systematics of amaryllidaceae based on cladistic analysis of plastid rbcl and trnl-f sequence data

1325

American Journal of Botany 86(9): 1325–1345. 1999.

SYSTEMATICS OF AMARYLLIDACEAE BASED ON

CLADISTIC ANALYSIS OF PLASTID RBCL AND TRNL-FSEQUENCE DATA1

ALAN W. MEEROW,2,3 MICHAEL F. FAY,4 CHARLES L. GUY,5

QIN-BAO LI,5 FARIDAH Q. ZAMAN,4 AND MARK W. CHASE4

3University of Florida, Fort Lauderdale Research and Education Center, 3205 College Avenue,Fort Lauderdale, Florida 33314;

4Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3DS, UK;5University of Florida, Department of Environmental Horticulture, 1545 Fifield Hall, Gainesville, Florida 32611

Cladistic analyses of plastid DNA sequences rbcL and trnL-F are presented separately and combined for 48 genera ofAmaryllidaceae and 29 genera of related asparagalean families. The combined analysis is the most highly resolved of thethree and provides good support for the monophyly of Amaryllidaceae and indicates Agapanthaceae as its sister family.Alliaceae are in turn sister to the Amaryllidaceae/Agapanthaceae clade. The origins of the family appear to be westernGondwanaland (Africa), and infrafamilial relationships are resolved along biogeographic lines. Tribe Amaryllideae, primarilySouth African, is sister to the rest of Amaryllidaceae; this tribe is supported by numerous morphological synapomorphiesas well. The remaining two African tribes of the family, Haemantheae and Cyrtantheae, are well supported, but their positionrelative to the Australasian Calostemmateae and a large clade comprising the Eurasian and American genera, is not yetclear. The Eurasian and American elements of the family are each monophyletic sister clades. Internal resolution of theEurasian clade only partially supports currently accepted tribal concepts, and few conclusions can be drawn on the rela-tionships of the genera based on these data. A monophyletic Lycorideae (Central and East Asian) is weakly supported.Galanthus and Leucojum (Galantheae pro parte) are supported as sister genera by the bootstrap. The American clade showsa higher degree of internal resolution. Hippeastreae (minus Griffinia and Worsleya) are well supported, and Zephyranthinaeare resolved as a distinct subtribe. An Andean clade marked by a chromosome number of 2n 5 46 (and derivatives thereof)is resolved with weak support. The plastid DNA phylogenies are discussed in the context of biogeography and characterevolution in the family.

Key words: Amaryllidaceae; cladistic analysis; molecular systematics; moncotyledons; phylogeny; plastid DNA.

The Amaryllidaceae J. St.-Hil., a cosmopolitan (pre-dominantly pantropical) family of petaloid monocots,represent one of the elements of the Linnaean Hexandriamonogynia (Linnaeus, 1753), the 51 genera of whichhave been variously classified since as liliaceous or ama-ryllidaceous. This basic dichotomy represents the gener-ally uncertain phylogenetic placement of many petaloidmonocots until the past two decades. Seven of the 51genera that Linneaus placed in Hexandria monogyniahave since been included within a common taxonomicunit, as section Narcissi (Adanson, 1763; de Jussieu,1789), family Amaryllideae (Jaume-St.-Hilaire, 1805;Brown, 1810), order Amaryllidaceae (Lindley, 1836;Herbert, 1837), tribe Amarylleae (Bentham and Hooker,1883), suborder Amarylleae (Baker, 1888), subfamilyAmaryllidoideae (Pax, 1888); family Amaryllidaceae

1 Manuscript received 10 August 1998; revision accepted 14 January1999.

Significant portions of this work were completed during the seniorauthor’s sabbatical leave at the Royal Botanic Gardens, Kew. Some ofthe sequences were generated at the DNA Sequencing Core of the In-terdisciplinary Center for Biotechnology Research at the University ofFlorida. Financial support was provided by NSF grants DEB-968787and IBN-9317450 to AWM and CLG, a Research Enhancement Awardfrom the Institute of Food and Agricultural Sciences of the Universityof Florida, and the Royal Botanic Gardens, Kew. The authors thankDavid L. Swofford for allowing the use of experimental versions of hisPAUP* software. Florida Agricultural Experiment Station Journal SeriesNumber R-03602.

2 Author for correspondence.

(Hutchinson, 1934, 1959), and subfamily Amarylloideae(Traub, 1963). Brown (1810) was the first to propose thatthe genera with superior ovaries be excluded from Ama-ryllidaceae, a restriction followed faithfully until Hutch-inson (1934). Herbert (1837) recognized that the Tacca-ceae were not allied to Amaryllidaceae, and Pax (1888)formally removed Velloziaceae as part of the family (Her-bert’s suborder Xerophyteae). Hutchinson’s (1934, 1959)classification was the first radical recircumscription ofAmaryllidaceae since Brown (1810). In defining the uni-fying character of the family to be ‘‘an umbellate inflo-rescence subtended by an involucre of one or more spa-thaceous bracts,’’ he segregated Agavaceae, Hypoxida-ceae, and Alstroemeriaceae and added tribes Agapan-theae, Allieae, and Gilliesieae (Alliaceae). Takhtajan(1969) recognized Amaryllidaceae in the narrowestsense, and maintained a distinct Alliaceae. Cronquist(1988) and Thorne (1976) included Amaryllidaceae with-in broad concepts of Liliaceae.

Concepts of familial and ordinal limits of the mono-cotyledons were radically challenged by Huber (1969),who emphasized less conspicuous characters, particularlyembryological characters, over gross floral or vegetativemorphology. Huber’s work highlighted the heterogeneitypresent in many traditional monocot families, especiallyLiliaceae. Much of this work was refined and placed intophylogenetic context by Dahlgren and coworkers (Dahl-gren and Clifford, 1982; Dahlgren and Rasmussen, 1983;

1326 [Vol. 86AMERICAN JOURNAL OF BOTANY

Dahlgren, Clifford, and Yeo, 1985). In Dahlgren, Clif-ford, and Yeo’s (1985) synthesis, Amaryllidaceae and Al-liaceae are both recognized as members of the order As-paragales, an order of 31 families that have evolved manytraits in parallel with Liliales. One of the most importantand consistent characters separating these two orders isthe presence of phytomelan in the seed coat of Aspara-gales (Huber, 1969). To date, phylogenetic analyses ofthe monocotyledons, based on both morphological andgene sequence matrices, have supported this classificationwith some amendment (Duvall et al., 1993; Stevensonand Loconte, 1995; Chase et al., 1995a, b), but the pre-cise relationship of Amaryllidaceae to other Asparagalesremained elusive until Fay and Chase (1996) used mo-lecular data to argue that Amaryllidaceae, Agapantha-ceae, and Alliaceae form a monophyletic group and thattogether they are related most closely to Hyacinthaceaes.s. and the resurrected family Themidaceae (the formertribe Brodiaeeae of Alliaceae).

Despite a lack of consensus on generic limits and tribaldelimitations within the Amaryllidaceae, cladistic analy-sis has only rarely been applied to problems in the family,such as by Nordal and Duncan (1984) for Haemanthusand Scadoxus, two closely related, baccate-fruited Afri-can genera, Meerow (1987a, 1989) for Eucrosia and Eu-charis and Caliphruria, respectively, and Snijman (1994)and Snijman and Linder (1996) for various taxa of tribeAmaryllideae. Applying phylogenetic studies for the en-tire family is difficult due to homoplasy for many con-spicuous characters within this highly canalized group(Meerow, 1987a, 1989, 1995). This led Meerow (1995)to conclude that ‘‘future reconstruction attempts willgreatly benefit from the inclusion of molecular data.’’

The four most recent infrafamilial classifications (Table1) of Amaryllidaceae are those of Traub (1963), Dahl-gren, Clifford, and Yeo (1985), Muller-Doblies and Mul-ler-Doblies (1996), and Meerow and Snijman (1998).Traub’s scheme included Alliaceae, Hemerocallidaceae,and Ixioliriaceae as subfamilies, following Hutchinson(1934, 1959) in part. Within his subfamily Amaryllo-ideae, he erected two informal taxa, ‘‘infrafamilies’’Amarylloidinae and Pancratioidinae, both of which werepolyphyletic (Meerow, 1995). Dahlgren, Clifford, andYeo (1985) dispensed with any subfamilial classificationabove the level of tribe, recognizing eight, and treated asAmaryllidaceae only those genera in Traub’s Amaryllo-ideae. Stenomesseae and Eustephieae were combined.Meerow (1995) resurrected Eustephieae from Stenomes-seae and suggested that two new tribes may need to berecognized, Calostemmateae and Hymenocallideae. Mul-ler-Doblies and Muller-Doblies (1996) recognized tentribes (among them Calostemmateae) and 19 subtribes,many of them monogeneric; Meerow and Snijman (1998)recognize 14 tribes, with two subtribes only in one ofthem (Table 1).

Fay and Chase (1996) recircumscribed Amaryllidaceaeto include Agapanthus, previously included in Alliaceae,as subfamily Agapanthoideae. This recircumscription wasbased on phylogenetic analysis of rbcL sequence data,with only four genera of Amaryllidaceae s.s. included inthe analysis. All the epigynous genera were treated asAmaryllidoideae. Bootstrap support for this treatmentwas weak (63%). The sampling within Amaryllidaceae

s.s. in Fay and Chase (1996) did not allow sufficient res-olution of the generic relationships within the family, andwe present here phylogenetic analyses of three plastidDNA sequence data sets for a much wider range of taxa.

The phylogenetic application of sequences of rbcL iswell documented (e.g., Chase et al., 1993; Olmstead andPalmer, 1994) and has been used to clarify relationshipsbetween and within a number of asparagoid families, in-cluding Themidaceae (Fay and Chase, 1996), Asphode-laceae (de Bruijn et al., unpublished data), Alliaceae (Fayet al., unpublished data), and Orchidaceae (Cameron etal., 1999). Within Amaryllidaceae, however, levels of res-olution obtained within some major clades, particularlythose from the Neotropics, were not sufficient to elucidatetribal relationships fully (Fay et al., 1995). For this rea-son, we chose to combine our rbcL matrix with two forthe trnL intron/trnL-F spacer region of noncoding plastidDNA, for which Taberlet et al. (1991) had developed‘‘universal’’ primers for amplification. Sequences of thisregion have been used in phylogenetic studies of Cras-sulaceae (Kim, t’Hart, and Mes, 1996; Mes, Van Bred-erode and t’Hart, 1996; Mes, Wijers, and t’Hart, 1997),Gentianaceae (Gielly and Taberlet, 1996; Gielly et al.,1996), Paeoniaceae (Sang, Crawford and Stuessy, 1997),Proteaceae (Maguire et al., 1997), Ranunculaceae (Kita,Ueda, and Kadota, 1995), among others, either alone orin combination with other loci. This region of the plastidgenome evolves more than three times faster, on average,than rbcL (Gielly and Taberlet, 1994) and can thereforepotentially add increased resolution to a phylogeny gen-erated by rbcL sequences.

Combining independent character matrices, whetherboth molecular or molecular and morphological, very of-ten increases the resolution of the ingroup and the boot-strap support of the internal nodes of the phylogenetictrees (Chase et al., 1995b; Olmstead and Sweere, 1994;Rudall et al., 1998; Soltis et al., 1998). In this paper wepresent the first family-wide phylogenetic analysis ofAmaryllidaceae using three plastid DNA sequences,alone and in combination, and comment on the evolu-tionary and bigeographic implications of the results.

MATERIALS AND METHODS

Plant materials—The sources of plant material and vouchers/acces-sions used in this analysis are listed in Table 2, along with GenBank orEMBL accession numbers for the sequences.

DNA extraction, gene amplification, and sequencing—Sequencesfor rbcL were generated at both RBG Kew and the University of Florida(Table 2); all trnL-F sequences were obtained at Kew.

RBG Kew—DNA was extracted from 1.0 g fresh, 0.2–0.25 g silicagel-dried leaves, or ;0.1 g material from herbarium sheets using the2X CTAB method of Doyle and Doyle (1987). All samples were thenpurified on cesium chloride/ethidium bromide gradients (1.55 g/mL den-sity). Gene amplification of the rbcL gene was carried out using forwardprimers that match the first 20 or 26 base pairs (bp) of the coding regionand reverse primers that correspond to 20-bp sequences that begin atposition 1352 or 1367 in the coding region (Table 3; Chase and al.,1995a). The trnL-trnF region was amplified using the c and f primersof Taberlet et al. (1991). Amplified products were purified using Magicmini columns (Promega, Madison, Wisconsin) or QIAquick (Qiagen,Valencia, California) columns, following manufacturers protocols. Stan-

September 1999] 1327MEEROW ET AL.—AMARYLLIDACEAE SYSTEMATICS

TABLE 1. Four most recent intrafamilial classifications of Amaryllidaceae s.s. The lines indicate subsequent segregation or inclusion of genera.

a As Dahlgren, Clifford, and Yeo (1985) did not consistently list the component genera in their tribal concepts, their exact generic compositionis inferred. Most of their delimitations are presumed to have followed Traub (1963).

dard dideoxy methods or modified dideoxy cycle sequencing with dyeterminators run on an ABI 373A or 377 automated sequencer (accordingto the manufacturer’s protocols; Applied Biosystems, Inc., Foster City,California) were used to sequence the amplification products directly.For rbcL, both strands were sequenced for 70–90% of the exon. Wehave ;1320 bp of rbcL sequence data for most taxa. The trnL-F regionis length variable; we sequenced both strands for 70–90% of the region,obtaining between 750 and 900 bp of sequence data for most taxa.

