Vi-
enna, Austria. ISBN 3-900051-07-0, UR L http://www.R-project.org/
Rambaut, A. 2002. Sequence alignment editor, version 2.0a11. Oxford:
University of Oxford. htt p://evolve.zoo.ac.uk
Rambaut, A. & Drummond, A. 2004. Tracer, version 1.3. Edinburgh:
University of Edinburgh.
Rendle, A.B. 1925. The classification of f lowering plants, vol. 2, Lon-
don: Cambridge University Press.
Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylo-
genetic inference under mixed models. B. M. C. Bioinf. 19: 1572–
1574. http://dx.doi.org/10.1093/bioinformatics/btg180
Sattarian, A. & Van der Maesen, L.J.G. 2006. Endocarp morpholog y
of African Celtis (Celtidaceae/Ulmaceae). Blumea 51: 389–397.
http://dx.doi.org/10.3767/000651906X622337
Sattarian, A., Van den Berg, R.G. & Van der Maesen, L.J.G. 2006.
Pollen morphology of African Celtis (Celtidaceae). Feddes Repert.
117: 34 40. h ttp://d x .doi.org /10.10 02/fedr.200 611087
Shimodaira, H. 2002. An approximately unbiased test of phylogenetic
tree selection. Syst. Biol. 51: 492–508.
ht tp://dx .d oi.org /10.1080/10 63515 0290 069913
Shimodaira, H. & Hasegawa, M. 2001. CONSEL: For assessing the
484
TAX ON
62 (3) • Ju ne 2013: 473–485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
484 Version of Record (identical to print version).
confidence of phylogenetic tree selection. B. M. C. Bioinf. 17:
1246 –1247. http://dx.doi.org /10.1093/bioin for matics/17.12.1246
Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in se-
quence-based phylogenetic analyses. Syst. Biol. 49: 3 69 –381.
http://dx.doi.org/10.1093/sysbio/49.2.369
Soepadmo, E. 1977. Ulmaceae. Pp. 31–76 in: Van Steenis, C.G.G.J.
(ed .) , Flora Malesiana, ser. 1, vol. 8. Alphen aan den Rijn: Sijthoff
& Noordhoff.
Song, B. & Li, F.Z. 2002. The utility of trnK int ron 5′ reg ion in phylo -
genetic analysis of Ulmaceae s. l. Acta Phytotax. Sin. 40: 125–132.
Song, B.H., Wang, X.Q., Li, F.Z. & Hong, D.Y. 2001. Further evidence
for paraphyly of the Celtidaceae f rom the chloroplast gene matK.
Pl. Syst. Evol. 228: 107–115.
ht tp://dx .d oi.org /10.100 7/s00 606 0170041
Sturms, R., Kakar, S., Trent, J., III & Hargrove, M.S. 2010. Tre ma
and Parasponia hemoglobins reveal convergent evolution of oxy-
gen transport in plants. Biochemistr y 49: 4085–4093.
ht tp://dx .d oi.org /10.1021/ bi100284 4
Swofford, D.L. 2003. PAUP*: Phylogenetic analysis using parsimony
(*and other methods), version 4.0b10. Sunderland, Massachusetts:
Sinauer.
Sytsma, K.J., Morawetz, J., Pires, J.C., Nepokroeff, M., Conti, E.,
Zjhra, M., Hall, J.C. & Chase, M.W. 2002. Urticalean rosids:
Circumscription, rosid ancest ry, and phylogenetics based on rbcL,
trnL-trnF, and ndhF sequences. Amer. J. Bot. 89: 1531–1546.
http://dx.doi.org/10.3732 /ajb.89.9.1531
Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. 1991. Universal prim-
ers for amplification of three non-coding regions of chloroplast
DNA. Pl. Molec. Biol. 17: 110 5 110 9.
http://dx.doi.org/10.1007/BF00037152
Takahashi, M. 1989. Pollen morphology of Celtidaceae and Ulmaceae:
A reinvestigation. Pp. 253–265 in: Crane, P.R. & Blackmore, S.
(ed s.), Evolution, systematics, and fossil histor y of the Hamamel-
idae, vol. 2, Higher Hamamelidae. Oxford: Clarendon Press.
Takaso, T. 1987. Ovule ontogeny and mor phology in Ulmaceae. Pp.
220 in: Greuter, W., Zimmer, B. & Behnke, H.D. (eds.), Abstracts,
XIV International Botanical Congress. Königstein: Koeltz Scien-
tific Books.
Takaso, T. & Tobe, H. 1990. Seed coat mor phology and evolution
in Celtidaceae and Ulmaceae (Urticales). Bot. Mag. (Tokyo) 103:
25–41. http://dx.doi.org/10.1007/BF02488408
Terabayashi, S. 1991. Vernation patterns in Celtidaceae and Ulmaceae
(Urticales), and their evolutionary and systematic implications.
Bot. Mag. (Tokyo) 104: 1–13.
http://dx.doi.org/10.1007/BF02493399
The Plant List 2 010. Version 1. Published on the Internet; http://www
.theplantlist.org/ (accessed 20 Mar. 2013).
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Hig-
gins, D.G. 1997. The CLUSTAL_X windows interface: Flexible
strategies for multiple sequence alignment aided by quality analysis
tools. Nucl. Acids Res. 25: 4876–4882.
http://dx.doi.org/10.1093/nar/25.24.4876
Todzia, C.A. 1993. Ulmaceae. Pp. 603–611 in: Kubitzki, K., Rohwer,
J.G. & Bittrich, V. (eds.), The families and genera of vascular
plants, vol. 2. Berlin: Spr inger.
