MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
GALLEY 400
PHYLOGENY OF THE TRIBE
HYMENOCALLIDEAE
(AMARYLLIDACEAE) BASED
ON MORPHOLOGY AND
MOLECULAR CHARACTERS 1
mobt 89_303 Mp_400
File # 03em
Alan W. Meerow, 2 Charles L. Guy, 3
Qin-Bao Li, 3 and Jason R. Clayton 4
ABSTRACT
The generic limits of Hymenocallis have been variously proposed by different taxonomic workers, often without
discussion or data. The genera Leptochiton, Ismene, Elisena, and Pseudostenomesson have been included with Hymenocallis, lumped together as the genus Ismene, or maintained as distinct genera. Recent cladistic analysis of plastid and
nrDNA for Amaryllidaceae support a distinct tribe Hymenocallideae. Cladistic analyses of morphology, and plastid
(trnL-F region) and nuclear ribosomal DNA (ITS) are presented alone and in combination for the tribe. Leptochiton is
sister to the rest of the genera in the tribe in all analyses. While Hymenocallis is always resolved as monophyletic,
Ismene is variably paraphyletic or monophyletic. The combined sequence data produce the most resolved and bestsupported phylogeny, wherein Hymenocallis and Ismene are monophyletic sister genera. These data support an origin
for the tribe in the Andes, with vicariant distribution of the largely Mesoamerican Hymenocallis. Formal recognition of
Ismene subg. Elisena and Pseudostenomesson is established.
Key words: Amaryllidaceae, cladistics, molecular systematics, phylogeny.
Systematics of the genus Hymenocallis Salisb.
(Amaryllidaceae) and its allies have defied precise
systematic understanding at both the specific and
generic levels (Flory, 1976; Meerow & Dehgan,
1985). The genera Hymenocallis and Ismene Salisb.
were established by Salisbury (1812) for the Neotropical species with fleshy seeds originally assigned to the Old World genus Pancratium L. The
zygomorphic-flowered Elisena was described by
Herbert (1837), who recognized Hymenocallis and
Ismene as distinct genera. Baker (1888) subsumed
Ismene within Hymenocallis but retained Elisena as
distinct, as did Pax (1890). While Stapf (1933)
treated H. quitoensis Herb. as a species of Pamianthe Stapf, Sealy (1937) considered the species to
exhibit sufficient morphological divergence to be
recognized as a monotypic genus, Leptochiton Sealy.
Hutchinson (1934, 1959) retained both Elisena and
Ismene (presumably including Leptochiton) as distinct. Velarde (1949) established the Peruvian genus Pseudostenomesson for a fleshy-seeded species
originally described as Stenomesson morissonii Vargas as well as one new species. Traub (1962) recognized all four erstwhile genera as subgenera of
Hymenocallis in his synoptic treatment: subg. Hy-
menocallis, subg. Ismene (Salisb.) Baker ex Traub
(including Leptochiton), subg. Elisena (Herb.)
Traub, and subg. Pseudostenomesson (Velarde)
Traub. Traub (1980) later reduced these subgenera
to the rank of section without explanation. Ravenna
(1980) in his description of H. heliantha (5 Leptochiton heliantha (Ravenna) Gereau & Meerow)
suggested that subgenera Ismene (including Lepidochiton), Elisena, and Pseudostenomesson should
probably be all recognized as the genus Ismene,
distinct from Hymenocallis. Meerow and Dehgan
(1985) suggested that Pseudostenomesson might
warrant recognition at the rank of genus due to its
extreme phenetic divergence (funnelform-tubular
perianth) versus the ‘‘pancratioid’’ flower of Leptochiton, Ismene subg. Ismene, and Hymenocallis.
‘‘Pancratoid’’ floral morphology refers to a large,
white, fragrant, crateriform flower with a conspicuous staminal cup (cf. Pancratium L.). This type of
flower appears to be adapted for sphingid moth pollination (Bauml, 1979; Grant, 1983; Morton, 1965).
Meerow (1990) treated Leptochiton as a distinct genus and recognized Hymenocallis and Ismene (including Elisena and Pseudostenomesson) as distinct,
a treatment followed by Gereau et al. (1993) and
This work was supported in part by NSF Grant DEB 9628787 to AWM and GLG.
USDA-ARS-SHRS, 13601 Old Cutler Rd., Miami, Florida 33158, U.S.A., and Fairchild Tropical Garden, 11931
Old Cutler Road, Miami, Florida 33156, U.S.A.
3
University of Florida-IFAS, Department of Environmental Horticulture, 1545 Fifield Hall, Gainesville, Florida
32611, U.S.A.
4
USDA-ARS-SHRS, 13601 Old Cutler Rd., Miami, Florida 33158, U.S.A.
1
2
ANN. MISSOURI BOT. GARD. 89: 400–413. 2002.
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
GALLEY 401
mobt 89_303 Mp_401
File # 03em
Volume 89, Number 3
2002
Meerow et al.
Phylogeny of Tribe Hymenocallideae
Meerow and Snijman (1998). No cladistic analysis
has focused exclusively on testing the validity of
this treatment, although at least one representative
of each subgenus was included in overall molecular
studies of Amaryllidaceae (Meerow et al., 1999,
2000a).
Hymenocallis and its allied segregate genera are
entirely Neotropical in distribution [a single West
African taxon, H. senegambica, was treated by Sealy (1954) as an early adventive introduction of H.
caribaea]. Hymenocallis sensus stricto, with 50 to
60 species, is chiefly Mesoamerican and extends
into the West Indies and the southeastern United
States. It is sparingly represented in northern South
America. Leptochiton Sealy (2 spp.), Ismene (ca. 10
spp.), Elisena Herb. (2 or 3 spp.), and Pseudostenomesson Velarde (2 spp.) are all endemic to the
Central Andean region of South America. Hymenocallis, Ismene, and Leptochiton are contrasted in
Table 1.
