Systematic Botany (2014), 39(3)
© Copyright 2014 by the American Society of Plant Taxonomists
DOI 10.1600/036364414X681554
Date of publication 05/27/2014
Molecular Phylogeny Estimation of the Bamboo Genus Chusquea (Poaceae: Bambusoideae:
Bambuseae) and Description of Two New Subgenera
Amanda E. Fisher,1,4 Lynn G. Clark,2 and Scot A. Kelchner3
1
Rancho Santa Ana Botanic Garden, Claremont Graduate University, Claremont, California 91711-3157, U. S. A.
Department of Ecology, Evolution and Organismal Biology, Iowa State University, Ames, Iowa 50011-1020, U. S. A.
3
Department of Biological Sciences, Idaho State University, Pocatello, Idaho 83209-8007 U. S. A.
4
Author for correspondence (afisher@rsabg.org)
2
Communicating Editor: Jimmy Triplett
Abstract—Chusquea is a diverse genus of American woody bamboos, accounting for almost half of the woody bamboo species in the
Neotropics. Previous analyses of molecular data have recovered four major lineages within Chusquea, but morphological synapomorphies
have been identified only for subgenus Rettbergia. This study estimates a chloroplast phylogeny of Chusquea with a focus on relationships
within the large and intractable Euchusquea clade. Phylogenetic analyses were conducted on 40% of the described species in Chusquea,
with data from five chloroplast regions and a preliminary survey of the nuclear internal transcribed spacer complex. Several results from
previous studies were corroborated, including the presence of two clades formerly comprising the genus Neurolepis and monophyly of
subgenus Rettbergia. The clades formerly in Neurolepis are named as Chusquea subgenus Platonia and Chusquea subgenus Magnifoliae
based on molecular support and potential morphological synapomorphies. We recovered two strongly supported and five weakly
supported clades within Euchusquea, but relationships among these lineages were not resolved and species composition of the clades
conflicts strongly with current taxonomic groupings based on morphology. Low resolution of the chloroplast phylogeny estimation, low
variability in nuclear data, character conflict, and geographical distribution of chloroplast lineages all suggest a recent radiation of the
Euchusquea clade. Given the present weak molecular support for relationships within Euchusquea and the lack of synapomorphic morphological characters to define clades, we recommend the use of the current morphology-based taxonomy as a practical means of assessing
and describing diversity in the Euchusquea clade.
Keywords—Bamboo, Neotropical radiation, chloroplast phylogeny, grass evolution, Shimodaira-Hasegawa test.
Chusquea Kunth is the most diverse genus of Neotropical
woody bamboos, with 169 described species and estimates
of as many as 220 species in total ( Judziewicz et al. 1999;
Fisher et al. 2009). It is the sole genus in subtribe Chusqueinae
Soderstr. & Ellis and is well supported by molecular data
as monophyletic (Kelchner and Clark 1997; Clark et al. 2007;
Fisher et al. 2009; Kelchner and Bamboo Phylogeny Group
2013). Detailed reviews of the nomenclatural and taxonomic
history of Chusquea can be found in Fisher et al. (2009) and
Clark (1989). Following the recent taxonomic submergence
of Neurolepis Meisn. within Chusquea (Fisher et al. 2009), diagnosable morphological characters for Chusquea are (i) spikelets with four glumes and a single, terminal, fertile floret
lacking a rachilla extension, and (ii) the presence of two papillae on subsidiary cells of the foliar stomatal complex. Lack
of aerial branching, a single bud per node, and strongly
nerved ligules appear to be plesiomorphic in the genus.
Chusquea species are found in mountainous regions of
the Neotropics and austral temperate zone (21! N-47! S), with
centers of diversity in Mexico, Central America, the northern
Andes, and southeastern Brazil. These bamboos primarily
occur in montane forests and high-elevation grasslands
and tend to occupy specialized habitats such as Andean
páramos, Brazilian campos de altitude, Araucaria forests,
Nothofagus forests, and Mexico’s pine-oak forests (McClure
1973; Soderstrom and Calderón 1978b; Veblen 1982; Clark
1992; Londoño 1996; Clark 1997; Judziewicz et al. 1999;
Safford 1999). The dominant vegetation in these highelevation grasslands often consists of Chusquea species such
as C. pinifolia (Nees) Nees and C. heterophylla Nees in
southeastern Brazil, C. subtessellata Hitchc. in Costa Rica’s
Talamanca range (Tol and Cleef 1994), C. spencei Ernst and
C. angustifolia (Soderstr. & C. E. Calderón) L. G. Clark in
Venezuelan páramos, C. tessellata Munro in Colombian
‘bamboo páramos,’ and C. stuebelii (Pilg.) L. G. Clark,
C. aristata Munro, C. rigida (L. G. Clark) L. G. Clark, and
C. nana (L. G. Clark) L. G. Clark in Ecuador (Laegaard
1992; Clark 1996). Many Chusquea species are able to aggressively colonize areas disturbed naturally or by humans
(Judziewicz et al. 1999) and some species can form dense
thickets in montane forests. Once established, Chusquea
bamboos may inhibit germination of surrounding plants
by creating a thick leaf litter (Tol and Cleef 1994) or by
limiting understory light (Widmer 1997; Holz and Veblen
2006; Giordano et al. 2009). In some cases, Chusquea species are primary drivers of tree regeneration cycles. For
example, mature stands of C. montana Phil., C. uliginosa
Phil. (as C. tenuiflora Phil.), and C. culeou E. Desv. in austral
Andean beech forests can out-compete shade intolerant
Nothofagus seedlings (Veblen 1982). Eventually, the monocarpic Chusquea flower and senesce, allowing a new cohort
of Nothofagus to grow at an accelerated rate (Holz and
Veblen 2006).
Excluding subg. Platonia and subg. Magnifoliae (Neurolepis
I and II clades), Chusquea characteristically have abundant
vegetative branching derived from multiple buds per node
and solid culms (McClure 1973; Clark 1989, 1997), but
otherwise exhibit considerable morphological variation
(Fig. 1). Species in montane forests are typically scandent
to clambering on surrounding vegetation, but species in
high-elevation grasslands have an erect and shrubby habit.
Chusquea have unusually complex vegetative branch complements due to the presence of up to 200 buds at each stem
node. These buds are commonly dimorphic, with distinctive
arrangements of a single large central bud and two to many
smaller subsidiary buds (Soderstrom and Calderón 1978a;
Clark 1989).
Previous work identified the sister group to Chusquea
(Clark et al. 2007; Fisher et al. 2009) and provided a preliminary molecular survey of the genus (Kelchner and Clark
1997). Chusquea collections were first described as Nastus
Juss. and were viewed by later authors as being closely
SYSTEMATIC BOTANY
[Volume 39
Fig. 1. Chusquea classification and morphological variation. Photographs of characteristic species for each group are arranged below each heading,
within the orange outline. Five subgenera (gray boxes) are listed, including the newly described subg. Platonia (Neurolepis I) and subg. Magnifoliae
(Neurolepis II). Within subg. Chusquea there are five sections (white boxes). With the exception of subg. Magnifoliae and subg. Platonia, Chusquea species tend to have abundant vegetative branching and they exhibit considerable variation in their bud complement, branch complement, and overall
habit. Photographs by L. G. Clark except C. subtessellata and C. pohlii by A. E. Fisher.
aligned with Nastus based on the shared character of a
single floret per spikelet (Kunth 1816; Desvaux 1831; Nees
von Esenbeck 1835; Munro 1868). Although Kelchner and
Clark (1997) found evidence that one sampled Neurolepis
species is sister to Chusquea s. s., morphological analyses
suggested a sister relationship between Chusquea and
Nastus + Hickelia A. Camus (Clark 1997). Clark et al.
(2007) explicitly tested these two hypotheses and found
that Chusquea + Neurolepis (Chusqueinae) is sister to an
Arthrostylidiinae + Guaduinae clade with chloroplast
rpl16 intron data, even though morphological data resolve
Chusqueinae as sister to the Ehrhartoideae (rice) subfamily
(Clark et al. 2007). Phylogenetic estimation based on a five
region chloroplast dataset by Kelchner and Bamboo Phylogeny Group (2013) corroborated the rpl16 intron data of
Clark et al. (2007), but refuted analyses based on morphology in that there is a moderately supported sister relationship between the Chusqueinae and an Arthrostylidiinae +
Guaduinae clade. Importantly, Clark et al.’s (2007) molecular
analysis of the one-flowered bamboos unexpectedly recovered two distinct lineages of Neurolepis paraphyletic to
Chusquea. Fisher et al. (2009) verified the paraphyletic chloro-
plast relationship of Neurolepis and Chusquea with increased
taxon sampling and statistical testing of alternative hypotheses. The two Neurolepis lineages are now submerged into
Chusquea as subg. Platonia and subg. Magnifoliae.
Three additional subgenera have been recognized (subg.
Chusquea, subg. Rettbergia (Raddi) L. G. Clark, subg.
Swallenochloa (McClure) L. G. Clark, Table 1) and are mainly
defined by differences in habit, branching pattern, and bud
arrangement (Clark 1989, 1992, 1997). Using sequence data
from the chloroplast rpl16 intron, Kelchner and Clark (1997)
found that the sampled species of subg. Rettbergia are monophyletic, whereas subg. Swallenochloa and subg. Chusquea are
polyphyletic within a larger, well-supported clade informally
referred to as the Euchusquea clade.
Despite these earlier analyses of the genus based on
molecular data and recent work across the Bambusoideae
(Sungkaew et al. 2009; Kelchner and Bamboo Phylogeny
Group 2013), little is known about evolutionary relationships
within Chusquea, especially within the species-rich Euchusquea.
The current study was devised to test the monophyly of
taxonomic groups within Euchusquea and to provide a more
complete phylogenetic framework for Chusquea. To that end,
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
Table 1. A summary of the genus Chusquea. Species are classified into five subgenera and then into sections or informal groups. For each taxon
the table lists the number of species, the morphological characteristics, and known distribution, habitat, and elevation preferences.
subg. Chusquea
sect. Chusquea
sect. Longifoliae
Species #
Characteristics
88
20
2–1 subsidiary buds; triangular or circular central bud
Extravaginal branching; several to many subsidiary
buds per node
Infravaginal branching; triangular central bud;
18–80 subsidiary buds per node; long,
narrow foliage leaf blades
Viny to arching or pendant habit; scabrid culm leaves;
infravaginal branching; subsidiary buds in a
usually tight circular cluster; elongated central
bud prophylls; glume II twice as long as glume I
Viny with long, scabrid internodes; infravaginal
branching; 2–12 subsidiary buds per node;
foliage leaf blades relatively long and wide
Crescent or fully encircling verticil of subsidiary
buds; thin foliage leaf blades; dorsally
compressed spikelets; reduced glumes I and II
Pseudopetiolate culm leaf blade that remains green
11
sect. Longiprophyllae
6
sect. Serpentes
6
sect. Verticillatae
18
C. ramosissima group
3
C. meyeriana group
5
subg. Rettbergia
11
subg. Swallenochloa
39
sect. Swallenochloa
39
C. culeou group
3
C. heterophylla group
2
C. mimosa group
4
subg. Platonia (Neurolepis I)
11
subg. Magnifoliae (Neurolepis II)
10
Spatheate bracts often subtending the synflorescence;
reflexed lower inflorescence branches; reduced
glumes I and II
Clambering habit; circular central bud; infravaginal
branching; arachnoid subsidiary branches;
spatheate bracts usually subtending the
synflorescence; connate lemma apices
Shrubby habit; erect culms; shortened waxy internodes;
triangular central bud; subsidiary buds forming
a crescent below or linearly flanking the central
bud; intravaginal or extravaginal branching;
tessellate leaf blades
Subsidiary buds forming a crescent below or linearly
flanking the central bud; intravaginal branching;
short, waxy internodes; thick, stiff foliage leaf
blades; narrowly paniculate synflorescences;
irregular flowering or short cycles.
