Content uploaded by Alan W Meerow
Author content
All content in this area was uploaded by Alan W Meerow on Mar 10, 2014
Content may be subject to copyright.
Generic relationships among the baccate-fruited Amaryllidaceae
(tribe Haemantheae) inferred from plastid and nuclear
non-coding DNA sequences
A. W. Meerow
1, 2
and J. R. Clayton
1
1
USDA-ARS-SHRS, National Germplasm Repository, Miami, Florida, USA
2
Fairchild Tropical Garden, Miami, Florida, USA
Received October 22, 2002; accepted September 3, 2003
Published online: February 12, 2004
ÓSpringer-Verlag 2004
Abstract. Using sequences from the plastid trnL-F
region and nrDNA ITS, we investigated the phy-
logeny of the fleshy-fruited African tribe Haeman-
theae of the Amaryllidaceae across 19 species
representing all genera of the tribe. ITS and a
combined matrix produce the most resolute and
well-supported tree with parsimony analysis. Two
main clades are resolved, one comprising the
monophyletic rhizomatous genera Clivia and Cryp-
tostephanus, and a larger clade that unites
Haemanthus and Scadoxus as sister genera to an
Apodolirion/Gethyllis subclade. One of four
included Gethyllis species, G. lanuginosa, resolves
as sister to Apodolirion with ITS. Relationships
among the Clivia species are not in agreement with
a previous published phylogeny. Biogeographic
analysis using the divergence/vicariance method
roots the tribe in Eastern South Africa, with several
subsequent dispersals to the winter rainfall Western
Cape region. Chromosomal change from an ances-
tral 2n¼22 (characteristic of Clivia) is associated
with each main clade. Reduction in number has
occurred in all but Cryptostephanus, which has
2n¼24 chromosomes. Increasing the sampling
across all of the species in the tribe will allow a
more detailed understanding of the biogeographic
patterns inherent in the parsimony topology, which
undoubtedly reflect Quaternary climatic changes in
Southern Africa.
Key words: Amaryllidaceae, Haemantheae, geo-
phytes, South Africa, monocotyledons, DNA,
phylogenetics, systematics.
Baccate fruits have evolved only once in the
Amaryllidaceae (Meerow et al. 1999), and
solely in Africa, but the genera possessing
them have not always been recognized as a
monophyletic group. Haemanthus L. and
Gethyllis L. were the first two genera of the
group to be described (Linneaus 1753). Her-
bert (1837) placed Haemanthus (including
Scadoxus Raf.) and Clivia Lindl. in the tribe
Amaryllidiformes, while Gethyllis was classi-
fied with Sternbergia L. in Oporanthiformes.
Salisbury (1866) recognized the distinct tribes
Haemantheae Salisb. and Gethyllideae Salisb.
Bentham and Hooker (1883) united Crypto-
stephanus Baker with Narcissus L. in their
subtribe Coronatae, while maintaining
Haemanthus,Clivia Lindl. and Apodolirion
Baker in subtribe Genuinae. Cryptostephanus
has perianthal appendages at the throat of the
flower that Bentham and Hooker (1883)
considered homologous to the corona of
Narcissus. Pax (Pax 1887) situated Haeman-
thus and Clivia in his subtribe Haemanthinae
Plant Syst. Evol. 244: 141–155 (2004)
DOI 10.1007/s00606-003-0085-z
Pax, placed Gethyllis and Apodolirion in
Zephyranthinae (on the basis of their fused
spathe bracts and single-flowered inflorescenc-
es), and Cryptostephanus within Narcissinae, a
treatment largely followed by Hutchinson
(1934), though Pax’s (1887) subtribes were
elevated to the rank of tribe. All of these
groups were polyphyletic, uniting genera from
disparate lineages within the family (see dis-
cussion by Nordal and Duncan 1984).
Traub (1963) was the first to recognize the
relationship between Clivia and Cryptosteph-
anus, but placed both as the sole genera in tribe
Clivieae Traub. Haemanthus was relegated to
the monotypic Haemantheae, while Gethyllis
and Apodolirion were placed alone in Gethylli-
deae, with the suggestion that the two genera
were likely indistinct. Melchior (1964) placed
both Clivia and Cryptostephanus in Haeman-
theae. Dahlgren et al. (1985) largely adopted
Traub’s (1963) classification, though Gethylli-
deae and Clivieae were subsumed in Haeman-
theae.
The two most recent formal classifications
of the Amaryllidaceae are those of Mu
¨ller-
Doblies and Mu
¨ller-Doblies (1996), and Mee-
row and Snijman (1998). Both recognized two
tribes for the baccate-fruited genera: Haeman-
theae (Haemanthus,Scadoxus,Clivia and
Cryptostephanus) and Gethyllideae (Gethyllis
and Apodolirion). Mu
¨ller-Doblies and Mu
¨ller-
Doblies (1996) further recognized two sub-
tribes in Haemantheae, Haemanthinae D. &
U.M.-D. (Haemanthus and Scadoxus) and
Cliviinae D. & U.M.-D. (Clivia and Crypto-
stephanus). Scadoxus was segregated from
Haemanthus by Friis and Nordal (1976). All
of the baccate-fruited genera are endemic to
Africa.
Meerow et al. (1999), using three plastid
DNA sequences, confirmed the monophyly of
Haemantheae, but indicated that Gethyllideae
was embedded within the former tribe, and
thus could not be recognized without render-
ing Haemantheae paraphyletic. The level of
sampling and the number of phylogenetically
informative base substitutions were insufficient
to resolve the relationships within the tribe in
that study beyond the well-supported sister
relationship of Apodolirion and Gethyllis which
together terminated a successive grade begin-
ning with Clivia, followed by Cryptostephanus,
Scadoxus and Haemanthus. However, boot-
strap support for each branch in the grade was
moderate to strong. Ito et al. (1999), using
plastid matK sequences also resolved a mono-
phyletic Haemantheae, though only three gen-
era were sampled. Haemanthus and Scadoxus
were sister taxa in their study with 98%
bootstrap support.