University of Florida—Genomic DNA was obtained from youngfresh leaf tissue using an extraction protocol developed for plant tissuesrich in soluble polysacchrides (Li, Cai, and Guy, 1994). Polymerasechain reaction (PCR) protocols were those of Li and Guy (1996). ThePCR reaction product was fractionated by electrophoresis on a 0.8%low melting point agarose gel, then the DNA was purified from gelslices with phenol and chloroform and dissolved in 10 mL of 10 mmol/LTris (pH 8.0) buffer. A 10-mL ligation reaction was prepared containing

1328 [Vol. 86AMERICAN JOURNAL OF BOTANYT

AB

LE

2.T

axa,

vouc

her

spec

imen

s,an

dG

enB

ank

acce

ssio

nnu

mbe

rsus

edin

the

plas

tid

DN

Ase

quen

ceph

ylog

eny

anal

yses

ofA

mar

ylli

dace

ae.

Tax

onV

ouch

er

Gen

Ban

kac

cess

ion

num

ber

rbcL

trnL

gene

trnL

-Fsp

acer

All

iace

aeA

lliu

msi

culu

mva

r.bu

lgar

icum

A.

subh

irsu

tum

L.

Gil

lies

iagr

amin

eaL

indl

.Ip

heio

nun

iflor

um(G

raha

m)

Raf

.L

euco

cory

nepa

ucifl

ora

R.

Phi

l.

M.

W.

Cha

se83

5(K

)M

.W

.C

hase

439

(K)

M.

W.

Cha

se45

0(K

)M

.W

.C

hase

627

(K)

UC

Irvi

neA

rbor

etum

8182

GB

AN

-Z69

200a

GB

AN

-Z69

205

GB

AN

-Z69

208

GB

AN

-AF

1169

92G

BA

N-A

F11

6998

GB

AN

-AF

1170

01G

BA

N-A

F11

7000

GB

AN

-AF

1170

18G

BA

N-A

F11

7021

GB

AN

-AF

1170

25

GB

AN

-AF

1170

57G

BA

N-A

F11

7058

GB

AN

-AF

1170

45G

BA

N-A

F11

7049

GB

AN

-AF

1170

53M

ilul

asp

icat

aP

rain

Gre

y-W

ilso

n&

Phi

llip

s75

2(K

)G

BA

N-A

F11

6991

GB

AN

-AF

1170

02G

BA

N-A

F11

7056

Not

hosc

ordu

mbi

valv

eB

ritt

onM

.W

.C

hase

247

(NC

U)

GB

AN

-Z69

202

GB

AN

-AF

1170

24G

BA

N-A

F11

7052

Pab

ello

nia

incr

assa

ta(P

hil.

)Q

ueza

daan

dM

arti

c.U

CI

Arb

oret

um82

47G

BA

N-Z

6920

9G

BA

N-A

F11

7026

GB

AN

-AF

1170

54So

lari

aat

ropu

rpur

ea(P

hil.

)R

av.

M.

W.

Cha

se69

3(K

)G

BA

N-Z

6920

7G

BA

N-A

F11

7022

GB

AN

-AF

1170

50St

emm

atiu

mna

rcis

soid

esP

hil.

Bec

kett

,C

hees

e&

Wat

son

4688

(NY

)N

/AG

BA

N-A

F11

7020

GB

AN

-AF

1170

48

Tri

stag

ma

biva

lve

(Lin

dl.)

Tra

ubM

.W

.C

hase

692

(K)

GB

AN

-Z69

206

GB

AN

-AF

1170

23G

BA

N-A

F11

7051

Tul

bagh

iavi

olac

eaH

arv.

M.

W.

Cha

se24

8(N

CU

)G

BA

N-Z

6920

3G

BA

N-A

F11

6999

GB

AN

-AF

1170

30

Aga

pan

thac

eae

Aga

pant

hus

afri

canu

sH

offm

.M

.W

.C

hase

627

(K)

GB

AN

-Z69

221

GB

AN

-AF

1170

28G

BA

N-A

F11

7060

Aga

pant

hus

cam

panu

latu

sF.

M.

Lei

ghto

nM

.W

.C

hase

1008

(K)

GB

AN

-Z69

220

GB

AN

-AF

1170

29G

BA

N-A

F11

7059

Am

aryl

lid

acea

eA

mar

ylli

sbe

llad

onna

L.

M.

W.

Cha

se61

2(K

)G

BA

N-Z

6921

9G

BA

N-A

F10

479

GB

AN

-AF

1047

44A

podo

liri

onla

nceo

latu

mB

enth

.an

dH

ook.

Kir

sten

bosc

h,N

BG

714/

88G

BA

N-A

F11

6944

GB

AN

-AF

1047

89G

BA

N-A

F10

4767

Boo

phon

edi

stic

ha(L

.f.

)H

erb.

M.

W.

Cha

se22

46(K

)G

BA

N-A

F11

6945

GB

AN

-AF

1048

01G

BA

N-A

F10

4726

Bru

nsvi

gia

com

pton

iiW

.F.

Bar

ker

M.

W.

Cha

se22

40(K

)G

BA

N-A

F11

6946

GB

AN

-AF

1048

13G

BA

N-A

F10

4722

Cal

iphr

uria

kors

akof

fii(T

raub

)M

eero

wM

.W

.C

hase

962

(K)

GB

AN

-AF

1169

47G

BA

N-A

F10

4810

GB

AN

-AF

1047

31C

alos

tem

ma

lute

aS

ims

M.

W.

Cha

se15

05(K

)G

BA

N-A

F11

6948

GB

AN

-AF

1047

90G

BA

N-A

F10

4740

Chl

idan

thus

frag

rans

Her

b.M

eero

w23

12(F

LA

S)

GB

AN

-AF

1169

49G

BA

N-A

F10

4770

GB

AN

-AF

1047

23C

livi

ano

bili

sL

indl

.M

.W

.C

hase

3080

(K)

GB

AN

-AF

1169

50G

BA

N-A

F10

4776

GB

AN

-AF

1047

63C

rinu

mye

men

seD

efler

sM

.W

.C

hase

1595

(K)

GB

AN

-AF

1169

51G

BA

N-A

F10

4784

GB

AN

-AF

1047

56C

rypt

oste

phan

usva

nson

iiV

erdo

omM

eero

w23

10(F

LA

S)

GB

AN

-AF

1169

52G

BA

N-A

F10

4804

GB

AN

-AF

1047

43C

yrta

nthu

sel

atus

(Jac

q.)

Tra

ubM

.W

.C

hase

1572

(K)

GB

AN

-AF

1169

53G

BA

N-A

F10

4818

GB

AN

-AF

1047

53E

ucha

ris

cast

elna

eana

(Bai

ll.)

Mac

br.

Sch

unke

1415

6(F

LA

S)

GB

AN

-AF

1169

54G

BA

N-A

F10

4798

GB

AN

-AF

1047

66E

ucro

sia

eucr

osio

ides

(Pax

)T

raub

Mee

row

1117

(FL

AS

)G

BA

N-A

F11

6955

GB

AN

-AF

1047

88G

BA

N-A

F10

4742

Eus

teph

iada

rwin

iiV

arga

sM

.W

.C

hase

559

(K)

GB

AN

-AF

1169

56G

BA

N-A

F10

4794

GB

AN

-AF

1047

27G

alan

thus

plic

atus

M.

Bie

b.M

.W

.C

hase

741

(K)

GB

AN

-Z69

218

GB

AN

-AF

1047

99G

BA

N-A

F10

4730

Get

hyll

isci

liar

is(T

hunb

.)T

hunb

.D

unca

n11

23(N

BG

)G

BA

N-A

F11

6957

GB

AN

-AF

1048

16G

BA

N-A

F10

4745

Gri

ffini

ahy

acin

thin

aK

erG

awle

rM

eero

w21

06(F

LA

S)

GB

AN

-AF

1169

58G

BA

N-A

F10

4771

GB

AN

-AF

1047

36H

abra

nthu

sm

arti

nezi

iR

aven

naM

.W

.C

hase

1023

(K)

GB

AN

-AF

1169

59G

BA

N-A

F10

4772

GB

AN

-AF

1047

38H

aem

anth

ushu

mil

isJa

cq.

M.

W.

Cha

se20

25(K

)G

BA

N-A

F11

6960

GB

AN

-AF

1047

81G

BA

N-A

F10

4721

Han

noni

ahe

sper

idiu

mB

raun

-Bla

nq.

and

Mai

reM

.W

.C

hase

2023

(K)

GB

AN

-AF

1169

61G

BA

N-A

F10

4812

GB

AN

-AF

1047

34H

esse

aze

yher

iB

aker

M.

W.

Cha

se22

38(K

)G

BA

N-A

F11

6962

GB

AN

-AF

1048

13G

BA

N-A

F10

4741

Hie

rony

mie

lla

mar

gina

ta(P

ax)

A.

T.H

unz.

M.

W.

Cha

se19

01(K

)G

BA

N-A

F11

6963

GB

AN

-AF

1048

07G

BA

N-A

F10

4757

Hip

peas

trum

papi

lio

(Rav

.)V

anS

chee

pen

Mee

row

2307

(FL

AS

)G

BA

N-A

F11

6964

GB

AN

-AF

1047

75N

/AH

ymen

ocal

lis

mar

gina

ta(P

ax.)

A.

T.H

unz.

M.

W.

Cha

se19

01(K

)G

BA

N-A

F11

6965

GB

AN

-AF

1047

96G

BA

N-A

F10

4719

Ism

ene

long

ipet

ala

(Lin

dl.)

Mee

row

M.

W.

Cha

se35

83(K

)G

BA

N-A

F11

6966

N/A

GB

AN

-AF

1047

68Is

men

ena

rcis

siflo

raJa

cq.

Mee

row

2306

(FL

AS

)G

BA

N-A

F11

6967

GB

AN

-AF

1047

87G

BA

N-A

F10

4725

Ism

ene

varg

asii

(Vel

arde

)G

erea

uan

dM

eero

wM

eero

w23

08(F

LA

S)

GB

AN

-AF

1169

68G

BA

N-A

F10

4802

GB

AN

-AF

1047

32L

apie

dra

mar

tine

zii

Lag

.M

.W

.C

hase

1528

(K)

GB

AN

-AF

1169

69G

BA

N-A

F10

4806

GB

AN

-AF

1047

50L

epto

chit

onqu

itoe

nsis

(Her

b.)

Sea

lyM

eero

w11

16(F

LA

S)

GB

AN

-AF

1169

70G

BA

N-A

F10

4779

GB

AN

-AF

1047

55L

euco

jum

autu

mna

leL

.M

.W

.C

hase

607

(K)

GB

AN

-Z77

256

GB

AN

-AF

1047

73G

BA

N-A

F10

4758

Lyc

oris

squa

mig

era

Max

im.

M.

W.

Cha

se20

14(K

)G

BA

N-A

F11

6971

GB

AN

-AF

1047

80G

BA

N-A

F10

4733

Nar

ciss

usel

egan

s(H

aw.)

Spa

chM

.W

.C

hase

617

(K)

GB

AN

-AF

1169

72G

BA

N-A

F10

4791

GB

AN

-AF

1047

46N

erin

ebo

wde

nii

Wil

l.W

ats.

M.

W.

Cha

se61

6(K

)G

BA

N-A

F11

6973

GB

AN

-AF

1047

69G

BA

N-A

F10

4751


TA

BL

E2.

Con

tinu

ed.

Tax

onV

ouch

er

Gen

Ban

kac

cess

ion

num

ber

rbcL

trnL

gene

trnL

-Fsp

acer

Pam

iant

hepe

ruvi

ana

[Sta

pf]

Mee

row

2304

(FL

AS

)G

BA

N-A

F11

6974

GB

AN

-AF

1048

14G

BA

N-A

F10

4759

Pan

crat

ium

cana

rien

sis

L.

Mee

row

1142

(FL

AS

)G

BA

N-A

F11

6975

GB

AN

-AF

1047

78G

BA

N-A

F10

4718

Par

amon

gaia

web

erba

ueri

Val

arde

Mee

row

2303

(FL

AS

)G

BA

N-A

F11

6976

GB

AN

-AF

1047

77G

BA

N-A

F10

4764

Pha

edra

nass

adu

bia

(HB

K)

Mac

br.

M.

W.

Cha

se18

34(K

)G

BA

N-A

F11

6977

GB

AN

-AF

1048

09G

BA

N-A

F10

4729

Pro

iphy

scu

nnin

gham

ii(A

it.

exL

indl

.)M

abb.

Mee

row

1118

(FL

AS

)G

BA

N-A

F11

6978

GB

AN

-AF

1047

85G

BA

N-A

F10

4762

Rau

hia

deco

raR

aven

naM

.W

.C

hase

1573

(K)

GB

AN

-AF

1169

79G

BA

N-A

F10

4805

GB

AN

-AF

1047

35R

hodo

phia

lam

oell

eri

(R.

Phi

l.)

Tra

ubM

.W

.C

hase

1908

(K)

GB

AN

-AF

1169

80G

BA

N-A

F10

4782

GB

AN

-AF

1047

20Sc

adox

usci

nnab

aeri

nus

(Dec

ne.)