Hymenocallis and allies have usually been allied
with Eucharis Planch. in the tribe Eucharideae
(Hutchinson, 1934, 1959; Traub, 1963; Dahlgren
et al., 1985; Müller-Doblies & Müller-Doblies,
1996). Meerow (1989, 1995) argued that the linkage of these genera, largely through the perception
that both lineages shared a fleshy seed, was misconstrued, and proposed that either subtribal or
tribal recognition of Hymenocallis and allies was
warranted. Müller-Doblies and Müller-Doblies
(1996) placed them in Eucharideae subtribe Hymenocallidinae, while Meerow and Snijman (1998)
recognized a distinct tribe, Hymenocallideae. Family-wide analysis of plastid sequences (Meerow et
al., 1999) and nrDNA analyses of the monophyletic
American clade of the family (Meerow et al., 2000a)
support a distinct Hymenocallideae as sister to the
newly recognized tribe Clinantheae (a segregate of
the former Stenomesseae), but complete resolution
of the intratribal relationships is not apparent in
these large analyses. Both tribes are subclades of
a well-supported, Andean, tetraploid clade of genera.
In this paper, we present phylogenetic analyses
of morphological and molecular data for the tribe
Hymenocallideae, and seek to clarify the relationships within the tribe.
were used for one species each of the three subgenera of Ismene, one species of Leptochiton, and
the outgroup Pamianthe peruviana (Table 2, Meerow et al., 1999). For ITS and the morphological
data matrix, we increased our sampling with an additional four species of Hymenocallis and two additional species of Ismene subg. Ismene (Table 2).
The aligned sequence matrices are available from
the first author (miaam@ars-grin.gov).
MATERIALS
AND
METHODS
SAMPLING
Sequences for the plastid trnL-F region were
newly obtained for H. eucharidifolia, which, along
with H. latifolia, was used as an exemplar taxon of
Hymenocallis (Table 2). Previously cited sequences
401
MORPHOLOGICAL DATA
Morphological and cytological character state
data were derived from the following sources: Traub
(1962, 1980), Sealy (1954), Flory (1976), Velarde
(1949), Bauml (1979), Meerow and Dehgan (1985);
from examination of living material in research collections at the USDA, Miami, Florida; field observations of Hymenocallis, Ismene, and Leptochiton
species; and examination of herbarium material.
The morphological matrix consists of 12 species
representing 4 genera and 23 characters (Tables 3,
4).
SEQUENCE DATA
The trnL-F (trnL intron and spacer region between trnL and trnF) matrix consisted of 6 taxa and
906 base positions. The nrDNA ITS sequence matrix (ITS1, 5.8s intron, ITS2) consisted of 12 taxa
and 636 bp.
DNA EXTRACTION, AMPLIFICATION, AND
SEQUENCING PROTOCOLS
Genomic DNA was extracted from silica gel
dried leaf tissue as described by Meerow et al.
(2000a). The trnL-trnF region was amplified using
the primers of Taberlet et al. (1991) as described
by Meerow et al. (1999). Amplification of the ribosomal DNA ITS1/5.8S/ITS2 region was accomplished using flanking primers (18S, 26S) AB101
and AB102 (Douzery et al., 1999), and the original
White et al. (1990) internal primers ITS2 and 3 to
amplify the spacers along with the intervening 5.8S
sequence, as described by Meerow et al. (2000a).
Amplified products were purified using QIAquick
(Qiagen, Valencia, California) columns, following
manufacturer’s protocols. All polymerase chain reactions (PCR) were performed on an ABI 9700 (Applied Biosystems, Foster City, California).
Cycle sequencing reactions were performed directly on purified PCR products on the ABI 9700,
using standard dideoxy cycle protocols for sequencing with dye terminators on either an ABI 377 or
ABI 310 automated sequencer (according to the
Comparison of the genera and subgenera of Amaryllidaceae tribe Hymenocallideae.
Genus or
subgenus
Number of
species
Distribution
Elongate
pseudostem
Floral morphology
Ovules per
locule
Phytomelan
on seed coat
Chromosome
number
Absent
Pancratioid, actinomorphic, large,
white or yellow, fragrant, sessile,
(sub)erect; tube long; staminal cup
large and striped green within; free
filament short and incurved.
16–20
Present
2n 5 34
Ismene subg. Ismene
5–7
Central Andes at low to
high elevations
Present
Pancratioid, actinomorphic large, white
or yellow, fragrant, 2–10, subsessile
to pedicellate, horizontal or declinate, tube 6 long, staminal cup
large, striped green within; free filament short and incurved.
2–4
Absent
2n 5 23–86, 46,
104–110
Ismene subg. Elisena
2–4
Peru and Ecuador at mid
to high elevations
Present
Zygomorphic, large, white, not fragrant,
2–10, subsessile, declinate; tube
short; staminal cup 6 large, deflexed from the tube, free filaments
long and declinate.
2
Absent
2n 5 46
Ismene subg. Pseudostenomesson
2
Peru, above 3000 m
Present
Funnelform-tubular, actinomorphic, 6
small, green, not fragrant, numerous,
pedicellate, pendulous; tube long;
staminal cup subcylindrical, free filament straight.
2
Absent
2n 5 46
Hymenocallis
ca. 50
SE U.S., West Indies, Mesoamerica
Absent
Pancratioid, actinomorphic, large,
white, fragrant, 1 to many, mostly
sessile, erect; tube long; staminal
cup large or small and variable in
shape, not striped within; free filament long and straight.
2–10
Absent
2n 5 46, 40
most common
but variable
mobt 89_303 Mp_402
File # 03em
SW Ecuador and NW
Peru, at low elevations
GALLEY 402
2
Annals of the
Missouri Botanical Garden
Leptochiton
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
402
Table 1.
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
Volume 89, Number 3
2002
Table 2.
Vouchers and new GenBank accession numbers for DNA sequences of Hymenocallideae. All vouchers are deposited at FTG unless otherwise indicated.