As treated here, Haemantheae consists of
six genera. Cryptostephanus (2 spp.) and Clivia
(5 spp.) are bulbless, rhizomatous perennials.
With the exception of the newly described
Clivia mirabilis Rourke, both genera are found
in summer rainfall regions. Clivia is adapted to
butterfly and sunbird pollination, and has
showy orange and yellow flowers. The species
are chiefly understory herbs of coastal and
Afro-montane forest. The two species of
Cryptostephanus are either savanna or forest
herbs. The small flowers have a paraperigone,
and it is the only genus in the tribe whose seeds
have a phytomelanous testa. Scadoxus (9–12
spp.) and Haemanthus (21 spp.) have long been
recognized intuitively as sister taxa (in the past
treated as a single genus; e.g. Bjo
¨rnstad and
Friis 1972). Both genera have brush-like inflo-
rescence morphology, in which the bracts often
form part of the pollinator attraction system.
Scadoxus are forest understory herbs, some
species of which do not form true bulbs. The
genus is most common in the African tropics.
Haemanthus, all species forming bulbs, is
strictly southern African, with species in both
the summer and winter rainfall regions of the
Cape (Snijman 1984). Finally, Gethyllis (ca. 35
species, Mu
¨ller-Doblies 1986) and Apodolirion
(ca. 6 species, Mu
¨ller-Doblies 1986) are two
closely related uni-flowered Cape endemics
that both retain the ovary inside the bulbs
until the large, fleshy, aromatic fruit matures.
They are differentiated by the capitate stigma
in Gethyllis (vs. tri-lobed in Apodolirion) and
the often numerous stamens in Gethyllis (vs.
six in Apodolirion). Gethyllis is most common
142 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
in the winter rainfall region of South Africa,
Apodolirion in the summer rainfall zone.
The purpose of this present study was to
establish baseline generic relationships of the
genera of the Haemantheae by increasing the
sampling for the plastid trnL-F region, and
adding sequences from the internally tran-
scribed spacer (ITS) of nuclear ribosomal
DNA.
Materials and methods
Sampling. Genomic DNA was extracted from
silica gel dried leaf tissue of the taxa listed in
Table 1 as described by Meerow et al. (2000).
DNA extraction, amplification and sequencing
protocols. The trnL-trnF region was amplified and
sequenced 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 intron as described
by Meerow et al. (2000). All polymerase chain
reaction (PCR) amplifications were performed on
an ABI 9700 (Perkin-Elmer Applied Biosystems,
Foster City, California, USA).
Amplified products were purified using
QIAquick (Qiagen, Valencia, California, USA)
columns, following manufacturers’ protocols. Cy-
cle 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 310 or ABI
3100 automated sequencer (according to the man-
ufacturer’s protocols; Applied Biosystems, Foster
City, California, USA).
Sequence alignment. Both the ITS and trnL-F
matrices were readily aligned manually using
Sequencher 4.1 (Gene Codes, Ann Arbor, Michi-
gan, USA). The alignment is accessible through
GenBank or from the first author (miaam@ars-
grin.gov).
Analyses. The ITS matrix consisted of 19 taxa
(two Apodolirion spp., four Clivia spp., two Cryp-
tostephanus spp., four Gethyllis spp., three Hae-
manthus spp., three Scadoxus spp., and one species
of Amaryllis, the latter designated as outgroup).
The plastid trnL-F matrix consisted of the same 19
taxa, plus the addition of Haemanthus humilis.
Amaryllis is the basal most genus within the tribe
Amaryllideae (Meerow and Snijman 2001) that in
turn is sister to the rest of the Amaryllidaceae. The
sister group relationships of Haemantheae are so
far unresolved (Meerow et al. 1999). We experi-
mented with a species of Cyrtanthus Herb.
(Cyrtantheae) and Calostemma R. Brown
(Calostemmateae) as outgroups, but found that
Amaryllis presented the least number of alignment
ambiguities and generated the shortest trees. Res-
olution of the sister relationships of Haemantheae
remain unclear (Meerow et al. 1999); however the
tribe Amaryllideae is sister to all other genera in the
family with high bootstrap support, even with as
highly conserved a gene as rbcL (Meerow et al.
1999). At present we are working to successfully
align ITS sequences across the entire Amaryllida-
ceae in order to resolve the basal polytomy that
resolves after the branching of tribe Amaryllideae
with all plastid sequences that have been applied to
the problem to date (Ito et al. 1999; Meerow et al.
1999, 2000). We feel it is better to use as outgroup
the most basal genus in a tribe that is indisputably
outside of the ingroup of interest.
Aligned matrices were analyzed using the
parsimony algorithm of PAUP* for Macintosh
(version 4.0b10; Swofford 1998), with the MUL-
PARS option invoked. Tree branches were retained
only if unambiguous support was available
(i.e. branches were collapsed if the minimum
length ¼0). Gaps were coded as missing characters
in the initial analyses, but a gap matrix was also
constructed from each alignment using the pro-
gram PAUPGAP (Anthony Cox, RBG Kew),
which applies a strict interpretation of gaps
(i.e. no partial homology). This binary matrix was
added to the sequence alignment and analyzed in
combination. For all matrices, a branch and bound
(Hendy and Penny 1982) search was conducted
under the Fitch (equal) weights (Fitch 1971)
criterion with furthest addition sequence.