Usually subequal, linearly arranged buds lacking
an obvious dominant central bud
Erect, fastigiate culms; subsidiary buds forming
a crescent below the dominant central bud;
extravaginal branching; small leaves (< 15 cm)
Erect culms that arch distally; deciduous culm and
foliage leaf sheaths; scarious culm and foliage
leaf sheath margins; dimorphic or trimorphic
bud complements
Lacking aerial branching; internodes 1–14 cm long
(relatively elongated); foliage leaf blades
(8-)16–180 cm L, 1–6.3(–8.3) cm W (except for
N. nobilis); absent to short pseudopetioles;
glumes usually with well developed awns
Lacking aerial branching; internodes 1–3 cm long
(relatively compressed); foliage leaf blades
(20–)30–300(–377) cm L, 1.2–24(–30) cm W;
pseudopetioles well developed; glumes
usually lacking awns
65 Chusquea species (40% of the total number of described
species) were sampled for five chloroplast regions and the
nuclear region ITS. Species sampling included representatives from all of the described taxonomic groups, spanned
the geographic distribution of the genus, and encompassed
prominent features of morphological diversity in Chusquea s. l.
(Fisher et al. 2009).
Materials and Methods
Taxon and Character Sampling—Leaf material for 65 species of
Chusquea and three outgroup taxa (Appendix 1) was collected from
Distribution and Habitat
Widespread
Mid- and high-elevation forests of Andes
Montane forests of southern Mexico,
Central America, and northern Andes
Montane forests of northern Andes, with
five spp. restricted to the cordilleras
of Colombia
Montane forests of Mexico, Central America,
and northern Andes
Mid- to low-elevation montane forests
of Mexico, Central America, and
northern Andes
Atlantic and lowland forests of Misiones,
Argentina, Paraguay, and Brazil
Brazil, especially Atlantic forest
Brazilian montane and Atlantic forests,
except C. arachniformis, which is
endemic to montane forests of
northwestern Colombia
Usually open habitats, often at high
elevation; widespread, but most
diverse in subpáramos and páramos
of the northern Andes, Central America,
and campos de altitude of eastern Brazil
Widespread distribution and habitat
preferences, as in subg. Swallenochloa
Open meadows and Nothofagus forest
understory in the austral
Andes (0–2,000m)
Brazilian campos de altitude
Mid- to high-elevation forests in Brazil;
C. nudiramea is also found in
low-elevation Atlantic forests
Upper montane forest and páramos
of Ecuador and Colombia;
C. fimbriligulata extends into
northern Peru
Central America, northern and central
Andes, Trinidad, Guyana, and
northern Brazil
young branches and immediately dried in silica gel. Vouchers are deposited at ISC, US, INB, and CR. Three outgroups were chosen from the
Arthrostylidiinae + Guaduinae Neotropical woody bamboo clade, the
lineage that is estimated to be sister to the Chusqueinae in recent
molecular phylogenetic studies (Clark et al. 2007; Sungkaew et al.
2009; Kelchner and Bamboo Phylogeny Group 2013).
Five chloroplast regions were used in the analysis: rpl16 intron, ndhA
intron, trnD-trnT intergenic spacer, trnT-trnL intergenic spacer, and ndhF
gene. The nuclear internal transcribed spacer complex (ITS1, 5.8S, ITS2)
was also surveyed for potential information. The individual chloroplast regions were chosen based on prior use in bamboos (Kelchner
and Clark 1997; Zhang 2000; Clark et al. 2007; Triplett and Clark 2010)
and variation of these regions across seed plants (Shaw et al. 2005; Shaw
et al. 2007).
SYSTEMATIC BOTANY
Although there can be serious issues with paralogy determination
when using ITS in phylogenetic studies (Alvarez and Wendel 2003), the
region was sampled in this study after repeated failure to amplify more
desirable low-copy nuclear loci (GBSSI and rpb2), presumably due to
degraded DNA in our dried leaf tissue. The ITS region was sequenced
with the goal of assessing its variation in Chusquea and as a potential
means of corroborating the topology derived from chloroplast data.
Because evidence indicates there are multiple copies of ITS present in
Chusquea’s genome and we did not clone PCR amplicons, we take a
cautious approach to the use and interpretation of our ITS data.
DNA Isolation, Amplification, and Sequencing—Total genomic DNA
was isolated from silica-gel-dried leaf samples with either Extract-N-Amp
kits (Sigma-Aldrich, St. Louis, Missouri) or Qiagen DNEasy plant mini kits
(Qiagen, Inc., Valencia, California) using the following modifications to
the Qiagen kit protocol: 35–50 mg of dry leaf material, 500 ml of lysis
buffer, and a final wash with cold EtOH. Samples extracted with the
Extract-N-Amp kit were subsequently cleaned with Qiagen PCR purification columns. Cleaned total genomic DNA served as the template for
PCR amplification and sequencing reactions for each molecular region
(protocols and primers as in Table 2). Three primers were trialed for ITS
(Table 2) in two pairwise combinations (Prince and Kress 2006). Direct
sequencing of 18SF-26SR amplicons resulted in multiple sequences in
the same chromatogram and this primer pair was abandoned owing to
inadequate resources for cloning. However, direct sequencing of ITSL26SR amplicons resulted in a single sequence except that, in a few cases,
ITSL-26SR amplicons were of two lengths. In these cases, the band of
length consistent with this ITS copy in other Chusquea species was excised
with a Wizard SV gel and PCR clean-up kit (Promega, Madison, Wisconsin)
according to the manufacturer’s instructions. Forty-nine Chusquea species
and one outgroup were sampled for the ITS nuclear region.
Sequencing was carried out on an Applied Biosystems 3130XL genetic
analyzer (Applied Biosystems, Foster City, California) at the Idaho State
University Molecular Research Core Facility or on an Automated 3730XL
DNA Analyzer (Applied Biosystems) at the Iowa State University DNA
Sequencing and Synthesis Facility. Sequence files were manually checked
for base-calling errors using the program 4Peaks 1.7.2 (Griekspoor and
Groothuis 2006).
Alignment and Data Exploration—Edited sequences were aligned by
hand in Se-Al 2.0 (Rambaut 2001). The criterion-based manual alignment
[Volume 39
of gaps followed Kelchner (2000), Graham et al. (2000), and Borsch et al.
(2003). Shared microstructural changes, including insertions, deletions,
and hairpin inversion events were coded as absent or present in a
0/1 matrix after probable biological origins for the events were inferred
from the surrounding sequence. Nucleotide characters of insertiondeletion regions were then excluded from the dataset in phylogenetic
analyses. Areas with ambiguous alignments, such as variation between
sequences due to mononucleotide repeats, were also excluded during
analyses. Sequences were searched against the National Center for Biotechnology Information (NCBI) nucleotide collection database (nr/nt) to
evaluate their similarity to the target regions in grasses.
Indices of nucleotide base frequencies, uncorrected-p distances, and
number of parsimony informative characters were assessed using PAUP*
v4b10 (Swofford 2003). Neighbor-net analyses of uncorrected-p distances
were performed in SplitsTree 4 (Huson and Bryant 2006) to assess character conflict in individual and combined datasets (Morrison 2010). In
addition, secondary structures of ITS sequences were modeled on the
Mfold server (http://mfold.rna.albany.edu; Zuker 2003) in order to
reduce the likelihood of including paralogous sequences (including
pseudogenes) or sequences of fungal origin (Feliner and Rosselló 2007).
Model Selection—Models of nucleotide evolution for individual data
sets and the combined chloroplast data were evaluated using dynamic
likelihood ratio tests (dLRT) and the Akaike information criterion (AIC)
method as implemented in jModelTest (Posada 2008). Likelihood scores
were computed within jModelTest using PhyML (Guindon and Gascuel
2003) with 88 candidate models, 11 substitution schemes, and options
for unequal base frequencies, proportion of invariant sites (I), and rate
variation among sites. Multiple models were considered best-fit for the
data if AIC differences were D £ 2.
Phylogenetic Tree Estimation—Phylogenetic analyses were performed on the chloroplast and ITS nuclear data separately, to account
for the possibility that the organellar and nuclear genomes track
separate evolutionary histories. Tree topologies and support values
resulting from different estimation frameworks (Bayesian inference,
maximum likelihood, and maximum parsimony) were compared for
each dataset to assess robustness of tree topology to changes in
model assumptions. Time intensive analyses that could be facilitated
by parallel processing were run on the Idaho State University EGG
Bioinformatics computing cluster.
Table 2. Molecular regions used in this study, PCR amplification primers, and thermocycler protocols. Primers labeled SEQ were favored for
sequencing, but were also used for amplification. Primer sequences were obtained from the following publications: ITS nuclear region 18SF, 26SR
(Prince and Kress 2006); ITS nuclear region ITSL (Hsiao et al. 1995); ndhA intron SAK26 & SAK28 (Watts et al. 2008); ndhF gene 972F, 2110R, 1318F
and 1603R (Olmstead and Sweere 1994); rpl16 intron F71, R1516 (Kelchner and Clark 1997); trnD-trnT intergenic spacer trnDF & trnTR (Demesure
et al. 1995); trnT-trnL intergenic spacer TABA, TABB (Taberlet et al. 1991). Touchdown PCR was used to amplify ndhF and trnD-trnT intergenic spacer
by reducing the annealing temperature of each cycle -1! C, until a target temperature is reached, followed with additional cycles at this annealing
temperature (Don et al. 1991).
1,000 bp
trnD-trnT
intergenic
spacer
1,100 bp
trnT-trnL
intergenic
spacer
880 bp
94! C, 1m; 10 (94! C, 1m 30s; touchdown
53–43! C, 2m; 72! C, 3m); 20 (94! C, 1m 30s;
43! C, 2m; 72! C, 3m); 72! C 10m.
+
rpl16 intron
+
1,140 bp
80! C, 5m; 35 (95! C, 1m; 50! C, 1m;
+15! C, 0.3! C/s; 65! C, 5m); 65! C, 4m.
+
ndhF gene
(3’ end)
(94! C, 1m 30s; 58! C, 2m; 72! C, 1m); 72! C, 7m.
80! C, 5m; 35 (95! C, 1m; 50! C, 1m;
+15! C, 0.3! C/s; 65! C, 5m); 65! C, 4m.
+
800 bp
30
94! C, 2m; 35 (94! C, 45s; touchdown
58–48.5! C, 1m; 72! C, 1m 15s); 72! C, 5m.
+
ndhA intron
18SF: CGATTGAATGGTCCGGTGAAG (54.4! C)
26SR: AGGACGCTTCTACAGACTACAA (53! C)
ITSL: TCGTAACAAGGTTTCCGTAGGTG(55.3! C)
SAK26: CAATATCTCTACGTGYGATTCG (! C)
SAK28: AACTGTTRGATAATCATAGTCG (47.4–49.2! C)
SAK43: TCTTTTTCAGGTGGTCTACGAG (53! C)
SAK44: ACTGTGCTTCAACTATATCAAC (49.9! C)
927F: GTCTCAATTGGGTTATATGATG (48.9! C)
2110R: CCCCCTAYATATTTGATACCTTCTCC (55.2! C)
SEQ:1318F: GGATTAACTGCGTTTTATATGTTTCG (55.2! C)
1603R: GCATAGTATTTCCCGTTTCATGAGG (56! C)
F71: GCTATGCTTAGTGTGTGACTCGTTG (57.7! C)
SAK7: GAACGACAGAACCTATGA (45.8! C)
SAK8: CCATCCCACCCAATGAAG (53.2! C)
R1516: CCCTTCATTCTTCCTCTATGTTG (53! C)
R1661: CGTACCCATATTTTTCCACCACGAC (57.9! C)
TrnDF: ACCAATTGAACTACAATCCC (48.5! C)
SAK9F: ACCAATTGAACTACAATCCC (50.2! C)
SAK10R: GCATAAGTCATCGGTTCAAATC (51.7! C)
trnTR: CCCTTTTAACTCAGTGGTA (48.8! C)
SEQ: SAK2R: TGCCCCTATCGTCTAGTGGT (53.8! C)
SAK1F: GGATTTGAACCAGCGTATACA (50.5! C)
TABA: CATTACAAATGCGATGCTCT (48.5! C)
TABB: TCTACCGATTTCGCCATATC (49.7! C)
AFIP1: TAAGGAGAACATAGAATCATAGC (50.5! C)
AFIP2: GCTATGATTCTATGTTCTCCT (49! C)
80! C, 5m; 35 (95! C, 1m; 50! C, 1m;
+15! C, 0.3! C/s; 65! C, 5m); 65! C, 4m.