GenBank accession no. or previous citation
Origin
trnL gene
trnL-F spacer
Mexico
Mexico
Florida, USA
Mexico
Trinidad
—
—
Meerow et al. (1999)
AF411078
—
—
—
Meerow et al. (1999)
AF411079
—
Meerow
Meerow
Meerow
Meerow
Meerow
Meerow 2452
Meerow 2441
Meerow 2306
Peru
Peru
Peru
—
—
Meerow et al. (1999)
—
—
Meerow et al. (1999)
AF411080
Meerow et al. (2000)
Meerow et al. (2000)
Sagastegui 15454
Peru
Meerow et al. (1999)
Meerow et al. (1999)
Meerow et al. (2000)
Meerow 2308
Meerow 1116
Meerow 2304
Peru
Ecuador
Peru
Meerow et al. (1999)
Meerow et al. (1999)
Meerow et al. (1999)
Meerow et al. (1999)
Meerow et al. (1999)
Meerow et al. (1999)
Meerow et al. (2000)
Meerow et al. (2000)
Meerow et al. (2000)
Meerow
Meerow
Meerow
Meerow
Meerow
2424
2433
2438
2439
2440
ITS
et
et
et
et
et
al.
al.
al.
al.
al.
(2000)
(2000)
(2000)
(2000)
(2000)
403
mobt 89_303 Mp_403
File # 03em
Meerow et al.
Phylogeny of Tribe Hymenocallideae
Hymenocallis acutifolia (Herb.) Sweet
H. glauca M. Roem.
H. latifolia (Mill.) Roem.
H. eucharidifolia Bak.
H. tubiflora Salisb.
Ismene subg. Ismene
Ismene amancaes (Ruiz & Pav.) Herb.
I. hawkesii (Vargas) Gereau & Meerow
I. narcissiflora Jacq.
Ismene subg. Elisena (Herb.) Meerow
I. longipetala (Lindl.) Meerow
Ismene subg. Pseudostenomesson (Velarde)
Meerow
I. vargasii (Velarde) Gereau & Meerow
Leptochiton quitoensis (Herb.) Sealy
Pamianthe peruviana Stapf
Voucher
GALLEY 403
Taxon
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
404
Annals of the
Missouri Botanical Garden
Table 3.
Characters and character states used in the cladistic analyses of Hymenocallideae based on morphology.
Character
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
mobt 89_303 Mp_404
File # 03em
GALLEY 404
Elongate pseudostem
Flower number
Flowers sessile/pedicellate
Flower habit
Tube length
Tube habit
Perianth morphology
Perianth symmetry
Flower color
Fragrance
Staminal cup shape
Staminal cup striping
Free filament
Free filament
Pollen grain size
Pollen grain
Exine reticulum
Ovules per locule
Seed per locule
Phytomelan on testa
Seed coat
Seed shape
Most common diploid chromosome number
States and coding
absent (0); present (1)
2–10 (0); solitary (1); .10 (2)
sessile (0); pedicellate (1)
erect (0); declinate/horizontal (1); pendent (2)
shorter than tepals (0); longer than or equal to tepals (1)
straight (0); curved (1)
pancratioid (0); funnelform-tubular (1); 6 funnelform (2)
actinomorphic (0); zygomorphic (1)
white (0); yellow (1); green (2)
present (0); absent (1)
rotate or funnelform (0); cylindrical (1)
present (0); absent (1)
incurved (0); straight (1); declinate (2)
longer than cup (0); shorter than cup (1)
very large (0); large (1); medium (2)
auriculate (0); not (1)
coarse (0); medium (1)
.20 (0); 16–20 (1); 2–10 (2); 2–4 (3)
numerous (0); 2–5 (1); 1 (2)
present (0); absent (1)
not fleshy (0); fleshy (1)
flat, winged (0); globose (1)
46 (0); 34 (1); 46, 40 (2)
manufacturer’s protocols; Applied Biosystems, Foster City, California).
SEQUENCE ALIGNMENTS
Both sequence matrices were readily aligned
manually using the program Sequencher (GeneCodes, Inc., Ann Arbor, Michigan) as few gaps
needed to be inserted.
CLADISTIC ANALYSES
Pamianthe (tribe Clinantheae) was used as outgroup for all analyses. In larger sequence analyses
(Meerow et al., 1999, 2000a), this genus resolves
as most closely related to the Hymenocallideae. Pamianthe and Leptochiton (the latter putatively the
least derived genus in the Hymenocallideae; see
discussion below) share two four-base sequence elements in the trnL-F region (bp325–328, 821–824)
Table 4. Character state matrix for cladistic analysis of 23 morphological characters in Hymenocallideae. Polymorphisms: 1 5 (0,1); * 5 (0,1,2).
Taxon
Character
Hymenocallis acutifolia
Hymenocallis eucharidifolia
Hymenocallis glauca
Hymenocallis latifolia
Hymenocallis tubiflora
Ismene amancaes
Ismene hawkesii
Ismene longipetala
Ismene narcissiflora
Ismene vargasii
Leptochiton quitoensis
Pamianthe peruviana
Matrix
1
2
12345678901234567890123
00001000000110000221112
0*001000000130000221112
0*001000000130000221112
0*001000000130000221112
0*001000000130000221112
10111100100001000321110
10111100000001000321110
10110021011120211321110
10111100000001000321110
10120110211110211321110
01001100100001000110111
10111000000001110000000
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
mobt 89_303 Mp_405
File # 03em
GALLEY 405
Volume 89, Number 3
2002
Meerow et al.