Ueda, K., Kosuge, K. & Tobe, H. 1997. A molecular phylogeny of
Celtidaceae and Ulmaceae (Urticales) based on rbcL nucleotide
sequences. J. Pl. Res. 110: 17 1–178 .
http://dx.doi.org/10.1007/ BF0250930 5
Van Velzen, R., Bakker, F.T., Sattarian, A. & Van der Maesen,
L.J.G. 2006. Evolutionary relationships of Celtidaceae. Pp.
7–30 in: Sattarian, A. (ed.), Contribution to the biosystematics
of Celtis L. (Celtidaceae) with special emphasis on the African
species. Disser tation, Wageningen University, Wageningen, the
Netherlands.
Wang, H.C., Moore, M. J., Soltis, P.S., Bell, C.D., Brockington, S.F.,
Alexandre, R., Davis, C.C., Latvis, M., Manchester, S.R. &
Soltis, D.E. 2009. Rosid radiation and the rapid rise of angiosperm-
dominated forests. Proc. Natl. Acad. Sci. U.S.A. 106: 3853–3858.
http://dx.doi.org/10.1073/p nas.081337610 6
Weyland, H. 1937. Beit räge zur Kenntnis der rheinischen Tertiärf lora.
II. Erste Ergänzungen und Berichtigungen zur Flora der Blätter-
kohle und des Polierschiefers von Rott im Siebengebirge. Palaeon-
tographica, Abt. B, Paläophytol. 83: 67–122.
Wiegrefe, S.J., Sytsma, K.J. & Guries, R.P. 1998. The Ulmaceae,
one family or two? Evidence from chloroplast DNA restriction
site mapping. Pl. Syst. Evol. 210: 249–270.
ht tp://dx .d oi.org /10.100 7/BF00 985671
Yesson, C., Russell, S.J., Parrish, T., Dalling, J.W. & Garwood, N.C.
2004. Phylogenetic framework for Tr ema (Celtidaceae). Pl. Syst.
Evol. 24 8: 85–109. ht t p://dx.doi.o rg /10 .1007/s 006 06 - 0 04- 0186-3
Zavada, M.S. & Dilcher, D.L. 1986. Comparative pollen morphology
and its relationship to phylogeny of pollen in the Hamamelidae.
Ann. Missouri Bot. Gard. 73: 348–381.
http://dx.doi.org/10.2307/2399117
Zavada, M.S. & Kim, M. 1996. Phylogenetic analysis of Ulmaceae.
Pl. Syst. Evol. 200: 13–20. http://dx.doi.org/10.1007/BF00984745
Zhang, S., Soltis, D.E., Yang, Y., Li, D. & Yi, T. 2011. Multi-gene anal-
ysis provides a well-supported phylogeny of Rosales. Molec. Phylo-
gen. Evol. 60: 21–28. ht tp://dx.doi.o rg/10.1016/j.ympev.2011.0 4.00 8
Zhou, Z. & Bartholomew, B. 2003. Cannabaceae. Pp. 74–75 in: Wu,
Z.Y., Raven & P.H., Hong, D.Y. (eds.), Flora of China, vol. 5. Bei-
jing: Science Press; St. Louis: Missouri Botanical Garden Press.
Zwickl, D.J. 2006. Genetic algorithm approaches for the phylogenetic
analysis of large biological sequence datasets under the maximum
likelihood criterion. Dissertation, The University of Texas at Aus-
tin, Texas, U.S.A.
Appendix 1.
List of taxa st udied: Taxon, locality, voucher specimen a nd GenBank accession numbers for rbcL, atpB-rbcL, trnL-trnF and rp s16. A dash (–)
indicates an unavailable sequence.
OUTGROUP TAXA: Broussonetia papyrifera Vent. (Mora ceae), China, Shaanxi; Yi & Zhang 080645 (KU N); JF317438; JN040322; JN040358; JN040284.
Cudrania tricuspidata Bureau ex Lavallée (Moraceae), China, Yunnan; Zhang 090045 (KUN); J F317440; JN040332; JN040371; JN040296. Debregeasia
saeneb (Forssk.) Hepper & J.R.I. Wood (Urticaceae), China, Yunnan; Yi 0 9052 (KUN); JF317441; JN040333; JN040372; JN040297. Ficus tikoua Bureau
(Moraceae), China, Hunan; Yi & Zhang 080142 (KUN); JF317445; JN040334; JN040373; JN040298. Pilea cadierei Gagnep. & Guillaumin (Urticaceae), China,
Yun nan ; Yi 0 9051 (KUN); JF317451; JN040342; J N040383; JN040307. Ulmus macrocarpa Hance (Ulmaceae), China, Shandong; Yi & Zhang 080312 (KUN);
JF317455; JN040354; JN0 40396; J N040320. INGROUP TAXA: Aphananthe aspera (Thunb.) Planch., China, Hubei; Sun 10001 (KU N); JN040397; JN0 40321;
JN040355; J N040281. Aphananthe monoica ( Hemsl.) J.-F. Leroy, Mexico; Fernández 3462 ( U); JN040398; – ; JN040356; JN0 40282. Aphananthe philippinensis
Planch., Australia, Queensland; Fo rs te r 6 657 (L); JN040399; –; JN040357; JN040283. Cannabis sativa L. 1, China, Yunnan; Yan g 0 03 (KUN); J N040400;
JN040323; JN040359; JN040285. Cannabis sativa L., China, Yunnan; Yi 10108 (KUN); JN040401; JN040324; JN040360; JN040286. Celtis biondii Pamp. 2,
China , Yunnan; Yi 1033 0 (KUN); JN040404; JN040327; JN040363; JN040289. Celtis ehrenbergiana (K lotzsch) Liebm. 1, Netherlands, Gelderland; Leeuw
481 (WAG); JN040406; JN040329; JN040365; JN040291. Celtis ehrenbergiana (Klotzsch) Liebm. 2, Belgium, Vlaams-Brabant; coll. unknown s.n. 19371993
(Meise Bot anical Garden); JN0 40408; JN040331; JN040367; JN040293. Celtis iguanaea (Jacq.) Sarg., Germa ny, Bayer n; coll. unknown s.n. (KU N); J N040402;
JN040325; JN040361; JN040287. Celtis madagascariensis Satta rian, Madagascar, Toliara; Phillipson 2938 (MO); JN040405; J N040328; JN040364; JN040290.
