author's personal copy -...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Molecular phylogeny of Cypripedium (Orchidaceae: Cypripedioideae) inferredfrom multiple nuclear and chloroplast regions

Ji-hong Li a,h, Zhong-jian Liu b, Gerardo A. Salazar c, Peter Bernhardt d, Holger Perner e, Yukawa Tomohisa f,Xiao-hua Jin a, Shih-wen Chung g, Yi-bo Luo a,b,⇑a State Key Laboratory of Systematic and Evolutionary Botany, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, Chinab The National Orchid Conservation Center, Shenzhen 518114, Chinac Departamento de Botánica, Instituto de Biología, Universidad Nacional Autónoma de México, Apartado Postal 70-367, 04510 Mexico, D.F., Mexicod Department of Biology, St. Louis University, St. Louis, MO 63103, USAe Huanglong Administration of National Scenic Spot, Huanglong Sichuan, Chinaf Tsukuba Botanical Garden, National Museum of Nature and Science, 1-1, Amakubo 4, Tsukuba 305-0005, Japang Department of Life Science, National Taiwan University, 1, Roosevelt Road, Sec. 4, Taipei 106, Taiwanh Graduate University of Chinese Academy of Sciences, Beijing 100039, China

a r t i c l e i n f o

Article history:Received 25 November 2010Revised 3 June 2011Accepted 8 June 2011Available online 22 June 2011

Keywords:CypripediumcpDNAnrDNA ITSPhylogenyMonophyleticParaphyletic

a b s t r a c t

A molecular analysis was performed on 56 taxa in the orchid genus Cypripedium using nrDNA ITS and fivechloroplast regions (trnH-psbA, atpI-atpH, trnS-trnfM, trnL-F spacer, and the trnL intron). The genus Cyp-ripedium was confirmed as monophyletic. Our data provided strong support for monophyletic groupingof eight infrageneric sections (Subtropica, Obtusipetala, Trigonopedia, Sinopedilum, Bifolia, Flabelinervia,Arietinum, and Cypripedium) defined in earlier taxonomic treatments, and paraphyletic grouping of twosections (Irapeana and Retinervi). Within the genus Cypripedium, the first divergent lineage consisted oftwo Mesomaerican species, and subsequently the Cypripedium debile lineage from eastern Asia was split.Our study did not support the notion that two Asian species (Cypripedium subtropicum and Cypripediumsingchii) were closely related to either Mesoamerican Cypripedium irapeanum or North American Cypripe-dium californicum, as indicated by previous interpretations based on morphological evidences. In addi-tion, one pair of vicariant species, Cypripedium plectrochilum (eastern Asia) and Cypripedium arietinum(North America), unique to section Arietinum, was confirmed. Furthermore, within the monophyleticsection Cypripedium two previously recognized subsections, Cypripedium and Macrantha, were shownto be paraphyletic. Our results suggested that this section split into two groups based on distribution(North America vs. Eurasia) instead of such previously used, morphological traits as flower color, andthe shape of the lips (labellum) and lateral petals.

� 2011 Elsevier Inc. All rights reserved.

1. Introduction

The genus Cypripedium has a far wider distribution and morevegetative and floral variations than the four remaining genera inthe subfamily Cypripediodeae, Orchidaceae (van der Pijl andDodson, 1966; sensu Cribb, 1997). Cypripedium consists of approx-imately 50 species of terrestrial herbs found in woodland and mea-dow habitats, from sea level to middle-montane elevations. Thesespecies are distributed through subtropical to temperate latitudesof the Northern Hemisphere excluding northern Africa (Cribb,1997; Averyanov, 2000; Chen and Cribb, 2005; Perner, 2008). East-ern Asia and North America represent two centers of diversity for

this genus. At least 38 species occur in Eastern Asian region, andabout 30 species in southwestern China (Cribb, 1997; Chen andCribb, 2005). North America (including Mexico) with about 16 spe-cies, is the second and smaller center of diversity (Cribb, 1997).

Historically, the morphological boundaries of this genus arebased on only two diagnostic characters. Cypripedium spp. arethe only ‘‘slipper’’ orchids in the subfamily bearing both plicateleaves and unilocular ovaries with parietal placentation (Coxet al., 1997; Cribb, 1997). While the genus Cypripedium is regardedas monophyletic based on the morphological data (above) and DNAsequences (Cox et al., 1997; Cribb, 1997; Eccarius, 2009), the pre-vious molecular phylogenies of this genus have left a number ofunresolved questions due to incomplete samples. As fresh speci-men material of pivotal species were unavailable for previous phy-logenetic analyses such studies were based on unbalancedsamplings from eastern Asia compared to more comprehensivecollections derived from the Western Hemisphere and Western

1055-7903/$ - see front matter � 2011 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2011.06.006

⇑ Corresponding author at: State Key Laboratory of Systematic and EvolutionaryBotany, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China.Fax: +86 010 82590843.

E-mail address: [email protected] (Y.-b. Luo).

Molecular Phylogenetics and Evolution 61 (2011) 308–320

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier .com/locate /ympev


Europe (Cox et al., 1997). In particular, earlier studies lacked infor-mation on the DNA sequence analysis of a key species, Cypripediumsubtropicum (Cox et al., 1997; Eccarius, 2009). This enigmatic and,until recently, unavailable species, was known from a single collec-tion in Tibet, China. Habit, inflorescence and column structure in C.subtropicum resembled three Mesoamerican-South American taxa,Mexipedium xerophyticum, Cypripedium irapeanum and Selenipedi-um spp. (Rosso, 1996; Chen and Lang, 1986; Cribb, 1997). Conse-quently, the uncertain position of C. subtropicum within thephylogeny blurred borders between genera within the subfamilyCypripedioideae (Chen, 1983; Chen and Lang, 1986; Cox et al.,1997; Cribb, 1997). Indeed, the taxonomic borders of nearly everygenus described within this subfamily have been debated since thesecond half of the 19th century (Atwood, 1984). The positions ofeastern Asian Cypripedium debile and North American Cypripediumcalifornicum within the genus Cypripedium also varied betweenstudies (Cox et al., 1997; Eccarius, 2009).

We argue that this uncertainty regarding the precise relation-ship between C. subtropicum and C. irapeanum in previous studieslikely inhibited accurate reconstruction of phylogenetic relation-ships within the genus Cypripedium, and between Cypripediumand other genera within the subfamily, Cypripedioideae. We alsoargue that incomplete sampling of species from eastern Asia raisesquestions about earlier interpretations by taxonomists that someEurasian species are sister species of some North American specieswith parallel morphologies (Chen, 1983; Chen and Lang, 1986; Coxet al., 1997; Cribb, 1997).

For this study, we have focused on improved sampling of Cypri-pedium to facilitate a more accurate reconstruction of phylogeneticrelationships. We have included the first published sequences ofthe key species, C. subtropicum and its close ally Cypripedium sing-chii (Liu and Chen, 2009), and also collected specimens from 46 ofthe 50 described Cypripedium spp. (excluding known recurrent hy-brids), with an emphasis on expanded representation of easternAsian taxa. Based on sequences of five cpDNA regions (trnH-psbA,atpI-atpH, trnS-trnfM, trnL-F spacer, and the trnL intron) and theinternal transcribed spacer region of nuclear ribosomal DNA(nrDNA ITS), we provide a well-supported phylogenetic resolutionfor the intra-generic placement of C. subtropicum and C. singchii,and a robust generic phylogeny for Cypripedium within theCypripedioideae.

2. Materials and methods

2.1. Ingroup sampling and outgroup selection

The specimens analyzed in this study were selected to maxi-mize all infrageneric groups following the classification by Cribb(1997) and Perner (2008), and to represent the full distributionalrange of the genus Cypripedium in the Northern Hemisphere. Spec-imens were obtained from either cultivated or wild-collectedplants. Taxonomy and nomenclature followed Cribb (1997) andPerner (2008). We obtained 67 samples representing 46 of the cur-rently recognized 50 species excluding the rare Cypripedium dickin-sonianum, Cypripedium ludlowii, Cypripedium elegans andCypripedium cordigerum. Based on previous molecular analyses onorchid taxa (see Cox et al., 1997; Cameron et al., 1999; Chaseet al., 2003; Górniak et al., 2010), five species assigned to theremaining four genera in the subfamily Cypripedioideae (Mexipedi-um, Paphiopedium, Phragmipedium and Selenipedium) were also in-cluded in this study, and additionally six species now classified asmembers of the subfamilies Apostasioideae (Apostasia, Neuwiedia)and Vanilloideae (Vanilla s.s.) were used to represent outgroups. Alist of the taxa analyzed, including information on voucher speci-mens and GenBank accessions, is given in Table 1.