Phylogeny of Tribe Hymenocallideae
405
that are absent from the rest of the Hymenocallideae. Phylogenetic analyses were run using PAUP*
version 4.0b8 beta (Swofford, 1998). An exhaustive
search of all possible tree topologies was conducted
for trnL-F. For ITS, the morphological, and all combined analyses, branch and bound searches were
conducted. Support for internal nodes of the trees
was determined with 5000 replicates of branch and
bound bootstrapping (Felsenstein, 1985) and by
calculation of Bremer (1988) decay indices (DI) using the program TreeRot (Sorenson, 1999). A
branch and bound search was implemented for
each constraint statement postulated by TreeRot. A
bootstrap value of 50–64% was considered weak,
65–74% moderate, and 75–100% strong support.
Combining independent character matrices,
whether both molecular or molecular and morphological, very often increases the resolution of the
ingroup and the bootstrap support of the internal
nodes of the phylogenetic trees (Olmstead &
Sweere, 1994; Chase et al., 1995; Yukawa et al.,
1996; Rudall et al., 1998; Soltis et al., 1998; Meerow et al., 1999). Nonetheless, there is controversy
about whether different data sets should be analyzed separately or together (De Queiroz et al.,
1995; Huelsenbeck et al., 1996). Congruence of the
independent matrices has generally been demonstrated before they are combined, but it has also
been argued that incongruence should not be a predetermined factor against doing so (Dubuisson et
al., 1998; Seelanan et al., 1997). Miyamoto and
Fitch (1995) argued that data sets should always
be analyzed independently, as underlying assumptions, constraints, or weighting strategies will vary
from data set to data set. Kluge (1989) and Nixon
and Carpenter (1996) argued that simultaneous
analysis of multiple data sets better maximizes parsimony and allows secondary signals to appear from
the combined data. Bull et al. (1993), Rodrigo et
al. (1993), and De Queiroz (1993) advocated combining data only after a statistical test of congruence, what Huelsenbeck et al. (1996) called ‘‘conditional combination.’’ Before combining the data
sets, we performed a partition homogeneity test
(Farris et al., 1994, 1995) on the variously combined matrices, using a branch and bound search.
trees, Hymenocallis is monophyletic (bootstrap 5
89%, DI 5 1), while Ismene is paraphyletic. Ismene
longipetala (subg. Elisena) and I. vargasii (subg.
Pseudostenomesson) are sisters in both trees. Leptochiton is sister to both Hymenocallis and Ismene
in one tree (Fig. 1A). The 6 apomorphies at the
ancestral node are an increase in pollen grain size,
auriculate pollen grains, reduction in ovule number
from more than 20 to 16 to 20; reduction in number
of seeds per locule; and evolution of globose, fleshy
seeds. Apomorphies for Hymenocallis (Fig. 1A) are
the absence of an elongate pseudostem, predominantly sessile and erect flowers, and 2n 5 46, 40
chromosomes. Other than Hymenocallis, the only
clade with strong bootstrap support is the sister relationship of Ismene subg. Elisena and subgenus
Pseudostenomesson (100%, DI 5 6), based on 7
apomorphies: perigone tube length reduction, nonpancratioid floral morphology, loss of floral fragrance, cylindrical staminal cup, and smaller nonauriculate pollen grains with less coarse exine
reticulum. If all of the characters are ordered as
irreversible, a single tree is found of length 5 48,
with CI 5 0.67 and RI 5 0.88 (Fig. 1B). There is
moderate bootstrap support for a monophyletic Ismene (65%; DI 5 2; apomorphies: elongate pseudostem, pedicellate and declinate/horizontal flowers, and 2–4 ovules per locule). There is weak
support for the sister relationship of Hymenocallis
and Ismene (56%, DI 5 1; apomorphies: reduction
in ovule and seed number, respectively; and the
loss of phytomelan from the testa). Leptochiton is
moderately supported as sister to both (65%, DI 5
1; apomorphies: reduction in ovule and seed number and the evolution of a fleshy seed). Ismene subg.
Ismene has a 91% bootstrap and DI 5 2. Ismene
subg. Elisena (I. longipetala) and subgenus Pseudostenomesson (I. vargasii) are again sister groups
with 100% bootstrap and a DI 5 9. A monophyletic
Hymenocallis receives 87% bootstrap support with
a DI 5 4. Hymenocallis latifolia, H. glauca, and
H. eucharidifolia form a monophyletic group with
bootstrap support of 60 and DI 5 1. This same tree
topology (Fig. 1B) is 40 steps long with CI 5 0.80
and RI 5 0.80 if a branch and bound search is run
with the topology as a constraint with all characters
unordered.
RESULTS
trnL-F
MORPHOLOGICAL MATRIX
PLASTID
SEQUENCES
With all characters unordered, two most parsimonious trees (Fig. 1A, one shown) were found of
length 5 37, consistency index (CI) 5 0.86, and
retention index (RI) 5 0.88. Sixteen of the 23 characters used were parsimony informative. In both
Using trnL-F sequences, which provide 7 parsimony-informative base substitutions, three equally
most parsimonious trees are found of length 5 82,
CI 5 0.99, and RI 5 0.88 (Fig. 2, one tree shown).
All three trees resolve a monophyletic Ismene with
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
406
GALLEY 406
Annals of the
Missouri Botanical Garden
mobt 89_303 Mp_406
File # 03em
Figure 1. Cladograms for Hymenocallideae based on morphological characters. —A. One of two most parsimonious trees found with all characters unordered. —B. Single most
parsimonious tree found if all characters are ordered as irreversible. Numbers above branches are branch lengths. Bootstrap percentages and decay indices (italic) are in parentheses.
Vertical lines and numbers below branches are apomorphies along that branch (see Table 3).
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
mobt 89_303 Mp_407
File # 03em
GALLEY 407
Volume 89, Number 3
2002
Meerow et al.
Phylogeny of Tribe Hymenocallideae
407
Figure 2. One of the three most parsimonious trees found by cladistic analysis of plastid trnL-F DNA sequences
for the Hymenocallideae. Numbers above branches are branch lengths; numbers below branches are bootstrap percentages, followed by decay indices (italic). The large arrow indicates a node that collapses in the strict consensus of
all three trees.