I.F

riis

and

I.N

orda

lM

.W

.C

hase

549

(K)

GB

AN

-AF

1169

81G

BA

N-A

F10

4783

GB

AN

-AF

1047

54Sp

reke

lia

form

osis

sim

a(L

.)H

erb.

M.

W.

Cha

se57

7(K

)G

BA

N-A

F11

6982

GB

AN

-AF

1048

08G

BA

N-A

F10

4728

Sten

omes

son

pear

cei

Bak

.M

.W

.C

hase

1591

(K)

GB

AN

-Z69

217

GB

AN

-AF

1048

00G

BA

N-A

F10

4739

Sten

omes

son

vari

egat

um(R

.an

dP.

)B

ak.

Mee

row

1159

(FL

AS

)G

BA

N-A

F11

6983

GB

AN

-AF

1048

11G

BA

N-A

F10

4724

Ster

nber

gia

lute

a(L

.)S

pren

g.M

.W

.C

hase

615

(K)

GB

AN

-AF

1169

84G

BA

N-A

F10

4793

GB

AN

-AF

1047

47St

rum

aria

trun

cata

Jacq

.S

nijm

an28

1(N

BG

)G

BA

N-A

F11

6985

GB

AN

-AF

1048

19G

BA

N-A

F10

4765

Tra

ubia

mod

esta

(R.

A.

Phi

l.)

Rav

enna

Mee

row

2301

(FL

AS

)G

BA

N-A

F11

6986

GB

AN

-AF

1047

92G

BA

N-A

F10

4748

Ung

erni

afla

vaB

oiss

.ex

Hau

ssk.

exB

oiss

.M

.W

.C

hase

3640

(K)

GB

AN

-AF

1169

87G

BA

N-A

F10

4797

GB

AN

-AF

1047

49V

agar

iapa

rvifl

ora

Her

b.M

.W

.C

hase

1066

(K)

GB

AN

-AF

1169

88G

BA

N-A

F10

4786

GB

AN

-AF

1047

60W

orsl

eya

rayn

eri

(Hoo

k.)

Tra

uban

dM

olde

nke

Mee

row

2302

(FL

AS

)G

BA

N-A

F11

6989

GB

AN

-AF

1047

74G

BA

N-A

F10

4761

Zep

hyra

nthe

sfil

ifol

iaH

erb.

exB

aker

M.

W.

Cha

se18

36(K

)G

BA

N-A

F11

6990

GB

AN

-AF

1048

15G

BA

N-A

F10

4737

An

ther

icac

eae

Ant

heri

cum

lili

ago

M.

W.

Cha

se51

5(K

)G

BA

N-Z

6922

5G

BA

N-A

F11

7005

GB

AN

-AF

1170

33E

chea

ndia

sp.

M.

W.

Cha

se60

2(K

)G

BA

N-Z

6922

5G

BA

N-A

F11

7014

GB

AN

-AF

1170

39L

euco

crin

umm

onta

num

Nut

t.ex

A.

Gra

yM

.W

.C

hase

795

(K)

GB

AN

-Z77

252

GB

AN

-AF

1170

03G

BA

N-A

F11

7031

Beh

nia

ceae

Beh

nia

reti

cula

taD

idr.

Gol

dbla

tt92

73(M

O)

GB

AN

-Z69

226

GB

AN

-AF

1170

07G

BA

N-A

F11

7035

Con

vall

aria

ceae

Asp

idis

tra

elat

ior

Blu

me

M.

W.

Cha

se83

3(K

)G

BA

N-Z

7726

9G

BA

N-A

F11

7016

GB

AN

-AF

1170

44L

irio

pepl

atyp

hyll

aF.

T.W

ang

&T.

Tan

gM

.W

.C

hase

1102

(K)

GB

AN

-Z77

271

GB

AN

-AF

1170

09G

BA

N-A

F11

7038

Pel

iosa

nthe

ssp

.M

.W

.C

hase

497

(K)

GB

AN

-Z77

272

GB

AN

-AF

1170

06G

BA

N-A

F11

7034

Pol

ygon

atum

hook

eri

Bak

erM

.W

.C

hase

847

(K)

GB

AN

-Z73

695

GB

AN

-AF

1170

10G

BA

N-A

F11

7036

Hem

eroc

alli

dac

eae

Gei

tono

ples

ium

cym

osum

Ade

laid

eB

.G

.88

0709

GB

AN

-AF

1169

97G

BA

N-A

F11

7027

GB

AN

-AF

1170

55

Hya

cin

thac

eae

Alb

uca

shaw

iiB

aker

M.

W.

Cha

se10

12(K

)G

BA

N-Z

6922

3G

BA

N-A

F11

7012

GB

AN

-AF

1170

42H

yaci

nthu

sor

ient

alis

L.

M.

W.

Cha

se15

03(K

)G

BA

N-A

F11

6995

GB

AN

-AF

1170

13G

BA

N-A

F11

7043

Orn

itho

galu

mlo

ngeb

ract

eatu

mJa

cq.

M.

W.

Cha

se15

07(K

)G

BA

N-Z

6922

4G

BA

N-A

F11

7008

GB

AN

-AF

1170

37

Lax

man

nia

ceae

Eus

trep

hus

lati

foli

usR

.B

r.A

dela

ide

B.

G.

8805

87G

BA

N-A

F11

6996

GB

AN

-AF

1170

04G

BA

N-A

F11

7032

Th

emid

acea

eB

esse

rael

egan

sS

chul

t.M

.W

.C

hase

626

(K)

GB

AN

-Z69

215

GB

AN

-AF

1170

15G

BA

N-A

F11

7040

Bro

diae

ajo

lone

nsis

Eas

tw.

M.

W.

Cha

se18

31(K

)G

BA

N-A

F11

6993

GB

AN

-AF

1170

17G

BA

N-A

F11

7046

Mil

lam

agni

fica

E.

Moo

reM

eero

w23

09(F

LA

S)

GB

AN

-AF

1169

94G

BA

N-A

F11

7011

GB

AN

-AF

1170

41M

uill

am

arit

ima

S.

Wat

s.M

.W

.C

hase

779

(K)

GB

AN

-Z69

213

GB

AN

-AF

1170

19G

BA

N-A

F11

7047

aT

hepr

efix

GB

AN

has

been

adde

dfo

rli

nkin

gth

eon

line

vers

ion

ofA

mer

ican

Jour

nal

ofB

otan

yto

Gen

Ban

kan

dis

not

part

ofth

eac

tual

Gen

Ban

kac

cess

ion

num

ber.


TABLE 3. PCR and sequencing primers for rbcL and trnL-F used inthis study.

Sequence Primer name

rbcLRoyal Botanic Gardens, Kew

59ATGTCACCACAAACAGAAAC39

59GCGTTGGAGAGAGATCGTTTTCT39

59TCGCATGTACCYGCAGTTGC39

59CTTTCCAAAATTTCACAAGCAGCA39

1F636F724R

(moncots)1368R

University of Florida59ATGTCACCACAAACAGAAACTAAAGCAAGT39

59AATTTGATCTCCTTCCATATTTCGCA39

59AAACTTTCCAAGGCCCGC39

59GCGACTTCGGTCTTTTTC39

59GGTAAACTGGAAGGGGAA39

59GCGGGCCTTGGAAAGTTT39

Zurawski’sZ-1

Zurawski’sZ-1375R

ALM1(5Z-427)

ALM2ALM3ALM4

trnL-F (Taberlet et al., 1991)59CGAAATCGGTAGACGCTACG39

59ATTTGAACTGGTGACACGAG39

59GGGGATAGAGGGACTTGAAC39

59GGTTCAAGTCCCTCTATCCC39

trnL-ctrnL-ftrnL-dtrnL-e

25 ng of pGEM-Tt vector (Promega, Madison, Wisconsin), 50–100 ngPCR products, 1 mL of 103 ligase buffer [300 mmol/L Tris-HCl, pH7.8, 100 mmol/L MgCl2, 100 mmol/L dithiothreitol, 5 mmol/L ATP, and1.5 U T4 DNA ligase (Promega, Madison, Wisconsin)]. The ligationswere incubated at 168C overnight. One hundred microlitres of compe-tent E. coli XL-1 Blue cells were transformed with 2.5 mL of eachligation mixture, and spread on a Luria-Bertani (LB) agar plate (100 315 mm) containing 50 mg/mL of ampicillin and 12.5 mL tetracycline.The plate was spread with 50 mL of 2% X-gal and 50 mL of 100 mmol/L isopropyl-beta-D-thiogalactopyranoside before using. The plate wasincubated at 378C overnight. Individual colonies were counted, andwhite colonies were selected to grow overnight at 378C in LB mediacontaining 50 mg/mL ampicillin. Plasmid DNA was digested with PstI and Sst II, and the restricted DNAs were fractionated on a 0.8%agarose gel to verify the presence of the cloned insert. Plasmid DNAcontaining the cloned, amplified insert was purified, and DNA sequenc-ing was accomplished using the Taq DyeDeoxyy Terminator CycleSequencing Kit (Applied Biosystems Inc., Foster City, California) onan automated sequencer (Applied Biosystems Inc., Foster City, Cali-fornia) by the DNA sequencing Core of the Interdisplinary Center forBiotechnology Research at the University of Florida. Sequencing wasaccomplished using vector T7 and Sp6 primers along with specific prim-ers for the rbcL gene received from G. Zurawski (Table 3). The com-plete sequence of both strands was determined using sets of syntheticprimers (Table 3).

Sequence alignment—Sequences of rbcL were easily aligned man-ually because no length variation was detected. For trnL-F, two methodswere employed. Sequences of several taxa representing the range ofprobable variation in the matrix were aligned using the Clustal optionin Sequence Navigator (Applied Biosystems, Inc.), followed by manualoptimisation and alignment of subsequent sequences. Alternatively, theprogram Sequencher (Gene Codes, Inc., Ann Arbor, Michigan) was usedto align sequences of closely related taxa with subsequent builds ofthese smaller alignments performed manually. Copies of the alignedmatrices are available from the senior author.

Analysis—Aligned matrices were analyzed using the parsimony al-

gorithm of the software package PAUP* for Macintosh (v4.0 d59-64,Swofford, 1998) with a successive weighting (SW; Farris, 1969) strat-egy. SW was employed to globally reduce the effect of highly homo-plasious base positions on the resulting topologies (Lledo et al., 1998;Wenzel, 1997). Whole category weights (codon or tranversion) exhibitbroad and overlapping ranges of consistency (Olmstead, 1997), whereasSW independently assesses each base position of the multiple alignmentbased on their consistency in the initial analysis. The initial tree searchwas conducted under the Fitch (equal weights; Fitch, 1971) criterionwith 1000 random sequence additions and SPR (subtree pruning-re-grafting) branch swapping but permitting only ten trees to be held ateach step to reduce the time spent searching trees at suboptimal levels.All trees collected in the 1000 replicates were swapped on to eithercompletion or an upper limit of 5000 trees. The characters were thenreweighted by the rescaled consistency index, and a further 50 repli-cations of random sequence additions were conducted with the weightedmatrix saving 15 trees per replication. These trees were then swappedon to completion or an upper limit of 5000 trees. The resulting treeswere then used to reweight the matix a second time by the rescaledconsistency index, and another 50 replications of random sequence ad-dition conducted, saving 15 trees per replication, with subsequent swap-ping on those trees. This cycle was repeated until two successive roundsfound trees of the same length. All analyses were run with the MUL-PARS option and ACCTRAN optimization. Branches with zero lengthwere collapsed if the maximum value 5 0 (‘‘amb 1’’). Internal supportwas determined by bootstrapping (5000 replicates) with the final rew-eighted character matrix and with the jackknife program (5000 repli-cates) of Farris et al. (1996) without SW weights applied. The cut-offbootstrap percentage is 50; minimum jackknife support percentage is63 (Farris et al., 1996). The rbcL matrix consisted of 81 taxa, 51 Amar-yllidaceae s.s. representing 48 genera, and 30 additional taxa repre-senting 28 genera of Agapanthaceae, Alliaceae, Anthericaceae, Behni-aceae, Convallariaceae, Hyacinthaceae, Laxmanniaceae, and Themida-ceae, with Geitonoplesium sp. (Hemerocallidaceae) used as outgroup.The trnL-F matrix includes these same with the addition of Stemmatiumnarcissoides (Alliaceae).

The trnL-F region consists of an intron, a short exon, and an inter-gene spacer (Taberlet et al., 1991). We combined the components oftrnL-F because they are nearly all noncoding, but each of the two largerregions was analyzed separately to determine whether they were con-gruent. Because they were congruent (results not shown), we lumpedthem together as the ‘‘noncoding matrix’’ to compare directly with rbcLbefore we combined all of them.