2.2. DNA extraction, PCR amplification, and sequencing

Total DNA was extracted from silica-gel dried leaves with theTIANamp plant DNA Kit (Tiangen, China) according to the instruc-tions of the manufacturer. The primers used for amplification andsequencing of each individual region were ITS5/ITS4 (Baldwin,1992) for ITS, c/d for the trnL intron and e/f for the trnL-F interge-neric spacer (Taberlet et al., 1991), trnH (Tate and Simpson, 2003)/psbA (Sang et al., 1997) for the trnH-psbA intergenic spacer, trnS/trnfM for the trnS-trnfM intergenic spacer (Demesure et al.,1995), and atpI/atpH for the atpI-atpH intergenic spacer (Shawet al., 2007). PCRs were performed in a total reaction (50 ll) con-taining 10 ng of template DNA, 5 ll of 10� reaction buffer withMgCl2, 50 mM of each dNTP, 0.5 U of Ex Taq DNA polymerase(Takara Biotechnology, Japan) and 0.2 mM of each primer (SangonBiotechnology, China). The thermal cycler programme consisted ofan initial denaturation step at 94 �C for 2 min, followed by 30 cy-cles of 30 s at 94 �C, 30 s at 52–55 �C (depending on the annealingtemperature of specific primers), 1 min at 72 �C and a final exten-sion at 72 �C for 10 min. The PCR products were purified by usingthe DNA Fragment Quick Purification/Recover Kit (Dingguo, China),following the manufacturer’s protocol prior to sequencing.Sequencing reactions were performed using the dye-terminatorcycle-sequencing ready-reaction kit following the manufacturer’sprotocol, and analyzed on an ABI 3730 DNA Sequencer (AppliedBiosystems, Foster City, CA, USA). Each fragment was sequencedfor both strands. For DNA sequences added newly in this study,their 50 and 30 ends were identified using Cypripediodeae orOrchidaceae sequences already available in Genebank (http://www.ncbi.nlm.nih.gov). New sequences have been deposited inGenBank under the Accession Numbers JF796853—JF797170,JF825972—JF825978, FR720327—FR720330, and FR851209—FR851227.

2.3. Phylogenetic analyses

All sequences were aligned using ClustalX v.1.83 (Thompsonet al., 1997). With the use of BioEdit (Hall, 1999) manual adjust-ments were made by inserting gaps to improve the alignments.The analyses excluded two difficult-to-align regions in ITS, repre-senting 51 sites, one poly T region and one difficult-to-align regionin trnL [UAA] 30 exon–trnF [GAA] intergenic spacer, encompassing84 positions, two difficult-to-align regions in trnH-psbA intergenicspacer, involving 168 sites, one poly A region in atpI-tpH intergenicspacer, including 33 sites, one difficult-to-align region in trnS-trnfMintergenic spacer, containing 20 positions. Some unavailable se-quences in combined data analyses were treated as missing. Se-quence alignments and Nexus formatted files are available uponrequest from the corresponding author.

The homogeneities across the five cpDNA fragments, andbetween nrDNA ITS data and the combined cpDNA dataset (trnL in-tron, trnL [UAA] 30 exon–trnF [GAA] intergenic spacer, and trnH-psbA, atpI-atpH, trnS-trnfM intergenic spacers) were tested usingthe incongruence length difference (ILD) test (Farris et al., 1995),as implemented in PAUP� v4.0b10 (Swofford, 2003). The ILD testwas conducted with 1000 replicates, each with 10 random additionsequence replicates, TBR branch swapping, and keeping no morethan 100 trees per random addition replicate. Following Cunning-ham (1997), a significance level of P = 0.01 was adopted for thistest. Also in the present study, congruence between datasets wasassessed by comparison of topology and support values of strictconsensus trees of data partitions. This ‘‘hard incongruence’’ testwas performed by directly comparing visually the support andresolution of each of the clades in the separate analyses with ahigher bootstrap percentage (BP) and posterior probability (PP)than BP > 75 and PP > 90 (Wiens, 1998; Norup et al., 2006).

J.-h. Li et al. / Molecular Phylogenetics and Evolution 61 (2011) 308–320 309

Author's personal copyTa

ble

1Li

stof

taxa

and

plan

tm

ater

ials

(cla

ssiS

ecti

onCy

prip

ediu

mfi

cati

onfo

llow

sCr

ibb,

1997

and

Pern

er,2

008)

.

Taxa

Vou

cher

sG

enB

ank

acce

ssio

nn

um

bers

ITS

trnH

-psb

Atr

nS-t

rnfM

atpI

-atp

Htr

nLin

tron

trnL

–Fsp

acer

Ingr

oup

Cypr

iped

ium

Sect

ion

subt

ropi

caC.

subt

ropi

cum

Jin

XH

N80

03(P

E),C

hin

aJF

7968

96JF

7969

34JF

7970

53JF

7971

17–c

JF79

6994

C.si

ngch

iiLi

uZJ

E880

0(N

OC

C),

Ch

ina

JF79

6897

JF79

6935

JF79

7054

JF79

7118

–cJF

7969

95C.

war

dii

LiJH

S610

1(N

OC

C),

Ch

ina

JF79

6924

–cJF

7970

98JF

7971

16JF

7968

53JF

7969

96

Sect

ion

Irap

eana

C.ca

lifor

nicu

mG

un

nla

ugs

onD

S114

4a ,USA

(Cu

lt.)

Z785

32b

JF79

6941

JF79

7060

JF79

7151

–cA

B05

6264

b

C.ir

apea

num

Sala

zar

GA

7234

(MEX

U),

Mex

ico

FR72

0328

FR85

1212

FR85

1227

FR85

1222

FR85

1217

C.m

olle

Sala

zar

GA

6883

(MEX

U),

Mex

ico

FR72

0327

FR85

1211

FR85

1226

FR85

1221

FR85

1216

Sect

ion

Cypr

iped

ium

Sect

ion

Cypr

iped

ium

subs

ecti

onCy

prip

ediu

mC.

calc

eolu

sZh

ang

YE8

044a ,C

hin

a(C

ult

.)Z7

8521

bJF

7969

82JF

7971

02JF

7971

40JF

7968

69A

Y55

7224

b

C.he

nryi

LiJH

S615

7(N

OC

C),

Ch

ina

JF79

6910

JF79

6956

JF79

7074

JF79

7150

JF79

6861

JF79

7011

C.se

gaw

iiG

un

nla

ugs

onD

S114

5a ,Tai

wan

(Cu

lt.)

Z785

20b

JF79

6974

JF79

7092

JF79

7135

JF79

6862

JF79

7012

C.sh

anxi

ense

Lin

gC

YS0

609

(PE)

,Ch

ina

JF79

6919

JF79

6975

JF79

7093

JF79

7136

JF79

6868

JF79

7018

C.pa

rvifl

orum

var.

parv

iflor

umTa

ylor

SLS0

516

(PE)

,USA

EF37

0089

bJF

7969

68JF

7970

86JF

7971

33JF

7968

63JF

7970

13C.

parv

iflor

umva

r.pu

besc

ens

Eden

sR

S013

3(P

E),U

SAZ7

8519

bJF

7969

40JF

7970

59JF

7971

31JF

7968

66JF

7970

16C.

cand

idum

Stee

leB

S129

9a ,USA

(Cu

lt.)