We also combined the two data matrices,
opting for the ‘‘total evidence’’ approach (Dubuis-
son et al. 1998, Seelanan et al. 1997). However,
before combining the ITS and trnL-F data sets (and
the gap matrices with the sequence alignments), we
performed partition homogeneity tests on the
matrices (Farris et al. 1994, 1995) to assess the
degree of congruence between them. Five hundred
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 143
Table 1. Species, voucher specimens and GenBank sequence accession numbers used in the phylogenetic analyses of Haemantheae. Vouchers are
deposited at NBI unless otherwise stated
Taxon Voucher Specimen
or Accession
GenBank Accession No. or
Literature Citation
Area Code
1
ITS trnL-F
(trnL, trnL-F spacer)
Amaryllis belladonna L. M. W. Chase 612 (K) Meerow and
Snijman (2001)
Meerow et al. (1999) A
Apodolirion cedarbergense D. M.-D. Dulse s. n. AY280344 AY278957, AY278971 A
A. lanceolatum (Thunb.) Baker NBG 714/88 AY280345 Meerow et al. (1999) A
Clivia caulescens R. A. Dyer Rourke 2167 AY280346 AY278958, AY278973 B
C. gardenii Hook. Rourke 2160 AY280357 AY278960, AY278974 B
C. miniata Regel Rourke 2143 AY280348 AY278961, AY278975 B
C. nobilis Lindl. M. W. Chase 3080 (K) AY280349 Meerow et al. (1999) B
Cryptostephanus haemanthoides Pax Koopowitz 11040 AY280350 AY278962, AY278976 D
C. vansonii Verdoorn Meerow 2310 (FTG) AY280351 Meerow et al. (1999) C
Gethyllis britteniana Baker Van Jaarsveld 5618 AY280352 AY278963, AY278977 A
G. ciliaris (Thunb.) Thunb. Duncan 1123 AY280353 Meerow et al., (1999) A
G. lanuginosa Marl. Van Jaarsveld 4377 AY280354 AY278964, AY278978 A
G. verticillata R. Br. ex Herb. Meerow 2350 (FTG) AY280355 AY278965, AY278979 A
Haemanthus albiflos Jacq. Meerow 2351 (FTG) AY280356 AY278966, AY278980 B
H. graniticus Snijman Snijman 308 AY280357 AY278967, AY278981 AB
H. humilis Jacq. M. W. Chase 2025 (K) – Meerow et al. (1999) B
H. pumilio Jacq. Snijman 668 AY280358 AY278968, AY278982 A
Scadoxus cinnabarinus (Decne.)
Friis & Nordal
M. W. Chase 549 (K) AY280360 Meerow et al. (1999) E
S. membranaceus (Bak.)
Friis & Nordal
NBG 708/88 AY280360 AY278969, AY278983 B
S. puniceus (L.) Friis & Nordal NBG 43/72 AY280361 AY278970, AY278984 BE
1
A = Western South Africa, B = Eastern South Africa, C = Zimbabwe, D = East Africa, E = Tropical Africa
144 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
heuristic searches were conducted for each test,
each with 10 random addition replications, saving
no more than 20 trees from each for TBR branch
swapping.
Internal support was determined by bootstrap-
ping (Felsenstein 1985; 5000 replicates with simple
addition) and calculation of Bremer (1988) decay
indices (DI) using the program TreeRot v. 2.1
(Sorenson 1996). The cut-off bootstrap percentage
is 50. A bootstrap value greater than 75% was
considered good support, 65–75% was designated
moderate support, and less than 65% as weak
(Meerow and Snijman 2001, Meerow et al. 2002).
Five hundred replicate heuristic searches were
implemented for each constraint statement postu-
lated by TreeRot, saving 10 trees per replicate. A
minimum DI ¼2 was considered to represent good
support for a clade.
The biogeographic patterns inferred from our
gene trees were assessed using both Fitch optimi-
zation (Maddison et al. 1992) with MacClade
version 4.03 (Maddison and Maddison 2001) and
the dispersal-vicariance method of analysis
(Ronquist 1997) using the program DIVA version
1.1 (Ronquist 1996). The program uses vicariance
(i.e allopatric speciation) in its optimization of
ancestral distributions but takes into consideration
dispersal and extinction events and indicates their
direction (Ronquist 1996, 1997). The most parsi-
monious reconstructions minimize such events.
Unlike Fitch optimization, DIVA does not restrict
widespread distributions to terminals or limit
ancestral distributions to single areas (Ronquist
1996). The single tree from the combined sequence
analysis was used for optimization of five coded
geographic areas (Table 1). Fitch optimization of
area data was performed on the same tree using a
single multistate character (Table 1). An exact
optimization (versus heuristic) was invoked in the
analysis by allowing the maximum number of
alternative reconstructions to be held at any node.
The maximum areas allowed at ancestral nodes was
set to the minimum (2) in order to reduce
ambiguities at the more basal nodes of the tree
(DIVA tends to optimize all possible areas at the
lower nodes of the tree if the maximum value is
used). Five biogeographic areas were coded for the
analysis (Table 1), based on the distributions of the
taxa in our sequence matrix. Western South Africa
is equivalent to the winter-rainfall region; Eastern
South Africa, to the summer-rainfall zone.
Results
Plastid trnL-F. Of the 975 characters included
in the analysis, 23 were parsimony informative.
Three equally parsimonious trees were found of
length ¼96 steps, consistency index (CI) ¼0.94,
and retention index (RI) ¼0.86 (Fig. 1A). The
tree is not well-resolved, and only two genera
form monophyletic groups, Cryptostephanus
(bootstrap ¼95%, DI ¼3) and Scadoxus (no
bootstrap support, DI ¼1). Apodolirion and
Gethyllis form a clade (bootstrap ¼81%,
DI ¼2), but neither genus is resolved as mono-
phyletic. When a binary gap matrix is added to
the sequence alignment, the number of parsi-
mony informative characters increases to 57, out
of a total of 1039. The gap matrix is mostly
incongruent with the sequence alignment
(P ¼0.136 in the partition homogeneity test).
Two trees were found (partially shown in
Fig. 1B), of length ¼199, CI ¼0.77 and
RI ¼0.65. Support for a monophyletic Crypto-
stephanus is increased (bootstrap ¼99%,
DI ¼5), Clivia is resolved in one of the two trees
as monophyletic without support (Fig. 1B), and
Apodolirion and Gethyllis are each resolved as
monophyletic sister clades (individually without
support), but with a lower bootstrap (65%). The
monophyly of Scadoxus is lost with the addition
of the gap matrix.