Celtis sinensis Pers., Finland , Helsink i; coll. unknow n s.n. (Helsinki Bot anical Garden); JN040407; JN040330; JN040366; JN040292 . Celtis tournefortii Lam.,
Italy, Catania; coll. un known s.n. (Catania Botanical Garden); JN040403; JN040326; JN040362; JN040288. Chaetachme aristata Planch. 1, Madagascar,
485
Yang & al.
• Phylogenetics and character evolution of Cannabaceae
TAX ON
62 (3) • Ju ne 2013: 473–485
485Version of Record (identical to print version).
Tol iar a; M cPh ers on 14 423 (WAG); JN040410; –; JN040369; JN040295. Chaetachme aristata Planch. 2, South Africa, Transvaal; Schij ff 3444 (L); JN040409;
–; JN040368; JN040294. Chaetachme aristata Planch. 3, Congo, Orient ale; Bamps 4325 (WAG); JN040411; –; JN040370; –. Gironniera celtidifolia Gaudich.,
Papua New Gui nea, Hans Meyer Range; Sa nd s 855 (L);
JN040412; JN040335; JN040374; JN040299. Gironniera subaequalis Planch., Chi na, Yunnan; DNA
Barcoding Group B GBOWS 1411 ( KUN); JN040 413; J N040336; JN0 40375; –. Humulus lupulus L., China, Beijing; Xie 002 (KUN); J N040416; JN040338;
JN040378; JN040302. Humulus scandens (Lou r.) Merr., U.S.A., Ma ryland; Windler 4046 (U); J N040415; –; J N040377; JN040301. Humulus yunnanensis Hu,
China , Yunnan; Yang 005 (KU N); JN040414; JN040337; JN040376; JN040300. Lozanella permollis Killip & C.V. Morton, Bolivia, Cochabamba; Solomon
18073 (U ); J N040417; JN040339; JN040379; JN040303. Lozanella trematoides Gre enm., Mexico, Hidalgo; Pringle 8983 (L); JN040418; JN040340; JN04 0380;
JN040304. Parasponia andersonii Planch., Polyne sia; Mey er 255 6 (L); JN040419; JN040341; JN040381; JN040305. Parasponia melastomatifolia J. J. Sm., Papua
New Guinea, Milne Bay; Pul len 7963 (L); JN040420; –; JN040382; JN040306. Pteroceltis tatarinowii Maxim. 1, China, Beijing; Xie 003 ( KUN); JN040421;
JN040343; JN040384; JN040308. Pteroceltis tatarinowii Maxim. 2, China, Guizhou; Y i 10 081 (KUN); JN040422; JN040344; JN040385; JN040309. Trema
angustifolia (Planch.) Blume, China, Guangxi; Gong zw2009082601 (KUN); JN0 40425; JN040347; JN040388; JN040312. Trema cannabina Lou r. 1, Hawaii,
Papaikou; Lorence 9381 (Hawaii Botanical Ga rden); JN040426; JN040348; JN040389; JN040313. Trema cannabina Lou r. 2, China, Guangxi; Jiang & Yang
09525 (KU N); JN040424; JN040346; JN040387; JN040311. Trema discolor Blume, Hawaii, Papaikou; Lorence 9329 (Hawaii Botanical Garden); JN040427;
JN040349; JN040390; JN040314. Trema domingense Urb., Dominican Republic, Duarte; Ek ma n 12293 (U); J N040428; J N040350; JN040391; J N040315.
Tre ma le v ig ata Hand.-Mazz., China, Yunnan; Tian 0020 (KUN); J N040429; JN040351; JN040392; JN040316. Trema micrantha (L.) Blume, Bolivia, Beni;
Chatrou 413 ( U); JN040430; JN0 40352; JN040393; JN040317. Trema orientalis (L.) Blume, Gabon; coll. unknown s.n.; JN040431; J N040353; JN040394;
JN040318. Trema tomentosa (Roxb.) H. Hara 1, China, Guangxi; Liu 0 096 (KU N); JN040432; –; JN040395; JN040319. Trema tomentosa (Roxb.) H. Hara 2,
China, Guangxi; Jiang & Yang 09524 (KUN); JN040423; JN040345; JN040386; JN040310.
Appendix 1.
Continued.
Appendix 2.
Accession numbers of additional sequences from GenBank in the order trnLtrnF, rbcL, ndhF, matK. A d ash (–) indicates an unavailable
sequence.