+
800 bp
+
ITS
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
Bayesian inference (BI) estimation was performed in MrBayes 3.1
(Huelsenbeck and Ronquist 2001) using the general time reversible
(GTR) model, with estimates of invariant characters (I) and substitution
rate variation (G) parameters as recommended by Huelsenbeck and
Rannala (2004). Analysis of the combined chloroplast data included
partitions for each region plus coded microstructural changes (with
a standard discrete model specified for the microstructural data). In
separate analyses of the partitioned chloroplast and ITS nuclear datasets
the following settings were used: 40 million generations, two runs, four
chains per run, sampling every 2,000 generations, heating temperature
of 0.2, and a burn-in of 10,000 trees. Priors were set to Dirichlet on base
frequencies and the rate matrix, exponential on branch lengths, and
uniform on the gamma shape parameter (a), proportion of invariable
sites (I), and topology. MrBayes output files for each run were combined to estimate sample size scores in Tracer 1.4.1 (Rambaut and
Drummond 2007). Convergence of the posterior distribution was considered to have occurred when the average standard deviation of split
frequencies fell below 0.01 and likelihood values appeared to be stable
over several thousand generations.
Maximum likelihood (ML) analyses were performed in PhyML
ver. 2.4.4 (Guindon and Gascuel 2003) using one of the best-fit models
determined in jModelTest. Nonparametric bootstrap (BS) analyses were
conducted with 1,000 replicates to evaluate support for branches.
Maximum parsimony (MP) analyses were conducted in PAUP* 4b10
using a two-step “long-thin” search process (Kelchner 2003). This method
was used to estimate a robust topology across frameworks and facilitate
extensive BS analyses and was not intended for recovery of the most
parsimonious tree(s) from the data set. Sequences were added randomly
for 1,000 repetitions of heuristic search, with no more than 100 trees
greater than or equal to length 1 kept in memory for each repetition.
The second step of searching used the subsequent pool of 100,000 trees
as starting trees and swapped branches to completion. Trees generated
in the second step were combined in a strict consensus topology.
Support for branches was evaluated in PAUP* v4b10 using two nonoptimal, heuristic approaches after a conventional BS search (search =
heuristic addseq = random nreps = 10,000) ran for more than two months
on an iMac G5 (Apple Inc., Cupertino, California). A lack of phylogenetic
information may have overwhelmed tree-bisection-reconnection branch
swapping during a conventional exhaustive PAUP* bootstrap search
(Sullivan 2005). Instead, we trialed a parsimony fast-step BS analysis and
an analysis using a single optimal tree held in memory with tree-bisectionreconnection branch swapping (parsimony single tree; Mort et al. 2000).
Alternative Hypothesis Testing—Relationships in the chloroplast
phylogeny estimation and the morphological classification of Chusquea
present competing hypotheses of evolution. Shimodaira-Hasegawa tests
(SH tests; 1999) were conducted to compare several sets of hypotheses.
The unconstrained maximized likelihood estimate (MLE) tree of combined chloroplast data was tested in a series of two-tree comparisons
against MLE trees with monophyletic constraints on each subgenus or
section previously described for Chusquea (three subgenera and seven
sections). Trees were considered significantly different given the chloroplast data and best-fit model if p < 0.05 for the test distribution. After
the initial phylogenetic analyses suggested that a geographic signal
might be present in the data, a second series of SH tests compared the
chloroplast unconstrained MLE tree to a number of alternative hypotheses based on geographical distribution. Species occurring in five separate
geographic regions were constrained to be monophyletic in individual
MLE analyses: austral Andes, Brazil, Brazilian species excluding subg.
Rettbergia, northern Andes, and central American + Mexican species. In
addition, a single SH test compared the ITS MLE tree to a topology with
a constraint on a monophyletic subg. Swallenochloa sect. Swallenochloa.
Isolation By Distance—To explore scenarios of evolution in subg.
Rettbergia and Euchusquea, genetic distances were compared with physical
distances (Slatkin 1993). An isolation by distance analysis was conducted
on a dataset that sampled across subg. Rettbergia and Euchusquea and
a dataset that sampled only within Euchusquea. Uncorrected-p distances
of the chloroplast data were calculated in PAUP* v4b10 and latitude and
longitude values were recorded from herbarium labels for the accessions
sampled in this study. A distance matrix was calculated with the software Geographic Distance Matrix Generator v. 1.2.3 (Ersts 2009).
Results
Chloroplast and Nuclear Sequence Data—The combined
chloroplast alignment contains 5,278 nucleotide characters
for 68 species (0.71% missing data) and a binary matrix
of 88 inferred microstructural characters (Appendix S1,
TreeBASE study number TB2:S14424). BLAST searches returned
sequences of the expected regions from Bambusoideae as
top matches. The combined chloroplast alignment includes
446 variable characters and 260 parsimonious informative
characters (dataset characteristics available in Appendix S2),
and uncorrected-p distances were small for both the chloroplast and nuclear regions (0.0064–0.0167). No region contains
a unique sequence for each species, although the combined
chloroplast data set does.
In Chusquea, the ndhA intron is the least variable region,
with 114 taxon pairs sharing identical sequences. The trnDtrnT region has the highest combined number of variable
nucleotides and microstructural changes in the chloroplast
dataset (144), but many of these changes are not informative (70). The trnT-trnL region is one of the shorter alignments,
but contains a moderate number of variable characters (78),
a high percentage of which are informative (79%). Although
it adds informative data to the analysis, trnT-trnL was difficult to amplify and sequence due to five regions of mononucleotide repeats over six bases long within a 250 bp section.
Two internal primers (AFIP1, AFIP2) were designed to overcome polymerase failure during trnT-trnL amplification,
requiring four sequencing runs to obtain the entire region
for most samples. The ndhF gene showed variation comparable to the non-coding regions and, in this study, it contained more parsimony informative nucleotide characters (52)
than the other chloroplast regions, except trnD-trnT (58).
Most of the parsimony informative characters supplied by
ndhF provide support for branches deep in the tree topology, particularly those leading to the outgroups and subg.
Platonia and subg. Magnifoliae.
Nucleotide sequences of the ITS region were generated
for 50 taxa. Because we were not able to clone amplicons,
these data are of limited value and we use them here only
as a preliminary survey of nuclear sequence variation in
Chusquea and to check for broad patterns of hierarchical signal
at the nuclear level. We do not consider the data sufficient
for phylogeny estimation of Chusquea nuclear relationships.
BLAST searches of ITS sequences generated in this study
returned bamboo ITS accessions as top hits. Alignment
length was 162 bp for ITS1, 221 bp (including a 15 bp indel
in some species) for 5.8S, and 356 bp for ITS2. Many
Euchusquea clade ITS sequences are identical and the variation that is present in the ITS dataset is concentrated in
deep branches of the tree. Five species (C. andina Phil.,
C. culeou, C. leonardiorum L. G. Clark, C. liebmannii E. Fourn.,
and C. pinifolia) have large deletions in ITS (spanning all three
regions) ranging from 121 bp in C. andina and C. culeou to
613 bp in C. liebmannii. These five sequences were removed
from the alignment before phylogenetic analyses to reduce
potential error associated with homology uncertainty.
Several characteristics of the Chusquea ITS sequences analyzed in this study suggest that they are not pseudogenes
(Feliner and Rosselló 2007). Specifically, i) the Chusquea ITS1
spacer was found to contain a conserved sequence motif
with AAGGAA 30 of a hairpin structure, ii) an AAGAA loop
and an EcoRV site (GATATC; 283–287 bp) are conserved
in the 5.8S gene, and iii) several ITS2 RNA sequence folding
structures estimated in Mfold show universally conserved
pyrimidine mismatches (Zuker 2003). There is no obvious
indication that paralogous or pseudogenized ITS sequences
SYSTEMATIC BOTANY
are present in the alignment, particularly because of the
overall lack of observed sequence variation.
Neighbor-net networks were used to evaluate character
conflict and the combinability of chloroplast data (Suppl. 2).
Except for the trnT-trnL and trnD-trnT intergenic spacers, networks for each region suggest a similar pattern in the data,
which is largely tree-like but lacking general resolution. The
trnT-trnL network includes a large reticulation caused by a
13 base sequence of undetermined origin that is shared by
C. gracilis McClure & L. B. Sm. and C. aristata. The region in the
alignment containing this ambiguous sequence was excluded
in phylogenetic analyses. The trnD-trnT network found a welldefined split between two groups of Chusquea species
(Suppl. 2) although this split is supported by only six characters.
Tree Estimation—The AIC and dLRT model choice
methods agree on adequate models for each dataset from
[Volume 39
our candidate pool (Suppl. 3). GTR + I + G was estimated
as one of the best-fit models (D £ 2) for all regions except
ITS. Less complex nucleotide substitution rate categories
are adequate for the ITS dataset (TrN or TIM + I + G).
A BI consensus tree was summarized from 10,001 trees
after a burn-in of 20 million generations (Fig. 2) and effective sample sizes were found to be adequate in Tracer for
the combined BI runs (> 200 samples, as recommended by
Drummond and Rambaut (2007)). The posterior distribution
appeared to reach convergence after 4.4 million MCMC iterations as indicated by the average standard deviation of split
frequencies reaching a value < 0.01 and the stability of likelihood values. ML analysis of the combined chloroplast data
resulted in a topology with -lnL = 14,152.99. ML and MP
bootstrap support (BS) values are lower than BI posterior
probability values on all branches (Fig. 2). Reports of support
Fig. 2. Bayesian inference cladogram of combined chloroplast data including support values (maximum parsimony faststep bootstrap/maximum
parsimony single tree bootstrap above the branch and maximum likelihood bootstrap/Bayesian posterior probability below the branch).
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
for relationships in the following paragraphs are not meant
to reflect total evidence for branches (i.e. branches reported
as strongly supported may not be clear in network analyses or
supported by SH tests), but are based on BI posterior probability (PP) and MP/ML BS values.
The combined chloroplast data recovered many clades
with consistent species compositions in BI, ML, and MP
phylogeny estimations (Fig. 2). The tree is rooted with
Rhipidocladum harmonicum, making Guadua angustifolia sister
to Otatea acuminata. Subgenus Magnifoliae and subg. Platonia
are the two earliest-diverging clades. Subgenus Rettbergia
is monophyletic and sister to a large Euchusquea clade.
Branches leading to some clades within Euchusquea change
position along the tree backbone depending upon phylogenetic estimation method (Fig. 3).
The major clades in a BI analysis of the ITS nuclear
region (Fig. 4) correspond to those found in the chloroplast
analyses (Fig. 3). The ITS topology is not as resolved as the
chloroplast estimate, but it does contain several small
groupings that are also found in the chloroplast tree.
Sampled Euchusquea species form a monophyletic group
in chloroplast phylogeny estimations (Fig. 3). Delineating relationships within Euchusquea is difficult, however,
due to low variation among nucleotide sequences. Short
branches and poorly supported branching order among
Euchusquea lineages creates a polytomy at the base of the
clade (Fig. 3). Five Euchusquea clades appear in all BI, ML,
and MP topologies with relatively high branch support.
These five clades do not agree with the taxonomic groupings within Chusquea. Instead, the chloroplast signal suggests strong geographic partitioning of Euchusquea species
relationships. The five supported Euchusquea clades are
described here in terms of their species compositions as
well as their distributions.
Fig. 3. Bayesian inference phylogram of a partitioned analysis of the chloroplast data. Weighted branches indicate maximum likelihood
bootstrap support > 90 and Bayesian inference posterior probability support > 0.95. Dashed-line branches change position in maximum likelihood
or maximum parsimony analyses. Sw and Ch indicate the species is classified as subg. Swallenochloa or subg. Chusquea, respectively. Chusquea
longispiculata, C. leptophylla, C. lorentziana, C. gracilis, and C. abietifolia are not currently placed in a subgenus. Numbered clades are referenced in the
text. Chusquea contains four major lineages: subg. Platonia (Neurolepis I), subg. Magnifoliae (Neurolepis II), subg. Rettbergia, and a Euchusquea clade.
Within Euchusquea there are five well supported lineages, I-V. Shapes at the end of branches correspond to geographic distribution areas as
indicated on the map. Illustrated species are indicated with asterisks. Chusquea maculata and C. exasperata are members of subg. Chusquea sect.