ITS. Of the 749 characters (ITS1, 5.8S
gene, ITS2) included in the analyses, 153 were
parsimony informative. The search found 12
equally most parsimonious trees of
length ¼513, CI ¼0.76 and RI ¼0.73
(Fig. 2A). The larger number of characters
results in a much more resolved tree topology
than that from trnL-F. Two main clades are
resolved. One unites Cryptostephanus and
Clivia with a bootstrap of 53% and DI ¼1.
Clivia is monophyletic with strong support
(bootstrap ¼96%, DI ¼6), but a monophy-
letic Cryptostephanus has no support and is not
resolved in all twelve trees.
The second clade, well supported with a
bootstrap ¼96% and DI ¼6, consists of two
subclades. One unites Apodolirion and Gethyllis
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 145
with a bootstrap of 97% and DI ¼6, though one
species of Gethyllis (G. lanuginosa) is embedded
within Apodolirion (bootstrap ¼93%, DI ¼4).
The remaining three Gethyllis species are united
with a bootstrap ¼93% and DI ¼3. The second
subclade is weakly supported (boot-
strap ¼56%, DI ¼1), and unites Haemanthus
and Scadoxus as sister groups, each strongly
supported (Fig. 2).
The ITS gap matrix and sequence align-
ment are largely incongruent (P ¼0.038), and
addition of the matrix resulted in 191 parsi-
mony informative characters of a total of 844.
Three equally parsimonious trees were found
of length ¼642, CI ¼0.77 and RI ¼0.73 (par-
tially shown in Fig. 2B). The only topological
changes from the trees generated by sequence
matrix alone are 1) the breakup of Crypto-
stephanus such that C. vansonii is resolved as
sister to the rest of the tribe (with no support),
2) Gethyllis ciliaris and G. verticillata switch
positions, and 3) higher resolution within
Fig. 1. Trees found by phylogenetic analysis of plastid trnL-F DNA sequences across 19 species of
Haemantheae, with Amaryllis belladonna used as outgroup. AOne of three most parsimonious trees found with
gaps coded as missing data. BIncreased resolution gained by adding a binary strict gap matrix to the sequence
alignment (two trees found). Numbers above branches are branch lengths. Numbers below branches are
bootstrap percentages and decay indices (italic), respectively. Dashed lines are zero-length branches. A white
bar across a branch signifies a collapsed node in the strict consensus of all trees
146 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
Clivia (Fig. 2B). C. caulescens and C. gardenii
become sister species (bootstrap ¼64%,
DI ¼1), and C. miniata is united with C.
nobilis (bootstrap ¼69, DI ¼2).
Combined analysis. The Pvalue ¼0.46
indicates a moderate level of congruence
between the ITS and trnL-F sequence matrices.
A single most parsimonious tree was found of
length ¼612, CI ¼0.80, RI ¼0.73 (Fig. 3). The
tree is fully resolved, with the same two main
clades resolved by ITS alone (Fig. 2). Crypto-
stephanus is resolved as monophyletic with
moderate support (bootstrap ¼69, DI ¼4),
and the support for its sister relationship to
Clivia also increases (bootstrap ¼73%,
DI ¼5). Apodolirion receives weak bootstrap
support, but Gethyllis is still resolved as
paraphyletic, due to the resolution of G.
lanuginosa as sister to Apodolirion. Support
for the sister relationship of Scadoxus and
Fig. 2. Trees found by phylogenetic analysis of nrDNA ITS sequences across 18 species of Haemantheae, with
Amaryllis belladonna used as outgroup. AOne of twelve most parsimonious trees found with gaps coded as
missing data. BIncreased resolution gained by adding a binary strict gap matrix to the sequence alignment
(three trees found). Numbers above branches are branch lengths. Numbers below branches are bootstrap
percentages and decay indices (italic), respectively. A white bar across a branch signifies a collapsed node in the
strict consensus of all trees
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 147
Haemanthus remains weak, however. If both
gap matrices are added to the combined
sequence data set, there is increased support
for all of the clades (Fig. 3), but no change in
tree topology.
Biogeographic analysis. DIVA hypothe-
sizes six vicariance events to account for the
optimal reconstruction of area on the com-
bined sequence topology (Fig. 4), and roots
the ancestral node of Haemantheae in Eastern
South Africa (equivocal with Fitch optimiza-
tion). Dispersal to the Western Cape occurred
twice, once for the ancestor of Apodolirion and
Gethyllis, and again within Haemanthus. The
Western Cape Clivia mirabilis, not included in
our analysis, would ostensibly represent a third
vicariance event. Moreover, the remaining four
Apodolirion species not included in our anal-
ysis are from the summer rainfall areas of the
Eastern Cape, as are two of the 32 described
species of Gethyllis, thus vicariance events
between the winter and summer rainfall
regions of South Africa are undoubtedly
greater than three.
Discussion
Our combined trnL-F and ITS analysis
(Fig. 3), the most completely resolved and best
supported tree for Haemantheae, divides the
tribe into two main clades. The smaller clade,
uniting Clivia and Cryptostephanus, represents
Fig. 3. Single most parsimonious
tree found by phylogenetic analysis
of combined nrDNA ITS and plas-
tid trnL-F sequences across 19 spe-
cies of Haemantheae with Amaryllis
belladonna used as outgroup, and
gaps coded as missing characters.
Numbers above branches are
branch lengths. Numbers below
branches are bootstrap percentages
and decay indices (italic), respec-
tively. If these changed when a strict
gap matrix was added to the
sequence alignment, the revised
bootstrap and DI are shown be-
tween parentheses
148 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
entirely rhizomatous genera that never form
bulbs (Fig. 5A). Meerow et al. (1999) consider
the bulbless state plesiomorphic for the family.
Cryptostephanus is also the only genus of the
tribe that retains the plesiomorphic character
of a phytomelanous testa (Fig. 5B; Meerow
and Snijman 1998; loss of phytomelan in the
tribe Amaryllideae is considered an indepen-
dent event). The second clade contains all of
the genera that form true bulbs (Fig. 5A),
though Scadoxus is polymorphic for this
character and has been misdiagnosed as being
entirely rhizomatous (Friis and Nordal 1976).