EF37

0092

bJF

7969

42JF

7970

61JF

7971

32JF

7968

64JF

7970

14C.

kent

ucki

ense

Gu

nn

lau

gson

DS1

146a ,U

SA(C

ult

.)JF

7969

11JF

7969

58JF

7970

76JF

7971

30JF

7968

70JF

7970

19C.

mon

tanu

mFr

ysta

kM

S061

3(P

E),U

SAEF

3700

91b

JF79

6965

JF79

7083

JF79

7129

JF79

6867

JF79

7017

C.fa

scio

latu

mLi

JHS6

156

(NO

CC

),C

hin

aZ7

8530

bJF

7969

47JF

7970

65JF

7971

45JF

7968

74JF

7970

23C.

farr

eri

LiJH

S618

1(N

OC

C),

Ch

ina

JF79

6904

JF79

6946

JF79

7064

JF79

7144

JF79

6873

JF79

7022

Sect

ion

Cypr

iped

ium

subs

ecti

onM

acra

ntha

C.m

acra

ntho

sLi

ng

CY

S903

2a(P

E),C

hin

aZ7

8522

bJF

7969

62JF

7970

80JF

7971

48JF

7968

77JF

7970

26C.

mac

rant

hos

aff.

Gu

nn

lau

gson

DS1

148a ,C

hin

a(C

ult

.)JF

7969

15JF

7969

66JF

7970

84JF

7971

49JF

7968

78JF

7970

27C.

fros

chii

Gu

nn

lau

gson

DS1

150a ,C

hin

a(C

ult

.)JF

7969

09JF

7969

54JF

7970

72JF

7971

47JF

7968

76JF

7970

25C.

tibe

ticu

mva

r.ti

beti

cum

LiJH

S613

2(N

OC

C),

Ch

ina

JF79

6922

JF79

6978

JF79

7096

JF79

7138

JF79

6881

JF79

7030

C.ti

beti

cum

var.

tibe

ticu

maf

f.G

un

nla

ugs

onD

S115

1a ,Ch

ina

(Cu

lt.)

JF79

6902

JF79

6944

JF79

7063

JF79

7143

JF79

6872

JF79

7021

C.ti

beti

cum

var.

corr

ugat

umG

un

nla

ugs

onD

S114

9a ,Ch

ina

(Cu

lt.)

JF79

6901

JF79

6943

JF79

7062

JF79

7142

JF79

6871

JF79

7020

C.ru

biti

nctu

mG

un

nla

ugs

onD

S115

2a ,Ch

ina

(Cu

lt.)

JF79

6918

JF79

6973

JF79

7091

JF79

7134

JF79

6879

JF79

7028

C.ca

lcic

ola

LiJH

S615

3(N

OC

C),

Ch

ina

JF79

6921

JF79

6977

JF79

7095

JF79

7137

JF79

6880

JF79

7029

C.yu

nnan

ense

LiJH

S660

1(N

OC

C),

Ch

ina

JF79

6925

–cJF

7970

99JF

7971

41JF

7968

83JF

7970

32C.

fran

chet

iiLi

JHS6

261

(NO

CC

),C

hin

aJF

7969

08JF

7969

53JF

7970

71JF

7971

46JF

7968

75JF

7970

24C.

him

alai

cum

Z785

23b

–c–c

–c–c

–c

Sect

ion

Enan

tiop

edilu

mC.

fasc

icul

atum

Kou

zes

RN

6160

(PE)

,USA

JF79

6927

JF79

6984

JF79

7104

JF79

7166

–cJF

7970

44

Sect

ion

Obt

usip

etal

aC.

flavu

mva

r.fla

vum

LiJH

S616

1(N

OC

C),

Ch

ina

Z785

17b

JF79

6948

JF79

7066

JF79

7160

JF79

6857

JF79

7005

C.fla

vum

var.

sem

ialb

aG

un

nla

ugs

onD

S115

4a ,Ch

ina

(Cu

lt.)

JF79

6906

JF79

6950

JF79

7068

JF79

7162

–cJF

7970

07C.

flavu

mva

r.lu

teum

Gu

nn

lau

gson

DS1

055a ,C

hin

a(C

ult

.)JF

7969

05JF

7969

49JF

7970

67JF

7971

61–c

JF79

7006

C.pa

sser

inum

Kev

anP

S062

2(P

E),C

anad

aZ7

8516

bJF

7969

69JF

7970

87JF

7971

63JF

7968

58JF

7970

08C.

regi

nae

form

are

gina

eSt

eele

BS1

158a ,U

SA(C

ult

.)Z7

8518

bJF

7969

70JF

7970

88JF

7971

64JF

7968

59JF

7970

09C.

regi

nae

form

aal

bola

bium

Stee

leB

S115

9a ,USA

(Cu

lt.)

JF79

6917

JF79

6971

JF79

7089

JF79

7165

JF79

6860

JF79

7010

Sect

ion

Aca

ulia

C.ac

aule

Gil

lD

ES0

810

(PE)

,USA

Z785

25b

JF79

6936

JF79

7055

JF79

7167

–cA

B05

6262

b

Sect

ion

Ret

iner

via

C.pa

lang

shan

ense

LiJH

S615

2(N

OC

C),

Ch

ina

JF79

6916

JF79

6967

JF79

7085

JF79

7152

JF79

6893

JF79

7045

C.de

bile

Zhon

gSW

S652

7(P

E),T

aiw

anJF

7969

29JF

7969

87JF

7971

07JF

7971

68–c

JF79

7048

310 J.-h. Li et al. / Molecular Phylogenetics and Evolution 61 (2011) 308–320


C.de

bile

LiJH

N61

60(P

E),C

hin

aJF

7969

26JF

7969

83JF

7971

03JF

7971

69–c

JF79

7046

C.de

bile

Tom

ohis

Y06

-6(T

NS)

,Jap

anJF

7969

28JF

7969

86JF

7971

06–c

–cJF

7970

47

Sect

ion

Bifo

liaC.

gutt

atum

LiJH

S662

0(N

OC

C),

Ch

ina

Z785

26b

JF79

6955

JF79

7073

JF79

7115

JF79

6894

JF79

7049

C.ya

tabe

anum

Tom

ohis

Ys.

n.(

TNS)

,Jap

anZ7

8527

bJF

7969

85JF

7971

05JF

7971

14–c

JF79

7050

Sect

ion

Flab

ellin

ervi

aC.

japo

nicu

mLi

JHE6

701

(NO

CC

),C

hin

aEF

3700

93b

JF79

6957

JF79

7075

JF79

7156

–cJF

7970

51C.

form

osan

umG

un

nla

ugs

onD

S101

0a ,Tai

wan

(Cu

lt.)

Z785

24b

JF79

6951

JF79

7069

JF79

7155

JF79

6890

JF79

7041

Sect

ion

Ari

etin

umC.

arie

tinu

mK

evan

PS6

611

(PE)

,Can

ada

JF79

6899

JF79

6938

JF79

7057

JF79

7153

JF79

6892

JF79

7043

C.pl

ectr

ochi

lum

LiJH

S630

1(N

OC

C),

Ch

ina

Z785

28b

JF79

6972

JF79

7090

JF79

7154

JF79

6891

JF79

7042

Sect

ion

Sino

pedi

lum

C.ba

rdol

phia

num

LiJH

S616

3(N

OC

C),

Ch

ina

JF79

6900

JF79

6939

JF79

7058

JF79

7119

–cJF

7970

02C.

bard

olph

ianu

mva

r.zh

ongd

iane

nse

LiJH

S610

7(N

OC

C),

Ch

ina

JF79

6933

JF79

6993

JF79

7113

JF79

7170

JF79

6895

JF79

7052

C.m

icra

nthu

mLi

JHS6

155

(NO

CC

),C

hin

aJF

7969

14JF

7969

64JF

7970

82JF

7971

20JF

7968

56JF

7970

04C.

forr

esti

iLi

JHS6

306

(NO

CC

),C

hin

aJF

7969

07JF

7969

52JF

7970

70JF

7971

21–c

JF79

7003

Sect

ion

Trig

onop

edia

C.m

arga

rita

ceum

LiJH

S630

2(N

OC

C),

Ch

ina

Z785

30b

JF79

6963

JF79

7081

JF79

7127

–cJF

7969

98C.

lichi

ange

nse

LiJH

S630

3(N

OC

C),

Ch

ina

Z785

29b

JF79

6960

JF79

7078

JF79

7124

–cJF

7970

01C.

wum

enge

nse

LiJH

S610

2(N

OC

C),

Ch

ina

JF79

6913

JF79

6961

JF79

7079

JF79

7125

–cJF

7970

33C.

sich

uane

nse

LiJH

S615

4(N

OC

C),

Ch

ina

JF79

6920

JF79

6976

JF79

7094

JF79

7122

JF79

6854

JF79

6999

C.fa

rges

iiPe

rner

HE8

160

(PE)

,Ch

ina

JF79

6903

JF79

6945

–cJF

7971

26–c

JF79

6997

C.le

ntig

inos

umLi

JHS7

704

(NO

CC

),C

hin

aJF

7969

12JF

7969

59JF

7970

77JF

7971

23JF

7968

55JF

7970

00

Ou

tgro

up

Sele

nipe

dium

aequ

inoc

tial

eG

órn

iak

M.A

405

(PE)

,s.n

.JF

8259

77JF

8259

74JF

8259

76JF

8259

72JF

8259

73JF

8259

75Pa

phio

pedi

lum

mal

ipoe

nse

Liu

ZJE7

030a ,C

hin

a(C

ult

.)Z7

8498

bJF

7969

81JF

7971

01JF

7971

58JF

7968

85JF

7970

35Ph

ragm

iped

ium

schl

imii

Pern

erH

S703

2a ,Sou

thA

mer

ica

(Cu

lt.)