RESULTS

The rbcL matrix alone—Of 1340 included base po-sitions in the analyses, 226 were parsimony informative.More than 5000 equally most parsimonious Fitch treeswere found (tree length 5 974) with a consistency index(CI) of 0.62 and a retention index (RI) of 0.71. SW pro-duced at least 5000 equally parsimonious trees with alength of 450819 (Fitch length 5 975), a CI 5 0.88 (Fitch5 0.62), and RI 5 0.89 (Fitch 5 0.70). The large numberof equally parsimonious trees is largely the result of theshort branch lengths that occur within most of the internalclades (Figs. 1–2) and the imposed constraints againstcollapsing zero-length branches. However, the strict con-sensus of the weighted trees is more resolved than thatof the Fitch trees. The additional step of the SW trees isessentially the ‘‘cost’’ of optimizing consistent charactersover highly homoplastic base positions (Lledo et al.,1998). The Amaryllidaceae are not resolved as mono-phyletic in the strict consensus of all 5000 SW trees; inthese Agapanthus and Amaryllidaceae tribe Amaryllideae


Fig. 1. One of 5000 equally parsimonious trees generated by cladistic analysis of the successively weighted rbcL sequence matrix for Ama-ryllidaceae and other Asparagalean genera. Numbers above branches are branch lengths. Bootstrap (plain) and jackknife (boldface) percentages arebelow branches supported by one or both. An asterisk below a branch signifies that both bootstrap and jacknife 5 100%. A white bar across abranch signifies lack of resolution in the strict consensus tree of the 5000 trees. ‘‘Agapanthus afr.’’ 5 A. africanus, ‘‘Agapanthus cam.’’ 5 A.campanulatus, ‘‘Allium sub.’’ 5 A. subhirsutum, ‘‘Allium sic.’’ 5 A. siculum var. bulgaricum. The tree is continued in Fig. 2.

form a polytomy with the rest of Amaryllidaceae sensustricto (s.s.). Moreover, these clades have no bootstrap orjackknife support.

The rbcL matrix (Fig. 1) resolves Hyacinthaceae/Themidaceae as sister to Anthericaceae/Behniaceae withmoderate bootstrap and jackknife support and positions

this clade as sister to Agapanthus/Amaryllidaceae butwith no jackknife or bootstrap support. The Alliaceae areresolved as an unsupported paraphyletic grade.

In many of the trees (Fig. 1), the African tribe Ama-ryllideae is sister to the rest of Amaryllidaceae s.s. Thismonophyletic group has high bootstrap and jackknife


Fig. 2. One of 5000 equally parsimonious trees generated by cladistic analysis of the successively weighted rbcL sequence matrix for Ama-ryllidaceae and other Asparagalean genera. Numbers above branches are branch lengths. Bootstrap (plain) and jackknife (boldface) percentages arebelow branches supported by one or both. A white bar across a branch signifies lack of resolution in the strict consensus tree of the 5000 trees.The tree is continued in Fig. 1.

support (Fig. 1). The rest of the family forms a polytomy(Fig. 2) that includes a baccate-fruited clade (Haeman-theae, including Gethyllideae), the Cyrtantheae (confinedto Africa), Calostemmateae (Australasia), and a mono-phyletic Eurasian/American group. Of these latter, onlythe Eurasian/American clade has any bootstrap (62) andjackknife support (67). Calostemmateae have a bootstrappercentage of 63 but no jackknife support. Within theHaemantheae, Apodolirion and Gethyllis (Gethyllideae)are resolved as sister taxa in the Fitch topologies, but notin the SW trees (Fig. 2).

Within the Eurasian/American clade, the Americangenera are monophyletic in all trees (Fig. 2) but lackbootstrap and jackknife support. The Eurasian genera

form a polytomous grade within this clade. These se-quences resolve the tribe Hippeastreae (excluding Grif-finia and Worsleya 5 Griffineae Ravenna emend.) andHymenocallideae. Tribe Hippeastreae are the only cladein Amaryllidaceae s.s. other than tribe Amaryllideae thatis supported by bootstrap and jackknife percentage great-er than 90% (Fig. 2). Worsleya appears as sister to Chli-danthus (Eustephieae), and Griffinia is unresolved (Fig.2). A subclade representing subtribe Zephyranthinae issupported by low bootstrap percentage, but the remainingrelationships within this clade are unresolved. The rest ofthe American tribes (Eucharideae, Eustephieae, and Sten-omesseae) are not resolved by rbcL, but one unexpectedclade with weak bootstrap support (55) encompasses all


Fig. 3. One of 5000 equally parsimonious trees generated by cladistic analysis of successively weighted trnL-F sequence matrix for Amaryl-lidaceae and other Asparagalean genera. Numbers above branches are branch lengths. Bootstrap (plain) and jackknife (boldface) percentages arebelow branches supported by one or both. An asterisk below a branch signifies that both bootstrap and jacknife 5 100%. ‘‘Agapanthus afr.’’ 5 A.africanus, ‘‘Agapanthus cam.’’ 5 A. campanulatus, ‘‘Allium sub.’’ 5 A. subhirsutum, ‘‘Allium sic.’’ 5 A. siculum var. bulgaricum. The tree iscontinued in Fig. 4.

the included petiolate-leafed Andean taxa with 2n 5 46chromosomes.

The trnL-F matrix alone—Of the 1389 base positions(including gaps) included in the analysis, 378 were par-simony informative. More than 5000 equally most par-simonious trees were found of length 5 1540 with CI 50.66 and RI 5 0.73. SW found more than 5000 equallyparsimonious trees of length 5 747723 (Fitch 5 1541)with CI 5 0.89 (Fitch 5 0.66) and RI 5 0.91 (Fitch 0.73),

the strict consensus of which is more resolved than theinitial Fitch consensus. The trnL-F matrix (Fig. 3) re-solves a monophyletic Amaryllidaceae s.s. (bootstrap andjackknife support . 80%) as sister to Alliaceae with lowbootstrap (56%) and somewhat higher jackknife support(71%). Agapanthus is sister to the Amaryllidaceae/Alli-aceae clade with supporting bootstrap and jackknife per-centages of 81 and 87%, respectively.

In all trnL-F topologies the well-supported Amarylli-deae is sister to the rest of Amaryllidaceae, with high


Fig. 4. One of 5000 equally parsimonious trees generated by cladistic analysis of successively weighted trnL-F sequence matrix for Amaryl-lidaceae and other Asparagalean genera. Numbers above branches are branch lengths. Bootstrap (plain) and jackknife (boldface) percentages arebelow branches supported by one or both. A white bar across a branch signifies lack of resolution in the strict consensus tree of the 5000 trees.The tree is continued in Fig. 3.

bootstrap and jackknife support. As with rbcL, the re-maining African tribes (Haemantheae, Gethyllideae, Cyr-tantheae) and Australasian Calostemmateae (itself, awell-supported clade) form an unresolved polytomy withthe American/Eurasian taxa in the strict consensus (Fig4). Unlike the rbcL topology, Gethyllideae resolves as awell-supported monophyletic subclade of Haemantheaethat is sister to Haemanthus (Fig. 4).

Hannonia and Lycorideae (Lycoris and Ungernia) are

outside of an otherwise monophyletic Eurasian clade inwhich Galantheae (Galanthus and Leucojum) are re-solved with high bootstrap (93%) and jackknife (92%)percentages (Fig. 4). The monophyletic Lycorideae forma weak clade with Griffinia and Hannonia as sister gen-era.

Compared to the rbcL topology, the American generaare less resolved by trnL-F; Griffinia and Worsleya ap-pear outside the clade comprising all other American taxa


Fig. 5. One of 5000 equally parsimonious trees generated by cladistic analysis of successively weighted combined rbcL and trnL-F sequencematrix for Amaryllidaceae and other Asparagalean genera. Numbers above branches are branch lengths. Bootstrap (plain) and jackknife (boldface)percentages are below branches supported by one or both. An asterisk below a branch signifies that both bootstrap and jacknife 5 100%. ‘‘Aga-panthus afr.’’ 5 A. africanus, ‘‘Agapanthus cam.’’ 5 A. campanulatus, ‘‘Allium sub.’’ 5 A. subhirsutum, ‘‘Allium sic.’’ 5 A. siculum var. bulgaricum.The tree is continued in Fig. 6.

(Fig. 4). The petiolate Andean clade, which appears in therbcL consensus, loses two members, Eucharis and Rau-hia. Hymenocallideae are not resolved, and Leptochitonand Pamianthe are resolved as sister genera with mod-erate bootstrap and strong jackknife support. Hippeas-treae (less Griffineae) appear with low bootstrap support(61) but with different internal resolution than with rbcL.Again, short branch lengths are characteristic of most ofthe internal nodes of the American clade (Fig. 4).

The combined matrix—More than 5000 equally mostparsimonious trees were found of length 5 2546 with CI5 0.64 and RI 5 0.71. SW found more than 5000 equallyparsimonious trees of length 5 1194297 (Fitch 5 2546)with CI 5 0.89 (Fitch 5 0.64) and RI 5 0.89 (Fitch 50.71). The strict consensus of the weighted trees is moreresolved than the initial Fitch consensus. Agapanthus issister to Amaryllidaceae in the combined topologies (Fig.5), albeit with low bootstrap support (60%). A monophy-


Fig. 6. One of 5000 equally parsimonious trees generated by cladistic analysis of successively weighted combined rbcL and trnL-F sequencematrix for Amaryllidaceae and other Asparagalean genera. Numbers above branches are branch lengths. Bootstrap (plain) and jackknife (boldface)percentages are below branches supported by one or both. The tree is continued in Fig. 5.

letic Alliaceae is sister to the former clade, with a boot-strap of 79% and jackknife of 77%. Both Amaryllideaeand Haemantheae are well-supported tribal clades (Figs.5–6), with higher bootstrap and jackknife percentagesthan in either of the separate analyses, and the formerresolves as sister to the rest of Amaryllidaceae s.s.

Within Amaryllideae, most of the included genera re-solve in a grade with Amaryllis and then Crinum as the

successive sister taxa to the rest (Fig. 5). Within Hae-mantheae, a well-supported, monophyletic Gethyllideaeis again sister to Haemanthus (Fig. 6). As in both theindividual analyses, Calostemmateae and Cyrtantheae re-main as part of the polytomy inclusive of Haemantheaeand the large Eurasian/American clade.

In the combined analysis, the Neotropical (American)and Eurasian genera are sister groups with strong boot-


strap and jackknife support (95, 98%), but of the two,only the American clade has weak bootstrap support (Fig.6). Galanthus/Leucojum, Hannonia/Vagaria, and Pan-cratium/Sternbergia are supported sister genera, but theremaining relationships, consistent in all trees, have nosupport.

Within the American clade (Fig. 6), a distinct Andeansubclade has weak bootstrap support (68). Eustephieaehave no consensus, bootstrap, or jackknife support. Awell-supported Hippeastreae are in a polytomy with Grif-finia, Worsleya, and the Andean clade. In Hippeastreae,a distinct Zephyranthinae and Rhodophiala/Traubia arewell supported. Within the Andean clade, the resolutionof Hymenocallideae, observed in the rbcL trees (Fig. 2),is lost, with Ismene, however, remaining monophyletic.Leptochiton and Pamianthe are weakly supported sistertaxa. Eucharis and Rauhia fail to join the rest of a weaklysupported petiolate-leafed subclade that is sister group toHymenocallis (marked by a single synapomorphy).

DISCUSSION

Suprafamilial relationships of Amaryllidaceae—Fayand Chase (1996) presented a smaller rbcL analysis andargued for the inclusion of Agapanthus as a monotypicsubfamily within Amaryllidaceae. The sister-group statusof Agapanthus to Amaryllidaceae s.s. is only weakly sup-ported by our combined matrix (bootstrap 5 60%). Basedon these data, it would be possible to argue for recog-nizing Amaryllidaceae in a modified Hutchinsonian(1934) sense, i.e., with three subfamilies, Allioideae,Agapanthoideae, and Amarylloideae.

Backlund and Bremer (1998) discussed the issue ofmonogeneric families and how best to treat them. Theygenerated a set of guiding principles for classification:(1) primary principle of monophyly and (2) a set of sec-ondary principles: (a) maximizing stability, (b) maximiz-ing phylogenetic information (5 minimizing redundan-cy), (c) maximizing support for monophyly, and (d) max-imizing ease of identification. Principle 1 is consideredmost important by Backlund and Bremer (1998), as it isby most modern systematists. However, the secondaryprinciples will vary in importance among different taxa.

Monophyly is maximized by either treating Agapan-thus as a monogeneric family or accepting Amaryllida-ceae in the Hutchinsonian sense. However, the supportfor a broad concept of Amaryllidaceae (including Alli-aceae and Agapanthaceae) is only moderate (bootstrap 579%, jackknife 5 77%). The only morphological char-acter that unites all three families is the pseudo-umbellateinflorescence (homoplasious with Themidaceae), whereasthe Alliaceae are readily marked by their solid styles andsulfonated compounds, and the Amaryllidaceae have in-ferior ovaries and unique alkaloid chemistry. Maximizingstability in this case seems rather moot, given that Aga-panthus has been maintained for years as part of Alli-aceae, a classification that violates the primary principleof monophyly, and Amaryllidaceae and Alliaceae havebeen united on and off again over the last two centuries.However, maximizing phylogenetic information and easeof identification are best served by treating Agapanthusas the sole genus of a separate family (Agapanthaceae

Voight), while maintaining the independent status of Al-liaceae.

The combined analysis supports most of the other re-lationships hypothesized by Fay and Chase (1996). Thesister-group status of Themidaceae and Hyacinthaceae isconfirmed with good support, and there is bootstrap sup-port in the combined analysis for this clade as sistergroup to Amaryllidaceae/Alliaceae/Agapanthaceae, al-though Antheridaceae/Behniaceae resolves outside of thisclade. It should be noted that, with this level of samplingof the broader Asparagales, these relationships are largelya matter of outgroup selection.