It is unclear whether bulbs form in Scadoxus
only under certain environmental conditions
or if bulb formation is limited to just certain
species.
The second clade contains two subclades
that can be characterized morphologically as
well (Fig. 5). The sister relationship of
Haemanthus and Scadoxus is only resolved by
ITS and by the combined analysis, but is well
supported by the morphological synapomor-
phy of the brush-like inflorescence (Fig. 5C),
facilitated by the reduction in perianth size (all
species) and the dominance of the spathe
bracts during anthesis [this occurs in at least
some of the species of each genus (Friis and
Nordal 1976, Nordal and Duncan 1984)].
Within Haemanthus, well-supported sister
clades are resolved that correspond to Eastern
Cape (H. albiflos,H. humilis) vs. Western Cape
(H. graniticus,H. pumilio) endemics (Snijman
1984). The gethyllid subclade is characterized
by a suite of morphological characters: uniflo-
ry, obsolete scape (Fig. 5D), and the long,
aromatic, cylindrical, many-seeded fruit of
both recognized genera, in contrast to the
one or few seeded berry of the other genera in
the tribe.
Chromosomal change appears to have
figured importantly in cladogenesis within
Haemantheae (Fig. 5E). Clivia has 2n¼22
chromosomes, the plesiomorphic number for
the family (Meerow 1995), while Cryptosteph-
anus has 2n¼24, which may have been derived
Fig. 4. Single most parsimonious tree found
by phylogenetic analysis of combined nrDNA
ITS and plastid trnL-F sequences across 19
species of Haemantheae with Amaryllis
belladonna used as outgroup showing
optimization of bigeographic data. Fitch opti-
mization is indicated by pattern; dispersal-
vicariance optimization is coded by small letters
at ancestral nodes
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 149
Fig. 5. Fitch optimization of selected characters on the single most parsimonious tree found by phylogenetic
analysis of combined nrDNA ITS and plastid trnL-F sequences across 19 species of Haemantheae with
Amaryllis belladonna used as outgroup. APresence of bulb. BPresence of phytomelan in the testa.
CInflorescence morphology. DScape development. ESomatic chromosome number. FLight environment
150 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
from the ancestral x¼11 (Gouws 1949). The
inferred trend is reduction in number from the
ancestral 2n¼22. Only 2n¼12 chromosomes
has been found in either genus of the erstwhile
Gethyllideae (Wilsenach 1965, Vosa 1986).
Scadoxus and Haemanthus have 2n¼18 and
16 chromosomes, respectively (Vosa and
Marchi 1980). Vosa and Marchi (1980) dem-
onstrated that two small teleocentric chromo-
somes in the karyotype of Scadoxus are
homologous to one large, metacentric chro-
mosome in the complement of Haemanthus.
Vosa and Marchi (1980) considered this to be
an incidence of disploid reduction, and iden-
tified the two short chromosomes of Scadoxus
that were likely homologous to a single long
chromosome in Haemanthus. Vosa and Snij-
man (1984) documented further recombination
events in the evolutionary history of Haeman-
thus that could be correlated with speciation
patterns. Translocations appeared to be fre-
quent occurrence in the genus.
Gouws (1949) noted the striking similarities
between the karyotype of Clivia and Crypto-
stephanus. The latter genus has more acrocen-
tric/sub-telocentric chromosomes than Clivia.
The ‘‘extra’’ chromosome in the haploid com-
plement of Cryptostephanus does not appear
typical of a supernumerary (‘‘B’’) chromosome
(Jones and Rees 1982). None of the chromo-
somes in the haploid karyotype of C. vansonii
are telocentric, and the three shortest, sub-
telocentric chromosomes all have apparent
homologs in the diploid complement (Gouws
1949). An alternative origin for 2n¼24 in
Cryptostephanus is a tetraploid derivation from
the 2n¼12 that demarcates the gethyllids.
However, our tree topology (Fig. 3) would
suggest the origin of this number in Apodoli-
rion and Gethyllis occurred after the divergence
of Clivia and Cryptostephanus.
As was suggested by previous plastid anal-
yses (Meerow et al. 1999), recognition of a
distinct tribe Gethyllideae for Apodolirion and
Gethyllis would render the Haemantheae
paraphyletic. The two genera do, however,
form a monophyletic subclade that is sister to
Haemanthus/Scadoxus (Fig. 3). Unlike all of
the other genera in the tribe which have a few-
seeded berry fruit, the fleshy fruit of both
Apodolirion and Gethyllis is a long, aromatic,
cylindrical, many-seeded structure (Meerow
and Snijman 1998). The seeds of these genera
are small and hard, in contrast to the larger,
water-rich, more or less fleshy seeds of the rest
of the genera in Haemantheae. The scape
remains inside the bulbs of Gethyllis and
Apodolirion, and both are uni-flowered with
fused spathe bracts. At least some species of
Gethyllis have 18 or more stamens. Traub
(1963) expressed doubt about maintaining
Apodolirion and Gethyllis as distinct genera,
an argument also taken up to some extent by
Hilliard and Burtt (1973). Wilsenach (1965)
found little variation among in the karyotypes
of representatives of both genera. While our
sampling is hardly complete, there are two well
supported clades resolved within the gethyl-
loids (Fig. 3); however G. lanuginosa is sister to
Apodolirion in the ITS and combined phylo-
geny. In the trnL-F sequence with gap matrix
(Fig. 1B), the two genera are resolved as
distinct sisters. This is clearly a question that
will benefit from a full sampling of all known
species of both genera, and ultimately, recog-
nition of a single genus may be warranted.
Our combined analysis resolves two sub-
clades within Clivia with moderate to strong
support (Fig. 3). These sister species relation-
ships are not in agreement with the maximum
likelihood topologies postulated by Ran et al.
(2001) in their study of Clivia using ITS and
5.8S nrDNA sequences, in which sister rela-
tionships are the reverse of those in Fig. 3.