Z785

14b

JF79

6980

JF79

7100

JF79

7157

JF79

6884

JF79

7034

Ph.e

xsta

min

odiu

mSa

laza

rG

A74

99(M

EXU

),M

exic

oFR

7203

29FR

8512

09FR

8512

24FR

8512

19FR

8512

14M

exip

ediu

mxe

roph

ytic

umSa

laza

rG

A56

27(M

EXU

),M

exic

oFR

7203

30FR

8512

10FR

8512

25FR

8512

20FR

8512

15V

.pom

pona

Sala

zar

GA

7053

(AM

O),

Mex

ico

GQ

8672

37b

–cFR

8512

23FR

8512

18FR

8512

13V

anill

asi

amen

sis

Luo

YB

E107

0(P

E),C

hin

aJF

8259

78JF

7969

92JF

7971

12–c

–cJF

7970

40V

.she

nzhe

nens

eLi

uZJ

E108

0a ,Ch

ina

(Cu

lt.)

JF79

6930

JF79

6988

JF79

7108

–cJF

7968

86JF

7970

36A

post

asia

shen

zhen

ense

Liu

ZJE1

090a ,C

hin

a(C

ult

.)–c

JF79

6991

JF79

7111

–cJF

7968

89JF

7970

39N

euw

iedi

am

alip

oens

eLi

uZJ

E110

0a ,Ch

ina

(Cu

lt.)

JF79

6931

JF79

6989

JF79

7109

JF79

7159

JF79

6887

JF79

7037

N.s

inga

pure

ana

Liu

ZJE1

060a ,C

hin

a(C

ult

.)JF

7969

32JF

7969

90JF

7971

10–c

JF79

6888

JF79

7038

Not

es:

All

spec

imen

sw

ere

coll

ecte

din

the

fiel

dw

ith

the

foll

owin

gex

cept

ion

s.a

Priv

ate

gard

enof

gree

nh

ouse

coll

ecti

on.

bM

ater

ial

obta

ined

from

Gen

ban

k.c

Un

avai

labl

ese

quen

ces.



Phylogenetic analyses for each matrix were carried out usingmaximum parsimony (MP) and Bayesian inference (BI) methodsin PAUP� v.4.0b10 (Swofford, 2003) and MrBayes v.3.0b4 (Ronquistand Huelsenbeck, 2003), respectively. For MP analyses, heuristicsearches were conducted with 1000 random addition replicatesfollowed by tree bisection-reconnection branch swapping. Allcharacters were unordered and equally weighted with gaps treatedas missing data. Topological robustness was assessed by 1000bootstrap replicates. For BI analyses, each DNA region was assignedits own model of nucleotide substitution, as determined by theAkaike information criterion (AIC) in Modeltest v.3.06 (Posadaand Crandall, 1998). Four chains of the Markov chain Monte Carlowere run, sampling one tree every 1000 generations for 3000,000,starting with a random tree. Majority rule (>50%) consensus treeswere constructed after removing the burn-in period samples (thefirst 25% of sampled trees).

3. Results

3.1. Analysis of nuclear ITS sequence data

For the nrDNA ITS dataset, the numbers of variable and parsi-mony informative sites, tree statistics for the MP analysis and thebest-fit model determined by Modeltest are given in Table 2. TheBayesian phylogram is shown in Fig. 1 with BP and PP. The treetopologies generated by BI were approximately congruent withthose of MP analysis.

Monophyletic Cypripedium was strongly recognized with BP98/PP100. The Cypripedium spp. from eastern Asia and North Americaformed one clade with strong support (BP85/PP100). This cladewas sister to a much smaller clade comprised of only the two Mes-oamerican species, C. irapeanum and Cypripedium molle, with a sup-port of BP100/PP100. Among the 12 sections recognized by Cribb(1997) and Perner (2008), Subtropica, Obtusipetala, Trigonopedia,Sinopedilum, Bifolia, Flabelinervia, Arietinum, and Cypripedium weremonophyletic with well support values. In contrast, Irapeana, Ret-inervia, and two subsections in section Cypripedium (Cypripediumand Macrantha) were paraphyletic.

3.2. Analysis of chloroplast sequence data

For the homogeneity test, the trnH-psbA, trnS-trnfM, atpI-atpH,trnL [UAA] 30 exon–trnF [GAA] intergenic spacer, and trnL intron,contained 138, 97, 103, 75, and 55 parsimony informative sites,respectively. The P values resulting from the partition homogeneitytest are listed in Table 3. There was no significant incongruenceacross these five cpDNA fragments at P < 0.01. The trnL [UAA] 30

exon–trnF [GAA] intergenic spacer was shown to be highly congru-ent with trnS-trnfM, trnL intron, and trnH-psbA. Moreover, the vi-sual node-by-node comparisons, revealed no major incongruencefor all the nodes of topology of each cpDNA fragment (data not

shown) with bootstrap support and a posterior probability higherthan BP > 75 and PP > 90, respectively (Wiens, 1998; Norup et al.,2006). According to this congruence for these regions evolved asa single locus we combined the five cpDNA regions into the wholeanalysis.

For the combined cpDNA dataset, the numbers of variable andparsimony informative sites, tree statistics for the MP analysisand the best-fit model determined by Modeltest are presented inTable 2. The Bayesian phylogram is shown in Fig. 2 with BP andPP. The tree topologies generated by BI were consistent with thoseof MP analysis. Monophyletic Cypripedium was strongly supportedwith BP99/PP100. This topologies generated by chloroplast se-quences mostly accords to the cladograms on the basis of nrDNAITS.

3.3. Combined analysis

As maternally inherited cpDNA and bi-parentally inherited nu-clear DNA may have different evolutionary histories we analyzeddata partitions to examine the potential difference among gen-omes. The partition homogeneity test for the nrDNA ITS and cpDNAdatasets resulted in P = 0.01 (Table 3) implying incongruencebetween the two datasets. Moreover, the visual node-by-nodecomparisons revealed that the placements of Cypripedium margari-taceum, Cypripedium passerinum, Cypripedium calcicola, Cypripe-dium farreri, and Cypripedium fasiolatum, were incongruentbetween ITS and cpDNA topologies with bootstrap support and aposterior probability higher than BP > 75 and PP > 90 (Wiens,1998; Norup et al., 2006), respectively. Conflict among topologies,derived from analyzing different character sets, may be the resultof recurrent hybridization, incomplete lineage sorting and/or intro-gression. When these five conflict species were excluded, the par-tition homogeneity test for the nrDNA ITS and cpDNA datasetsshowed that P = 0.06 (Table 3). This indicated no significant incon-gruence between the two datasets. Therefore, we chose to combinethe two datasets excluding the five conflict species.

For this combined nrDNA ITS and cpDNA dataset, the numbersof variable and parsimony informative sites, tree statistics for theMP analysis and the best-fit model determined by Modeltest aregiven in Table 2. The Bayesian phylogram is shown in Fig. 3 withBP and PP. The tree topologies produced by BI were nearly thesame as those of MP analysis. The combined analysis resulted ingreater resolution as compared to the separate analyses. Moreover,when specimens representing the Apostasioideae and/or the Vanil-loideae were excluded as outgroups from the combined analyses,the results remained the same (data not shown), which indicatedthat the phylogeny of Cypripedium was stable and reliable in thisstudy.