Aphananthe aspera –; –; AF500366; AF345320; Broussonetia papyrifera –; –; AY289269; AF345326; Cannabis sativa 1 –; –; AY289250; AF345317; Can-
nabis sativa 2 –; –; AY289250; –; Celtis iguanaea –; –; –; JQ589979; Celtis africana –; –; –; JF270686; Celtis australis –; H E963395; –; HE967374; Celtis
latifolia –; JF738634; –; –; Celtis occidentalis –; –; –; AY257535; Celti s schippii –; JX987578; –; G Q981961; Celtis sinensis –; –; –; AF345316; Celt is tetrandra
–; JF317439; JF317479; JF317420; Celtis tournefortii 2 A J575060; –; –; –; Chaetachme aristata –; –; –; JF270688; Cudrania tricuspidata –; –; AY289272;
JF 317421; Debregeasia saeneb –; –; –; JF317422; Ficus tikoua –; –; –; JF317426; Gironniera subaequalis –; –; –; AF345319; Humulus scandens –; –; –;
JQ773628; Humulus lupulus –; –; AY289251; AF345318; Lozanella enantiophylla AF501595; AF500341; AF500367; –; Parasponia parviflora AF50159 6;
AF500342; AF500368; –; Paras ponia rigida AY488675; U59820; –; –; Parasponia simulans AY488674; –; –; –; Pilea cadierei –; –; –; JF317431; Pteroceltis
tatarinowii –; –; AF500369; AF345324; Trema micrantha –; –; –; JQ589372; Trema orientali s –; –; –; JF270972; Trema angustifolia –; –; –; JF317434; Tr em a
aspera AY488681; –; –; –; Trema integerrima AY488718; –; –; –; Trema lamarckiana AY488698; –; –; –; Trema politor ia AY488676 ; –; –; –; Trema tomen tosa
–; –; –; AF345325; Ulmus americana –; –; AF500365; –; Ulmus parvifolia –; –; –; AF345321.
Appendix 3.
Morphological data matrix.
Taxon/Character
Sexual system
Leaf arrangement
Pollen aperture
number
Aestivation
Fruit type
Seed coat
morphology
Perianth at
fruiting time
Stipule arrange-
ment
Aphananthe aspera 01000111
Aphananthe monoica 01000111
Aphananthe philippinensis 01000111
Broussonetia papyrifera 11101111
Cannabis sativa 1 32011011
Cannabis sativa 2 32011011
Celtis biondii 41010001
Celtis ehrenbergiana 1 21010001
Celtis ehrenbergiana 2 21010001
Celtis iguanaea 21010001
Celtis madagascariensis 21010001
Celtis sinensis 41010001
Celtis tournefortii 01010001
Chaetachme aristata 1 31000000
Chaetachme aristata 2 31000000
Chaetachme aristata 3 31000000
Cudrania tricuspidata 1 1 0&1 1 1 0 1 1
Debregeasia saeneb 1 1 ? 0 1 ? 1 0
Ficus tikoua 011011?1
Gironniera celtidifolia 31010111
Gironniera subaequalis 31010111
Appendix 3.
Continued.
Taxon/Character
Sexual system
Leaf arrangement
Pollen aperture
number
Aestivation
Fruit type
Seed coat
morphology
Perianth at
fruiting time
Stipule arrange-
ment
Humulus lupulus 32011112
Humulus scandens 32011112
Humulus yunnanensis 32011112
Lozanella permollis 10010100
Lozanella trematoides 10010100
Parasponia andersonii 01110110
Parasponia melastomatifolia 01110110
Pilea cadierei 10101?10
Pteroceltis tatarinowii 1 01012011
Pteroceltis tatarinowii 2 01012011
Trema angustifolia 31100111
Trema cannabina 1 41100111
Trema cannabina 2 41100111
Trema discolor 01100111
Trema domingense 01100111
Trema levigata 01100101
Trema micrantha 41100111
Trema orientalis 01100111
Trema tomentosa 1 01100111
Trema tomentosa 2 01100111
Ulmus macrocarpa 01212111
... [8] and Trema Lour. [9] [10,11,12,13,14,15] and ca. 117 species [16]. ...
... Celtis, the largest genus of the family, is represented with about 73 species, with a large number of synonyms, distributing themselves in the tropical and temperate regions of the world [18,14]. The genus can be recognized as trees to shrubs, monoic or polygamous, armed with spines or inermous; leaves alternate and distichous, subtriplinerved; domatia present mostly in the axils and or furcations of secondary veins; stipules free, caduceus; inflorescences solitary or in pairs, bisexual or unisexual, deciduous bracteate; staminate flower often in glomerate; bisexual flowers solitary often uniflorous, 5-stamens; stigmas bifurcate or 2-lobed; drupes with endocarp calcareous [18]. ...
... Celtis does not have accessions yet as from Neotropical region to the published phylogenies [19,20,10,11,21,14,16]. Yang [14] includes three accessions of two species occurring in the Neotropical region of Celtis (C. ...
Thesis
Full-text available
Master's thesis. Celtis L. (Cannabaceae) from Brazil. Review of Celtis species from Brazil.
... La famille des Cannabaceae comporte 10 genres et plus de 110 espèces (Zhang et al., 2018). Yang et al. (2013) ont étudié l'évolution de différents traits d'histoire de vie au sein de cette famille, notamment les différents systèmes sexuels ( Figure 4). Parmi les 42 feuilles utilisées pour leur phylogénie, 16 (38%) sont monoïques strictes, 6 (14%) sont dioïques strictes, 4 (10%) sont andromonoïques, 11 (26%) sont monoïques et dioïques, et enfin 5 (12%) sont hermaphrodites Yang et al. (2013). ...
... Yang et al. (2013) ont étudié l'évolution de différents traits d'histoire de vie au sein de cette famille, notamment les différents systèmes sexuels ( Figure 4). Parmi les 42 feuilles utilisées pour leur phylogénie, 16 (38%) sont monoïques strictes, 6 (14%) sont dioïques strictes, 4 (10%) sont andromonoïques, 11 (26%) sont monoïques et dioïques, et enfin 5 (12%) sont hermaphrodites Yang et al. (2013). Les auteurs concluent que la monoécie est l'état ancestral de la famille des Cannabaceae le plus probable. ...