81% bootstrap support (DI 5 2), and Leptochiton
as sister to the rest of the tribe but without support.
A monophyletic Hymenocallis is resolved as sister
to Ismene in one tree (Fig. 2), but Hymenocallis and
Ismene form a clade in all three (73% bootstrap, DI
5 1). Ismene subg. Ismene (I. narcissiflora) and Elisena (I. longipetala) are resolved as sister groups
in all three trees with a bootstrap of 70% (DI 5 1).
ITS SEQUENCES
ITS provides 50 parsimony-informative characters, and 9 trees of length 5 209, CI 5 0.73, and
RI 5 0.77 were found (Fig. 3). In all of the trees,
Leptochiton is resolved as sister to both Hymenocallis and Ismene (Fig. 3A), but without significant
support. Hymenocallis is monophyletic (bootstrap 5
97%, DI 5 5), but Ismene is monophyletic in only
2 of the 9 trees (Fig. 3B, one shown). However,
Ismene subg. Ismene (I. amancaes, I. hawkesii, I.
narcissiflora) is monophyletic with weak bootstrap
support (59%) and DI 5 1 (Fig. 3B).
COMBINED
trnL-F
AND ITS SEQUENCES
The P value from the partition homogeneity test
5 0.93, indicating that the trnL-F and ITS sequence matrices were highly congruent. Six most
parsimonious trees were found of length 5 292, CI
5 0.92, and RI 5 0.77 (Fig. 4). In all trees, Hymenocallis and Ismene are monophyletic sister genera with bootstrap support of 94% and a DI 5 3.
Leptochiton is sister to both, but without significant
support. Bootstrap support for a monophyletic Hymenocallis is 98% (DI 5 5), but only 68% (DI 5
1) for a monophyletic Ismene. The only other inter-
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
408
GALLEY 408
Annals of the
Missouri Botanical Garden
mobt 89_303 Mp_408
File # 03em
Figure 3. Cladograms for Hymenocallideae based on nrDNA ITS sequences. —A. Strict consensus of nine equally parsimonious trees. —B. One of two trees in which Ismene is
a monophyletic sister group to Hymenocallis. Numbers above branches are branch lengths; numbers below branches are bootstrap percentages followed by decay indices (italic).
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
GALLEY 409
Volume 89, Number 3
2002
mobt 89_303 Mp_409
File # 03em
Meerow et al.
Phylogeny of Tribe Hymenocallideae
409
Figure 4. One of six most parsimonious trees found by cladistic analysis of combined plastid trnL-F and nrDNA
ITS sequences. Numbers above branches are branch lengths; numbers below branches are bootstrap percentages followed by decay indices (italic). The larger arrow indicates a node that collapses in the strict consensus of all six trees.
nal resolution within Ismene that receives bootstrap
support is a sister relationship between I. narcissiflora and I. hawkesii (both within subg. Ismene) at
84% with DI 5 2.
COMBINED SEQUENCE AND
MORPHOLOGICAL MATRICES
The P value of the partition homogeneity test was
0.0003, indicating significant incongruence between the morphological and DNA sequence data
matrices. Much of the apparent incongruence can
be attributed to the weak resolution of the morphologically based topologies, and we felt that it would
still be informative to combine the two matrices in
a single analysis. Of the 1565 characters included,
76 were parsimony informative. A single tree was
found of length 5 332, CI 5 0.92, and RI 5 0.79
(Fig. 5A). Hymenocallis is monophyletic with 100%
bootstrap support (DI 5 8), but Ismene is paraphyletic. Bootstrap support for the monophyly of Ismene
subg. Ismene rises to 81% (DI 5 2), but Ismene
subg. Elisena (I. longipetala) and Pseudostenomesson (I. vargasii) are sister groups (bootstrap 5 97%,
DI 5 4) weakly supported (bootstrap 5 57%, DI
5 1) as sister to Hymenocallis. Leptochiton is again
sister to the other members of Hymenocallideae but
without support. If trees one step longer were also
retained in the search, in addition to the single
shortest tree (Fig. 5A), a single, fully resolved tree
of length 5 333, CI 5 0.90, and RI 5 0.77 was
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
410
GALLEY 410
Annals of the
Missouri Botanical Garden
mobt 89_303 Mp_410
File # 03em
Figure 5. Cladograms for Hymenocallideae based on combined morphological and DNA sequence matrices. —A. Single most parsimonious tree. Numbers above branches are
branch lengths; numbers below branches are bootstrap percentages followed by decay indices (italic). —B. Single additional tree found if trees one step longer than tree pictured in
5A are retained. Numbers above branches are branch lengths.
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
GALLEY 411
mobt 89_303 Mp_411
File # 03em
Volume 89, Number 3
2002
Meerow et al.
Phylogeny of Tribe Hymenocallideae
found (Fig. 5B). In this tree (Fig. 5B), both Hymenocallis and Ismene are monophyletic sister genera,
as are Ismene subg. Elisena and Pseudostenomesson.
Amaryllidaceae (Meerow et al., 2000a), with a vicariance event that gave rise to the largely North
American Hymenocallis. Leptochiton, with 16 to 20
ovules per locule and a phytomelanous testa, occupies a relict position in the tribe with links to
the non-fleshy seeded Andean endemic Clinantheae. However, it is the genus Ismene that reflects
the patterns of floral morphological diversity that
occur in the Eustephieae, Clinantheae, and Stenomesseae (sensu Meerow et al., 2000a). Ismene subg.
Ismene retains the plesiomorphic pancratioid floral
morphology of Leptochiton, Pamianthe, and Hymenocallis, while the smaller Ismene subg. Elisena and
subg. Pseudostenomesson express floral novelties.