Longiprophyllae and are morphologically similar, yet they are not closely related in the tree topology. Their closest relatives according to chloroplast
data are shown for comparison. Line drawing of C. serpens is reproduced from Clark 1985; C. maclurei is from Clark 1986; C. maculata and
C. exasperata are from Clark 1990.
SYSTEMATIC BOTANY
[Volume 39
Fig. 4. ITS Bayesian phylogram (left) and cladogram (right). Scale-bar on phylogram is 0.01 changes. Support values are listed as maximum
parsimony bootstrap above the branch; maximum likelihood bootstrap/Bayesian posterior probability below the branch. Lettered nodes are referenced
in the text.
I. CHUSQUEA CULEOU CLADE: AUSTRAL ANDES—A C. culeou
clade was recovered in chloroplast analyses with strong
support (I, Fig. 3; 93% ML BS, 1.00 PP) and contains species
restricted to austral Andean montane forests (Triplett and
Clark 2003). C. andina, C. cumingii Nees, and C. quila Kunth
form a well-supported group and C. gigantea Demoly +
C. uliginosa Phil. is also supported (88% ML BS, 1.00 PP).
The widely distributed C. culeou appears to be sister to the
rest of this clade.
II. CHUSQUEA RAMOSISSIMA CLADE: PARAGUAY, URUGUAY,
BRAZIL, ARGENTINA—The C. ramosissima clade consists of
C. ramosissima Lindm., C. tenella Nees, and C. longispiculata
L. G. Clark and is moderately supported as monophyletic
(II, Fig. 3; 90% ML BS, 1.00 PP). Chusquea ramosissima is
sister to C. tenella with strong support. Chusquea ramosissima
and C. tenella are widespread species in low elevation
areas of Paraguay, Uruguay, eastern Brazil, and the state
of Misiones, Argentina, while C. longispiculata is restricted
to southeastern Brazil.
III. CHUSQUEA UNIFLORA CLADE: NORTHERN ANDES AND CENTRAL
AMERICA—The C. uniflora clade is composed of species
found in low to middle elevation montane and cloud
forests in the northern Andes, along with C. scabra
Soderstr. & C. E. Calderón, a low elevation species from
Costa Rica. There is strong support for monophyly (III,
Fig. 3; 96% ML BS, 1.00 PP), but relationships within the
clade are unclear, with BI, ML, and MP bootstrap topologies recovering distinct relationships.
IV. CHUSQUEA SERPENS CLADE: NORTHERN AND CENTRAL ANDES,
CENTRAL AMERICA—The C. serpens clade is well supported
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
(IV, Fig. 3; 98% ML BS, 1.00 PP) and contains four species,
C. patens L. G. Clark, C. pohlii L. G. Clark, C. serpens L. G.
Clark, and C. maculata L. G. Clark. Chusquea pohlii and
C. patens are restricted to Costa Rica and Panama. Chusquea
maculata is found only in Colombia and Venezuela, while
C. serpens is distributed through those countries, as well
as Ecuador.
V. CHUSQUEA SCANDENS CLADE: NORTHERN AND CENTRAL ANDES,
CENTRAL AMERICA—The C. scandens clade is strongly supported with chloroplast data (V, Fig. 3; 95% ML BS, 1.00 PP)
and includes 21 species in BI, ML, and MP analyses. There are
several relationships within the C. scandens clade with some
branch support in chloroplast analyses. The only strongly
supported sister relationship exists between C. foliosa L. G.
Clark + C. longifolia Swallen. Many species in the C. scandens
clade (e.g. C. leonardiorum and C. subtessellata) are found in
high-elevation habitats such as cloud forest, subpáramo,
and páramo (Soderstrom and Calderón 1978b; Clark 1992).
This group is mainly northern Andean, with a more recently
diverging group of Central American species. Several of the
Central American species are also found in southern Mexico
(C. foliosa and C. longifolia), and C. lorentziana Griseb. from
northern Argentina and Bolivia also appears in this clade in
all chloroplast analyses (although an SH test cannot reject
the possibility that it is a member of the C. culeou clade;
see below). In the ITS topology a group of six species from
the chloroplast C. scandens clade are recovered as monophyletic with weak support (E, Fig. 4). Within this clade
there are supported relationships between C. subtessellata +
C. talamancensis and C. costaricensis, C. paludicola, C. tonduzii
and C. vulcanalis.
Alternative Hypothesis Testing—The possibility that morphologically defined taxonomic groups might be supported
by the chloroplast data given the best-fit model was investigated with a series of two-tree comparison SH tests.
The SH test rejected most alternative hypotheses with
three exceptions: monophyletic constraint of subg. Chusquea
sect. Longiprophyllae, subg. Swallenochloa C. culeou group, and
subg. Swallenochloa C. nudiramea group (Table 3). These
Table 3. Results of two-tree SH tests comparing the maximum likelihood estimate tree with alternative topology hypotheses as described in
the text. Chloroplast data was used, unless ITS data is indicated. The
sampling column refers to the number of species sampled in this study
out of the total number of species in the group.
Hypothesis of monophyly
Species
sampled
SH test
p value
subg. Rettbergia + C. ramosissima
subg. Swallenochloa
subg. Swallenochloa sect. Swallenochloa
subg. Chusquea sect. Longifoliae
subg. Chusquea sect. Serpentes
subg. Chusquea sect. Verticillatae
subg. Chusquea sect. Chusquea
subg. Swallenochloa C. culeou group
subg. Swallenochloa C. nudiramea group
subg. Chusquea sect. Longiprophyllae
Brazilian Euchusquea species
ITS data subg. Swallenochloa sect. Swallenochloa
austral Andean species
Brazilian species
northern Andean species
Central American species
7/13
21/38
14/29
8/11
4/6
7/18
8/19
3/3
3/4
2/6
NA
10/29
NA
NA
NA
NA
0.024*
0.000*
0.000*
0.000*
0.028*
0.007*
0.000*
0.378
0.303
0.182
0.675
0.060
0.057
0.000*
0.000*
0.000*
taxonomic groups contain a small number of species or
were sparsely sampled in this study, as in the case of
sect. Longiprophyllae.
An SH test rejected a constrained topology containing a
monophyletic subg. Swallenochloa sect. Swallenochloa as an
alternative to the unconstrained MLE topology (p = 0.000*).
The sampled species of subg. Swallenochloa sect. Swallenochloa
are not monophyletic in any of the chloroplast optimal
topologies. However, an SH test with the ITS data and the
best-fit model failed to reject a topology constraining a group
of ten sect. Swallenochloa species as monophyletic (p = 0.06),
although we note again the low level of character variation
present in the ITS data set.
Hypotheses of relationships predicted by geographic
proximity were also tested. The C. culeou clade is composed of all sampled taxa from high-elevation habitats
and western slopes of the austral Andes. C. ramosissima
and C. tenella are mainly found in southeastern Brazil, but
extend slightly into Argentina and Paraguay. An SH test
failed to reject a topology containing a monophyletic C. culeou
clade + C. lorentziana + C. ramosissima + C. tenella (p =
0.057). Another SH test assessed the three Brazilian clades
in the chloroplast topology: subg. Rettbergia, except
C. arachniformis (known only from the northern Andes),
the C. leptophylla clade, and the C. ramosissima clade. Given
the chloroplast data and the model, an SH test rejected a
topology in which all of the Brazilian species are monophyletic (p = 0.000*). A second SH test of Brazilian species that
excluded subg. Rettbergia failed to reject the possibility of
a topology containing a monophyletic Brazilian clade (p =
0.675). An SH test of a topology constraining northern
Andean taxa to monophyly (species present in subg.
Rettbergia, C. uniflora, C. serpens, and C. scandens clades in the
chloroplast topology) is rejected given the data and the
model (p = 0.000*). Finally, nineteen Central American species are present in the C. pittieri, C. uniflora, C. serpens, and
C. scandens clades in the chloroplast topology and an SH test
rejected a topology that constrained these groups to be
monophyletic (p = 0.000*).
Isolation by distance—Comparisons of genetic distances
to the proximity of collected individuals were conducted
on two groups of Chusquea species: (i) subg. Rettbergia +
Euchusquea (Fig. 5A) and (ii) the Euchusquea clade (Fig. 5B).
In both analyses there is no apparent relationship between
genetic distance and the location of any two individuals.
Genetically similar species might be found in the same location (C. talamancensis and C. paludicola, p distance = 0, physical distance = 0) or > 6,000 km apart (C. ramosissima and
C. bilimekii, p distance = 0.001, physical distance = 6,943 km).
Species that are genetically less similar to one another might
also be in the same location (C. virgata and C. longifolia,
p distance = 0.01, physical distance = 67 km) or they might
be distant (C. nudiramea and C. muelleri, p distance = 0.012,
physical distance = 7,350 km). Across the complete dataset,
two distinct patterns are seen: large genetic distances
between the Euchusquea and subg. Rettbergia species and
small genetic distances within Euchusquea.
Discussion
Our study recovered four major lineages in Chusquea with
strong support (Fig. 3). A monophyletic Chusquea includes
subg. Magnifoliae (Neurolepis II) sister to subg. Platonia
SYSTEMATIC BOTANY
Fig. 5. Isolation by distance analyses comparing chloroplast genetic
distance on the x axis with physical distance of accessions sampled in
this study on the y axis for A) species in Euchusquea and subg. Rettbergia
and B) Euchusquea species only.
(Neurolepis I) + (subg. Rettbergia + Euchusquea). Relationships
among the four lineages corroborate those recovered by
Kelchner and Clark (1997), Clark et al. (2007), and Fisher et al.
(2009). The ITS nuclear dataset shows little variation (Fig. 4),
but is consistent with the chloroplast dataset in resolving the
four major Chusquea lineages.
We recovered five supported lineages in the Euchusquea
clade (Fig. 3) and chloroplast species relationships within
Euchusquea correspond to some degree with geographic
areas, although the pattern is complicated by more than
one Chusquea lineage present in most geographic regions
(i.e. northern Andes, Central America, Brazil, possibly the
austral Andes). Given geographic patterns, low variation
in Euchusquea, and conflicts between molecular and morphological hypotheses of relationships, the data suggest
that Chusquea may be a young and actively radiating group
of bamboos.
Several limitations of the study prevent unambiguous
interpretation of the data and leave many relationships
unresolved. The chloroplast data contains little variation
compared to similar sized datasets for other grass genera, a
common problem in bamboo molecular phylogeny studies
(Triplett and Clark 2010; Yang et al. 2010). The limited
molecular variation may be causing problems with analyses
(i.e. tree topology estimation, branch support values, SH
tests, isolation by distance) by introducing sampling error.
Additionally, the ITS data should be viewed as exploratory because amplicons were not cloned and the sequences
exhibit low levels of sequence variation.
Supported Lineages—Subgenus Rettbergia is a well-supported
lineage that is primarily distributed in Brazil but also
includes the northern Andean C. arachniformis. This species
was only recently described from Colombia (Clark and
Londoño 1998) and exhibits the infravaginal branching,
spatheate capitate inflorescences, and connate lemma tips
characteristic of subg. Rettbergia species. Kelchner and Clark
(1997) sampled the rpl16 intron for this species (as Chusquea
[Volume 39
sp. A) and placed it well within the subg. Rettbergia clade.
Several other Chusquea species were included in subg.
Rettbergia until recently: C. anelythra Nees, C. anelytroides
Rupr. ex Döll, C. barbata L. G. Clark, C. pallida Munro,
C. ramosissima, C. tenuiglumis Döll, and C. wilkesii Munro
(Clark 1993; Judziewicz et al. 1999). This study sampled
C. ramosissima and found no support for its inclusion in
subg. Rettbergia in combined chloroplast MP, MLE, or BI
trees, and an SH test rejected inclusion of this taxon in
the subgenus. Chusquea ramosissima was also found to be
more closely related to Euchusquea than subg. Rettbergia
in a parsimony analysis of rpl16 intron data (Kelchner and
Clark 1997, Clark et al. 2007), but this result is contradicted by results of a morphological analysis (Clark et al.
2007) and also of analysis of the trnD-trnT intergenic
spacer (data not shown).