While our C. caulescens sequence appears
congruent with that of Ran et al. (2001),
downloaded from GenBank, our sequence of
C. nobilis is most congruent with Ran et al.’s
(2001) C. gardenii. Our C. miniata and
C. gardenii sequences are also considerably
divergent from those of Ran et al.’s (2001).
Koopowitz (2002) points out that C. miniata
overlaps with C. nobilis in the southern part of
the former’s range, and with C. gardenii at its
more northerly limits. Though the three spe-
cies are somewhat ecologically specialized,
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 151
mixed populations of C. miniata and either of
the other two species have been observed
(Koopowitz 2002). Though Koopowitz (2002)
concludes that natural hybridization is rare
among Clivia species, some degree of historical
introgression of genes of one species into
another can not be ruled out. Conrad and
Reeves (2002), using four non-coding plastid
regions (a total of 17 phylogenetically infor-
mative base substitutions), resolved yet a third
topology for Clivia, in which C. miniata and C.
caulescens are sister species, with C. gardenii,
C. nobilis, and finally C. mirabilis Rourke
[recently described from the Western Cape
(Rourke 2002)] forming a successive grade.
Understanding the genetic relationships
among Clivia species would clearly benefit
from a population genetic marker approach
such as microsatellite DNA.
A detailed biogeographic analysis of
Haemantheae is premature given that
Apodolirion,Gethyllis,Haemanthus and
Scadoxus were not fully sampled, especially
in regard to species of both Apodolirion and
Gethyllis from Eastern South Africa. However,
both Fitch optimization of biogeographic data
onto our combined tree and divergence/vicar-
iance analysis (Fig. 4) indicates that the prob-
able origins of the tribe are in eastern South
Africa. Only two genera occur outside of
South Africa. Scadoxus is found from the
Arabian peninsula west to Senegal and south
to the East Cape (Friis and Nordal 1976,
Nordal and Duncan 1984), but only three of
the nine recognized species occur in South
Africa (S. membranaceus,S. multiflorus (Mar-
tyn) Raf. and S. puniceus). Cryptostephanus is
absent from South Africa completely, and is
distributed from South Central to East Africa
(Koopowitz 2002).
Prior to the Pliocene, Africa’s southwestern
region was a more mesic, subtropical environ-
ment (Coetzee 1978, 1983, 1986; Hendey 1983;
Scholtz 1985). The earliest evidence of modern
semi-arid, winter-rainfall pattern dates to the
Late Pliocene [though Coetzee (1993) hypoth-
esizes an earlier establishment of the Benguella
current, as well as replacement of forest by
savanna and grassland in the mid-Tertiary], but
it may not have been fully established until the
Early Pleistocene (Hendey 1983, Tankard and
Rogers 1978). Moreover, the winter-rainfall
region of southern Africa experienced a more
recent pattern of expansion and contraction
with concurrent wetter and drier conditions
during glacial and interglacial periods of the
Quaternary (Tankard 1976, van Zinderen Bak-
ker 1976, Tyson 1986, Crockcroft et al. 1987).
Divergence of the three main clades within
Haemantheae may thus have occurred during
the Pliocene, while speciation within their
component genera might have been engendered
by more recent paleoclimatic events. However,
a detailed history of these late Pleistocene and
Quaternary events in the Cape region is elusive
(Cowling et al. 1999). In all of the African tribes
of Amaryllidaceae, most genera have species in
both the winter and summer rainfall regions;
only in the tribe Amaryllideae does Western
Cape endemism occur at the generic level
(Snijman and Linder 1996).
Patterson and Givnish (2002) discussed the
correlation of rhizomatous growth habit, bac-
cate fruits and net-veined leaves with coloni-
zation of low light habitats in Liliales. In
Haemantheae we can see partial support for a
similar scenario (Fig. 5F), insofar as three
genera of the tribe (Clivia,Cryptostephanus
and Scadoxus) are predominately plants of
low-light habitats and lack bulbs (completely
or in part; Fig. 5A). However, net venation
only occurs in Scadoxus, and at least one
species of Haemanthus (H. albiflos) has sec-
ondarily colonized shady habitats (Fig. 5F).
Given the position of Clivia and Cryptosteph-
anus as sister to the rest of the tribe, there is a
least a reasonable possibility that the bulbless
condition and the evolution of berry fruits, in
conjunction with a forest understory habitat,
was a basal event in the divergence of the
Haemantheae from the rest of the family.
Taxonomically, our sequence phylogeny
would support the recognition of three subtribes
in Haemantheae: Cliviinae D. & U. M.-D.,
Haemanthineae Pax and Gethyllidinae. Only
the latter has yet to be formally named.
152 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
Amaryllidaceae J. St.-Hil., tribe Haeman-
theae, subtribe Gethyllidinae (Dumort.) Mee-
row, stat. nov. Tribe Gethyllideae Dumort.,
Anal. Fam., Pl. 58, 1829. Type: Gethyllis L.,
1753.
In conclusion, our combined analysis
results in a well-resolved, well-supported phy-
logenetic framework for the Haemantheae that
can be augmented further by increasing the
depth of sequence sampling for Apodolirion,
Gethyllis, Haemanthus and Scadoxus. Such
a project is underway in South Africa
(G. Reeves, personal communication), and
will hopefully clarify some the questions
remaining about generic and species relation-
ships within the tribe, as well as illuminate the
phytogeographic history of the baccate-fruited
amaryllids.
This work was partially supported by National
Science Foundation Grants DEB-968787 and
0129179. We thank Drs. H. Koopowitz and D. A.
Snijman for providing leaf material of several
species, and D. A. Snijman for information about
species distributions within Haemantheae.
References
Bentham G., Hooker J. D. (1883) Genera planta-
rum, vol. 3. L. Reeve, London.
Bjo
¨rnstad I. N., Friis I. B. (1972) Studies on the
genus Haemanthus L. (Amaryllidaceae) II. A
revision of the Section Demeusea (De Wild. &
Th. Dur.) Pax & Hoffm. emend. I. Bjo
¨rnstad &
I. Friis. Norw. J. Bot. 19: 207–222.
Bremer K. (1988) The limits of amino acid
sequence data in angiosperm phylogenetic recon-
struction. Evolution 42: 198–213.
