1017
American Journal of Botany 92(6): 1017?1024. 2005.
EVOLUTION OF CYRTANDRA (GESNERIACEAE)
IN THE PACIFIC OCEAN:
THE ORIGIN OF A SUPERTRAMP CLADE1
QUENTIN C. B. CRONK,2 MICHAEL KIEHN,3 WARREN L. WAGNER,4 AND
JAMES F. SMITH5,6
2Botanical Garden and Center for Plant Research, University of British Columbia, 6804 SW Marine Drive, Vancouver, British
Columbia V6T 1Z4, Canada; 3Department of Systematic and Evolutionary Botany, University of Vienna, Rennweg 14, A-1030
Wien, Austria; 4Department of Botany, MRC 166, P.O. Box 37102, Smithsonian Institution, Washington, D.C. 20013-7102 USA;
5Department of Biology, Boise State University, 1910 University Drive, Boise, Idaho 83725 USA
Cyrtandra comprises at least 600 species distributed throughout Malesia, where it is known for many local endemics and in Polynesia
and Micronesia, where it is present on most island groups, and is among the most successfully dispersing genera of the Pacific. To
ascertain the origin of the oceanic Pacific island species of Cyrtandra, we sequenced the internal transcribed spacers of nuclear
ribosomal DNA of samples from throughout its geographical range. Because all oceanic Pacific island species form a well-supported
clade, these species apparently result from a single initial colonization into the Pacific, possibly by a species from the eastern rim of
SE Asia via a NW-to-SE stepping stone migration. Hawaiian species form a monophyletic group, probably as a result of a single
colonization. The Pacific island clade of Cyrtandra dispersed across huge distances, in contrast to the apparent localization of the SE
Asian clades. Although highly vagile, the Pacific clade is restricted to oceanic islands. Individual species are often endemic to a single
island, characteristic of the ??supertramp?? life form sensu Diamond (1974, Science 184: 803?806). The evolution of fleshy fruit within
Cyrtandra provided an adaptation for colonization throughout the oceanic Pacific via bird dispersal from a single common ancestor.
Key words: biogeography; Gesneriaceae; Hawaiian Islands; internal transcribed spacer; long-distance dispersal; molecular phy-
logeny; nuclear ribosomal DNA; oceanic islands.
Pacific plant geography?The patterns of distribution,
means of dispersal, and evolution of plant life on the islands
of the Pacific Ocean have interested biologists for centuries.
Close relationships with Asia and Australia have been pro-
posed for some taxa (Knox et al., 1993; Wright et al., 2000;
Jaramillo and Manos, 2001; Howarth et al., 2003), whereas
other groups have been hypothesized to originate from the
Americas (Baldwin et al., 1991; DeJoode and Wendel, 1992;
Howarth et al., 1997; Alice and Campbell, 1999; Ganders et
al., 2000; Lindqvist and Albert, 2002; Wagner et al., in press).
An intriguing question has been the number of times various
taxa expanded their ranges onto the oceanic Pacific islands via
long-distance dispersal. Some groups obviously have a single
origin because they are limited to a single endemic species. In
contrast, other taxa seem to have experienced a much greater
expansion into the Pacific, and (1) are widespread among
many islands, such as Pandanus tectorius (Pandanaceae; Wag-
ner et al., 1990), (2) have undergone diversification and evo-
lution resulting in numerous island endemics, such as Psycho-
tria (Rubiaceae; Nepokroeff et al., 2003), or (3) are a com-
bination of a few widespread species and endemic island spe-
1 Manuscript received 11 August 2004; revision accepted 15 March 2005.
We thank the National Tropical Botanical Garden for providing leaf ma-
terial for Cyrtandra paliku, C. urvillei, C. kusaimontana, C. tahuatensis, and
C. samoensis used in this study, and in particular David Lorence, Ken Wood,
and Tim Flynn for kindly making available material for this study. Additional
material was supplied by Anton Weber (Institute of Botany, University of
Vienna), Max M. J. von Balgooy (National Herbarium Nederland, University
of Leiden), and Saula Vodonaivalu (herbarium, University of the South Pa-
cific, Sula, Fiji). We also thank Jill Preston, David Denton, and Michele Kres-
ge for invaluable assistance in the laboratory, and Diana Percy, Jean-Yves
Meyer, and Joel Lau for assistance in the field. Mark Mort and an anonymous
reviewer provided helpful suggestions on the manuscript.