Relationships within Amaryllidaceae—Within Ama-ryllidaceae s.s., several groups are well supported withinall of the analyses, some of which correspond to tradi-tionally accepted tribes of the family. The most unex-pected resolution concerns the sister status of the Eur-asian/American clades. This is only supported in thecombined analysis, and resolution of this group in rela-tion to the remaining African and Australasian clades isstill elusive because of short branch lengths in this por-tion of the trees (Figs. 2, 4, 6).

In a survey of internal morphology of American andAfrican Amaryllidaceae, Arroyo and Cutler (1984) notedseveral characters that separated American genera fromAfrican. All American species surveyed have scapes withcollenchyma, a one-layered rhizodermis, and obvolutebracts. All Amaryllideae (entirely African with the ex-ception of pantropical Crinum) have schlerenchyma inthe scape, a multilayered rhizodermis, and equitantbracts. Haemanthus and Cyrtanthus exhibit scape androot anatomy of the American species but the equitantbracts of Amaryllideae (Arroyo and Cutler, 1984). Ca-lostemmateae (Calostemma and Proiphys), which werenot discussed by Arroyo and Cutler (1984), have equitantbracts. Many of the Eurasian genera have fused spathebracts, which obscures the pattern of their coherence, butboth Lycoris and Pancratium species with free bractsshow the equitant condition. Obvolute bracts may thusbe a synapomorphy of the American clade.

Two American subclades are found in the consensusof the combined analysis (Fig. 5), with both Griffinia andWorsleya forming a polytomy with them. The moreweakly supported Andean subclade (tribes Eucharideae,Stenomesseae and Eustephieae) is characterized by 2n 546 chromosomes, which has been interpreted as a tetra-ploid derivation from an ancestral 2n 5 22 (Meerow,1985, 1987a, c, 1989). The strongly supported Hippeas-treae is characterized for the most part by x 5 6 or 11,with diploid chromosome numbers of 22, 24 or less. Theshort branch lengths and numerous polytomies in the An-dean group (Fig. 6) may indicate that they are a relativelyyoung clade with an evolutionary history tied closely tothe geologically recent Andean uplift (Meerow, 1987c).

Four recognized tribes of Amaryllidaceae are consis-tently resolved by the plastid DNA sequences, and allreceive strong bootstrap and jackknife support in at leastthe combined analysis. These are the Amaryllideae, Hae-mantheae, Calostemmateae, and Hippeastreae.

Amaryllideae—This tribe, with much of its generic di-versity confined to South Africa is sister to the rest of


the Amaryllidaceae and has high bootstrap and jackknifesupport. Compared to other tribes in Amaryllidaceae,Amaryllideae are marked by a large number of synapo-morphies (Snijman and Linder, 1996): extensible fibersin the leaf tissue, bisulculate pollen with spinulose exines,scapes with a sclerenchymatous sheath, unitegmic orategmic ovules, and nondormant, water-rich, nonphyto-melanous seeds with chlorophyllous embryos. A few ofthe genera extend outside of South Africa proper, butonly Crinum, with seeds well adapted for oceanic dis-persal (Koshimizu, 1930), ranges through Asia, Australia,and America. Snijman and Linder’s (1996) phylogeneticanalysis of the tribe based on morphological, seed ana-tomical, and cytological data resulted in recognition oftwo monophyletic subtribes: Crininae (Boophone, Cri-num, Ammocharis, and Cybistetes) and Amaryllidinae(Amaryllis, Nerine, Brunsvigia, Crossyne, Hessea, Stru-maria, and Carpolyza). Muller-Doblies and Muller-Dob-lies (1996) recognized four subtribes with little discus-sion and no phylogenetic analysis: Crininae, Boophoni-nae, Amaryllidinae, and Strumariinae, the latter two con-taining several segregate genera from Hessea andStrumaria (Table 1). Our sampling of this tribe is incom-plete, and therefore we feel it is premature to attach agreat deal of confidence to the generic sister relationshipsseen here. Four genera of Snijman and Linder’s (1996)Amaryllidinae do form a weakly supported clade (Fig. 5)with Amaryllis as sister to the rest of the tribe in the rbcLand combined analyses. Identical positioning of Amaryl-lis occurred in Snijman’s (1992) cladistic analyses if tribeHippeastreae was used as the outgroup, with Haeman-theae as outgroup (Snijman and Linder, 1996), and alsoboth outgroups used (Snijman, 1992). Amaryllis resolvesas sister to a clade containing the other genera they ul-timately placed, with Amaryllis, in subtribe Amaryllidi-nae. Muller-Doblies and Muller-Doblies’ (1996) conceptof Amaryllidinae [Amaryllis, Nerine, and Namaquanula(5 Hessea)] would make their subtribe Strumariinae par-aphyletic and Amaryllidinae polyphyletic.

Haemantheae—This baccate-fruited tribe is anothermorphologically well-marked group with strong molec-ular support. The limits of the tribe, however, have beencontroversial. Muller-Doblies and Muller-Doblies (1996)insisted on retaining Cyrtanthus in the tribe, albeit as amonotypic subtribe, Cyrtanthinae. The basis for unitingCyrtanthus with the Haemantheae has always been weak,chiefly the shared chromosome number with Haemanthus(2n 5 16; Ising, 1970; Vosa and Snijman, 1984) and itsstrictly African range. This diploid number also occursin some Hippeastreae (Flory, 1977; Grau and Bayer,1991). Uniting Cyrtanthus with Haemantheae has no mo-lecular support in our analyses, and we believe that Cyr-tanthus, the only solely African genus with the flattened,winged, phytomelanous seed so common in the Americanclade, should be recognized as a monotypic tribe (Traub,1963; Dahlgren, Clifford, and Yeo, 1985; Meerow andSnijman, 1998). Recognition of Gethyllideae as a distincttribe (Muller-Doblies and Muller-Doblies, 1996; Meerowand Snijman, 1998), however, is not supported by themolecular data. Although the large, elongate, baccatefruits and small hard seeds of Apodolirion and Gethyllisare a departure from the berries and large succulent seeds

of the rest of Haemantheae, the two genera, though re-solved as sister taxa (Figs. 4, 6), are firmly embeddedwithin Haemantheae. Recognizing them as a distinct tax-on would render the rest of Haemantheae paraphyletic.Haemantheae are the only tribe of Amaryllidaceae thatcontain rhizomatous genera (Cryptostephanus and Sca-doxus in part), a condition that occurs in the sister familyAgapanthaceae. This has generally been conceived as aplesiomorphy within the family (Nordal and Duncan,1981; Meerow, 1995, 1997; Muller-Doblies and Muller-Doblies, 1996). In the rbcL and combined consensustrees, the three ‘‘bulbless’’ genera form a grade at thebase of the Haemantheae, which would support this hy-pothesis, although Cryptostephanus, the only member ofthe tribe with the ancestral state of a phytomelanous testa,is not the first branch in the grade. Haemanthus and Sca-doxus, which have been treated as one genus in the past(e.g., Hutchinson, 1934, 1959; Traub, 1963), are sistergenera only in the rbcL topologies (Fig. 2). The positionof Scadoxus, the only genus of the tribe polymorphic forthe rhizomatous state, as the final terminal taxon in the‘‘bulbless’’ grade seems reasonable. In any event, allthree matrices render recognition of a subtribe Cliviinaefor Clivia and Cryptostephanus by Muller-Doblies andMuller-Doblies (1996) as paraphyletic. Any further in-sight on the internal relationships within Haemantheaerequires additional sampling.

Calostemmateae—Calostemmateae, treated as part of apolyphyletic Eucharideae by Hutchinson (1934, 1959),Traub (1963), and Dahlgren, Clifford, and Yeo (1985),were first suggested as a distinct lineage by Meerow(1989) and formally recognized by Muller-Doblies andMuller-Doblies (1996). The tribe consists of two Austra-lasian genera (Proiphys, forest understory herbs of Ma-laysia, Indonesia, the Philippines and tropical Australia,and Calostemma, endemic to Australia). A few speciesof Crinum, with the broadest distribution of any genus inthe family, are the only other members of Amaryllidaceaepresent in Australia. The indehiscent capsules of bothgenera are similar in appearance to the unripe berry-fruitsof Scadoxus and Haemanthus (Haemantheae), but earlyin the development of the seed, the embryo germinatesprecociously, and a bulbil forms within the capsule andfunctions as the mature propagule (Rendle, 1901). Thetwo genera exhibit the equitant bract condition of theAfrican and Eurasian genera.

Hippeastreae—All but two of the genera treated byMeerow and Snijman (1998) as part of Hippeastreae areresolved as a well-supported monophyletic clade in allthe analyses (Figs. 2, 4, 6). The two genera that lie out-side of this clade are Worsleya and Griffinia, both Bra-zilian endemics, exhibiting the rare character of blue-range pigmentation in the flowers. The variable position-ing of these two in the various analyses is interesting initself. In the trnL-F topologies (Fig. 4), Worsleya is partof the basal polytomy within the Eurasian/Americanclade, whereas Griffinia weakly resolves as sister to theMediterranean Hannonia (the latter on a long terminalbranch). In the rbcL consensus (Fig. 2), Worsleya re-solves as sister to Chlidanthus (Eustephieae), whereasGriffinia remains unresolved along with the rest of Eus-


tephieae. In the combined analysis (Fig. 6), both are po-sitioned within the American clade, but unresolved witheither Hippeastreae s.s. or the weakly supported tetraploidAndean clade. The failure of Worsleya or Griffinia toresolve as part of Hippeastreae in any of the analysescasts doubt on Muller-Doblies and Muller-Doblies’(1996) submergence of Worsleya in Hippeastrum andweakens Meerow and Snijman’s (1998) retention of bothgenera in tribe Hippeastreae.

Another unexpected indication of relationship occurswithin the tetraploid Andean clade, where a distinct pet-iolate-leafed subclade is resolved in the rbcL topologies(Fig. 2). This resolution is not retained by the trnL-F andcombined analyses in which Eucharis and Rauhia arepulled from this group. Nonetheless, a core of petiolategenera remain monophyletic, with weak support in thecombined analysis. Despite the fact that petiolate leaveshave evolved independently several times elsewhere inthe Amaryllidaceae (Amaryllideae, Calostemmataceae,Haemantheae, Hymenocallideae, and Hippeastreae), themolecular data begin to indicate that it may be a syna-pomorphy for this group. Hymenocallideae as a distincttribe receive weak support (55%, one synapomorphy) inthe rbcL matrix only, and the rest of Stenomesseae ispoorly resolved by all matrices.

Within the Eurasian clade of the combined analysis(Fig. 6), Lycorideae appears as sister to the rest, althoughwithout support. This tribe represents the more or less tem-perate Asian component of the family, with Lycoris rang-ing from Korea, through China, Myanamar, and Japan, andUngernia restricted to the mountains of central Asia. Mul-ler-Doblies and Muller-Doblies (1978) described similari-ties in the bulb anatomy of Ungernia and Sternbergia, apossible synapomorphy between Lycorideae and the restof this clade. One genus of the Eurasian clade, Pancrati-um, is represented throughout Africa, tropical Asia, as wellas Mediterranean Europe and the Middle East, a distri-bution that could signify a more ancestral position withinthe Eurasian clade. The current data, however, do not sup-port this resolution for Pancratium, with only a single spe-cies from the Canary Islands represented in the analyses.The presence of Pancratium in Africa may thus be sec-ondary, though it is the only genus outside of the Africantribes Amaryllideae and Haemantheae with external tri-chomes (Bjornstad, 1973).

Sister relationships of Hannonia and Vagaria receivegood bootstrap and jackknife support, as does the tradi-tional alliance of Galanthus and Leucojum. Crespo et al.(1995), using ITS sequences, refuted Muller-Doblies andMuller-Doblies’ placement of Lapiedra in Pancratieae,and our data support a closer relationship with Galan-theae or Narcisseae for this genus. However, concepts ofGalantheae, Narcisseae, and Pancratieae presented inMuller-Doblies and Muller-Doblies (1996) or Meerowand Snijman (1998) are not resolved in any of the threeanalyses, and we believe that caution should be used be-fore categorical statements are made about tribal lineageswithin this group.

The low internal branch lengths throughout the Ama-ryllidaceae, except in some of the deepest branches, area striking contrast to the other asparagalean families in-cluded in the analysis (Figs. 1, 3, 5). The significance ofthis is not clear. It could mean that a great deal of the

modern diversity in the family is of relatively recent oc-currence (as is likely, for example, within the Andeanclade), or else base substitution rates in the chloroplastgenome are lower within the family than for other As-paragales.

Character state evolution in the Amaryllidaceae—Byoptimizing morphological or other ‘‘traditional’’ charac-ters onto a gene tree, one is able to gain insight aboutputative transformation series or state polarities that havecharacterized the evolution of the group under study. Thiscan be useful for constructing a character state matrix foran ingroup in which rampant homoplasy in such char-acters confounds the endeavor. Certain characters thathave been used to justify older intrafamilial classifica-tions of Amaryllidaceae do show stability within someof the clades resolved by the combined analysis, whileothers appear extremely homoplasious (Fig. 7).