In this study, Euchusquea is monophyletic, but expands
to include members of the polyphyletic subg. Swallenochloa
and subg. Chusquea. This finding corroborates other evidence for Euchusquea monophyly (Kelchner and Clark
1997; Clark et al. 2007; Fisher et al. 2009), supports Clark’s
(1989) inclusion of Swallenochloa within Chusquea, and
supports Clark’s (1997) hypothesis of a close relationship
between subg. Chusquea and subg. Swallenochloa. Potentially,
the Euchusquea clade could be designated a subgenus as it
is supported as monophyletic by molecular data; there are,
however, no obvious morphological synapomorphies for
the clade (Clark 1997).
Relatively dense species sampling allowed us to detect a
number of clades within the Euchusquea clade that were
robust to changes in methodological framework and corroborate relationships found in other studies.
Chloroplast data supports the C. culeou clade (I, Fig. 3)
as a monophyletic lineage in Chusquea although species in
this clade are morphologically different. Most of the species
of Chusquea found in Chile have extravaginal branching and
are classified in subg. Chusquea sect. Chusquea. They are represented here by C. quila and C. uliginosa. A single species in
Argentina (C. deficiens) exhibits characteristics that place it
in sect. Verticillatae, but it was not sampled in this study.
The high-elevation C. montana is found in both Argentina
and Chile and is a shrubby member of subg. Swallenochloa
sect. Swallenochloa. Other austral Andean members of subg.
Swallenochloa are C. culeou, C. gigantea, and C. andina (the
C. culeou group); they exhibit overlapping variation in overall height and branch dimorphism, which is considered to
be a result of elevation, latitude, and shade effects (Veblen
1982; Pearson et al. 1994; Triplett and Clark 2003).
Both Kelchner and Clark (1997) and Clark et al. (2007)
sampled C. ramosissima, but neither study was able to
resolve its placement beyond inclusion in the Euchusquea
clade. Our results corroborate the exclusion of C. ramosissima
from subg. Rettbergia and instead suggest a close relationship
with the sympatric C. tenella (II, Fig. 3). Both C. tenella and
C. ramosissima exhibit a pseudopetiolate culm leaf blade that
often remains green after the plant has reached maturity.
They share this trait with C. tenuiglumis (not sampled).
Species in the C. uniflora clade (III, Fig. 3) are from lowelevation montane or cloud forest habitats in the northern
Andes and Central America and are classified in subg.
Chusquea sect. Verticillatae, sect. Longifoliae, sect. Longiprophyllae,
and subg. Swallenochloa sect. Swallenochloa. This clade has been
sampled in previous studies (represented by C. exasperata;
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
Kelchner and Clark 1997; Clark et al. 2007), but its position
within Euchusquea has not been resolved.
The C. serpens clade (IV, Fig. 3) contains species from the
northern Andes and Central America that are classified in
subg. Chusquea sect. Longiprophyllae, sect. Longifoliae, and
sect. Serpentes. There is strong support for monophyly of
the clade, but no apparent morphological synapomorphy.
Chusquea patens and C. pohlii are both found only in Costa
Rica and Panama, while C. serpens extends into Colombia,
Ecuador, and Venezuela and C. maculata is found in Colombia
and Venezuela. Given the relationships of these species in
the phylogeny estimation and their distributions, this clade
may represent a dispersal of Chusquea from Central America
to the Andes.
Species in the C. scandens clade (V, Fig. 3) are drawn
from subg. Swallenochloa sect. Swallenochloa, subg. Chusquea
sect. Chusquea, sect. Longifoliae, sect. Verticillatae, and sect.
Serpentes. These species are found in the northern Andes
and Central America, with the exception of C. lorentziana
(Argentina and Bolivia) and the extended distribution of
C. foliosa and C. longifolia into Mexico. The close relationship between C. foliosa and C. longifolia is interesting in that
both species are classified in subg. Chusquea sect. Longifoliae
and are found in oak cloud forests. The C. scandens clade
also contains a group of closely related species with a center
of diversity in Costa Rica: C. costaricensis, C. talamancensis,
C. tonduzii, C. paludicola, C. vulcanalis, C. foliosa, C. longifolia,
C. subtilis, C. tomentosa, and C. subtessellata.
Several other interesting relationships are worth discussing, but are in need of further study. This is the first
molecular evidence for the placement of C. abietifolia
within Euchusquea (Fig. 3). Chusquea abietifolia is the only
Chusquea species found in the West Indies (with collections
from Jamaica, Hispaniola, Cuba, and Puerto Rico; Tropicos).
The BI and ML analyses place it (unsupported) as sister
to the C. uniflora group + the remainder of Euchusquea,
while the MP analysis places it as sister to the C. uniflora
group (unsupported). Chusquea abietifolia’s relationship within
Euchusquea might be difficult to estimate due to a long
branch (13 changes in this dataset) distinguishing it from
the Euchusquea backbone.
Our analysis also recovered an unsupported group of
Chusquea species that are primarily found in Brazil (Fig. 3).
Resolution within this group is also unsupported and is
limited to a sister relationship between C. nudiramea L. G.
Clark and C. windischii L. G. Clark and a lineage containing
C. gracilis, C. juergensii Hack., C. leptophylla Nees, C. mimosa
McClure & L. B. Sm., and C. pinifolia. The majority of these
species are restricted to the South Atlantic forest (Espı́rito
Santo to southern Santa Catarina), although C. leptophylla
also occurs in Minas Gerais (Tropicos). C. juergensii, C. mimosa,
and C. pinifolia are fairly widely distributed, while
C. nudiramea, C. windischii, and C. microphylla are narrow
endemics (Clark 1997). Three species in this group (C. pinifolia,
C. windischii, and C. microphylla) are found in campos de altitude habitats (Clark 1992). Chusquea juergensii and C. mimosa
australis are often associated with Araucaria formations.
Chusquea Evolution—One explanation for the small
amount of variation found in Euchusquea chloroplast and
nuclear sequences is that Chusquea exhibits slow evolutionary rates, possibly compounded by long generation
times (Wilson et al. 1990; Gaut et al. 1997). Yet, chloroplast
substitution rates in Chusquea are faster than those seen in
most bamboo lineages, with the exception of the herbaceous
bamboos (Kelchner and Bamboo Phylogeny Group 2013).
Low sequence variation within Euchusquea may also stem
from recent origin of the clade and a lack of time to incur
substitutions (Ruiz-Sanchez 2011). We prefer a recent origin
explanation based on the small amounts of variation found
in both the chloroplast and ITS datasets and the geographic
patterning of Euchusquea species.
If species in Euchusquea are actively radiating, then
our molecular datasets might be complicated by processes
that are common in recently diverged lineages, including
incomplete lineage sorting and hybridization that result
in chloroplast sharing. Hybridization and incomplete lineage sorting have been discussed in many phylogenetic
studies (Maddison and Knowles 2006; Rokas and Carroll
2006; Holland et al. 2008; Degnan and Rosenberg 2009;
King and Roalson 2009; Polihronakis 2010) because these
non-treelike processes of genetic inheritance can result in
erroneous inferences about the evolution of a group when
tree diagrams are used for such data in phylogenetic analyses
(Huson and Bryant 2006; Bapteste et al. 2013). These processes could lead to a case where tree-based estimates of
DNA phylogeny do not predict morphology. This might be
occurring in Euchusquea, where the chloroplast topology
contradicts the current taxonomic classification based on
hypotheses of morphological evolution (although morphological characters may also evolve in a manner that leads to
a homoplasious interpretation of phylogeny). For example,
the sampled species in subg. Chusquea sect. Longiprophyllae
have prophylls as long as 10 cm, buds arranged in a tight
circle at the node, relatively narrow, lanceolate leaf blades,
and a first glume that is twice as long as the second glume
(Fig. 3, Clark 1990); but the two species of sect. Longiprophyllae
sampled in this study (C. maculata and C. exasperata) are
not sister. Instead, our chloroplast DNA topology places
C. maculata as sister to C. serpens, a species with no elongation of the prophyll, buds arranged in a crescent at the
node, wide leaves, and subequal lower glumes. Chusquea
exasperata is closely related to C. maclurei L. G. Clark and
C. scabra according to our chloroplast estimate, although
the three species are different from each other in branching
pattern and leaf shape.
In another example of conflict between the chloroplast
topology and the taxonomic classification of Chusquea,
16 species of subg. Chusquea sect. Verticillatae exhibit distinctive thin leaf blades, dorsally compressed spikelets,
reduced first and second glumes, and often a verticillate
arrangement of buds (Fig. 1, Soderstrom and Calderón 1978a;
Clark 1989). In our analysis of chloroplast data, the sampled
members of sect. Verticillatae are recovered in several different Euchusquea clades and an SH test rejects a topology
that constrains these species to monophyly.
Hybridization has not often been implicated as an important process in bamboo evolution because of the infrequent flowering of most bamboo species (Clark et al. 1989).
Bamboos exhibit gregarious monocarpy, a reproductive
strategy known as semelparous, hapaxanthic, or plietesial
in other plant groups. Bamboos with this life cycle might
remain vegetative for decades before a flowering event
occurs (Janzen 1974) and the asynchronous nature of flowering among bamboo species is expected to decrease the
potential for hybridization. Recent molecular work, however,
has shown that hybridization was involved in the formation
SYSTEMATIC BOTANY
of temperate bamboo species in the genera Hibanobambusa
Maruy. & H. Okamura, Sasa Makino & Shibata, Sasamorpha
Nakai, and Pleioblastus Nakai (Triplett and Clark 2010).
Hybridization also seems to occur among species in
Arundinaria Michx. (Triplett et al. 2010) and Chimonocalamus
(Yang et al. 2013) and likely occurs among the Paleotropical
bamboos as well (Goh et al. 2011; Wong and Low 2011).
Within Chusquea, hybridization appears to have occurred
between frequently flowering species of subg. Swallenochloa
sect. Swallenochloa (Clark et al. 1989; Pohl 1991). Clark et al.
(1989) identified morphologically intermediate plants hypothesized to represent C. amistadensis L. G. Clark, Davidse & R. P.
Ellis x C. subtessellata and C. vulcanalis x C. subtessellata in
several populations in Panama and Costa Rica and C. spencei x
C. tessellata in Colombian populations. Based on low nucleotide sequence variation within Euchusquea, the presence of
morphological intermediates, a large number of sympatric
species in Central America and the northern Andes, frequent
flowering of some high-elevation species, and the conflict
between molecular and morphological based estimates of
relatedness, hybridization between Euchusquea species is
likely to occur more than previously thought, and merits further investigation. The Chusquea ITS tree topology (Fig. 4) is
not based on cloned sequences which unfortunately restricts
our ability to test hypotheses of hybridization in Euchusquea
in this study.
Although hybridization probably occurs between some
Euchusquea species, incomplete lineage sorting might also
affect tree-based chloroplast and nuclear phylogenetic estimates in Chusquea. Incomplete lineage sorting occurs during
the divergence of nascent species when descendent populations retain multiple ancestral haplotypes before alleles
coalesce to a single haplotype (Maddison 1997). Although
several studies have identified coalescent-based approaches
to investigate the effects of incomplete lineage sorting on phylogeny estimations of closely related species (Maddison and
Knowles 2006; Carstens and Knowles 2007; King and Roalson
2009; Polihronakis 2010), the current Chusquea dataset does
not include adequate independent loci nor the population
level sampling necessary to use these approaches effectively.
If the common ancestor of Euchusquea contained polymorphisms that have not yet had time to fix in diverging lineages, then one would expect to see a random pattern of
ancestral haplotypes in progeny species (Comes and Abbott
2001; Maddison and Knowles 2006), with no morphological
or geographic signal (Morando et al. 2004; Jabaily and Sytsma
2010). Chloroplast genetic relationships among Euchusquea
species seem to be largely unlinked from morphological traits
and a test of isolation by distance did not uncover a simple
pattern between chloroplast genetic relatedness and geographic
distance (Fig. 5). There are apparent geographic patterns in
Euchusquea chloroplast relationships, but those patterns are
complicated by multiple lineages co-occurring geographically.
Since Euchusquea may have recently undergone a species
radiation or may be radiating at present, it is reasonable to
infer that incomplete lineage sorting can partly explain the
observed incongruence between chloroplast and morphological hypotheses of relationships.