The monophyletic status of the genus Cypripedium was con-firmed again and supported with BP98/PP100. All sections withinthe genus, except for paraphyletic Irapeana and Retinervia,

Table 2Analyses of datasets.

Information ITS trnH-psbA trnS-trnfM atpI-atpH trnL intron trnL–F spacer cpDNA Combined ITS and cpDNA(conflict taxa excluded)

No. ingroups 56 53 54 54 37 55 55 51No. outgroups 10 10 11 7 10 11 11 11Aligned length 806 1574 1212 1047 879 638 5350 6156No. variable characters 365 232 138 169 106 83 627 1017No. parsimony-informative characters 268 138 97 103 75 55 360 665Tree length (steps) 950 319 340 460 310 159 811 1907Consistency index (CI) 0.567 0.892 0.858 0.813 0.884 0.866 0.845 0.783Retention index (RI) 0.834 0.854 0.896 0.867 0.821 0.856 0.886 0.818Rescaled consistency index (RC) 0.473 0.703 0.777 0.705 0.733 0.656 0.749 0.571Model GTR + I + G GTR + I + G TIM + G TVM + G TVM + G K81uf + G TVM + I + G TIM + I + G



Fig. 1. Phylogram obtained from Bayesian inference analysis of the nrDNA ITS data. Numbers at the nodes are bootstrap percentages and Bayesian posterior probabilities(>50%), respectively. ‘‘–’’ indicates that the node was not supported in MP analysis. The section names in bold are indicated as monophyletic while two section names in grayare implied as paraphyletic. Within section Cypripedium, the taxa with an asterisk (�) represent subsection Cypripedium while the rest of the taxa remained members ofsubsection Macrantha. EA, NA, and MA represent eastern Asia, North America and Mesoamerica separately.



appeared monophyletic with strong support values. Within thegenus Cypripedium, the first divergent lineage consisted of C. ira-peanum and C. molle (endemic to Mesoamerica) with strong sup-port (BP100/PP100), and subsequently the C. debile lineage fromeastern Asia (BP100/PP100) was split. The remaining taxa whichwere distributed in eastern Asia and North America formed oneclade with a support of BP73/PP100. Cypripedium arietinum (NorthAmerica) and Cypripedium pletrochilum (eastern Asia) in the sectionArietinum (BP100/PP100) were sister species. Within monophyleticsection Subtropica from eastern Asia (BP100/PP100), C. subtropicumand its sister species, C. singchii, were the collective sister of Cypri-pedium wardii. Obviously, section Subtropica (eastern Asia) was notallied to the first divergent lineage composing of two Mesoameri-can species. Within monophyletic section Obtusipetala (BP100/PP100), all three variants of Cypripedium flavum (eastern Asia) weresister to the two color forms of the North American species, Cypri-pedium reginae. Two species found through Asia into the Alaskanpanhandle (Cypripedium guttatum and Cypripedium yatabeanum)comprised the monophyletic section Bifolia (BP100/PP100). Theeastern Asian Cypripedium palangshanense and North AmericanCypripedium fasciculatum (section Enantiopedilum) formed a cladewith weak support (BP50/PP77). Two eastern Asian monophyleticsections, consisting of Trigonopedia (BP97/PP100) with five taxaand Sinopedilum (BP100/PP100) with four taxa, were sister cladeswith strong support (BP100/PP100). Within section Trigonopedia,Cypripdium fargesii and Cypripedium sichuanense formed one clade(BP100/PP100). The second clade (BP100/PP100) was comprisedof Cypripdium lichiangense and Cypripdium wumengense as sisters(BP100/PP100) and Cypripdium lentiginosum. Within section Sino-pedilum, both variants of Cypripdium bardolphianum were includedin the same clade (BP100/PP100). The second clade (BP62/PP77)consisted of Cypripdium forrestii and Cypripdium micranthum. Twoeastern Asian species, Cypripdium formosanum and Cypripdiumjaponicum, formed the monophyletic section Bifolia (BP100/PP100).

Section Cypripedium including five North American taxa and 14Eurasian taxa was the most species rich section and representedthe most diversification in the genus. This lineage split into a dis-tinct North American vs. Eurasian group with a support of BP100/PP100 and BP68/PP97, respectively. Within North American taxa,Cypripedium kentuckiense and two variants of Cypripedium parviflo-rum were in the same clade with a support of BP94/PP100. Withinthe Eurasian species, Cypripedium segawai was sister to Cypripediumhenryi with a support of BP100/PP100, and C. shanxiense (easternAsian) was the sister species of Cypripedium calceolus (broadly Eur-asian) with a support of BP100/PP100. Cypripedium himalaicum wassister to the rest of the taxa with a support of BP85/PP99. Among therest of taxa, Cypripedium franchetii was sister to Cypripediumyunnanense with a support of BP96/PP100; Cypripedium froschii,Cypripedium rubintinctum and the three variants of Cypripediumtibeticum, formed a clade with a support of BP96/PP100.

4. Discussion

4.1. An updated phylogeny of the genus Cypripedium

In this study an updated phylogeny of Cypripedium was pro-posed with a more comprehensive sampling of taxa and a largerdataset of nrDNA ITS and five cpDNA fragments combined. As inprevious molecular and morphological phylogenies (Cox et al.,1997; Cribb, 1997; Eccarius, 2009), we reconfirmed that the genusCypripedium is monophyletic. Moreover, our results suggested thatthe sections within this genus require some reinterpretation andreorganization.

4.1.1. Section Irapeana CribbThe clade comprising two Mesoamerican species, C. irapeanum

and C. molle, is sister to the rest taxa of the genus Cypripedium(Fig. 1–3). The two taxa share the same vegetative habit, flowerlip morphology, purple spots on the involute lip margin andmultiple-flower inflorescences found in the slipper orchid genus,Selenipedium (Rosso, 1966; Chen and Lang, 1986; Cribb, 1997). At-wood (1984) suggested that Selenipedium graded into Cypripediumvia C. irapeanum, and postulated that reduction in the sheer sizeand number of vegetative organs and flower number was a recur-rent trend in the evolution of the subfamily Cypripediodeae due,most probably, to differences in latitude, elevation and climateover time. However, section Irapeana sensu Atwood (1984) is para-phyletic in this study, which is in accordance with two previousmolecular treatments based on nrDNA ITS sequences (Cox et al.,1997; Eccarius, 2009). C. californicum did not share the same cladewith C. irapeanum and C. molle as previous morphological studiesimplied (Cribb, 1997; Eccarius, 2009).

C. californicum is one of the most restricted species in its distri-bution, confined to the mountains of southwestern Oregon andnorthern California (Cribb, 1997). This species is far closer to sec-tion Cypripedium in our BI analysis (Fig. 3) and shares some recur-rent morphological features with many Eurasian and NorthAmerican species in section Cypripedium. This includes pubescenceon the outer surfaces of its sepals, tapering petals, and a convexand broad staminode (Perner, unpubl. data). Compared to thetwo, large, Mesoamerican species, C. californicum (North America)remains readily distinguishable by its multi-flowered raceme,much smaller flowers, complete absence of any yellow pigmenta-tion on any floral organ and synchronous flowering (Cribb, 1997)(see Plate 1).

C. dickinsonianum, a member of section Irapeana (Cribb, 1997),was not available for this study. This species is known from fiveor six small populations and is distinguished from both C. irapea-num and C. molle by its much smaller, mechanically self-pollinating(autogamous) flowers. The fact that C. dickinsonianum is found onlyin two disjunctive areas at the northern and southern limits of thedistribution of C. irapeanum (with which it is marginally sympatric)strongly suggests it may represent an autogamous form or variantof C. irapeanum (Soto and Solano, 2007). Once again, specimenmaterial must be obtained to elucidate relationships.

4.1.2. Sections Retinervia S.C. Chen and Enantiopedilum Pfitzer vs.section Irapeana Cribb

Three samples representing C. debile formed a clade (withBP100/PP100), which was a sister to all the other taxa from easternAsia and North America with strong support in this study (Fig. 3).In contrast, the topology provided by Eccarius (2009) found thatthe lineage including only species C. debile was firstly split fromthe genus Cypripedium with a support of BP74. C. debile is a locallycommon species in mainland China, Taiwan, and Japan (Cribb,1997). It occurs in shaded, montane sites under conifers but in

Table 3P values from partition-homogeneity tests.