The most notable correlation between evolutionarydepth as resolved by plastid DNA sequences and mor-phological synapomorphies is found in the Amaryllideae(Fig. 7). All of the characters listed are synapomorphousfor the tribe, which terminates the longest internal branchwithin Amaryllidaceae on our gene trees (Figs. 1, 3, 5).

Presence or absence of bulbs—The bulbless conditionoccurs in the sister group to Amaryllidaceae, the mono-generic Agapanthaceae. It is also the character state forthe only South African subfamily of Alliaceae, Tulbagh-oideae (Fay and Chase, 1996). In Amaryllidaceae, theabsence of bulbs characterizes only three genera, Clivia,Cryptostephanus, and Scadoxus, but the latter also in-cludes species that form a true bulb. If this is a symple-siomorphy as most have interpreted it (Nordal and Duncan,1984; Muller-Doblies and Muller-Doblies, 1996; Meerowand Snijman, 1998), the bulbous state has evolved at leastthree times in the family, in Amaryllideae, Haemantheae,and within the ancestral stock for the rest of the family.

Petiolate leaves—Petiolate or, more accurately, pseu-dopetiolate leaves are widepread throughout the Aspara-gales, and this character exhibits a great deal of homo-plasy within Amaryllidaceae (Meerow and Snijman,1998). At the extreme, one-to-few petiolate species occurin otherwise lorate-leafed genera (e.g., Crinum, Hymen-ocallis). The state may occur throughout a genus, butrenders a tribe polymorphic (Calostemmateae, Haeman-theae, Griffineae). In the tetraploid Andean clade, a sub-clade is defined by the synapomorphy of a petiolate leafin the rbcL trees, but Eucharis and Rauhia pull away withtrnL-F and in the combined analyses.

Mesophyll palisade—It has been suggested that thepresence or absence of a distinct palisade layer in the leafmesophyll may have systematic significance (Arroyo andCutler, 1984; Artyushenko, 1989). Petiolate-leafed taxanever have palisade chlorenchyma (Meerow and Snij-man, 1998). It is characteristic of Amaryllideae (Crinumis polymorphic), but absent in Haemantheae (the state inunknown for Gethyllis and Apodolirion). In Calostem-mateae, it is present in Calostemma but absent in thepetiolate Proiphys (Meerow, unpublished data). Palisadealmost universally occurs in the Eurasian clade. It is ab-


Fig. 7. Amaryllidaceae/Agapanthaceae clade from one of 5000 equally parsimonious trees generated by cladistic analysis of the successivelyweighted combined rbcL and trnL-F sequence matrix for Amaryllidaceae and other Asparagalean genera with selected morphological and karyo-logical states optimized on the tree. A ‘‘P’’ to the right of a character bar or box refers to the state being polymorphic in the adjacent taxon; ifsuperimposed on the character bar itself, polymorphy is widespread among all adjacent taxa. ‘‘Agapanthus afr.’’ 5 A. africanus, ‘‘Agapanthuscam.’’ 5 A. campanulatus, Ismene E 5 I. subg. Elisena, Ismene I 5 I. subg. Ismene, Ismene P 5 I. subg. Pseudostenomesson, Stenomesson p. 5S. pearcei, Stenomesson var. 5 S. variegatum.

sent in Leucojum (Artyushenko, 1989), and Galanthus ispolymorphic (Davis and Barnett, 1997). Within theAmerican clade, it is wholly characteristic of the Euste-phieae (Arroyo and Cutler, 1984; Meerow and Snijman,1998) but occurs only sporadically within the Hippeas-

treae (Arroyo and Cutler, 1994). The state of Worsleya isnot known. Outside of Eustephieae, a distinct palisade isabsent from the Andean tetraploid clade (Meerow, 1987a,1989). The inference based on the distribution of thischaracter state on our topology (Fig. 7) is that a distinct


palisade is plesiomorphic within the family, though thestate within Agapanthaceae, sister to Amaryllidaceae, hasnot to our knowledge been reported.

Pubescence—The presence of trichomes on the exter-nal parts of Amaryllidaceae is common only in someAmaryllideae and Haemantheae and one African speciesof Pancratium (Arroyo and Cutler, 1984; Meerow andSnijman, 1998). It is completely unknown in the Amer-ican clade, Cyrtantheae, and Calostemmateae. It mayhave evolved independently in the three clades withinwhich it occurs.

Scape characters—Solid scapes are the predominantcondition in Amaryllidaceae as occurs in Agapanthaceaeas well. Hollow scapes are almost universally character-istic of Hippeastreae, and thus appears to be a synapo-morphy for that tribe. The only other genera within whichhollow scapes occur are Leucojum and Cyrtanthus bothof which are polymorphic for the character (Traub, 1963;Reid and Dyer, 1984). As discussed previously, obvolutespathe bracts seem to be apomorphic for the Americanclade, and the presence of schlerenchyma in the scape isan autapomorphy for Amaryllideae.

Floral symmetry—Zygomorphic and actinomorphicflowers occur in the Amaryllidaceae, and several genera(Crinum, Cyrtanthus, Phycella) are polymorphic. Snij-man and Linder (1996) consider actinomorphy the apo-morphic condition in Amaryllideae. The flowers of Aga-panthaceae are zygomorphic. Within Haemantheae, onlyClivia is zygomorphic. In the American clade, zygomor-phy is the rule in the ‘‘hippeastroid’’ subclade. Pyrolirionand Zephyranthes (including Haylockia) are the only gen-era characterized exclusively by actinomorphic flowers,while Phycella (not included in the sequence analyses) ispolymorphic. In the Andean subclade, only Eucrosia,Plagiolirion, and Ismene subgenus Elisena are exclusive-ly zygomorphic; Rauhia is polymorphic. The Eurasianclade is on the whole actinomorphic; only Lycoris ischaracterized by zygomorphic flowers. The mosaic oc-currence of actinomorphy throughout the family (Fig. 7)and the occurrence of polymorphic genera suggest thattransformations between the two states of floral symme-try may be easily modified by pollinator-mediated selec-tion, and perhaps controlled by one or few genes.

Paraperigone—The ‘‘paraperigone’’ is an anomaloussecondary outgrowth of the perianthal meristem withramifying vasculature (Arber, 1939; Singh, 1972), not tobe confused with a similar-looking structure formed bystaminal connation (see below). It is most well developed(and typified) by the corona of Narcissus. Such a well-developed paraperigone occurs in only one other genus,the Chilean endemic Placea (Hippeastreae). However, ahomologous series of fimbrae, scales, or a continuous cal-lose ring occurs in Cryptostephanus (Haemantheae), oneor two species of Cyrtanthus, and variably thoughout Ly-corideae and Hippeastreae. It has thus probably evolvedat least three times (Haemantheae, Cyrtantheae, and theEurasian/American clade), but from a meristematic po-tential that is deep rooted in the family. Polymorphism

for this character within genera may suggest that it iseasily lost.

Staminal connation—The fusion of the staminal fila-ments was the single most important character withwhich Traub (1957, 1963) justified recognizing his ‘‘in-frafamily’’ Pancratioidinae, a subfamilial taxon that infact was glaringly polyphyletic. Though staminal con-nation is a widespread character state within the Andeantetraploid clade (Fig. 7), it is paralleled elsewhere in thefamily, particularly in Amaryllideae subtribe Amarylli-dinae (Snijman and Linder, 1996), the Calostemmateae,in Gethyllis, and some species of Cyrtanthus (Reid andDyer, 1984). In the Eurasian clade it occurs in Pancra-tium, the flower morphology in general of which bearsstriking resemblance to several Andean genera (Hymen-ocallideae pro parte, Paramongaia, and Pamianthe).Meerow and Dehgan (1985) attemepted to link these so-called ‘‘pancratioid’’ genera by pollen morphology, but amore parsimonious explanation may be convergence forpollinator specificity (Morton, 1965; Bauml, 1979; Grant,1983). However, Pancratium and these Andean generaare monophyletic in a larger sense (as part of the Eur-asian/American clade), and the exact position of Pancra-tium within the Eurasian subclade is still not stronglyresolved (Fig. 6).

Fruit and seed characters—Fruit and seed morphologyhave been an important focus of experimentation withinthe family. Baccate fruits have apparently evolved onlyonce, despite the difference in gross morphology betweenthe long, aromatic fruit of Gethyllis and Apodolirion andthe berries of the rest of Haemantheae. Phytomelan [theancestral state for all Asparagales (Huber, 1969)] hasbeen lost from the testa as many as five times in theAmaryllidaceae: in Amaryllideae, Griffineae, Hymeno-callideae, Haemantheae, and Calostemmateae [in Calos-temmateae a true seed never forms, but an integumentaryrudiment is present (Rendle, 1901)]. In both Haeman-theae and Hymenocallideae, phytomelan is found aroundthe seeds of one genus each (Cryptostephanus and Lep-tochiton, respectively). The loss of one integument [orboth, as been controversially reported for some Crinum(Prillieux, 1858; von Schlimbach, 1924; Tomita, 1931;Markotter, 1936, but see Snijman and Linder, 1996)] issynapomorphic for Amaryllideae.

A flattened, winged seed, which occurs in Agapantha-ceae, is very common in the American clade, but other-wise occurs only in Ungernia (Lycorideae) and Cyrtan-theae. The most similar type of seed to this is the D-shaped seed of Worsleya and some Pancratium. A dry,hard, wedge-shaped or irregularly round seed is charac-teristic of most of the Eurasian clade (except Lycorideae),frequently with an elaiosome at the chalazal end. Amongall genera of the family, Pancratium is the most poly-morphic for seed type (Werker and Fahn, 1975).

Characterization of certain seeds of Amaryllidaceae asfleshy (regardless of whether phytomelan is present) hasled, in the past, to false homologies (see discussion inMeerow, 1989). Truly fleshy seeds occur in Amaryllideae(in which case the bulk of the seed volume is endosperm;Rendle, 1901), Hymenocallideae (the fleshy portion is in-tegumentary; Whitehead and Brown, 1940), and some


water-rich Haemantheae. But an ‘‘intermediate’’ state oc-curs in a number of genera in which the seed is round,turgid, but not really fleshy (the seed will burst underpressure rather than give way), and contains copious, oilyendosperm. This type of morphology is found in Cryp-tostephanus (Haemantheae), Lycoris (Lycorideae), Eu-charideae sensu Meerow (1989), Griffinia, and a singlespecies of Hippeastrum.

Chromosome number—A chromosome number of 2n5 22 is considered plesiomorphic in Amaryllidaceae dueto the broad occurrence in many of the tribes of the fam-ily (Goldblatt, 1976; Flory, 1977; Meerow, 1984, 1987b).Andean-centered genera in the tribes Eucharideae, Eus-tephieae, Hymenocallideae, and Stenomesseae are char-acterized by a somatic chromosome number of 2n 5 46or presumptive derivations thereof (Di Fulvio, 1973, Flo-ry, 1977; Williams, 1981; Meerow, 1987a, b). Resolutionof these genera as a clade in our plastid DNA trees sup-ports the interpretation of a monophyletic polyploid ori-gin for these tribes from an ancestor with 2n 5 22 viachromosome fragmentation or duplication and subse-quent doubling or vice versa (Sato, 1938; Lakshmi, 1978;Meerow, 1987b).

Biogeographic implications—Raven and Axelrod(1974) postulated a western Gondwanaland origin forAmaryllidaceae sensu Huber (1969), and this is support-ed by the plastid DNA phylogeny. The deepest branchesof the topology originate in Africa, including the sistergroup of the family Agapanthus. Africa has also been thesite of considerable innovation in the family’s history aswell, as typified by the Afrocentric tribes Amaryllideae,Haemantheae, and Cyrtantheae. Most of the diversitywithin those three tribes is, however, centered in SouthAfrica, and thus may reflect radiation engendered by themore recent paleoclimatic and geological history of Af-rica encompassing Neogene and later times (Axelrod,1972; Raven and Axelrod, 1974). The increased aridityof the African climate and the uplift of the continentalmass beginning near the end of the Oligocene, furtherabetted by Quaternary climatic fluctuations, were cata-strophic to many elements of the African flora, but it mayhave been a selective pressure for diversity among groupsof geophytes capable of adapting to increasing drought.The geophyte richness of South Africa is well docu-mented (Goldblatt, 1978), and the Cape region has beensuggested as a possible refuge for certain African plantand animal groups as the tropical flora of the continentwas impoverished (Raven and Axelrod, 1974). However,the three basal genera of the baccate-fruited Haemantheaeaccording to our combined analysis (Fig. 6), Clivia,Cryptostephanus, and Scadoxus, are all forest understorytaxa, do not form bulbs, and are at least in part (Scadoxus,Cryptostephanus) elements of tropical vegetation farthernorth. Cryptostephanus does not occur in South Africa atall, and this is the only genus of Haemantheae in whichthe plesiomorphic state of a phytomelanous testa occurs.

The Calostemmateae, the only exclusively Australasianelement of the family, may have been isolated from theAfrican lineages as Australia separated from westernGondwanaland (Raven and Axelrod, 1974). Direct mi-gration between Africa and Australia may have persisted

up through the close of the early Cretaceous, althoughIndia and Madagascar may have provided a less directcorridor up until the late Cretaceous (Raven and Axelrod,1974). That the Calosternmateae remains within the un-resolved grade of otherwise African tribes would suggestrelative antiquity for the lineage. Crinum is the onlyamaryllid that is known to occur on Madagascar, despitethe island’s probable role as a refuge for taxa decimatedby the Neogene African extinctions, whereas indigenousIndian amaryllids are restricted to Crinum and two tothree species of Pancratium. The adaptations of Crinumfor long-distance dispersal have been demonstrated (Ko-shimizu, 1930), and Pancratium may have been able todirectly enter India from either Africa or Eurasia duringthe late Cretaceous or early Eocene (Raven and Axelrod,1974).