Contrasting with results of the isolation by distance analysis,
the phylogeny estimation for Euchusquea shows a pattern of
chloroplast lineages that correspond to geographic distribution. Considering the pattern present in the chloroplast analysis, it is possible that gene flow among Euchusquea species
[Volume 39
within a geographical region led to chloroplast sharing and
development of a regional chloroplast signal. There are few
supported relationships within clades (Fig. 3) and it is therefore possible that chloroplasts in extant Euchusquea have not
completed lineage sorting at the level of morphologically distinguishable species, but instead retain polymorphisms that
are shared with other species in their geographic area. The
geographic signal in the Euchusquea is most likely due to a
combination of the effects of a recent radiation of Euchusquea
species, incomplete lineage sorting, and hybridization-mediated
chloroplast sharing. Euchusquea is probably a young group
that is actively radiating in several geographic areas, particularly the northern Andes, Central America, and southeastern Brazil and reproductive isolation barriers may be
weak or non-existent.
Implications for Taxonomy—Species in Chusquea have
previously been placed in three subgenera, numerous sections, and informal groups based on unique combinations of
morphological characters. Molecular evidence supports the
existence of a monophyletic subg. Rettbergia, a Euchusquea
clade containing polyphyletic subg. Swallenochloa and subg.
Chusquea, and two early-diverging clades comprised of species formerly in the genus Neurolepis that have been informally referred to as Neurolepis I and Neurolepis II.
Chloroplast data suggest a consistent pattern of lineages
within Euchusquea, but most of these lineages are not strongly
supported and alternative sets of relationships, including morphologically defined taxonomic groups, cannot be ruled out
by SH tests. In addition, Euchusquea chloroplast lineages lack
obvious morphological synapomorphies that would serve in
a new subgeneric or sectional classification. There remains a
need to organize and reference the large amount of morphological variation present in Euchusquea for floristic, systematic, ecological, and physiological studies (Bortolus 2008).
Until there is compelling evidence for the existence of evolutionary lineages that are also morphologically definable,
we propose that maintaining the current subgeneric classification is the most practical and biologically relevant course in
Euchusquea, with the caveat that subg. Chusquea and subg.
Swallenochloa appear to be polyphyletic with chloroplast data.
In contrast to the lineages of the Euchusquea clade, earlydiverging lineages of Chusquea represent well-supported
primary divisions within the genus and several potential
morphological synapomorphies have been identified. We
propose elevating the Neurolepis I and Neurolepis II clades
to subgeneric rank. Symplesiomorphies for the Neurolepis
clades are a lack of aerial branching (an unusual character
shared by only two other woody bamboo genera, Glaziophyton
Franchet and Greslania Balansa) and inner ligules with evident nerves (Judziewicz et al. 1999). Morphological differences distinguishing the Neurolepis clades were identified in
Fisher et al. (2009), but the large-leaved C. nobilis was in a
seemingly incongruous placement in the otherwise smallleaved Neurolepis I clade. After re-examination, C. nobilis has
been found to exhibit characters such as awned glumes that
are consistent with Neurolepis I species, with the exception of
its large leaves.
The only previously published subgeneric name for Neurolepis
is subg. Platonia (Kunth) Nees. Nees von Esenbeck (1835)
created subg. Platonia on the occasion of submerging Platonia
Kunth (a basionym of Neurolepis and a later homonym of
Platonia Mart.) and other bamboo genera with one-flowered
spikelets in Chusquea. Following Kunth’s concept, the type
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
of subg. Platonia (Kunth) Nees is C. elata, a species resolved
within the Neurolepis I clade in the analyses of Fisher et al.
(2009). Therefore, subg. Platonia follows C. elata and is the
correct name of the Neurolepis I clade at the subgenus level.
As there are no additional subgeneric names for these clades,
Chusquea subg. Magnifoliae is published here as a subgenus
novum. The epithet describes the large leaves characteristic
of species in the Neurolepis II clade and correspondingly,
C. magnifolia L. G. Clark is the type species. The descriptions
below follow Fisher et al. (2009), with the addition of a distinguishing reproductive character (presence or absence of
awned glumes).
Chusquea subgenus Platonia Fisher & L. G. Clark, nom.
et stat. nov. Platonia Kunth, Rev. Gram 1: 139. 1829.,
non Raf. (1810), nec Mart. (1832).—TYPE: P. elata Kunth.
Neurolepis Meisner, Pl. Vasc. Gen. 1: 426. 1843.—TYPE:
N. elata (Kunth) Pilg.
Internodes 1–14 cm long, solid or hollow. Culm leaves
poorly or not differentiated. Bud complement, if developed, consisting of a single bud triangular in outline and
oriented vertically. Aerial branching absent. Foliage leaf
blades usually 8–164(-180) cm long and (0.6-)1–6.3(-8.3) cm
wide but 162–300(-400) cm long and 3.5–10(-12) cm wide
in C. nobilis; pseudopetioles absent to short (1–6 mm long),
but 1.5–17.5 cm in C. asymmetrica. Glumes usually with well
developed awns, sometimes acuminate to subulate. Lemma
with apex free.
Chusquea acuminatissima (Munro) L. G. Clark, C. aristata
Munro, C. asymmetrica (L. G. Clark) L. G. Clark, C. elata
(Kunth) L. G. Clark, C. fimbriligulata (L. G. Clark) L. G.
Clark, C. laegaardii (L. G. Clark) L. G. Clark, C. nana (L. G.
Clark), L. G. Clark, C. nobilis (Munro) L. G. Clark, C. rigida
(L. G. Clark) L. G. Clark, C. stuebelii (Pilg.) L. G. Clark,
C. villosa (L. G. Clark) L. G. Clark.
Chusquea subgenus Magnifoliae L. G. Clark & Fisher, subg.
nov.—TYPE: Chusquea magnifolia L. G. Clark.
Internodes 1–3 cm long, solid. Culm leaves poorly or
not differentiated. Bud complement, if developed, consisting
of a single bud triangular in outline and oriented vertically.
Aerial branching absent. Foliage leaf blades (20-)30 – 300
(-377) cm long and 1.2 – 24(-30) cm wide; pseudopetioles
well developed; lacking blade tissue, (1-)3– 50 cm long.
Glumes usually lacking awns. Lemma with apex free.
C. angusta (Swallen) L. G. Clark, C. cylindrica L. G. Clark,
C. diversiglumis (Soderstr.) L. G. Clark, C. glomerata (Swallen)
Dorr, C. magnifolia L. G. Clark, C. mollis (Swallen) L. G.
Clark, C. petiolata (Davidse & L. G. Clark) L. G. Clark,
C. silverstonei (Davidse & L. G. Clark) L. G. Clark, C. spectabilis
L. G. Clark, C. tovari L. G. Clark.
Acknowledgments. This research was completed as a partial fulfillment of a doctoral dissertation by Fisher at Idaho State University.
Funding for research supplies and sequencing was primarily through
NSF Grant DEB-0515818 to Kelchner and additional funding was provided by the Idaho State University Molecular Research Core Facility
and NSF Grant DEB-0515712 to Clark. This work would not have been
completed without the help of the staff of the Costa Rica National
Museum and INBio, specifically Nelson Zamora and Alvaro Herrera,
for assistance with permits, access to herbaria, and logistical help in the
field. Chris Tyrrell collaborated on fieldwork and Mayra Montiel Longhi
gave logistical support in Costa Rica. Fieldwork in Costa Rica (2008) by
Fisher and Clark was supported by NSF Grant DEB-0515818 to Kelchner
and DEB-0515712 to Clark, respectively. Rusty Russell and Paul Peterson
facilitated Fisher’s access to Chusquea specimens at US. Jimmy Triplett,
Chris Tyrrell, Amy Denton, and staff at the Idaho State University Molecular Research Core Facility offered important advice on molecular protocols. Idaho State University undergraduates Alexandra Meier, Wendy
Newbold, Tsedey Kassaye, and Blackfoot High School (Blackfoot, Idaho)
student Paige Casperson assisted Fisher with labwork related to this project.
Lucinda McDade and committee members Nancy Huntly and Michael
Thomas made helpful comments on earlier drafts of the manuscript.
Literature Cited
Alvarez, I. and J. F. Wendel. 2003. Ribosomal ITS sequences and plant
phylogenetic inference. Molecular Phylogenetics and Evolution 29:
417–434.
Bapteste, E., L. van Iersel, A. Janke, S. Kelchner, S. Kelk, J. O. McInerney,
D. A. Morrison, L. Nakhleh, M. Steel, and L. Stougie. 2013. Networks:
Expanding evolutionary thinking. Trends in Genetics 29: 439–441.
Borsch, T., K. W. Hilu, D. Quandt, V. Wilde, C. Neinhuis, and W. Barthlott.
2003. Noncoding plastid trnT-trnF sequences reveal a well resolved
phylogeny of basal angiosperms. Journal of Evolutionary Biology
16: 558–576.
Bortolus, A. 2008. Error cascades in the biological sciences: the unwanted
consequences of using bad taxonomy in ecology. AMBIO: A Journal
of the Human Environment 37: 114–118.
Carstens, B. C. and L. L. Knowles. 2007. Estimating species phylogeny from gene-tree probabilities despite incomplete lineage
sorting: an example from Melanoplus grasshoppers. Systematic Biology
56: 400–411.
Clark, L. G. 1989. Systematics of Chusquea section Swallenochloa, section
Verticillatae, section Serpentes, and section Longifoliae (PoaceaeBambusoideae). Systematic Botany Monographs 27: 1–127.
Clark, L. G. 1990. Chusquea sect. Longiprophyllae (Poaceae: Bambusoideae):
A new Andean section and new species. Systematic Botany 15:
617–634.
Clark, L. G. 1992. Chusquea sect. Swallenochloa (Poaceae: Bambusoideae)
and allies in Brazil. Brittonia 44: 387–422.
Clark, L. G. 1993. Five new species of Chusquea (Poaceae: Bambusoideae)
and a new combination. Novon 3: 228–238.
Clark, L. G. 1996. Six new species of Neurolepis (Poaceae: Bambusoideae:
Bambuseae) from Ecuador and Peru. Novon 6: 335–350.
Clark, L. G. 1997. Diversity, biogeography and evolution of Chusquea.
P. 33–44 in The bamboos. ed. G. P. Chapman. London: Academic Press.
Clark, L. G. and X. Londoño. 1998. Eight new taxa and two new reports
of Bambuseae (Poaceae: Bambusoideae) from Colombia. Novon 8:
408–428.
Clark, L. G., G. Davidse, and R. P. Ellis. 1989. Natural hybridization
in bamboos: Evidence from Chusquea sect. Swallenochloa (Poaceae:
Bambusoideae). National Geographic Research 5: 459–476.
Clark, L. G., S. Dransfield, J. K. Triplett, and J. G. Sanchez-Ken. 2007.
Phylogenetic relationships among the one-flowered, determinate
genera of Bambuseae (Poaceae: Bambusoideae). Aliso 23: 315–332.
Comes, H. P. and R. J. Abbott. 2001. Molecular phylogeography, reticulation, and lineage sorting in Mediterranean Senecio sect. Senecio
(Asteraceae). Evolution 55: 1943–1962.
Degnan, J. H. and N. A. Rosenberg. 2009. Gene tree discordance, phylogenetic inference and the multispecies coalescent. Trends in Ecology &
Evolution 24: 332–340.
Demesure, B., N. Sodzi, and R. J. Petit. 1995. A set of universal primers
for amplification of polymorphic noncoding regions of mitochondrial and chloroplast DNA in plants. Molecular Ecology 4: 129–131.
Desvaux, A. N. 1831. Mémoires de la Société d’Agriculture. Sciences et Arts
d’Angers 1: 211.
Don, R. H., P. T. Cox, B. J. Wainwright, K. Baker, and J. S. Mattick. 1991.
’Touchdown’ PCR to circumvent spurious priming during gene
amplification. Nucleic Acids Research 19: 4008.
Drummond, A. and A. Rambaut. 2007. BEAST: Bayesian evolutionary
analysis by sampling trees. BMC Evolutionary Biology 7: 214.
Ersts, P. J. 2009. Geographic distance matrix generator, v. 1.2.3. New York:
American Museum of Natural History, Center for Biodiversity and
Conservation. Available from http://biodiversityinformatics.amnh
.org/open_source/gdmg.
Feliner, G. N. and J. A. Rosselló. 2007. Better the Devil You Know?
Guidelines for insightful utilization of nrDNA ITS in species-level
evolutionary studies in plants. Molecular Phylogenetics and Evolution
44: 911–919.