... De plus, bien que tous les systèmes sexuels ne soient pas connus au sein des Cannabaceae, il y a probablement une surestimation de la fréquence des Cannabaceae dioïques dans leur échantillonage (2/5 soit 40%) (Zhang et al., 2018). Il n'y a donc toujours pas de consensus sur le système sexuel de l'ancêtre commun des Cannabaceae mais il n'est pas impossible que ce dernier fut dioïque (Yang et al., 2013;Jin et al., 2020). En revanche, il est probable que l'ancêtre commun des genres Humulus et Cannabis était dioïque (Yang et al., 2013). ...
Thesis
La trajectoire décrivant l’évolution d’une paire de chromosomes sexuels à longtemps été proposée comme étant universelle pour tous les systèmes, cependant des propositions alternatives ont récemment nuancé ce «modèle» unique. D’après ce modèle, il y aurait dans un premier temps l’émergence d’une région non-recombinante (XY ou ZW), puis, une expansion de celle-ci. Simultanément, l’absence de recombinaison induit ce que l’on appelle la dégénérescence du chromosome Y (ou W). La dégénérescence est supposée augmenter et, après un certain temps évolutif, devrait conduire à un système dans lequel le chromosome Y (ou W) serait plus petit que le chromosome X (ou Z), voire disparaît. Cependant, seulement une trentaine de paires de chromosomes sexuels de plantes ont été étudiées avec des données empiriques, parmi plus de 15 000 espèces dioïques (i.e. plantes à sexes séparés). Il en résulte que certaines étapes du systèmes sont mieux supportées que d’autres. Plus précisément, la formation de la région non-recombinante a essentiellement été étudiée de manière théorique, tandis qu’une forte dégénérescence avec un chromosome Y (ou W) plus petit que le chromosome X (ou Z) n’a été décrite que chez les animaux. Afin de mieux décrire la première étape du modèle, l’émergence de la région non recombinante, le premier axe de cette thèse représente une étude de Silene acaulis ssp exscapa, la seule sous-espèce dioïque du complexe Silene acaulis. En effet, ceci laisse supposer que ce système sexuel est un caractère dérivé, donc probablement récent. Le mécanisme du déterminisme du sexe n’étant pas connu, j’ai voulu savoir si une région non-recombinante typique d’une paire de chromosomes sexuels est présente chez cette sous-espèce. Pour cela, j’ai utilisé un outil récemment publié basé sur l’analyse de fréquences génotypiques et phénotypiques de mâles et de femelles au sein d’une population. Deux jeux de données RNA-seq provenant de deux populations différentes ont permis d’identifier 27 gènes potentiellement XY, et suggèrent que la paire de chromosomes sexuels serait récente. Des analyses complémentaires sont tout de même nécessaires pour confirmer ces résultats. Deuxièmement, afin de tester l’existence d’une paire ancienne de chromosomes sexuels avec une forte dégénérescence chez les plantes, le deuxième axe de cette thèse est une étude de deux espèces dioïques de la famille des Cannabaceae, Cannabis sativa et Humulus lupulus. En effet, l’ancêtre commun de ces deux espèces, qui divergent depuis plusieurs dizaines de millions d’années, était probablement dioïque. De plus, des analyses cytologiques ont identifié des paires de chromosomes sexuels qui pourraient être anciennes. Pour caractériser l’âge et le niveau de dégénérescence de ces paires de chromosomes sexuels, des données RNA-seq d’un croisement ont été générées pour chacune des deux espèces. Un outil probabiliste analysant les ségrégations alléliques au sein d’un croisement a permis d’identifier la première paire de chromosomes sexuels homologue entre deux genres chez les plantes. De plus, ces chromosomes sexuels sont parmi les plus vieux et les plus dégénérés actuellement décrits chez les plantes. Par ailleurs, la détection de séquence Y-spécifiques pourrait permettre d’améliorer la culture de ces deux espèces puisque seules les femelles ont un intérêt économique et que le dimorphisme sexuel est faible. J’ai développé des amorces PCR qui montrent des résultats prometteurs. Plus généralement, ces résultats apportent de nouvelles informations concernant les étapes les moins bien décrites de l’évolution des chromosomes sexuels chez les plantes. Premièrement, nous montrons qu’une paire de chromosomes sexuels a probablement émergé récemment dans une espèce, et confirmons l’intérêt de continuer à l’étudier. Deuxièmement, nous confirmons que des chromosomes sexuels vieux et fortement dégénérés existent chez les plantes.
... Cannabaceae was created as a separate family in 1925, by dividing the Moraceae family. It originally included only two genera, Cannabis and Humulus, but later it was significantly expanded to include most members of the Celtidaceae family (Yang et al. 2013). Most Cannabaceae plants have a tree-like or shrubby form, except for the genera Cannabis (herbs) and Humulus (lianas). ...
... Most Cannabaceae plants have a tree-like or shrubby form, except for the genera Cannabis (herbs) and Humulus (lianas). The family is widespread world-wide, the genera Aphanthe, Celtis (the largest genus of this family with more than 80 species), and Trema are abundant, especially in tropical and subtropical climatic regions (Yang et al. 2013). In addition to hemp, the mainly representatives with economic value are hop (Humulus lupulus), which is used for the preparation of beer, and the plant Pterocelis tatarinowii, phloem fibers of which are used for the production of traditional Chinese Xuan paper (Cao 1993). ...
Article
Full-text available
Cannabis spp. are some of the most controversial medicinal plants in the world. They contain great amounts of biologically active secondary metabolites, including the typical phenolic compounds called cannabinoids. Because of their low toxicity and complex biological activities, cannabinoids can be useful in the therapy of various diseases, but adverse psychological effects (of Δ9-THC in particular) raise concerns. This review summarizes the current knowledge of selected active C. indica compounds and their therapeutic potential. We summarize the main compounds contained in cannabis, the mechanisms of their effects, and their potential therapeutic applications. Further, we mention some of the clinical tests used to evaluate the efficacy of cannabinoids in therapy.