The sister relationship of the Eurasian/Mediterraneanclade to the American genera raises the interesting ques-tion of when and where the Amaryllidaceae, in the main,entered the New World. It should be noted that this prob-ably occurred at least twice, as the arrival of Crinum inthe Americas via oceanic dispersal was undoubtedly anunrelated event (Arroyo and Cutler, 1984). Although mi-gration between Eurasia and North America has beenpossible throughout most of angiosperm history (Ravenand Axelrod, 1974), the hypothesized pathways havebeen for plants of temperate forest biota and not consid-ered to be important for plants of subhumid or semiaridvegetation (Raven, 1971, 1973). However, eastern NorthAmerica and western Europe may have shared a warm,seasonally dry climate from the late Cretaceous to theearly Eocene (Axelrod, 1973, 1975), which might haveallowed east/west movement of species, with islandchains of the Mid-Atlantic ridge providing steppingstones. Such a Madrean-Tethyan hypothesis would havethe initial entry of the Amaryllidaceae into the NewWorld through North America. Although there are mem-bers of the family in Mexico and the southern UnitedStates, they are, with the exception of the ubiquitous Cri-num, components of terminal subclades (Zephyranthes,Habranthus, Hymenocallis) in an overall American phy-logeny based on nuclear DNA ITS sequences (Meerow,Guy, and Li, 1998), all of which are linked to more basaltaxa endemic to South America. The validity of the Mad-rean-Tethan hypothesis has more recently been ques-tioned by various studies using isozyme, plastid DNArestriction fragment length polymorphisms (RFLPs), orcladistic analyses of taxa considered emblematic of thedisjunction: Buxus (Kohler and Bruckner, 1989), Datisca(Liston, Rieseberg, and Hanson, 1992), Lavatera (Ray,1994), Quercus (Manos, 1992; Nixon, 1993), Pinus (Lit-tle and Crutchfield, 1969; Miller, 1993), and Styrax(Fritsch, 1996). In these cases, the hypothesized Mad-rean-Tethyan linkage is not resolved as monophyletic, thetaxon itself is not monophyletic, or the estimated time ofdivergence does not fit the Madrean-Tethyan hyothesis.

Given the extant distribution of Amaryllidaceae inNorth America and the generic richness south of theequator, a northern latitude entry into the New World forthe family would necessitate massive extinction in NorthAmerica sometime after migration to South America tookplace. Glaciation would be the likely factor involved. Lit-tle migration of plants from North America to South


America probably took place before the Eocene (Ravenand Axelrod, 1974). All indications are that the move-ment of extant Amaryllidaceae has been northward fromSouth America (e.g., Meerow, 1987b, 1989). This doesnot necessarily preclude an earlier, initial arrival in NorthAmerica, migration to South America, and a more recent,but secondary, return of some elements of the family toNorth America long after glaciation extirpated the found-er populations. However, if a North American entry ishypothesized, this begs the question of why the Eurasiansister clade has been so successful in adapting to tem-perate habitats, which constitute the majority of the spe-cies in tribes Galantheae, Narcisseae, and Lycorideae,whereas the American clade is relatively depauperate oftemperate climate adaptation. There is nothing in our datato prove or disprove an initial New World entry of theAmaryllidaceae into North America, and the issue is forthe present unresolved.

In conclusion, our combined analysis of plastid DNAsequences rbcL and trnL-F provide good support for themonophyly of the Amaryllidaceae and indicate Agapan-thaceae as its likely sister family. The Alliaceae are inturn sister to the Amaryllidaceae/Agapanthus clade. Theorigins of the family are African. The phylogenetic re-lationships with Amaryllidaceae s.s. resolve stronglyalong biogeographic lines. The tribe Amaryllideae, pri-marily South African and well supported by numerousmorphological synapomorphies, is sister to the rest ofAmaryllidaceae. The remaining two African tribes of thefamily, Haemantheae and Cyrtantheae, are well support-ed, but their position relative to the Australasian Calos-temmateae and a large clade comprising the Eurasian/American genera, is not yet clear. The Eurasian elementsof the family and the American genera are monophyleticsister clades. Internal resolution of the Eurasian cladeonly partially supports currently accepted tribal concepts,and few conclusions can be drawn on the relationshipsof the genera based on these data. A monophyletic Ly-corideae (Central and East Asian) is weakly supported.Galanthus and Leucojum (Galantheae pro parte) are sup-ported as sister genera by the Bootstrap. The Americanclade shows a higher degree of internal resolution. Amonophyletic Hippeastreae (less Griffinia and Worsleya)is well supported, and a distinct subtribe, Zephyranthinae,is resolved as well. A distinct Andean clade marked bya chromosome number of 2n 5 46 and derivations there-of is resolved with weak support, and a distinct petiolateAndean subclade composed of elements of the tribes Eu-charideae and Stenomesseae is partially resolved withweak support. The lack of resolution of Griffinia andWorsleya in the overall American clade, and of Euste-phieae in the Andean subclade, may indicate that thesegenera represent more isolated elements of the Americanlineage.

LITERATURE CITED

ADANSON, M. 1763. Famille des plantes. Paris.ARBER, A. 1937. Studies in flower structure III. On the corona and

androecium in certain Amaryllidaceae. Annals of Botany II, 1: 293–304.

ARROYO, S. C., AND D. F. CUTLER. 1984. Evolutionary and taxonomicaspects of the internal morphology in Amaryllidaceae from SouthAmerica and Southern Africa. Kew Bulletin 39: 467–498.

ARTYUSHENKO, Z. T. 1989. Aspects of research on Amaryllidaceae Jau-me. Herbertia 45: 131–137.

AXELROD, D. I. 1972. Edaphic aridity as a factor in angiosperm evo-lution. American Naturalist 106: 311–320.

———. 1973. History of the Mediterranean ecosystem in California.In F. di Castri and H. A. Mooney [eds.], Mediterranean type eco-systems: origin and structure, 225–227. Springer-Verlag, Berlin.

———. 1975. Evolution and biogeography of Madrean-Tethyan scleo-phyll vegetation. Annals of the Missouri Botanical Garden 62:280–334.

BACKLUND, A., AND K. BREMER. 1998. To be or not to be—principlesof classification and monotypic plant families. Taxon 47: 391–400.

BAKER, J. G. 1888. Handbook of the Amaryllideae. Goerge Bell andSons, London.

BAUML, J. A. 1979. A study of the genus Hymenocallis (Amaryllida-ceae) in Mexico, M.S. thesis, Cornell University, Ithaca, NY.

BENTHAM, G., AND J. D. HOOKER. 1883. Genera plantarum, vol. 3. L.Reeve, London.

BJORNSTAD, I. N. 1973. A revision of the genus Pancratium L. (Ama-ryllidaceae) in Africa south of the Sahara. Norwegian Journal ofBotany 20: 281–291.

BROWN, R. 1810. Prodomus 1: 296–293. London.BRUIJN, A. Y. DE, M. W. CHASE, A. V. COX, P. J. RUDALL, M. A. T.

JOHNSON, AND L. E. EGUIARTE. In review. Phylogenetics of Aspho-delaceae (Asparagales): an analysis of plastid rbcL and trnL-FDNA sequences. Submitted to Annals of Botany.

CAMERON, K. M., M. W. CHASE, W. M. WHITTEN, P. J. KORES, D. C.JARRELL, V. A. ALBERT, T. YUKAWA, H. G. HILLS, AND D. H. GOLD-MAN. 1999. A phylogenetic analysis of the Orchidaceae: evidencefrom rbcL nucleotide sequences. American Journal of Botany 86:208–224.

CHASE, M. W., ET AL. 1993. Phylogenetics of seed plants: an analysisof nucleotide sequences from the plastid gene rbcL. Annals of theMissouri Botanical Garden 80: 528–580.

———, M. R. DUVALL, H. G. HILLS, J. G. CONRAN, A. V. COX, L. E.EGUIARTE, J. HARTWELL, M. F. FAY, L. R. CADDICK, K. M. CAM-ERON, AND S. HOOT. 1995a. Molecular phylogenetics of Lilianae.In P. J. Rudall, P. J. Cribb, D. F. Cutler, and C. J. Humphries [eds.],Monocotyledons: systematics and evolution, vol. 1, 109–137. Roy-al Botanic Gardens, Kew.

———, D. W. STEVENSON, P. WILKIN, AND P. J. RUDALL. 1995b. Mono-cot systematics: a combined analysis. In P. J. Rudall, P. J. Cribb,D. F. Cutler, and C. J. Humphries [eds.], Monocotyledons: system-atics and evolution, vol. 2, 685–730. Royal Botanic Gardens, Kew.

CRESPO, M. B. M., D. LLEDO, M. F. FAY, AND M. W. CHASE. 1996.Molecular phylogeny of Leucojum based on ITS sequences. Amer-ican Journal of Botany 83 (suppl.): 149 (Abstract).

CRONQUIST, A. 1988. The evolution and classification of floweringplants, 2d ed. New York Botanical Garden, Bronx, NY.

DAHLGREN, R. M. T., AND H. T. CLIFFORD. 1982. The monocotyledons:a comparative study. Academic Press, London.

———, AND F. N. RASMUSSEN. 1983. Monocotyledon evolution: char-acters and phylogenetic analysis. Evolutionary Biology 16: 255–395.

———, H. T. CLIFFORD, AND P. F. YEO. 1985. The families of mono-cotyledons: structure, evolution, and taxonomy. Springer-Verlag,Berlin.

DAVIS, A. P., AND J. R. BARNETT. 1997. The leaf anatomy of the genusGalanthus L. (Amaryllidaceae J. St.-Hil.). Botanical Journal of theLinnean Society 123: 333–352.

DE JUSSIEU, A. L. 1789. Genera plantarum. Paris.DI FULVIO, T. E. 1973. Contribucıon al conocimiento cariologico de

Amaryllidaceae. Estudio cromosomico en Hieronymiella y otrasgeneros afines. Kurtziana 7: 117–131.

DOYLE, J. J., AND J. L. DOYLE. 1987. A rapid DNA isolation procedurefrom small quantities of fresh leaf tissue. Phytochemistry Bulletin19: 11–15.

DUVALL, M. R., M. T. CLEGG, M. W. CHASE, W. D. CLARK, W. J. KRESS,H. G. HILLS, L. E. EGUIARTE, J. F. SMITH, B. S. GAUT, E. A. ZIMMER,G. H. LEARN, Jr. 1993. Phylogenetic hypotheses for the monocot-yledons constructed from rbcL data. Annals of the Missouri Botan-ical Garden 80: 607–619.


FARRIS, J. S. 1969. A successive approximations approach to characterweighting. Systematic Zoology 18: 374–385.

———, V. A. ALBERT, M. KALLERSJO, D. LIPSCOMB, AND A. G. KLUGE.1996. Parsimony jackknifing outperforms neighbor-joining. Cla-distics 12: 99–124.

FAY, M. F., AND M. W. CHASE. 1996. Resurrection of Themidaceae forthe Brodiaea alliance, and recircumscription of Alliaceae, Ama-ryllidaceae and Agapanthoideae. Taxon 45: 441–451.

———, A. W. MEEROW, J. HARTWELL, AND M. W. CHASE. 1995. Mo-lecular systematics and biogeography of the Amaryllidaceae. Amer-ican Journal of Botany 82 (suppl.): 129 (Abstract).

FITCH, W. M. 1971. Toward defining the course of evolution: minimumchange for a specific tree topology. Systematic Zoology 20: 406–416.

FLORY, W. S. 1977. Overview of chromosomal evolution in the Amar-yllidaceae. Nucleus 20: 70–88.

FRITSCH, P. 1996. Isozyme analysis of the intercontinental disjunctswithin Styrax (Styracaceae): implications for the Madrean-Tethyanhypothesis. American Journal of Botany 83: 342–355.

GIELLY, L., AND P. TABERLET. 1994. The use of chloroplast DNA toresolve plant phylogenies: noncoding versus rbcL sequences. Mo-lecular Biology and Evolution 11: 769–777.

———, AND ———. 1996. A phylogeny of the European gentiansinferred from chloroplast trnL (UAA) intron sequences. BotanicalJournal of the Linnean Society 120: 57–75.

———, Y. M. YUAN, P. KUPFER AND P. TABERLET. 1996. Phylogeneticuse of noncoding regions in the genus Gentiana L.: chloroplast trnL(UAA) intron versus nuclear ribosomal internal transcribed spacersequences. Molecular Phylogenetics and Evolution 5: 460–466.

GOLDBLATT, P. 1976. Chromosome cytology of Hessea, Strumaria, andCarpolyza (Amaryllidaceae). Annals of the Missouri Botanical 63:314–320.

———. 1978. An analysis of the flora of southern Africa: its charac-teristics, relationships, and origins. Annals of the Missouri Botan-ical Garden 65: 369–436.

GRANT, V. 1983. The systematic and geographical distribution of hawk-moth flowers in the temperate North American flora. Botanical Ga-zette 144: 439–449.