SYSTEMATIC BOTANY
Fisher, A. E., J. K. Triplett, C. S. Ho, A. D. Schiller, K. A. Oltrogge, E. S.
Schroder, S. A. Kelchner, and L. G. Clark. 2009. Paraphyly in the
bamboo subtribe Chusqueinae (Poaceae: Bambusoideae) and a
revised infrageneric classification for Chusquea. Systematic Botany
34: 673–683.
Gaut, B. S., L. G. Clark, J. F. Wendel, and S. V. Muse. 1997. Comparisons
of the molecular evolutionary process at rbcL and ndhF in the grass
family (Poaceae). Molecular Biology and Evolution 14: 769–777.
Giordano, C. V., R. A. Sanchez, and A. T. Austin. 2009. Gregarious
bamboo flowering opens a window of opportunity for regeneration
in a temperate forest in Patagonia. The New Phytologist 181: 880–889.
Goh, W. L., S. Chandran, K. Kamiya, and K. M. Wong. 2011. A natural
hybrid between Dendrocalamus pendulus and Gigantochloa scortechinii
(Poaceae: Bambusoideae: Bambuseae) in Peninsular Malaysia. Gardens’.
Bulletin Singapore 62: 223–238.
Graham, S. W., P. A. Reeves, A. C. E. Burns, and R. G. Olmstead. 2000.
Microstructural changes in noncoding chloroplast DNA: Interpretation, evolution, and utility of indels and inversions in basal
Angiosperm phylogenetic inference. International Journal of Plant
Sciences 161: S83–S96.
Griekspoor, A. and T. Groothuis. 2006. 4Peaks, v. 1.7.2. Nucleobytes.com.
Guindon, S. and O. Gascuel. 2003. A simple, fast, and accurate algorithm
to estimate large phylogenies by maximum likelihood. Systematic
Biology 52: 696–704.
Holland, B. R., S. Benthin, P. J. Lockhart, V. Moulton, and K. T. Huber.
2008. Using supernetworks to distinguish hybridization from lineagesorting. BMC Evolutionary Biology 8: 202.
Holz, A. C. and T. T. Veblen. 2006. Tree regeneration responses to
Chusquea montana bamboo die-off in a subalpine Nothofagus forest
in the southern Andes. Journal of Vegetation Science 17: 19–28.
Hsiao, C., N. J. Chatterton, K. H. Asay, and K. B. Jensen. 1995. Molecular
phylogeny of the Pooideae (Poaceae) based on nuclear rDNA (ITS)
sequences. Theoretical and Applied Genetics 90: 389–398.
Huelsenbeck, J. P. and F. Ronquist. 2001. MRBAYES: Bayesian inference
of phylogeny. Bioinformatics 17: 754–755.
Huelsenbeck, J. P. and B. Rannala. 2004. Frequentist properties of
Bayesian posterior probabilities of phylogenetic trees under simple
and complex substitution models. Systematic Biology 53: 904–913.
Huson, D. H. and D. Bryant. 2006. Application of phylogenetic networks
in evolutionary studies. Molecular Biology and Evolution 23: 254–267.
Jabaily, R. S. and K. J. Sytsma. 2010. Phylogenetics of Puya (Bromeliaceae):
Placement, major lineages, and evolution of Chilean species. American
Journal of Botany 97: 337–356.
Janzen, D. H. 1974. Why bamboos wait so long to flower. Annual Review
of Ecology and Systematics 7: 347–391.
Judziewicz, E. J., L. G. Clark, X. Londoño, and M. J. Stern. 1999. American
bamboos. Washington, D. C.: Smithsonian Institution.
Kelchner, S. A. 2000. The evolution of non-coding chloroplast DNA and
its application in plant systematics. Annals of the Missouri Botanical
Garden 87: 482–498.
Kelchner, S. A. 2003. Phylogenetic structure, biogeography, and evolution
of the Myoporaceae. Ph. D. thesis, Canberra: The Australian National
University.
Kelchner, S. A. and L. G. Clark. 1997. Molecular evolution and phylogenetic utility of the chloroplast rpl16 intron in Chusquea and
the Bambusoideae (Poaceae). Molecular Phylogenetics and Evolution
8: 385–397.
Kelchner, S. A. and Bamboo Phylogeny Group. 2013. Higher level phylogenetic relationships within the bamboos (Poaceae: Bambusoideae)
based on five plastid markers. Molecular Phylogenetics and Evolution
67: 404.
King, M. G. and E. H. Roalson. 2009. Discordance between phylogenetics
and coalescent based divergence modelling: exploring phylogeographic patterns of speciation in the Carex macrocephala species
complex. Molecular Ecology 18: 468–482.
Kunth, K. S. 1816. Pp. 201 in Nova Genera et Species Plantarum 1, Ed. 4,
eds. F. W. H. A. Humboldt, A. J. A. Bonpland, and K. S. Kunth.
Paris: Graeco-Latino-Germanico.
Laegaard, S. 1992. Influence of fire in the grass paramo vegetation of
Ecuador. Pp. 151–170 in Paramos: An Andean ecosystem under human
influence, ed. H. Balslev and J. L. Luteyn. San Diego, California:
Harcourt Brace Jovanovich Publishers.
Londoño, X. 1996. Diversity and distribution of new world bamboos,
with special emphasis on the Bambuseae. INBAR Working Paper 8.
New Delhi: INBAR.
Maddison, W. P. 1997. Gene trees in species trees. Systematic Biology
46: 523.
[Volume 39
Maddison, W. P. and L. L. Knowles. 2006. Inferring phylogeny despite
incomplete lineage sorting. Systematic Biology 55: 21–30.
McClure, F. A. 1973. Genera of bamboos native to the new world (Gramineae:
Bambusoideae), vol. 9. Washington, D. C.: Smithsonian Institution Press.
Morando, M., L. J. Avila, J. Baker, and J. W. J. Sites. 2004. Phylogeny
and phylogeography of the Liolaemus darwinii complex (Squamata:
Liolaemidae): Evidence for introgression and incomplete lineage
sorting. Evolution 58: 842–861.
Morrison, D. A. 2010. Using data-display networks for exploratory
data analysis in phylogenetic studies. Molecular Biology and Evolution 27: 1044–1057.
Mort, M. E., D. E. Soltis, P. S. Soltis, and M. L. Mabry. 2000. Comparison
of three methods for estimating internal support on phylogenetic
trees. Systematic Biology 49: 160–171.
Munro, W. 1868. Monograph of the Bambusaceae. Transactions of the
Linnean Society 26: 52–73.
Nees von Esenbeck, C. 1835. Bambuseae Brasilienses: Recensuit, et Alias
in India Orientali Provenientes Adjecit. Linnaea 33: 1–308.
Olmstead, R. G. and J. A. Sweere. 1994. Combining data in phylogenetic
systematics: an empirical approach using three molecular data sets
in the Solanaceae. Systematic Biology 43: 467–481.
Pearson, A. K., O. P. Pearson, and I. Gomez-Gallopin. 1994. Biology of
the bamboo Chusquea culeou (Poaceae: Bambusoideae) in southern
Argentina. Vegetatio 111: 93–126.
Pohl, R. W. 1991. Blooming history of the Costa Rican bamboos. Revista
de Biologia Tropical 39: 111–124.
Polihronakis, M. 2010. The interface between phylogenetics and population genetics: investigating gene trees, species trees, and population dynamics in the Phyllophaga fraterna species group. Evolution
64: 1048–1062.
Posada, D. 2008. jModelTest: Phylogenetic model averaging. Molecular
Biology and Evolution 25: 1253–1256.
Prince, L. and W. Kress. 2006. Phylogeny and biogeography of the prayer
plant family: getting to the root problem in Marantaceae. Pp. 645–659
in Monocots: Comparative biology and evolution, ed. J. T. Columbus, E. A.
Friar, J. M. Porter, L. Prince, and M. G. Simpson. Claremont, California:
Rancho Santa Ana Botanic Garden.
Rambaut, A. 2001. Se-Al: Sequence alignment editor, v. 2.0a11. Oxford,
U. K.: University of Oxford. Available from: http://tree.bio.ed.ac
.uk/software/seal.
Rambaut, A. and A. J. Drummond. 2007. Tracer, v. 1.4. Oxford, U. K.:
University of Oxford. Available from: http://tree.bio.ed.ac.uk/
software/tracer.
Rokas, A. and S. B. Carroll. 2006. Bushes in the tree of life. PLoS Biology
4: e352.
Ruiz-Sanchez, E. 2011. Biogeography and divergence time estimates
of woody bamboos: Insights in the evolution of the Neotropical
bamboos. Boletı́n de la Sociedad Botánica de México 88: 67–75.
Safford, H. D. 1999. Brazilian páramos I. An introduction to the physical
environment and vegetation of the campos de altitude. Journal of
Biogeography 26: 693–712.
Shaw, J., E. B. Lickey, E. E. Schilling, and R. L. Small. 2007. Comparison of
whole chloroplast genome sequences to choose noncoding regions
for phylogenetic studies in angiosperms: The tortoise and the hare
III. American Journal of Botany 94: 275–288.
Shaw, J., E. B. Lickey, J. T. Beck, S. B. Farmer, W. Liu, J. Miller, K. C.
Siripun, C. T. Winder, E. E. Schilling, and R. L. Small. 2005. The
tortoise and the hare II: Relative utility of 21 noncoding chloroplast
DNA sequences for phylogenetic analysis. American Journal of Botany
92: 142–166.
Shimodaira, H. and M. Hasegawa. 1999. Multiple comparisons of loglikelihoods with applications to phylogenetic inference. Molecular
Biology and Evolution 16: 1114–1116.
Slatkin, M. 1993. Isolation by distance in equilibrium and non-equilibrium
populations. Evolution 47: 264–279.
Soderstrom, T. R. and C. E. Calderón. 1978a. The species of Chusquea
(Poaceae: Bambusoideae) with verticillate buds. Brittonia 30: 154–164.
Soderstrom, T. R. and C. E. Calderón. 1978b. Chusquea and Swallenochloa
(Poaceae: Bambusoideae): Generic relationships and new species.
Brittonia 30: 297–312.
Sullivan, J. 2005. Maximum likelihood methods for phylogeny estimation. Methods in Enzymology 395: 757–779.
Sungkaew, S., C. M. Stapleton, N. Salamin, and T. R. Hodkinson. 2009.
Non-monophyly of the woody bamboos (Bambuseae; Poaceae):
A multi-gene region phylogenetic analysis of Bambusoideae s. s.
Journal of Plant Research 122: 95–108.
2014]
FISHER ET AL.: PHYLOGENY ESTIMATION OF CHUSQUEA
Swofford, D. L. 2003. PAUP* Phylogenetic analysis using parsimony
(*and other methods), v. 4.0 beta 10. Sunderland: Sinauer Associates.
Taberlet, P., L. Gielly, G. Pautou, and J. Bouvet. 1991. Universal primers
for amplification of three non-coding regions of chloroplast DNA.
Plant Molecular Biology 17: 1105–1109.
Tol, G. J. and A. M. Cleef. 1994. Above-ground biomass structure of
a Chusquea tessellata bamboo páramo, Chingaza National Park,
Cordillera Oriental, Colombia. Vegetatio 115: 29– 39.
Triplett, J. K. and L. G. Clark. 2003. Ambiguity and an American bamboo:
The Chusquea culeou species complex. Bamboo Science and Culture 17:
21–27.
Triplett, J. K. and L. G. Clark. 2010. Phylogeny of the temperate bamboos (Poaceae: Bambusoideae: Bambuseae) with an emphasis on
Arundinaria and allies. Systematic Botany 35: 102–120.
Triplett, J. K., K. A. Oltrogge, and L. G. Clark. 2010. Phylogenetic relationships and natural hybridization among the North American
woody bamboos (Poaceae: Bambusoideae: Arundinaria). American
Journal of Botany 97: 471–492.
Veblen, T. T. 1982. Growth patterns of Chusquea bamboos in the understory of Chilean Nothofagus forests and their influences in forest
dynamics. Bulletin of the Torrey Botanical Club 109: 474–487.
Watts, C. D., A. E. Fisher, C. D. Shrum, W. L. Newbold, S. Hansen, C. Liu,
and S. A. Kelchner. 2008. The D4 set: Primers that target highly
variable intron loops in plant chloroplast genomes. Molecular Ecology
Resources 8: 1344–1347.