... Differences in the morphology and molecular phylogeny occurred in asexual fungal pathogen [84], phyllostomid bats [85], marsupials [86], Hedyosmum (Chloranthaceae) [4], Cannabaceae [87], Phyllanthus sensu lato (Phyllanthaceae) [88], Alangiaceae [89], Acorus (Acoraceae) [90], Restionaceae, Anarthriaceae, and Centrolepidaceae [5]. Traditional classifications of Apiaceae have relied almost exclusively on fruit characters such as fruit shape, the degree and direction of mericarp compression, modifications of the pericarp ribs (e.g., wings or spines), and the shape of mericarp commissural faces [7,11]. ...
Article
Full-text available
Traditional classification based on morphological characters suggests that the genus Ostericum is closely related to Angelica, but molecular phylogenetic studies suggest that the genus Ostericum is related to Pternopetalum rather than Angelica. In this study, the plastomes of nine Ostericum species and five Angelica species were used to conduct bioinformatic and comparative analyses. The plastomes of Ostericum and Angelica exhibited significant differences in genome size, gene numbers, IR junctions, nucleotide diversity, divergent regions, and the repeat units of SSR types. In contrast, Ostericum is more similar to Pternopetalum rather than Angelica in comparative genomics analyses. In total, 80 protein-coding genes from 97 complete plastomes and 112 ITS sequences were used to reconstruct phylogenetic trees. Phylogenies showed that Angelica was mainly located in Selineae tribe while Ostericum was a sister to Pternopetalum and occurred in the Acronema clade. However, morphological analysis was inconsistent with molecular phylogenetic analysis: Angelica and Ostericum have similar fruit morphological characteristics while the fruits of Ostericum are quite different from the genus Pternopetalum. The phylogenetic relationship between Angelica and Ostericum is consistent with the results of plastome comparisons but discordant with morphological characters. The cause of this phenomenon may be convergent morphology and incomplete lineage sorting (ILS).
... The genera Cannabis L. and Humulus L. belong to the Cannabaceae family in which eight more genera (Celtis, Pteroceltis, Aphananthe, Chaetachme, Gironniera, Lozanella, Trema, and Parasponia spp.) have recently been included based on phylogenetic studies [1,2]. ...
Article
Full-text available
In this study, essential oils (EOs) and hydrolates (Hys) from Italian hemp (Cannabis sativa L. Kompolti cv.) and hop (Humulus Lupulus L., Chinook cv.) supply chains were chemically characterized and tested to investigate their apoptotic potential for the first time. Headspace–Gas Chromatography–Mass Spectrometry (HS-GC-MS) techniques were performed to describe their volatile chemical profile, highlighting a composition rich in terpene derivatives such as monoterpenes and sesquiterpenes among which β-myrcene, limonene, β-caryophyllene and α-humulene were the main constituents of EOs; in contrast, linalool, cis-p-menth-2,8-dien-1-ol, terpinen-4-ol, α-terpineol, caryophyllene oxide, and τ-cadinol were found in the Hys. The cytotoxicity activity on human leukemia cells (HL60), human neuroblastoma cells (SH-SY5Y), human metastatic adenocarcinoma breast cells (MCF7), human adenocarcinoma breast cells (MDA), and normal breast epithelial cell (MCF10A) for the EOs and Hys was studied by MTT assay and cytofluorimetric analysis and scanning and transmission electron microscopy were performed to define ultrastructural changes and the mechanism of cells death for HL 60 cells. An induction of the apoptotic mechanism was evidenced for hemp and hop EOs after treatment with the corresponding EC50 dose. In addition, TEM and SEM investigations revealed typical characteristics induced by the apoptotic pathway. Therefore, thanks to the integration of the applied methodologies with the used techniques, this work provides an overview on the metabolomic profile and the apoptotic potential of hemp and hop EOs and, for the first time, also of Hys. The findings of this preliminary study confirm that the EOs and Hys from Cannabis and Humulus species are sources of bioactive molecules with multiple biological effects yet to be explored.
... Parasponia represents five nodulating tropical tree species growing on volcanic islands of Indonesia and Papua New Guinea [10,11]. Parasponia is closely related to the genus Trema, which includes 18 species that do not nodulate [4,12]. Comparative analysis of Trema and Parasponia species showed that loss of the EPR3-type receptor EPR is specific to Parasponia species [4]. ...
Article
Full-text available
Background Nodule symbiosis with diazotrophic Frankia or rhizobium occurs in plant species belonging to ten taxonomic lineages within the related orders Fabales, Fagales, Cucurbitales, and Rosales. Phylogenomic studies indicate that this nitrogen-fixing nodulation trait has a single evolutionary origin. In legume model plants, the molecular interaction between plant and rhizobium microsymbiont is mapped to a significant degree. A specific LysM-type receptor kinase, LjEPR3 in Lotus japonicus and MtLYK10 in Medicago truncatula , was found to act in a secondary identity-based mechanism, controlling intracellular rhizobium infection. Furthermore, LjEPR3 showed to bind surface exopolysaccharides of Mesorhizobium loti , the diazotrophic microsymbiont of L. japonicus . EPR3 orthologous genes are not unique to legumes. Surprisingly, however, its ortholog EXOPOLYSACCHARIDE RECEPTOR ( EPR ) is pseudogenized in Parasponia, the only lineage of non-legume plants that nodulate also with rhizobium. Results Analysis of genome sequences showed that EPR3 orthologous genes are highly conserved in nodulating plants. We identified a conserved retrotransposon insertion in the EPR promoter region in three Parasponia species, which associates with defected transcriptional regulation of this gene. Subsequently, we studied the EPR gene of two Trema species as they represent the sister genus of Parasponia for which it is assumed it lost the nitrogen-fixing nodulation trait. Both Trema species possess apparently functional EPR genes that have a nodulation-specific expression profile when introduced into a Parasponia background. This indicates the EPR gene functioned in nodulation in the Parasponia-Trema ancestor. Conclusion We conclude that nodule-specific expression of EPR3 orthologous genes is shared between the legume and Parasponia-Trema lineage, suggesting an ancestral function in the nitrogen-fixing nodulation trait. Pseudogenization of EPR in Parasponia is an exceptional case in nodulating plants. We speculate that this may have been instrumental to the microsymbiont switch -from Frankia to rhizobium- that has occurred in the Parasponia lineage and the evolution of a novel crack entry infection mechanism.