Coetzee J. A. (1978) Climatic and biological
changes in south-western Africa during the late
Cainozoic. In: Zinderen Bakker E. M. van,
Coetzee J. A. (eds.) Palaeoecology of Africa 10.
Balkema, Rotterdam, pp. 13–29.
Coetzee J. A. (1983) Intimations on the Tertiary
vegetation of southern Africa. Bothalia 14: 345–
354.
Coetzee J. A. (1986) Palynological evidence for
major vegetation and climatic change in the
Miocene and Pliocene of the south-western
Cape. S. African J. Sci. 82: 71–72.
Coetzee J. A. (1993) African flora since the terminal
Jurassic. In: Goldblatt P. (ed.) Biological rela-
tionships between Africa and South America.
Yale University Press, New Haven, pp. 37–61.
Conrad F., Reeves G. (2002) Molecular systematics
of the genus Clivia. Clivia 4: 20–23.
Cowling R. M., Cartwright C. R., Parkington J. E.,
Allsopp J. C. (1999) Fossil wood charcoal
assemblages from Elands Bay Cave, South
Africa: implications for Late Quaternary vege-
tation and climates in the winter-rainfall fynbos
biome. J. Biog. 26: 367–378.
Crockcroft M. J., Wilkinson M. J., Tyson P. D.
(1987) The application of a present-day climatic
model to the late Quaternary in southern Africa.
Clim. Change 10: 161–181.
Dahlgren R. M. T., Clifford H. T., Yeo P. F. (1985)
The families of monocotyledons: structure, evo-
lution, and taxonomy. Springer, Berlin.
Douzery J. P., Pridgeon A. M., Kores P., Kurzweil
H., Linder P., Chase M. W. (1999) Molecular
phylogenetics of Diseae (Orchidaceae): a
contribution from nuclear ribosomal ITS
sequences. Amer. J. Bot 86: 887–899.
Dubuisson J. Y., Hebant-Mauri R., Galtier J. (1998)
Molecules and morphology: conflicts and con-
gruence within the fern genus Trichomanes (Hy-
menophyllaceae) Mol. Phylog. Evol. 9: 390–397.
Farris J. S., Ka
¨llersjo
¨M., Kluge A. G., Bult C.
(1994) Testing significance of incongruence.
Cladistics 10: 315–319.
Farris J. S., Ka
¨llersjo
¨M., Kluge A. G., Bult C.
(1995) Constructing a significance test for incon-
gruence. Syst. Biol. 44: 570–572.
Felsenstein J. (1985) Confidence limits on phylog-
enies: an approach using the bootstrap. Evolu-
tion 39: 783–791.
Fitch W. M. (1971) Toward defining the course of
evolution: Minimum change for a specific tree
topology. Syst. Zool. 20: 406–416.
Friis I., Nordal I. (1976) Studies on the genus
Haemanthus (Amaryllidaceae) IV. Division of
the genus into Haemanthus s. str. and Scadoxus
with notes on Haemanthus s. str. Norw. J. Bot.
23: 63–77.
Gouws J. B. (1949) Karyology of some South
African Amaryllidaceae. Pl. Life 5: 54–81.
Hendey Q. B. (1983) Palaeoenvironmental implica-
tions of the Late Tertiary vertebrate fauna of the
fynbos region. In: Deacon H. J., Hendey Q. B.,
Lambrechts J. J. N. (eds.) Fynbos palaeoecology:
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 153
a preliminary synthesis. So. Afr. Nat. Sci. Prog.
Rep. No. 75. CSIR, Pretoria, pp. 100–115.
Hendy M. D., Penny D. (1982) Branch and bound
algorithms to determine minimal evolutionary
trees. Mathem. Biosci. 59: 277–290.
Herbert W. (1837) Amaryllidaceae. J. Ridgeway
and Sons, London.
Hilliard O. M., Burtt B. L. (1973) Notes on some
plants of southern Africa. III. Notes Roy. Bot.
Gard. Edinburgh 32: 303–387.
Hutchinson J. (1934) The families of flowering
plants. II. Monocotyledons. MacMillan, Lon-
don.
Ito M., Kawamoto A., Kita Y., Yukawa T., Kurita
S. (1999) Phylogenetic relationships of Amaryl-
lidaceae based on matK sequence data. J. Pl.
Res. 112: 207–216.
Jones R. N., Rees H. (1982) B chromosomes.
Academic Press, New York.
Koopowitz H. (2002) Clivias. Timber Press, Port-
land, OR.
Linneaus C. (1753) Species plantarum. Stockholm.
Maddison D. R., Maddison W. P. (2001) Mac-
Clade, version 4.03. Sinauer Associates, Suther-
land.
Maddison D. R., Ruvolo M., Swofford D. L.
(1992) Geographic origins of human mitochon-
drial DNA: phylogenetic evidence from control
region sequences. Syst. Biol. 41:111–124.
Meerow A. W. (1995) Towards a phylogeny of the
Amaryllidaceae. In: Rudall P. J., Cribb P. J.,
Cutler D. F., Humphries C. J. (eds.) Monocot-
yledons: systematics and evolution. Royal
Botanic Gardens, Kew, pp. 169–179.
Meerow A. W., Snijman D. A. (1998) Amaryllid-
aceae. In: Kubitzki K. (ed.) Families and genera
of vascular plants, vol. 3. Springer, Berlin, pp.
83–110.
Meerow A. W., Snijman D. A. (2001) Phylogeny of
Amaryllidaceae tribe Amaryllideae based on
nrDNA ITS sequences and morphology. Amer.
J. Bot. 88: 2321–2330.
Meerow A. W., Guy C. L., Li Q-B., Yang S-Y.
(2000) Phylogeny of the American Amaryllida-
ceae based on nrDNA ITS sequences. Syst. Bot.