Ismene subg. Pseudostenomesson, occurring at the
highest elevations of any member of the tribe, might
be the youngest element of the polymorphic Ismene,
since the Andes likely did not extend above 1000
m elevation before the Pliocene (10 MYBP; Van der
Hammen, 1974, 1979). Analogous patterns of floral
diversity are found throughout the tetraploid Andean clade of the American Amaryllidaceae. In the
Clinantheae, the low- to mid-elevation genera Pamianthe and Paramongaia Velarde have pancratioid floral morphology, while the mostly high-elevation Clinanthus Herb. has colorful, putatively
ornithophilous flowers. In the more distantly related
petiolate-leafed Stenomesseae, Eucharis has the
pancratioid flower; Plagiolirion resembles a miniature Ismene subg. Elisena; and Stenomesson and
Urceolina exhibit colorful, putatively ornithophilous
flowers. Finally, in the Eustephieae, which is sister
to rest of the Andean clade (Meerow et al., 2000a),
the full range of variation is evident in a single
genus, Hieronymiella Pax (Hunziker, 1969). This
recurrent pattern suggests a scenario of rapid mosaic evolution (sensu Stebbins, 1984) within this
monophyletic, tetraploid group (Meerow, 1987).
The relatively low number of phylogenetically informative base substitutions in our sequence analyses of non-coding regions (7 for trnL-F; 50 for ITS)
supports a hypothesis of a relatively recent radiation within the Hymenocallideae tied to the rise of
the Andes. This seems most significant relative to
Ismene, the most polymorphic of the three hymenocallid genera, and the only one that has adapted
to high elevation.
Hymenocallis is most speciose in Mexico (Bauml,
1979), with a secondary area of diversity in the
southeastern United States (Smith & Flory, 1990,
2001; Smith et al., 2001). Only three described
species have been reported from South America:
the broadly and coastally distributed H. littoralis,
H. pedalis, and H. tubiflora. The genus does not
occur at all in the Andes, and H. tubiflora is the
DISCUSSION
Both plastid (Meerow et al., 1999) and ITS (Meerow et al., 2000a) sequences strongly support the
position of the tribe Hymenocallideae as a monophyletic group within the Andean tetraploid clade
of the endemic American Amaryllidaceae that is
sister to the newly recognized tribe Clinantheae
Meerow (Meerow et al., 2000a). The seeds of the
Clinantheae are uniformly dry, flat, winged, and
with phytomelanous testas. There are links between
Leptochiton and Pamianthe that Stapf implicitly
recognized, most notably the plesiomorphic presence of phytomelan in the testa of Leptochiton’s
seed [of which Meerow & Dehgan (1985) were unaware], but also the numerous ovules of this genus
(plesiomorphic as well). In the ITS phylogeny presented by Meerow et al. (2000a), support for Pamianthe as sister to the rest of Clinantheae (vs. a
sister group relationship to Hymenocallideae or an
unresolved position) was considerably weaker when
the aligned matrix was not successively weighted.
This is not surprising given that both genera occupy
a basal phylogenetic position in their respective
clades herein.
The difficulty of relying on morphological characters alone to generate phylogenies in Amaryllidaceae has been discussed (Meerow, 1995; Meerow
et al., 2000b), given a high degree of homoplasy
for many morphological characters in the family.
Our analysis (Fig. 1) generates trees with relatively
high CI and RI, but parsimony is still not able to
resolve Ismene nor consistently place Leptochiton in
the basal position within the tribe with unordered
morphological characters alone, in contrast to sequence data (Figs. 2–4), which also provide (in the
combined trnL-F and ITS matrix), over three times
the number of phylogenetically informative characters of morphology alone. The combined plastid
and nuclear sequence matrix produces the most fully resolved shortest trees. To ‘‘force’’ this topology
upon any of the other conflicting data matrices requires either ordering characters or accepting longer trees (albeit only one step longer in the combined sequence and morphological analysis).
When biogeographic information is optimized
upon the combined plastid and nrDNA tree (Fig.
4), the gene phylogeny supports an origin for the
tribe in the central Andes, inarguably a locus of
diversity for the Andean tetraploid clade of the
411
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
GALLEY 412
mobt 89_303 Mp_412
File # 03em
412
Annals of the
Missouri Botanical Garden
only species of the three that is restricted to northern South America (including Trinidad-Tobago).
The known distribution of the Hymenocallideae
suggests two possible hypotheses, either a long-distance dispersal event from the Andean center of
origin, or extinction of intervening populations of a
proto-Hymenocallis ancestor. The fleshy seed of Hymenocallis is the largest of all the endemic American Amaryllidaceae, exhibits no dormancy, and
germinates within 3–4 weeks after release, whether
or not in substrate (Whitehead & Brown, 1940;
pers. obs.). The relatively heavy seed does not immediately seem amenable to long-distance dispersal, and no dispersal agent other than water has even
been suggested for the genus. Thus ancestral extinction is a more convincing hypothesis, but without a better understanding of the historical biogeography of Hymenocallis and a well-resolved
phylogeny of the genus a likely explanation for its
distribution cannot be determined.
In summary, combined trnL-F and ITS sequences support the Meerow and Snijman (1998) treatment of Hymenocallideae with three genera: Hymenocallis, Ismene, and Leptochiton. Leptochiton is
sister to the Hymenocallis/Ismene clade and retains
two plesiomorphic characters of the Andean tetraploid clade: 16 to 20 ovules per locule and a phytomelanous seed coat. The central Andean endemism of Ismene and Leptochiton and the absence of
Hymenocallis from this region further suggest a vicariance event at some point subsequent to the origin of the tribe. It is thus appropriate to formalize
the recognition of the two new subgeneric combinations within Ismene.
Bull, J. J., J. P. Huelsenbeck, C. W. Cunningham, D. L.
Swofford & P. J. Waddell. 1993. Partitioning and combining data in phylogenetic analysis. Syst. Biol. 42:
384–397.
Chase, M. W., D. W. Stevenson, P. Wilkin & P. J. Rudall.
1995. Monocot systematics: A combined analysis. Pp.
685–730 in P. J. Rudall, P. J. Cribb, D. F. Cutler & C.