... Cannabis sativa L. (hemp or cannabis) belongs to the small family of Cannabaceae, which comprises of ten genera with Humulus (hop, 3 species) as sister genus (Yang et al. 2013). The genus Cannabis is monotypic (Small and Cronquist 1976) and it has been distributed globally as one of the oldest known crop plants. ...
Article
Full-text available
Chloroplast markers (cp markers) are the useful instrument for cannabis (syn. hemp, Cannabis sativa L.) to study relationships of accessions between different geographic origins. In an alignment of three published plastomes 38 chloroplast polymorphisms were identified from which 8 cp markers were used to study relationships of 53 cannabis accessions by high-resolution-melting analyis (HRMA). The marker set could distinguish six haplotypes (‘A’ to ‘F’) in the cannabis collection, where haplotypes ‘A’ and ‘F’ dominated with 34% and 50% of the individuals, respectively. A majority of populations (37) were homogeneous regarding the haplotype, 12 accessions were constituted of two haplotypes and 4 accessions of three haplotypes. Most of the European fibre cultivars consisted of the ‘F’-type (e.g. ‘Fibrimon’, Fibrimon 21’, Juso 14’, ‘Fasamo’ and ‘Schurig’), some were mixed ‘A/F’-types (e.g. ‘Fibrimon 21’, ‘Superfibra’, ‘Lorrin 110’, ‘Futura’, ‘Havelländische’). The Italian ‘Carmagnola in Selezione’ was exceptional in being a pure ‘A’-type. In the heterogenous populations, expected heterozygosity ranged from 0.06 to 0.41. The populations were well differentiated by this marker set locating 79% of the variation among populations (AMOVA). By comparison with plastomes from the closest related genus Humulus , haplotype ‘B’ could be identified as haplotype of the common ancestor of both genera. The haplotype ‘B’ is rare with a frequency of only 4% in the populations analysed. Unfortunately, the true geographic origin of most samples was unclear. However, amongst all published plastomes, only two were classified as haplotype ‘B’, both pointing independently back to Yunnan province (China), indicating Yunnan as the region of origin of the genus Cannabis .
Article
Most species of Urticaceae, the nettle family, have small and inconspicuous, diclinous flowers, in which the perianth, androecium and gynoecium tend to vary in number. Our objective was to study the morphology of the developing flowers of seven species of Urticaceae to understand the pathways that lead to the different patterns of floral reduction and the complex development of pseudomonomerous gynoecia. Buds and flowers were prepared for electron and light microscopy. Vascularization was studied via high resolution X-ray computed tomography micro-CT. Only one whorl of perianth organs is initiated, except for Phenax sonneratii, the flower of which is achlamydeous; variation in perianth merosity results from absence of organs from inception; dicliny results from the absence of stamens from inception (pistillate flowers) and from pistil abortion at intermediate developmental stages (staminate flowers). The gynoecium results from a primordium that divides partially forming two congenitally united primordia (most species) or from a single primordium that apparently does not divide. The gynoecium is served by a single (four species), or two vascular bundles. This second condition is expected for a pseudomonomerous gynoecium. Pistillode or rudimentary carpels occur in staminate flowers. The comparison among species shows that the developmental processes acting in the floral construction in Urticaceae is diverse.
Article
Tapa (barkcloth) is a non-woven textile made from the inner bark of some plant species. Tapa manufacture was once widespread throughout the Pacific and tapa from the eighteenth and nineteenth century form part of Pacific collections in many museums. Here we examined the feasibility of DNA identification of the plants used to make tapa artefacts by developing and testing a DNA reference database of chloroplast trnL intron P6 loop sequences from many of the plant species used to make tapa, as well as other New Zealand textile plants. This database enabled identification to genus for most species but many species shared identical sequences. Despite the lack of species-level resolution, this technique will still aid with identifying the origins of tapa artefacts made from plants with restricted distributions, such as endemic New Zealand and Hawaiian Islands plants. A second aim was to test a number of DNA extraction methods, including non-destructive methods of interest to the heritage sector, on tapa samples. Only one of the non-destructive sampling methods produced amplifiable DNA. However, we did find variation in the success of the destructive methods tested, with the Qiagen DNeasy Plant Mini Kit having the highest success rate.
Article
Full-text available
Universal primers for amplifying and sequencing a noncoding spacer between the atpB and rbcL genes ofthe chloroplast DNA were constructed from the published sequences of Marchantia (a liverwort), tobacco, and rice. Our results indicate an evolutionary trend of increasing spacer size from liverworts, through mosses, to vascular plants. This atpB-rbcL spacer is AT-rich, consistent with other chloroplast noncoding spacers. Due to weak functional constraints, the spacer is evolving rapidly. A sequence identity of 92.2% was observed between spacers of two closely related moss species, Rhytidiadelphus loreus and R. triquestrus. Insertion/deletion events are common in the evolution of this spacer. A 23 bp deletion occurrs in R. loreus. Variation is found between two populations of Amorphophallus henryi (Araceae) and between individuals in a populations of Pasania formosana (Fagaceae). We suggest that this spacer will be useful for molecular systematics at the subspecific, specific, and generic levels and, in some species, for population genetics studies.