Widmer, Y. 1997. Life history of some Chusquea species in old-growth
oak forest in Costa Rica. Pp. 17–31 in The bamboos, ed. G. P. Chapman.
London: Linnean Society of London.
Wilson, M. A., B. Gaut, and M. T. Clegg. 1990. Chloroplast DNA
evolves slowly in the palm family (Arecaceae). Molecular Biology
and Evolution 7: 303–314.
Wong, K. M. and Y. W. Low. 2011. Hybrid zone characteristics of the
intergeneric hybrid bamboo x Gigantocalamus maplenensis (Poaceae:
Bambusoideae) in peninsular Malaysia. Gardens’. Bulletin Singapore
63: 375–383.
Yang, H. M., Y. X. Zhang, J. B. Yang, and D. Z. Li. 2013. The monophyly
of Chimonocalamus and conflicting gene trees in Arundinarieae
(Poaceae: Bambusoideae) inferred from four plastid and two nuclear
markers. Molecular Phylogenetics and Evolution 68: 340–356.
Yang, J. B., H. Q. Yang, D. Z. Li, K. M. Wong, and Y. M. Yang. 2010.
Phylogeny of Bambusa and its allies (Poaceae: Bambusoideae) inferred
from nuclear GBSSI gene and plastid psbA-trnH, rpl32-trnL and rps16
intron DNA sequences. Taxon 59: 1102–1110.
Zhang, W. 2000. Phylogeny of the grass family (Poaceae) from rpl16 intron
sequence data. Molecular Phylogenetics and Evolution 15: 135–146.
Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Research 31: 3406–3415.
Appendix 1. Taxon sampling, voucher information, and GenBank
accession information for specimens used in this study. All vouchers are
deposited at ISC. The information is listed as: taxon, voucher collection
number; GenBank accessions: ndhF gene, rpl16 intron, ndhA intron,
trnD-trnT intergenic spacer, trnT-trnL intergenic spacer, and ITS complex (AC = Andre Mauricio Carvalho; AF = Amanda Fisher; EG =
Elizabeth Gordon; ER = Eduardo Ruiz; GSK = Gabriel Sánchez-Ken;
FZ = Fernando Zuloaga; IM = I. Montecinos; JT = Jimmy Triplett;
LC = Lynn Clark; LSS= Luis Sergio Sarahyba; MS = Margaret
Stern; PA = Patricio Asimbaya; PT = Pedro Tenorio; PW = Paulo
Windisch; SK = Scot Kelchner; SL = Simon Laegaard; WO = Walter
de Oliviera; WZ = Weiping Zhang; XL = Ximena Londoño).
Rhipidocladum harmonicum (Parodi) McClure. LC & WO 1103;
KF945324, KF945624, KF945453, KF945223, KF945387,–. Otatea acuminata
(Munro) C. Calderón & Soderstr. LC et al. 1312; AF182350, AF133473,
KF945452, FJ751732, KC020559, KF945270. Guadua angustifolia Kunth.
LC & XL 931; FJ751641, FJ751664, KF945451, FJ751734, FJ644154,
AY993946. Chusquea aristata Munro. LC et al. 1084; FJ751654, AY912196,
KF945450, FJ751747, KF945386, KF945268. Chusquea spectabilis L. G.
Clark. XL & LC 905; AF182355, U62793, KF945449, FJ751752, KC020550,
KF945269. Chusquea abietifolia Griesebach. LC et al. 1509; KF945271,
KF945591, KF945388, KF945170, KF945325,–. Chusquea albilanata L. G.
Clark & Londoño. LC et al. 1125; KF945272, KF945592, KF945389,
KF945171, KF945326, KF945224. Chusquea andina Philippi. LC & ER 980;
KF945273, KF945593, KF945390, KF945172, KF945327, KF945225. Chusquea
antioquensis L. G. Clark & Londoño. LC & XL 1222; KF945274, U62783,
KF945391, KF945173, KF945328, KF945226. Chusquea arachniformis L. G.
Clark & Londoño. LC & XL 1228; FJ751646, U62787, KF945392, KF945329,
FJ751739,–. Chusquea bambusoides (Raddi) Hack. LC 1029; FJ751649,
AY912194, KF945393, FJ751742, KC020548,–. Chusquea bilimekii Fournier.
LC et al. 1311; KF945275, U54757, KF945394, KF945174, KF945330,–.
Chusquea bradei L. G.Clark. AC 4377; FJ751650, FJ751666, KF945395,
FJ751743, KF945331,–. Chusquea coronalis Soderstr. & C. Calderón. SK
19; FJ751642, U54759, KF945396, FJ751735, KF945332, KF945227. Chusquea
costaricensis L. G. Clark & R. March. SK 15; KF945276, JQ352318,
KF945397, KF945175, KF945333, KF945228. Chusquea culeou Desvaux.
LC & ER 976; KF945277, AY912195, KF945398, KF945176, KF945334,
KF945229. Chusquea cumingii Nees.— s. n. dried sample; KF945278,
KF945596, KF945399, KF945177, KF945335,–. Chusquea exasperata
L. G. Clark. LC et al. 1093; KF945279, U62784, KF945597, KF945400,
KF945178, KF945336,–. Chusquea aff. fendleri Munro. LC & XL 1220;
KF945280, U62780, KF945401, KF945179, KF945337, KF945230. Chusquea
foliosa L. G. Clark. AF 31; KF945281, JQ352306, KF945402, KF945180,
KF945338, KF945231. Chusquea gigantea Demoly. s. n. Brest, France;
KF945282, KF945598, KF945403, KF945181, KF945339, KF945232. Chusquea
glauca L. G. Clark. LC et al. 1310; KF945283, KF945599, KF945404,
KF945182, KF945340, KF945233. Chusquea gracilis McClure & Smith.
LC 1034; KF945284, KF945600, KF945405, KF945183, KF945341, KF945234.
Chusquea juergensii Hackel. LC & PW 1067; KF945285, U62792, KF945406,
KF945184, KF945342, KF945235. Chusquea latifolia L. G. Clark. LC &
XL 417;–,–,–,–,–, AF019788. Chusquea lehmannii Pilger. LC & PA 1386;
KF945286, JQ352307, KF945407, KF945185, KF945343, KF945236.
Chusquea leonardiorum L. G. Clark. LC et al. 1081; KF945287, JQ352308,–,
KF945186, KF945344, KF945237. Chusquea leptophylla Nees. LC &
XL 1043; KF945288, KF945602, available on request (<200b), KF945187,
KF945345,–. Chusquea liebmannii Fournier. AF 1; KF945289, KF945603,
KF945408, KF945188, KF945346, available on request (<200b). Chusquea
longifolia Swallen. AF 41; KF945290, JQ352309, KF945409, KF945189,
KF945347, KF945238. Chusquea longispiculata L. G. Clark. LC & XL
1026; KF945291, KF945604, KF945410, KF945190, KF945348, KF945239.
Chusquea lorentziana Grisebach. LC & FZ 1017; KF945292, JQ352310,
KF945411, KF945191, KF945349, KF945240. Chusquea loxensis L. G. Clark.
LC et al. 1114; KF945293, JQ352311, KF945412, KF945192, KF945350,
KF945241. Chusquea maclurei L. G. Clark. LC & SL 1127; KF945294,
KF945605, KF945413, KF945193, KF945351, KF945242. Chusquea maculata
L. G. Clark. LC & XL 1226; KF945295, JQ352312, KF945414, KF945194,
KF945352, KF945243. Chusquea mayrae Fisher, Tyrrell & L. G. Clark.
AF 32;–, JQ352313,–,–,–, KF945244. Chusquea microphylla (Doell in
Martius) L. G. Clark. s. n. airdried; KF945296, KF945606, KF945415,
KF945195, KF945353,–. Chusquea mimosa subsp. australis L. G. Clark.
LC & XL 1040; KF945297, KF945607, KF945416, KF945196, KF945354,–.
Chusquea montana R.A. Philippi. LC & ER 993; KF945298, KF945608,
KF945417, KF945197, KF945355,–. Chusquea muelleri Munro. LC et al.
1140; KF945299, KF945609, KF945418, KF945198, KF945356,–. Chusquea
nudiramea L. G. Clark. LC & PW 1068; KF945300, U62792, KF945419,
KF945199, KF945357, KF945245. Chusquea oligophylla Ruprecht. LC &
XL 1031; FJ751651, U62785, KF945420, FJ751744, KF945358,–. Chusquea
oxylepis (Hack.) Ekman. LC & PW 1069; FJ751647, U62786, KF945421,
FJ751740, KF945359, KF945246. Chusquea paludicola L. G. Clark. SK 11;
KF945301, JQ352320, KF945422, KF945200, KF945360, KF945247. Chusquea
patens L. G. Clark. AF 2; KF945302, JQ352314, KF945423, KF945201,
KF945361, KF945248. Chusquea perligulata (Pilger) McClure. LC et al.
1074; KF945303, JQ352315, KF945424, KF945202, KF945362,–. Chusquea
perotensis L. G. Clark, Cortés & Cházaro. LC & PT 952; KF945304,
KF945611, KF945425, JQ352304, KF945363, KF945249. Chusquea pinifolia
(Nees) Nees. LC & PW 1056; KF945305, U54756, KF945426, KF945204,
KF945364, KF945250. Chusquea pittieri Hackel. AF 35; KF945306,
KF945613, KF945427, KF945205, KF945365, KF945251. Chusquea pohlii
L. G. Clark. AF 6; KF945307, JQ352316, KF945428, KF945206, KF945366,
KF945252. Chusquea quila Kunth. LC & IM 970; KF945308, KF945614,
KF945429, KF945207, KF945367, KF945253. Chusquea ramosissima
Lindman. AC 4358; KF945309, AF133472, KF945430, KF945208, KF945368,–.
Chusquea robusta L. G. Clark & Losure. LC et al. 1132; KF945310,
JQ352317, KF945431, KF945209, KF945369, KF945254. Chusquea scabra
Soderstrom & Calderón. AF 10; KF945311, KF945615, KF945432, KF945210,
KF945370, KF945255. Chusquea scandens Kunth. LC & XL 1235; FJ751643,
U62781, KF945433, FJ751736, KC020549,–. Chusquea serpens L. G.Clark.
LC & XL 1253; FJ751645, U54754, KF945434, FJ751738, KF945371,
KF945256. Chusquea spencei Ernst. LC et al. 1269; KF945312, U62788,
KF945435, KF945211, KF945372, KF945257. Chusquea subtessellata
Hitchcock. SK 14; KF945313, U54753, KF945436, KF945212, KF945373,
KF945258. Chusquea subtilis Widmer & L. G. Clark.AF 19; KF945314,
JQ352322, KF945437, KF945213, KF945374, KF945259. Chusquea
SYSTEMATIC BOTANY
talamancensis Widmer & L. G. Clark. SK 16; KF945315, U62789,
KF945438, KF945214, KF945375, KF945260. Chusquea tenella Nees.
LC & XL 1032; KF945316, KF945618, KF945439, KF945215, KF945376,–.
Chusquea tessellata Munro. LC et al. 1267; FJ751644, U54752, KF945440,
FJ751737, KF945377, KF945261. Chusquea tomentosa Widmer & L. G.
Clark. AF 20; KF945317, JQ352321, KF945441, KF945216, KF945378,
KF945262. Chusquea tonduzii Hackel. SK 13; KF945318, JQ352319,
KF945442, KF945217, KF945379, KF945263. Chusquea uliginosa R.A.
Philippi. LC & ER 992; KF945319, KF945621, KF945443, KF945218,
[Volume 39
KF945380, KF945264. Chusquea uniflora Steudel. LC & PA 1387;
KF945320, KF945622, KF945444, KF945219, KF945381, KF945265.
Chusquea urelytra Hackel. LC & XL 1052; FJ751648, FJ751665,
KF945445, FJ751741, KF945382,–. Chusquea virgata Hackel.AF 34;
KF945321, KF945623, KF945446, KF945220, KF945383, KF945266.
Chusquea vulcanalis (Soderstrom & Calderón) L. G. Clark. SK 12;
KF945322, U62790, KF945447, KF945221, KF945384, KF945267. Chusquea
windischii L. G. Clark. LC & XL 1047; KF945323, U54755, KF945448,
KF945222, KF945385,–.