J. Humphries (editors), Monocotyledons: Systematics
and Evolution, Vol. 2. Royal Botanic Gardens, Kew.
Dahlgren, R. M. T., H. T. Clifford & P. F. Yeo. 1985. The
Families of the Monocotyledons. Springer-Verlag, Berlin.
De Queiroz, A. 1993. For consensus (sometimes). Syst.
Biol. 42: 368–372.
, M. J. Donoghue & J. Kim. 1995. Separate versus
combined analysis of phylogenetic evidence. Ann. Rev.
Ecol. Syst. 26: 657–681.
Douzery, J. P., A. M. Pridgeon, P. Kores, H. Kurzweil, P.
Linder & M. W. Chase. 1999. Molecular phylogenetics
of Diseae (Orchidaceae): A contribution from nuclear
ribosomal ITS sequences. Amer. J. Bot. 86: 887–899.
Dubuisson, J. Y., R. Hebant-Mauri & J. Galtier. 1998.
Molecules and morphology: Conflicts and congruence
within the fern genus Trichomanes (Hymenophyllaceae)
Molec. Phylogenet. Evol. 9: 390–397.
Farris, J. S., A. G. Kluge & C. Bult. 1994. Testing significance of incongruence. Cladistics 10: 315–319.
, M. Källersjö, A. G. Kluge & C. Bult. 1995. Constructing a significance test for incongruence. Syst. Biol.
44: 570–572.
Felsenstein, J. 1985. Confidence limits on phylogenies:
An approach using the bootstrap. Evolution 39: 783–
791.
Flory, W. S. 1976. Distribution, chromosome number and
types of various species of Hymenocallis. The Nucleus
19: 204–227.
Gereau, R. E., A. W. Meerow & L. Brako. 1993. New
Combinations in Hippeastrum, Ismene, and Leptochiton
(Amaryllidaceae) for the Flora of Peru. Novon 3: 28–
30.
Grant, V. 1983. The systematic and geographical distribution of hawkmoth flowers in the temperate North
American flora. Bot. Gaz. 144: 439–449.
Hammen, T. Van der. 1974. The Pleistocene changes of
vegetation and climate in tropical South America. J.
Biogeogr. 1: 2–26.
. 1979. History of the flora, vegetation and climate
in the Colombian Cordillera Oriental during the last five
million years. Pp. 25–32 in H. Larsen & L. B. HolmNielsen (editors), Tropical Botany. Academic Press,
London.
Herbert, W. 1837. Amaryllidaceae. James Ridgeway and
Sons, London.
Huelsenbeck, J. P., J. J. Bull & C. W. Cunningham. 1996.
Combining data in phylogenetic analysis. Trends Ecol.
Evol. 11: 152–158.
Hunziker, A. T. 1969. Estudios sobre Amaryllidaceae. III.
Sinopsis provisional de Hieronymiella, y novedades Argentinas sobre Zephyranthes. Kurtziana 5: 343–367.
Hutchinson, J. 1934. Families of Flowering Plants, Vol. 2.
Monocotyledons, 1st ed. MacMillan, London.
. 1959. Families of Flowering Plants, Vol. 2. Monocotyledons, 2nd ed. Clarendon Press, Oxford, U.K.
Kluge, A. G. 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates
(Boidae, Serpentes). Syst. Zool. 38: 7–25.
Meerow, A. W. 1987. The identities and systematic rela-
Ismene subg. Elisena (Herbert) Meerow, comb.
nov. Basionym: Elisena Herb., Amaryllidaceae, 75, 201. 1837. TYPE: Ismene ringens
(Ruiz & Pav.) Gereau & Meerow, Novon 3: 29.
1993.
Ismene subg. Pseudostenomesson (Velarde)
Meerow, comb nov. Basionym: Pseudostenomesson Velarde. Rev. Cienc. (Lima) 51: 47–51.
1949. TYPE: Ismene vargasii (Velarde) Gereau
& Meerow, in L. Brako & J. Zarucchi, Monogr.
Syst. Bot. Missouri Bot. Gard. 45: 1253. 1993.
Literature Cited
Baker, J. G. 1888. Handbook of the Amaryllideae. George
Bell and Sons, London.
Bauml, J. A. 1979. A Study of the Genus Hymenocallis
(Amaryllidaceae) in Mexico. M.S. Thesis, Cornell University, Ithaca, New York.
Bremer, K. 1988. The limits of amino acid sequence data
in angiosperm phylogenetic reconstruction. Evolution
42: 198–213.
MISSOURI BOTANICAL GARDEN
Thursday Jul 25 2002 12:59 PM
Allen Press • A V i i O N System
GALLEY 413
mobt 89_303 Mp_413
File # 03em
Volume 89, Number 3
2002
Meerow et al.
Phylogeny of Tribe Hymenocallideae
413
tionships of Mathieua Klotzsch and Plagiolirion Baker
(Amaryllidaceae). Taxon 36: 566–572.
. 1989. Systematics of the Amazon lilies, Eucharis
and Caliphruria (Amaryllidaceae). Ann. Missouri Bot.
Gard. 76: 136–220.
. 1990. 202, Amaryllidaceae. In G. Harling (editor), Flora of Ecuador No. 41. University of Göteborg.
. 1995. Towards a phylogeny of the Amaryllidaceae. Pp. 169–179 in P. J. Rudall, P. J. Cribb, D. F.
Cutler & C. J. Humphries (editors), Monocotyledons:
Systematics and Evolution, Vol. 2. Royal Botanic Gardens, Kew.
& B. Dehgan. 1985. The auriculate pollen grain
of Hymenocallis quitoensis Herb. (Amaryllidaceae) and
its systematic implications. Amer. J. Bot. 72: 540–547.
& D. A. Snijman. 1998. Amaryllidaceae. Pp. 83–
110 in K. Kubitzki (editor), Families and Genera of
Vascular Plants, Vol. 3. Springer-Verlag, Berlin.