GRAU, J., AND E. BAYER. 1991. Zur systematischen Stellung der Gat-tung Traubia Moldenke (Amaryllidaceae). Mitteilungen der Botan-ischen Staatssammlung Munchen 30: 479–484.

HERBERT, W. 1837. Amaryllidaceae. J. Ridgeway and Sons, London.HUBER, H. 1969. Die Sammenmerkmale und verwandtschafts verhalt-

nisse der Liliiflorae. Mitteilungen der Botanischen StaatssammlungMunchen 8: 219–538.

HUTCHINSON, J. 1934. Families of flowering plants, vol. 2. Monocoty-ledons, 1st ed. MacMillan, London.

———. 1959. Families of flowering plants, vol. 2. Monocotyledons,2nd ed. Clarendon Press, Oxford.

ISING, G. 1970. Evolution of karyotypes in Cyrtanthus. Hereditas 65:1–28.

JAUME-SAINT-HILAIRE, J. H. 1805. Exposition de familles naturales. 1.Paris.

KIM, J. H., H. T’HART, AND T. H. MES. 1996. The phylogenetic positionof East Asian Sedum species (Crassulaceae) based on chloroplastDNA trnL (UAA)-trnF (GAA) intergenic spacer sequence varia-tion. Acta Botanica Neerlandica 45: 309–321.

KITA, Y., K. UEDA, AND Y. KADOTA. 1995. Molecular phylogeny andevaluation of the Asian Aconitum subgenus Aconitum (Ranuncu-laceae). Journal of Plant Research 108: 429–442.

KOHLER, E., AND P. BRUCKNER. 1989. The genus Buxus (Buxaceae):aspects of its differentiation in space and time. Plant Systematicsand Evolution 162: 267–283.

KOSHIMIZU, T. 1930. Carpobiological studies of Crinum asiaticum L.var. japonicum Bak. Memoirs of the College of Sciences, KyotoImperial University, Series, B., Biology 5: 183–227.

LAKSHMI, N. 1978. Cytological studies in two allopolyploid species ofthe genus Hymenocallis. Cytologia 43: 555–563.

LI, Q.-B., Q. CAI, AND C. L. GUY. 1994. A DNA extraction methodfor RAPD analysis from Plants rich in soluble polysaccharides.Plant Molecular Biology Reporter 12: 215–220.

———, AND C. L. GUY. 1996. Prolonged final extension time increasescloning efficiency of PCR products. BioTechniques 21: 192–196.

LINDLEY, J. 1836. The vegetable kingdom, 2nd ed. Bradbury and Evans,London.

LINNEAUS, C. 1753. Species plantarum. Stockholm.LITTLE, E. L., AND W. B. CRITCHFIELD. 1969. Subdivisions of the genus

Pinus (Pinus). United States Department of Agriculture Forest Ser-vice Miscellaneous Publication 1144.

LISTON, A., L. H. REISEBERG, AND M. A. HANSON. 1992. Geographicpartitioning of chloroplast SNA variation in the genus Datisca (Da-tiscaceae). Plant Systematics and Evolution 181: 121–132.

LLEDO, M. D., M. B. CRESPO, K. M. CAMERON, M. F. FAY, AND M. W.CHASE. 1998. Systematics of plumbaginaceae based upon cladisticanalysis of rbcL sequence data. Systematic Botany 23: 21–29.

MAGUIRE, T. L., J. G. CONRAN, G. G. COLLINS, AND M. SEDGLEY. 1997.Molecular analysis of interspecific and intergeneric relationships ofBanksia using RAPDs and non-coding chloroplast DNA sequences.Theoretical and Applied Genetics 95: 253–260.

MANOS, P. 1992. Cladistic analysis of molecular variation of ‘‘higher’’Hamamelididae and Fagaceae, and systematics of Quercus sectionProtobalanus. Ph.D. dissertation, Cornell University, Ithaca, NY.

MARKOTTER, E. I. 1936. Die lewensgeskiedenis van sekere geslagte vandie Amaryllidaceae. Annals of the University of StellenboschXIV.A.2: 1–84.

MEEROW, A. W. 1984. Karyotype evolution in the Amaryllidaceae.Herbertia 40: 139–154.

———. 1985. The evolutionary significance of pancratioid floral mor-phology in the Amaryllidaceae. American Journal of Botany 72:962 (Abstract).

———. 1987a. A monograph of Eucrosia (Amaryllidaceae). System-atic Botany 12: 429–461.

———. 1987b. Biosystematics of tetraploid Eucharis (Amaryllida-ceae). Annals of the Missouri Botanical Garden 74: 291–309.

———. 1987c. Chromosome cytology of Eucharis, Caliphruria, andUrceolina. American Journal of Botany 74: 1559–1575.

———. 1989. A monograph of the Amazon lilies, Eucharis and Cal-iphruria (Amaryllidaceae). Annals of the Missouri Botanical Gar-den 76: 136–220.

———. 1995. Towards a phylogeny of the Amaryllidaceae. In P. J.Rudall, P. J. Cribb, D. F. Cutler, and C. J. Humphries [eds.], Mono-cotyledons: systematics and evolution, 169–179. Royal BotanicGardens, Kew.

———. 1997. Response to Muller-Doblies & Muller-Doblies (1996).Feddes Repertorium 108 (7–8): S.c. 1–3.

———, AND B. DEHGAN. 1985. The auriculate pollen grain of Hymen-ocallis quitoensis Herb. (Amaryllidaceae) and its systematic impli-cations. American Journal of Botany 72: 540–547.

———, C. L. Guy, and Q.-B. Li. 1998. Phylogeny of neotropicalAmaryllidaceae based on nrDNA ITS sequences. American Journalof Botany 85 (Supplement): 144 (Abstract).

———, AND D. A. SNIJMAN. 1998. Amaryllidaceae. In K. Kubitzki[ed.], Families and genera of vascular plants, vol. 3, 83–110.Springer-Verlag, Berlin.

MES, T. H. M., G. J. WIJERS, AND H. T’HART. 1997. Phylogenetic re-lationships in Monanthes (Crassulaceae) based on morphological,chloroplast and nuclear DNA variation. Journal of EvolutionaryBiology 10: 193–216.

———, J. VAN BREDERODE, AND H. T’HART. 1996. Origin of thewoody macaronesian Sempervivoideae and the phylogenetic posi-tion of the East African species of Aeonium. Botanica Acta 109:477–491.

MILLER, C. I. 1993. Impact of the Eocene on the evolution of Pinus L.Annals of the Missouri Botanical Garden 80: 471–498.

MORTON, J. K. 1965. The experimental taxonomy of the West Africanspecies of Pancratium L. (Amaryllidaceae). Kew Bulletin 19: 337–347.

MULLER-DOBLIES, D., AND U. MULLER-DOBLIES. 1978. Bulbs and mor-phology: Ungernia. Lagascalia 8: 13–23.

———, AND ———. 1996. Tribes and subtribes and some speciescombinations in Amaryllidaceae J. St.-Hil. emend. R. Dahlgren andal. 1985. Feddes Repertiorum 107 (5–6): S.c. 1–9.

NIXON, K. C. 1993. Infrageneric classification of Quercus (Fagaceae)and typification of sectional names. Annals of Science and Forestry50 (supplement 1): 25–34.


NORDAL, I., AND T. DUNCAN. 1984. A cladistic analysis of Haemanthusand Scadoxus. Nordic Journal of Botany 4: 145–153.

OLMSTEAD, R. G. 1997. Is there a specific gene for a specific problem?Advances in Plant Molecular Systematics, First International Sym-posium of the Systematics Association, University of Glasgow (Ab-stract), Systematics Association, Glasgow.

———, AND J. D. PALMER. 1994. Chloroplast DNA systematics: areview of methods and data analysis. American Journal of Botany81: 1205–1224.

———, AND J. A. SWEERE. 1994. Combining data in phylogenetic sys-tematics—an empirical approach using 3 molecular data sets in theSolanaceae. Syst. Biol. 43: 467–481.

PAX, F. 1888. Amaryllidaceae. In A. Engler and K. Prantl [eds.], Dienaturlichen pflanzenfamilien II. Teil 5, 97–124. Verlag von Wil-helm Engelmann, Leipzig.

PRILLIEUX, E. 1858. De la structure et du mode formation des grainesbulbiformes de quelques Amaryllidees. Annales de Sciences Na-turales 4: 97–104.

RAVEN, P. H. 1971. The relationships between ‘‘mediterranean’’ floras.In P. H. Davis, P. C. Harper, and I. G. Hedge [eds.], Plant life ofsouth-west Asia, 119–134. Botanical Society, Edinburgh.

———. 1973. Plant species disjunctions: a summary. Annals of theMissouri Botanical Garden 59: 234–246.

———, AND D. I. AXELROD. 1974. Angiosperm biogeography and pastcontinental movements. Annals of the Missouri Botanical Garden61: 539–673.

RAY, M. F. 1994. Biogeography and systematics of Lavatera and Malva(Malvaceae) species. American Journal of Botany 81: 199 (Ab-stract).

REID, C., AND R. A. DYER. 1984. A review of the Southern Africanspecies of Cyrtanthus. American Plant Life Society, La Jolla, CA.

RENDLE, A. B. 1901. The bulbiform seeds of certain Amaryllidaceae.Journal of the Royal Horticultural Society 26: 89–96.

RUDALL, P. J., M. W. CHASE, D. F. CUTLER, J. R. RUSBY, AND A. Y. DE

BRUIJN. 1998. Anatomical and molecular systematics of Asteli-aceae and Hypoxidaceae. Botanical Journal of the Linnean Society127: 1–42.

SANG, T, D. J. CRAWFORD, AND T. F. STUESSY. 1997. Chloroplast DNAphylogeny, reticulate evolution, and biogeography of Paeonia(Paeoniaceae). American Journal of Botany 84: 1120–1136.

SATO, D. 1938. Karyotype evolution and phylogeny. IV. Karyotype inAmaryllidaceae with special reference to SAT chromosomes. Cy-tologia 9: 203–242.

SINGH, V. 1972. Floral morphology of the Amaryllidaceae. 1. Subfam-ily Amaryllidoideae. Canadian Journal of Botany 50: 1555–1565.

SNIJMAN, D. A. 1992. Systematic studies in the tribe Amaryllideae(Amaryllidaceae), Ph.D. dissertation, University of Cape Town,Capetown.

———. 1994. Systematics of Hessea, Strumaria and Carpolyza (Ama-ryllideae: Amaryllidaceae). Contributions from the Bolus Herbar-ium 16. Capetown.

———, AND H. P. LINDER. 1996. Phylogenetic relationships, seed char-acters, and dispersal system evolution in Amaryllideae (Amarylli-daceae). Annals of the Missouri Botanical Garden 83: 362–386.

SOLTIS, D. E., P. S. SOLTIS, M. E. MORT, M. W. CHASE, V. SAVOLAINEN,S. B. HOOT, AND C. M. MORTON. 1998. Inferring complex phylog-enies using parsimony: an empirical approach using three largeDNA data sets for angiosperms. Systematic Biology 47: 32–42.

STEVENSON, D. W., AND H. LOCONTE. 1995. Cladistic analysis of mono-cot families. In P. J. Rudall, P. J. Cribb, D. F. Cutler, and C. J.Humphries [eds.], Monocotyledons: systematics and evolution, vol.2, 685–730. Royal Botanic Gardens, Kew.

SWOFFORD, D. L. 1998. Phylogenetic analysis using parsimony, v. 4.01beta. Sinauer, Sunderland, MA.

TABERLET, P., L. GIELLY, G. PAUTOU, AND J. BOUVET. 1991. Universalprimers for amplification of three non-coding regions of chloroplastDNA. Plant Molecular Biology 17: 1105–1110.

TAKHTAJAN, A. 1969. Flowering plants: origin and dispersal. Oliver andBoyd, London.

THORNE, R. F. 1976. A phylogenetic classification of the Angiospermae.Evolutionary Biology 9: 35–106.

TOMITA K. 1931. Uber die Entwicklung des nackten Embryos von Cri-num latifolium L. Science Report of Tohoku Imperial UniversityIV, Biology 6: 163–169.

TRAUB, H. P. 1963. Genera of the Amaryllidaceae. American Plant LifeSociety, La Jolla, CA.

VON SCHLIMBACH, H. 1924. Beitrage zur Kenntnis der AfrikanischenFlora V. Bulletin de la Herbarium Boissier 4: 809–846.

VOSA, C. G., AND D. A. SNIJMAN. 1984. The cytology of the genusHaemanthus L. (Amaryllidaceae). Journal of South African Botany50: 237–259.

WENZEL, J. W. 1997. When is a phylogenetic test good enough? Me-moires de la Musee Nationale del’Histoire Naturelle 173: 31–45.

WERKER, E., AND A. FAHN. 1975. Seed anatomy of Pancratium speciesfrom three different habitats. Botanical Gazette 136: 396–403.

WHITEHEAD, M. R., AND C. A. BROWN. 1940. The seeds of the spiderlily Hymenocallis occidentalis. American Journal of Botany 27:199–203.

WILLIAMS, M. D. 1981. Chromosome count for Paramongaia weber-baueri Velarde. Plant Life 37: 83–89.

Documents

systematics of amaryllidaceae based on cladistic analysis of plastid rbcl and trnl-f sequence data