25:708–726.
Meerow A. W., Guy C. L., Li Q-B, Clayton J. R.
(2002) Phylogeny of the tribe Hymenocallideae
(Amaryllidaceae) based on morphology and
molecular characters. Ann. Missouri Bot. Gard.
89: 400–413.
Meerow A. W., Fay M. F., Guy C. L., Li Q-B.,
Zaman F. Q., Chase M. W. (1999) Systematics
of Amaryllidaceae based on cladistic analysis of
plastid rbcL and trnL-F sequence data. Amer. J.
Bot. 86: 1325–1345.
Melchior H. (1964) Liliiflorae. In: Melchior H. (ed.)
Engler’s Syllabus der Pflanzenfamilien 2. Berlin,
pp. 513–543.
Mu
¨ller-Doblies D. (1986) De Liliifloris notulae. 3.
Enumeratio specierum generum Gethyllis et
Apodolirion (Amaryllidaceae). Willdenowia 15:
465–471.
Mu
¨ller-Doblies D., Mu
¨ller-Doblies U. (1996) Tribes
and subtribes and some species combinations in
Amaryllidaceae J. St.-Hil. emend. R. Dahlgren
et al. 1985. Feddes Repert. 107 (5-6): s.c. 1–9.
Nordal I., Duncan T. (1984) A cladistic analysis of
Haemanthus and Scadoxus. Nord. J. Bot. 4:
145-153.
Patterson T. B., Givnish T. J. (2002) Phylogeny,
concerted convergence, and phylogenetic niche
conservatism in the core Liliales: insights from
rbcL and ndhF sequence data. Evolution: 56:
233–252.
Pax F. (1887) Amaryllidaceae. In: Engler A., Prantl
K. (eds.) Die Natu
¨rlichen Pflanzenfamilien II, 5.
W. Engelmann, Leipzig, pp. 97–124.
Ran Y., Hammett R. W. K., Murray B. G. (2001)
Phylogenetic analysis and karyotype evolution in
the genus Clivia (Amaryllidaceae). Ann. Bot. 87:
823–830.
Ronquist F. (1996) DIVA version 1.1. Computer
program and manual available by anonymous
FTP from Uppsala University (ftp.uu.se or
ftp.sysbot.uu.se).
Ronquist F. (1997) Dispersal-vicariance analysis: a
new approach to the quantification of historical
biogeography. Syst. Biol. 46: 195–203.
Rourke J. P. (2002) Clivia mirabilis (Amaryllida-
ceae: Haemantheae), a new species from North-
ern Cape, South Africa. Bothalia 32: 1–7.
Salisbury R. A. (1866) Genera plantarum. London.
Scholtz A. (1985) The palynology of the upper
lacustrine sediments of the Arnot Pipe, Banke,
Namaqualand. Ann. So. Afr. Mus. 95: 1–109.
Seelanan T., Schnabel A., Wendel J. (1997) Con-
gruence and consensus in the cotton tribe
(Malvaceae). Syst. Bot. 22: 259–290.
Snijman D. (1984) A revision of the genus
Haemanthus L. (Amaryllidaceae). S. African J.
Bot. Suppl. 12.
154 A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae
Snijman D. A., Linder H. P. (1996) Phylogenetic
relationships, seed characters, and dispersal sys-
tem evolution in Amaryllideae (Amaryllidaceae).
Ann. Missouri Bot. Gard. 83: 362–386.
Sorenson M. D. (1996) TreeRot.University of
Michigan, Ann Arbor.
Swofford D. L. (1998) Phylogenetic Analysis Using
Parsimony, v. 4.0 beta. Sinauer Associates,
Sutherland.
Taberlet P., Gielly L., Pautou G., Bouvet J. (1991)
Universal primers for amplification of three non-
coding regions of chloroplast DNA. Pl. Mol.
Biol. 17: 1105–1110.
Tankard A. J. (1976) Stratigraphy of a coastal cave
and its palaeoclimatic significance. In: Zinderen
Bakker E. M. van (ed.) Palaeoecology of Africa
9. Balkema, Rotterdam, pp. 151–159.
Tankard A. J., Rogers J. (1978) Late Cenozoic
palaeoenvironments on the west coast of south-
ern Africa. J. Biog. 5: 319–337.
Traub H. P. (1963) Genera of the Amaryllidaceae.
American Plant Life Society, La Jolla.
Tyson P. D. (1986) Climatic change and variability
in southern Africa. Oxford University Press,
Cape Town.
Vosa C. G. (1986) Chromosome studies in the
genus Gethyllis (Amaryllidaceae). Caryologia 39:
251–257.
Vosa C. G., Marchi P. D. (1980) Chromosome
analysis of Haemanthus and Scadoxus. Plant.
Syst. Evol. 135: 119–126.
Vosa, C. G., Snijman D. A. (1984) The cytology of
the genus Haemanthus L. (Amaryllidaceae). S.
African J. Bot. 50: 237–259.
White T. J., Bruns T., Lee S., Taylor J. (1990)
Amplification and direct sequencing of fungal
ribosomal RNA genes for phylogenetics. In:
Innis M., Gelfand D., Sninsky J., White T.
(eds.) PCR Protocols: a guide to methods and
applications. Academic Press, Orlando, pp.
315–322.
Wilsenach R. (1965) On the caryology and phylog-
eny of some genera of Amaryllidaceae. Pl. Life
21: 82–83.
Zinderen Bakker E. M. van (1976) The evolution of
Late-Quaternary palaeoclimates of southern
Africa. In: Zinderen Bakker E. M. van (ed.)
Palaeoecology of Africa 9. Balkema, Rotterdam,
pp. 160–202.
Address of the authors: Alan W, Meerow
(e-mail: miaam@ars-grin.gov), Jason R. Clayton,
USDA-ARS-SHRS, National Germplasm Reposi-
tory, 13601 Old Cutler Road, Miami, Florida,
33156, USA.
A. W. Meerow and J. R. Clayton: Molecular systematics of Haemantheae 155