, M. F. Fay, C. L. Guy, Q-B. Li, F. Q. Zaman &
M. W. Chase. 1999. Systematics of Amaryllidaceae
based on cladistic analysis of plastid rbcL and trnL-F
sequence data. Amer. J. Bot. 86: 1325–1345.
, C. L. Guy, Q. Li & S-L. Yang. 2000a. Phylogeny
of the American Amaryllidaceae based on nrDNA ITS
sequences. Syst. Bot. 25: 708–726.
, M. Fay, M. W. Chase, C. L. Guy, Q. Li, D. Snijman & S-L. Yang. 2000b. Phylogeny of the Amaryllidaceae: Molecules and morphology. Pp. 368–382 in K.
Wilson & D. Wallace (editors), Monocots: Systematics
and Evolution. CSIRO Press, Sydney.
Miyamoto, M. M. & W. M. Fitch. 1995. Testing species
phylogenies and phylogenetic methods with congruence. Syst. Biol. 44: 64–76.
Morton, J. K. 1965. The experimental taxonomy of the
West African species of Pancratium L. (Amaryllidaceae). Kew Bull. 19: 337–347.
Müller-Doblies, D. & U. Müller-Doblies. 1996. Tribes and
subtribes and some species combinations in Amaryllidaceae J. St.-Hil. emend. R. Dahlgren et al. 1985. Feddes Rep. 107(5–6): S.c. 1–9.
Nixon, K. C. & J. M. Carpenter. 1996. On simultaneous
analysis. Cladistics 12: 221–241.
Olmstead, R. G. & J. A. Sweere. 1994. Combining data
in phylogenetic systematics—An empirical approach
using three molecular data sets in the Solanaceae. Syst.
Biol. 43: 467–481.
Pax, F. 1890. Amaryllidoideae–Narcisseae–Eucharidinae.
In A. Engler & K. Prantl (editors), Die natürlichen
Pflanzenfamilien II. Teil 5: 110–11. Wilhelm Engelmann, Leipzig.
Ravenna, P. 1980. A new yellow-flowered Hymenocallis
(Amaryllidaceae) from north Peru. Bot. Not. 133: 97–
98.
Rodrigo, A. G., M. Kelly-Borges, P. R. Bergquist & P. L.
Bergquist. 1993. A randomization test of the null hypothesis that two cladograms are sample estimates of a
parametric phylogenetic tree. New Zealand J. Bot. 31:
257–268.
Rudall, P. J., M. W. Chase, D. F. Cutler, J. R. Rusby &
A. Y. De Bruijn. 1998. Anatomical and molecular systematics of Asteliaceae and Hypoxidaceae. Bot. J. Linn.
Soc. 127: 1–42.
Salisbury, R. A. 1812. On the cultivation of rare plants.
Trans. Hort. Soc. London 1: 261–366.
Sealy, J. R. 1937. Lepidochiton quitoensis. Bot. Mag. t.
9491.
. 1954. Review of the genus Hymenocallis. Kew
Bull. 2: 201–240.
Seelanan, T., A. Schnabel & J. Wendel. 1997. Congruence
and consensus in the cotton tribe (Malvaceae). Syst.
Bot. 22: 259–290.
Smith, G. L. & W. S. Flory. 1990. Studies on Hymenocallis
henryae (Amaryllidaceae). Brittonia 42: 212–220.
&
. 2001. Hymenocallis (Amaryllidaceae)
in Texas, with a new varietal combination. Novon 11:
229–232.
, L. C. Anderson & W. S. Flory. 2001. A New
Species of Hymenocallis (Amaryllidaceae) in the Lower
Central Florida Panhandle. Novon 11: 233–240.
Soltis, D. E., P. S. Soltis, M. E. Mort, M. W. Chase, V.
Savolainen, S. B. Hoot & C. M. Morton. 1998. Inferring
complex phylogenies using parsimony: An empirical approach using three large DNA data sets for angiosperms. Syst. Biol. 47: 32–42.
Sorenson, M. D. 1999. TreeRot, version 2. University of
Michigan, Boston University, Boston, Massachusetts.
Stapf, O. 1933. Pamianthe peruviana. Bot. Mag. t. 9315.
Stebbins, G. L. 1984. Mosaic evolution, mosaic selection
and angiosperm phylogeny. Bot. J. Linn. Soc. 88: 149–
164.
Swofford, D. L. 1998. PAUP: Phylogenetic Analysis Using
Parsimony, v. 4.0 beta. Sinauer, Sutherland, Massachusetts.
Taberlet, P., L. Gielly, G. Pautou & J. Bouvet. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Pl. Molec. Biol. 17: 1105–
1110
Traub, H. P. 1962. Key to the subgenera, alliances and
species of Hymenocallis. Pl. Life 18: 55–72.
. 1963. Genera of the Amaryllidaceae. American
Plant Life Society, La Jolla, California.
. 1980. Sections and alliances, genus Hymenocallis Salisb. Pl. Life 36: 46–49.
Velarde, M. O. 1949. Pseudostenomesson, un nuevo genero
de Amarilidáceas. Rev. Cienc. (Lima) 51: 47–51.
White, T. J., T. Bruns, S. Lee & J. Taylor. 1990. Amplification and direct sequencing of fungal ribosomal RNA
genes for phylogenetics. Pp. 315–322 in M. Innis, D.
Gelfand, J. Sninsky & T. White (editors), PCR Protocols: A Guide to Methods and Applications. Academic
Press, Orlando, Florida.
Whitehead, M. R. & C. A. Brown. 1940. The seed of the
spider lily, Hymenocallis occidentalis. Amer. J. Bot. 27:
129–198.
Yukawa, T., H. Ohba, K. M. Cameron & M. W. Chase.
1996. Chloroplast DNA phylogeny of subtribe Dendrobiinae (Orchidaceae): Insights from a combined analysis
based on rbcL sequences and restriction site variation.
J. Pl. Res. 109: 169–176.