Article
Full-text available
Data on pollen morphological features from 200 species in 20 families commonly included in the Hamamelidae and particular species in the Anacardiaceae and Salicaceae are presented in this paper. The basic descriptive analyses presented are derived from observations by light microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). Thirty pollen characters showed some variability, and each of the species was scored for these characters. These data were analyzed and similarity cluster analyses were generated. Both an unweighted pair group and a complete linkage strategy dendrogram were produced. Three major clusters of families were defined, based on these analyses. Group I consists of Trochodendraceae, Cercidiphyllaceae, Eupteleaceae, Platanaceae, Hamamelidaceae (including Altingioideae), Eucommiaceae, and Myrothamnaceae. The Liquidambaroideae, Eucommiaceae, and Myrothamnaceae, while closest to Group I, can be viewed as intermediate between Groups I and II in complete linkage strategy and between Groups II and III in unweighted pair group strategy. Group II-consisting of Daphniphyllaceae, Leitneriaceae, Barbeyaceae, and Fagaceae (excluding Nothofagus)-has a closer phenetic relationship to Group I than Group III. Group III is the largest of these groups: it consists of Ulmaceae, Cannabaceae, Juglandaceae, Rhoipteleaceae, Betulaceae, Casuarinaceae, and Myricaceae. The Balanopaceae and Nothofagus are somewhat isolated and peripheral entities but hold together in both linkage strategies. Thirty pollen characters of 78 taxa were analyzed using PAUP to produce a cladistic tree. The outgroup used was Tetracentron. Three phylogenetically related groups sorted out, which are the same as those already recognized in the Groups I, II, and III mentioned above. Group I occurs at the base of the tree (primitive), and Group II occurs as intermediate between Groups I and III (derived). In general, these data support the relationships suggested by Barabe for the Hamamelidae, based upon vegetative and floral features and the classification of Cronquist.
Chapter
Trees or shrubs, bisexual or unisexual, with watery sap. Leaves alternate, rarely opposite, simple, usually distichous, often oblique at the base, pinnately or pinnipalmately veined, entire or dentate, petiolate; stomates anomocytic or with poorly differentiated supporting cells; cystoliths commonly present, especially in the epidermis; stipules paired, lateral, interpetiolar, or intrapetiolar, free or fused, caducous. Inflorescences axillary, cymose, racemose, paniculate, fasciculate, or the female flowers solitary. Flowers perfect of imperfect, actinomorphic to slightly zygomorphic; perianth subcampanulate, with a single whorl with (2−)4–6(−9) tepals, free or united below, persistent; petals absent. Stamens as many as the sepals and opposite them, or rarely twice as many, or up to 16, free or with the filaments arising from the calyx tube, erect in bud, anthers 2-thecous and dehiscing longitudinally, dorsifixed and often somewhat versatile, a pistillode usually persent in male flowers, staminodes present or absent in female flowers. Gynoecium composed of 2 connate carpels; ovary superior, sessile or stipitate, 1-locular or rarely 2-locular (in Ulmus spp.); style branches 2, linear, simple or bifurcate, with decurrent stigmas; ovule 1, pendulous near the apex of the locule, anatropous or amphitropous, bitegmic, crassinucellar. Fruit a drupe or samara; seeds with straight or curved, dicotyledonous embryo and scanty or no endosperm.
Article
Two hundred sixty-three species of the Magnoliidae distributed over the main genera of all its families were investigated with the TEM for ultrastructural characters from their sieve elements. Details from sieve-element plastids (overall diameter, as well as qualitative and quantitative measurements of their starch and protein contents) provided the main data set for a putative phylogenetic arrangement of the families of Magnoliiflorae. The recording of the presence or absence of nondispersive protein bodies and nacreous wall thickenings yielded additional information. Within the Magnoliiflorae sieve-element plastids are proposed to have evolved from large-sized S-type plastids with high amount of starch to small-sized, starch-deficient P-type or, at the other extreme, to So-plastids. Along this line a basic group containing Austrobaileyaceae, Illiciaceae, Schisandraceae, Chloranthaceae, Myristicaceae, and Winteraceae is distinguished from four parallel and slightly advanced groups, the Annona-, Aristolochia-, Magnolia-, and Monimia-groups. The Aristolochia-group is held to represent one of the core groups of the Magnoliiflorae, incorporating the whole range of the different plastid forms from S-type to So-plastids or by several steps to form-Pc and the monocotyledon form-P2c plastids. Among the remaining five the Monimia-group, only, shows a comparably broad range, while all other taxa are confined to one or two plastid forms (i.e., S-type and/or form-Psc resp. -Pcs plastids).
Article
The family Ulmaceae is most often treated as a single family with two subfamilies: The Ulmeae (Ulmoideae) and Celteae (Celtidoideae) or, more recently, as two separate families: the Ulmaceae and the Celtidaceae (sensu Grudzinskaya). A flavonoid survey of 80 species of Ulmaceae shows that each of the 19 genera is characterized by the production of flavonols (Ulmoid) or glycoflavones (Celtoid), but not both. Further, the arrangement of genera based on this flavonoid dichotomy is remarkably compatible with the generic assignments in Grudzinskaya's bifamilial concept of the Ulmaceae. The only exceptions are Ampelocera, Aphananthe, and Gironniera (in part), which are normally considered Celtoid but possess Ulmoid (flavonols) chemistry. However, recent anatomical and morphological studies of these three genera indicate that their relationship to the Celtoid line may not be as direct as has been supposed, a point also suggested by the flavonoid chemistry.