Data set P value

trnH-psbA vs. trnS-trnfM 0.090trnH-psbA vs. atpI-atpH 0.100trnH-psbA vs. trnL intron 0.020trnH-psbA vs. trnL-F spacer 0.570trnS-trnfM vs. atpI-atpH 0.120trnS-trnfM vs. trnL intron 0.170trnS-trnfM vs. trnL-F spacer 0.690atpI-atpH vs. rnL intron 0.100atpI-atpH vs. trnL-F spacer 0.080trnL intron vs. trnL-F spacer 0.610ITS vs. cpDNA 0.01ITS vs. cpDNA (conflict taxa excluded) 0.06



Fig. 2. Phylogram obtained from Bayesian inference analysis of the cpDNA data. Numbers at the nodes are bootstrap percentages and Bayesian posterior probabilities (>50%),respectively. ‘‘–’’ indicates that the node was not supported in MP analysis. The section names in bold are indicated as monophyletic while two section names in gray areimplied as paraphyletic. Within section Cypripedium, the taxa with an asterisk (�) represent subsection Cypripedium while the remaining taxa are members of subsectionMacrantha. EA, NA, and MA represent eastern Asia, North America and Mesoamerica separately.



Fig. 3. Phylogram obtained from Bayesian inference analysis of the combined nrDNA ITS and cpDNA data excluded the conflict taxa. Numbers at the nodes are bootstrappercentages and Bayesian posterior probabilities (>50%), respectively. ‘‘–’’ indicates that the node was not supported in MP analysis. The section names in bolds are indicatedas monophyletic while two section names in gray are implied as paraphyletic. Within section Cypripedium, the taxa with an asterisk (�) represent subsection Cypripediumwhile the remaining taxa are members of subsection Macrantha. EA, NA, and MA represent eastern Asia, North America and Mesoamerica separately.



mainland China it may also be found in deciduous forests (Li, un-publ. data). Considering the highly disjunctive distribution of C. de-bile, Dressler (1993) would have classified it as a ‘‘Middle AgedHopper’’ indicative of long-distance dispersal in the mid-Tertiaryas this species is shared by three entirely, land-locked provinces(Gansu, Hubei and Sichuan) in China but is also relatively commonon the island of Taiwan, and the Japanese archipelago (Chen et al.,1999; Cribb, 1997). This species may be pivotal to understandingthe origin and colonization of the genus Cypripedium as C. debile(eastern Asia) split off from the rest of the North American (non-Mesoamerican) and eastern Asian Cypripedium species. As an earlydivergent lineage, C. debile failed to fit the evolutionary trendwithin the genus Cypripedium as stated by Atwood (1984) and

confirmed by Cox et al. (1997). These authors argued that evolutionin the genus Cypripedium proceeded from multi-flowered andmulti-leafed plants with tall stems to reduced stems with much re-duced leaf and flower numbers. C. debile lacks all of these ancestralcharacters, which still expressed in the two Mesoamerican species,C. irapeanum and C. molle, assigned to section Irapeana. Instead, C.debile has the derived short-stem, two-leaved condition with asmall, solitary flower terminating each scape. Based on morpholog-ical characters, C. debile more closely resembles taxa with reducedvegetative and reproductive characters as in sections Enantiopedi-lum, Trigonopedia, Sinopedilum and Arietinum.

In fact, while the lineage consisted of two Mesoamerican spe-cies (C. irapeanum and C. molle) and the C. debile lineage were both

Plate 1. The pictures of Cypripedium.



recovered as ‘‘primitive’’ lineages in this study neither lineage con-formed to earlier interpretations of the systematics of the genusCypripedium based solely on morphological/morphometric traits(see Chen and Lang, 1986; Cribb, 1997). We suspect that, part ofthe problem with interpretations based on morphological traits isthat many familiar floral traits, in particular, may be unsuitablein phylogenetic reconstructions because they may reflect littlemore than recent, short-term, shifts driven by pollinator-mediatedselection (Bateman et al., 1997; Cozzolino et al., 2001). Previousontogenetic studies (Tucker, 1997; Tucker and Bernhardt, 2000;Luo and Chen, 2000) showed repeatedly that floral traits associatedwith floral attractants (pigmentation patterns), rewards (nectarglands), anther dehiscence and stigmatic receptivity were the lastto develop in floral buds in large lineages. That is the most proba-ble reason why, as usual, some of the most frequently adoptedmorphological characters used in orchid taxonomy failed to paral-lel our molecular analyses throughout this study.

Note also that C. palangshanense (eastern Asia) and C. fascicula-tum (North America) formed a clade. Although this clade had weaksupport, the two species distributed on separate continents sharedsimilarities in the shapes of their labella, sepals, petals and stami-nodia (Cribb, 1997). Both C. fasciculatum and C. palangshanense lifttheir paired leaves above the ground. Their floral colors are similar,even though those of C. palangshanense are usually darker andmore reddish. In addition, C. palangshanense has a smaller habitwith a shorter, pubescent stem fewer flowers perinflorescence. C.elegans (eastern Asia), another tiny species placed previously with-in section Retinervia, shares similar traits in ovary, bract and leafwith C. palangshanense (Perner, unpubl. data). It was unfortunatethat live specimens of C. elegans were unavailable for this molecu-lar analysis as its position needs further exploration.

4.1.3. Section Arietinum C. MorrenThe monophyletic section Arietinum is nested within the Cypri-

pedium genus in this study. Therefore, we disagree with the earliertreatment by Atwood (1984) who promoted the two species totheir own genus based, in part, on pollen characters. The ratherthick imperforate lectum and granulate infratectal layer found inthe exine wall of C. arietinum is not unique and is also found inother Cypripedium species (Xi and Chen, 1991). Our molecular datasupport the section classification of Arietinum within Cypripediumas suggested by previous morphological analysis of Chen (1983).Despite their highly isolated distributions, C. arietinum (NorthAmerica) and Cypripedium plectrochilum (eastern Asia) sharedmany important traits. These include keeled labella. In addition,the lateral sepals of both species fail to fuse together, as in themajority of Cypripedium species. This lack of floral connation wasinterpreted as an ancestral character by Atwood (1984). Further-more, comparative photos or illustrations of these two vicariantspecies (Chen, 1983; Cribb, 1997) show only minor differences infloral characters such as pigmentation pattern and staminodeshape. Finally, it is interesting to note that both species appear tobe pollinated exclusively by bees in the same huge (c. 1100 spe-cies) and pandemic genus, Lasioglossum (Li et al., 2008).

4.1.4. Section Subtropica S.C. Chen and K.Y. LangThe so-called enigmatic C. subtropicum was not a sister taxon to

section Obtusipetala in this study. Neither was it closely allied tothe two Mesoamerican species in section Irapeana despite sharedmorphological traits (Chen and Lang, 1986; Cribb, 1997; Eccarius,2009). C. subtropicum also resembles in habit, inflorescence andcolumn structures such ‘‘primitive’’ slipper orchid genera as Mexip-edium and Selenipedium (Chen and Lang, 1986; Cribb, 1997), butthese plesiomorphic characters did not support its basal positionin the genus Cypripedium in our molecular analysis. Unlike thesetwo Neotropical genera, C. subtropicum has a unilocular ovary,

fruits that lack a vanilla-like scent and produces seeds that arefusiform and have a ‘‘tin’’ testa (Chen and Lang, 1986; Cribb,1997). C. subtropicum and its close relative C. singchii, were sisterto C. wardii in our analysis. In fact, all three exclusively easternAsian species in our section Subtropica share the most unusualcharacter of a vestigial and hooked staminode that is smaller thanthe stigma in the same flower (Cribb, 1997; Chen and Cribb, 2005).

4.1.5. Section Obtusipetala CribbC. passerinum was sister to C. flavum with strong support based

on the nrDNA ITS data (Fig. 1). In turn, C. flavum was sister toC. reginae with well support values inferred from the cpDNA data(Fig. 2). Therefore, it is questionable whether we should continueto regard C. flavum (eastern Asia) and C. reginae (North America)as a second example of an eastern Asian/North American, vicariantspecies pair, as in section Arietinum (see above), despite sharedvegetative and reproductive traits (Chen, 1983; Cribb, 1997). Basedon our unpublished field observations, we found specimens ofC. flavum with short-stems and small, whitish flowers reminiscentof C. passerinum. Its only the large-stemmed and large floweredvariants of C. flavum that resemble C. reginae.

4.1.6. Section Trigonopedia FranchWithin the monophyletic section Trigonopedia the species have

dramatic hairy-dark spots on their leaves and larger flowers thanthose found in section Sinopedilum (below). C. fargesii has longwhite hairs on its petals, and C. sichuanense is distinguished fromC. fargesii by its entirely glabrous sepals and petals and its smallerlip. C. sichuanense is discriminated, in turn, from C. wumengense,C. lentiginosum and C. lichiangense by its elongated staminode, itsnarrower and acuminate petals and the darker color of its flowers(Perner, 2002). Similar, overlapping modes of foliage and floral pre-sentation in this section may be related to recent studies on sexualreproduction in C. fargesii. The species mentioned above in this sec-tion may employ the same mode of pollination-by-deceit in whichleaf and floral characters combine to lure flies that feed on the exu-dates of fungus infected foliage (Ren et al., 2011).

4.1.7. Section Sinopedilum Perner vs. section Trigonopedia FranchAs a monophyletic section, our results confirmed the establish-

ment of section Sinopedilum by Perner (2008). These three Asianspecies (C. micranthum, C. forrestii and Cypripedium bardophianum)were assigned previously to section Trigonopedia. While all mem-bers of sections Sinopedilum and Trigonopedia are endemic to wes-tern China, we must make important morphological distinctionsbetween species in these two sections and between species in Sin-opedilum and other sections in the genus Cypripedium. First, the an-thers of most Cypripedium spp. release amorphous pollen masses(massulae) that are collected or raked off insect bodies by stigmaswith hook-like papillae (Dressler, 1993). In contrast, the species insection Sinopedilum (sensu Perner, 2008) have anther sacs that re-lease entire club-shaped pollinia terminating in a viscid ‘‘plug’’composed of a rubbery material, probably pollenkitt (Perner,2008). The stigmas of species placed in section Sinopedilum alsohave a median stigma lobe that is partly detached from the othertwo lobes and forms a stigmatic cavity for whole pollinium depo-sition instead of the usual hooked surface. The pollinia and stigmalobe characters present in Sinopedilum are completely absent inspecies of the section Trigonopedia (Perner, 2008).

4.1.8. Section Bifolia C. Morren and section Flabelinervia CribbThe two monophyletic sections, Bifolia and Flabelinervia, are

reconfirmed here. The two species in section Bifolia have spottedlips and lateral petals. These lip petals are urn-shaped without anincurved apical margin (Cribb, 1997; Chen and Cribb, 2005).C. guttatum remains one of the most widespread species in the



genus as it is distributed from the Urals across Asia to Siberia, Kor-ea and Manchuria, south through China to the Himalayas andacross the Pacific onto the Alaskan panhandle (Cribb, 1997; Chenand Cribb, 2005). C. yatabeanum is much more restricted in its dis-tribution than C. guttatum but C. yatabeanum is also found in theWestern Hemisphere on the Kodiak Islands off Alaska, as well asin Asia.

Section Flabelinervia from eastern Asia exclusively is distin-guished by opposite, fan-shaped leaves with radiating venation.The two, insular (Japan and Taiwan) species in this section alsoshare the distinctive ‘‘L-shaped’’ lip (in profile) which has narrowlypandurate dorsal entrance, crimped apex, and lacks side lobes(Cribb, 1997; Chen and Cribb, 2005).

4.1.9. Section Cypripedium LThis section contains most of the species in the genus Cypripe-

dium. Our study supported the monophyletic grouping of this sec-tion previously based on morphological traits and ITS sequences(Cribb, 1997; Cox et al., 1997; Eccarius, 2009). However, our resultssuggested that two subsections Cypripedium and Macrantha as pro-posed by Cribb (1997) were paraphyletic within this section. In thisstudy, segregation between two groups of species appears to bebased on distribution (North America vs. Eurasia) rather than ontraits of flower color and the shapes of lips and lateral petals asused by Cribb (1997). In addition, C. himalaicum is now linked toformer ‘‘subsection’’ Cypripedium (narrow, often twisted petals,usually yellow staminodes, narrower lips, yellow–brown-greenishcolor patterns) with the species in former ‘‘subsection’’ Macrantha(shorter and broader petals, globose to subglobose lips, whitestaminodes and flowers reddish).

Sister species C. henryi and C. segawai share the same pubescentovary, staminode structure, and linear, not twisted or scarcelytwisted petals and small lip. Most members of this section sharesuch morphological characters as a parallel-veined, ovoid or obo-void lip with marked side lobes, flat staminode (Cribb, 1997; Chenand Cribb, 2005), especially in the C. tibeticum complex. This com-plex consists currently of five taxa that remain notoriously difficultto distinguish by morphological characters; C. tibeticum var. tibeti-cum, C. tibeticum var. tibeticum aff., C. tibeticum var. corrugatum,Cypripedium rubitinctum, and C. froschii. Based on shared floral pre-sentation members of this complex may depend exclusively oninfrequent visits from large bumblebee (Bombus spp.) gynes and/or workers to effect pollination (see Bernhardt and Edens-Meier,2010). We also acknowledge parallel evolution within section Cyp-ripedium as several highly isolated species all reflected the sametrend towards pollination by a broad range of small-bodied, soli-tary bees (see Bernhardt and Edens-Meier, 2010). Unfortunately,this study lacked access to specimens of C. cordigerum and C. lud-lowii for clarification.

4.2. Conclusions and future work in phylogeny of the genusCypripedium

With a larger dataset of nrDNA ITS and five cpDNA regions andcomprehensive species sampling, this study was able to update thephylogeny of the genus Cypripedium. Within this present phylog-eny two early lineages, the first comprising two Mesoamericanspecies, and the second including three C. debile samples, havebeen clarified and resolved well, but evolutionary relationshipsamong sections representing Asian and North American species re-main unresolved or unclear. In particular, additional work remainswithin section Cypripedium at the upper part of our topology.Choosing additional gene regions with suitable evolutionary ratesis desirable for further work to better resolve phylogenetic rela-tionships. For the monotypic section Enantiopedilum we must findand include C. elegans in any future studies. We acknowledge that

rapid rates of extinction and/or rapid and comparatively recentperiods of diversification may help explain the unresolved phylo-genetic relationships among these clades. Nevertheless, additionof new datasets by sampling missing species unavailable for thisstudy should permit further progress clarifying sectional delimita-tion within the genus Cypripedium. The benefit of resolving theseproblems will permit us to map the evolution of such traits as pol-lination syndromes, pollinia structure, and flower and inflores-cence architecture with far greater confidence.

Acknowledgments

We would like to thank Drs. Retha Edens-Meier, Peter Kevan,Nan Vance, Yu Zhang, Douglas E. Gill, Sandra L. Taylor, and DarcyGunnlaugson, Bill Steele, Mark Frystak, Ross Kouzes, Jay Vannini,Toshiharu Mitsuhashi, Marcin Górniak, Chun-ying Ling, and mem-bers of the national orchid conservation center of China for theirhelp with sample collection and good suggestions. The first authorthanks particularly Darcy Gunnlaugson for his great help in provid-ing kindly leaf samples of more than 20 taxa and improving theEnglish of this manuscript. We are grateful to Drs. Xiao-guo Xiang,Qiang Zhang, and Yan-yan Guo for assistance in analyzing data. Weare particularly grateful to Dr. Phillip Cribb for providing the pho-tos of C. irapeanum and C. molle, and Drs. Jan Barber and Hong Liufor updating scientific terminology with grammatical usage withinthe manuscript. We are indebted to anonymous reviewers for theircomments and suggestions. The present study was supported bythe National Science Foundation of China (Grants No. 30970203,30770379).

References

Atwood, J.T., 1984. The relationship of the slipper orchids (SubfamilyCypripediodeae), Orchidaceae. Selbyana 7, 129–247.

Averyanov, L.V., 2000. The genus Cypripedium (Orchidaceae) in Russia. Lindleyana15, 197–221.

Baldwin, B.G., 1992. Phylogenetic utility of the internal transcribed spacers ofnuclear ribosomal DNA in plant: an example from the Compositae. Mol.Phylogenet. Evol. 1, 3–16.

Bateman, R.M., Pridgeon, A.M., Chase, M.W., 1997. Phylogenetics of subtribeOrchidinae (Orchidoideae, Orchideceae) based on nuclear ITS sequences II:infrageneric relationships and reclassification to achieve monophyly of Orchissensu stricto. Lindleyana 12, 113–141.

Bernhardt, P., Edens-Meier, R., 2010. What we think we know vs. what we need toknow about orchid pollination and conservation: Cypripedium L. as a modellineage. Bot. Rev. 76, 204–219.

Cameron, K.M., Chase, M.W., Whitten, W.M., Kores, P.J., Jarrell, D.C., Albert, V.A.,Yukawa, T., Hills, H.G., Goldman, D.H., 1999. A phylogenetic analysis of theOrchidaceae: evidence from rbcL nucleotide sequences. Am. J. Bot. 86, 208–224.

Chase, M.W., Freudenstein, J.V., Cameron, K.M., Barrett, R.L., 2003. DNA data andOrchidaceae systematics: a new phylogenetic classification. In: Dixon, K.W.,Kell, S.P., Barrett, R.L., Cribb, P.J. (Eds.), Orchid Conservation. Natural HistoryPublications, Kota Kinabalu, pp. 69–89.

Chen, S.C., 1983. Two pairs of vicariant species of Cypripedium (Orchidaceae) fromeastern Asian and eastern North America. Brittonia 35, 159–163.

Chen, S.C., Cribb, P.J., 2005. Cypripedium. In: Wu, Z.Y., Raven, P.H., Hong, D.Y. (Eds.),Flora of China, vol. 25. Science Press/Missouri Botanical Garden Press, Beijing/St.Louis., pp. 22–33.

Chen, S.C., Lang, K.Y., 1986. Cypripedium subtropicum sp. nov. — a new speciesrelated closely to Selenipedium. Acta Phytotax. Sin. 24, 317–322.

Chen, S.C., Tsi, Z.H., Luo, Y.B., 1999. Native Orchids of China in Colour. Science Press,Beijing, China.

Cox, A.V., Pridgeon, A.M., Albert, V.A., Chase, M.W., 1997. Phylogenetics of theslipper orchids (Cypripedioideae, Orchidaceae): nuclear rDNA ITS sequences.Plant Syst. Evol. 208, 197–223.

Cozzolino, S., Aceto, S., Caputo, P., Widmer, A., Dafni, A., 2001. Speciation processesin eastern Mediterranean Orchis s. l. species: molecular evidence and the role ofpollination biology. Israel J. Plant Sci. 49, 91–103.

Cribb, P.J., 1997. The Genus Cypripedium. Timber Press, Portland.Cunningham, C.W., 1997. Can three incongruence tests predict when data should be

combined? Mol. Biol. Evol. 14, 733–740.Demesure, B., Sodzi, N., Petit, R.J., 1995. A set of universal primers for amplification

of polymorphic non-coding regions of mitochondrial and chloroplast DNA inplants. Mol. Ecol. 4, 129–131.

Dressler, R.L., 1993. Phylogeny and Classification of the Orchid Family. DioscoridesPress, Portland.



Eccarius, W., 2009. Die Orchideengattung Cypripedium. Echinomedia publishers,Buergel, Germany.

Farris, J.S., Källersjö, M., Kluge, A., Bult, G.C., 1995. Constructing a significance testfor incongruence. Syst. Biol. 44, 570–572.

Górniak, M., Paun, O., Chase, M.W., 2010. Phylogenetic relationships withinOrchidaceae based on a low-copy nuclear coding gene, Xdh: congruence withorganellar and nuclear ribosomal DNA results. Mol. Phylogenet. Evol. 56, 784–795.

Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor andprogram for Window 95/98/NT. Nucleic Acid Symp. Ser. 41, 95–98.

Li, P., Luo, Y.B., Bernhardt, P., Kou, Y., Perner, H., 2008. Pollination of Cypripediumplectrochilum (Orchidaceae) by Lasioglossum spp. (Halictidae): the roles ofgeneralist attractants versus restrictive floral architecture. Plant Biol. 10, 220–230.

Liu, Z.J., Chen, L.J., 2009. Cypripedium singchii, a new species of Orchidaceae fromChina. J. Fariylake Bot. Gard. 8, 1–3.

Luo, Y.B., Chen, S.C., 2000. The floral morphology and ontogeny of some Chineserepresentatives of orchid subtribe Orchidinae. Bot. J. Linn. Soc. 134, 529–548.

Norup, M.V., Dransfield, J., Chase, M.W., Barfod, A.S., Fernando, E.S., Baker, W.J.,2006. Homoplasious character combinations and generic delimitation: a casestudy from the Indo-Pacific arecoid palms (Arecaceae: Areaceae). Am. J. Bot. 93,1065–1080.

Perner, H., 2002. Cypripedium sichuanense Perner. Die Orchidee 53, 89–94.Perner, H., 2008. Sinopedilum — a new section of the genus Cypripedium. Die

Orchidee 59, 35–51.Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of DNA substitution.

Bioinformatics 14, 817–818.Ren, Z.X., Li, D.Z., Bernhardt, P., Wang, H., 2011. Flowers of Cypripedium fargesii

(Orchidaceae) fool flat-footed flies (Platypezidae) by faking fungus-infectedfoliate. Proc. Natl. Acad. Sci. USA 108, 478–7480.

Ronquist, F., Huelsenbeck, J.P., 2003. MrBaye: Bayesian phylogenetic inferenceunder mixed models. Bioinformatics 19, 1972–1974.

Rosso, S.W., 1996. The vegetative anatomy of the Cypripedioideae (Orchidaceae). J.Linn. Soc. Bot. 59, 309–341.

Sang, T., Crawford, D.J., Stuessy, T.F., 1997. Chloroplast DNA phylogeny, reticulateevolution, and biogeography of Paeonia (Paeniaceae). Am. J. Bot. 84, 1120–1136.

Shaw, J., Lickey, E.B., Schillng, E.E., Small, R.L., 2007. Comparison of wholechloroplast genome sequences to choose noncoding regions for phylogeneticstudies in Angiosperms: the tortoise and the hare III. Am. J. Bot. 94, 275–288.

Soto, M.A., Solano, R., 2007. Ficha técnica de Cypripedium dickinsonianum. Informaciónactualizada sobre las especies del PROY-NOM-059-ECOL-2000 <http://www.conabio.gob.mx/conocimiento/ise/fichasnom/Cypripediumdickinsonianu>.

Swofford, D.L., 2003. Paup� Phylogenetic Analysis using Parisimony (and othermehods). Version 4, Sinauer Associates, Sunderland, Massachusetts.

Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers foramplification of three non-coding regions of chloroplast DNA. Plant Mol. Biol.17, 1105–1109.

Tate, J.A., Simpson, B.B., 2003. Paraphyly of Tarasa (Malvaceae) and diverse originsof polyploidy species. Syst. Bot. 28, 723–737.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. TheClustal X windows interface. Flexible strategies for multiple sequencealignment aided by quality analysis tools. Nucleic Acids Res. 24, 4876–4882.

Tucker, S., 1997. Floral evolution, development and convergence: the hierarchical-significance hypothesis. Int. J. Plant Sci. 158, 143–161.

Tucker, S., Bernhardt, P., 2000. Floral ontogeny, pattern formation and evolution inHibbertia and Adrastaea (Dilleniaceae). Am. J. Bot. 87, 1915–1936.

Van der Pijl, L., Dodson, C.H., 1966. Orchid Flowers: Their Pollination and Evolution.University of Miami Press, Miami, FL.

Wiens, J.J., 1998. Combining data sets with different phylogenetic histories. Syst.Biol. 47, 568–581.

Xi, Y.Z., Chen, S.C., 1991. A palynological study of the genus Cypripedium(Orchidaceae). Cathaya 3, 73–91.


author's personal copy -...

Documents