6 Author for correspondence (e-mail: jfsmith@boisestate.edu)
cies, such as Metrosideros (Myrtaceae) or Scaevola
(Goodeniaceae; Howarth et al., 2003). Among these latter ex-
amples, there is still uncertainty whether the invasion of the
Pacific has been the result of one or multiple introductions. In
the case of Metrosideros, the phylogenetic results imply an
origin in New Zealand of the Pacific island clade, but whether
one or several migrations occurred is still ambiguous (Wright
et al., 2000, 2001). On the other hand, Scaevola certainly ap-
pears to have expanded its range into the Pacific several times
(Howarth et al., 2003).
One region that has been of particular interest and received
a great deal of attention is the Hawaiian Islands. These islands
are some of the most remote land areas on the planet and have
a diverse array of endemic species. Ever since Fosberg?s es-
timate of 272 original colonists accounting for all of the native
angiosperms, speculation on the number of times various taxa
have colonized the Hawaiian Islands (Sakai et al., 1995; Wag-
ner et al., 2001) has ranged from a single to many introduc-
tions. With an increasing number of phylogenetic analyses us-
ing morphological and especially molecular data, more robust
hypotheses have resolved this question, often with unexpected
results. Some taxa that now appear to be the result of a single
introduction and subsequent adaptive radiation had been
thought to represent multiple introductions and were often
placed in different genera, for example the lobelioids (Givnish
et al., 1996) and Psychotria (Nepokroeff et al., 2003). In other
instances, phylogenetic analyses have supported previous hy-
potheses of multiple introductions, such as Scaevola (Howarth
et al., 2003). In one case, Rubus, once thought to be the result
of a single introduction, has been shown to result from two
separate events (Howarth et al., 1997; Alice and Campbell,
1999).
Island biotas are often noted as being disharmonic. Some
1018 [Vol. 92AMERICAN JOURNAL OF BOTANY
groups (poor dispersers) are under-represented, while others
(good dispersers) are over-represented (Carlquist, 1970; Gres-
sitt, 1982). Carlquist (1970) distinguished a set of innovations
promoting long-distance dispersal in plants. These traits vari-
ously characterize families, genera, or single species (tramp
species). Diamond (1974) coined the term ??supertramp?? for
species of high dispersability that are only found on small or
isolated islands?evidence that, in these cases, high dispersa-
bility compensates for ecological specialization (as marked, for
instance, by physiological characteristics) (McNab, 2002).
Certain lineages of plants (supertramp clades), although
widely found on oceanic islands, are not present on continental
regions. This is likely to be due to ecological factors rather
than low dispersability. The New Zealand and Pacific island
clade of Metrosideros (Wright et al., 2000; Wright et al., 2001)
is one such supertramp clade. Notably, very few examples of
back-colonization from oceanic islands to continents are
known, even by groups of plants that are widely distributed
around the Pacific Ocean. The phylogenetic investigation of
taxa that are widely dispersed on oceanic islands (in relation
to their continental relatives), is therefore of interest to shed
light on the origin of these potential supertramp clades and to
determine rates if any, of back-colonization and the frequency
with which oceanic island colonization has occurred. For
groups with large numbers of species distributed widely over
the Pacific, there are two contrasting possibilities of origin.
First, wide dispersal may result from frequent independent col-
onizations from the continental progenitor group. In this case,
the adaptations for dispersal must reside widely within the
??springboard?? continental group (e.g., Scaevola: Howarth et
al., 2003). Secondly, a single colonization may result in the
wide distribution and evolutionary radiation from within a sin-
gle oceanic clade. In this case, a key innovation is implicated
that results in higher dispersability in a single supertramp lin-
eage. The genus Cyrtandra (Gesneriaceae), which is common
in the Pacific, provides an excellent case for investigation of
patterns in Pacific-wide species. Phylogenetic methods have
the potential to determine whether Cyrtandra is a springboard
genus that has led to multiple colonizations of the Pacific, or
whether a single supertramp clade has resulted in all the Cyr-
tandra species of Micronesia and Polynesia.
The genus Cyrtandra?Cyrtandra J. R. & G. Forst. (Ges-
neriaceae) is a genus of at least 600 species (the description
of new species, especially in SE Asia and New Guinea, is still
in progress). It is easily recognizable by a combination of
??cyrtandroid?? characters: its two-stamened flowers (anoma-
lous five-stamened forms are very rare; Gillett, 1970), often
somewhat fleshy leaves, and generally ellipsoidal indehiscent
fruits. However, the genus is extremely variable in life history
traits. It ranges from dwarf herbs through climbers, epiphytes,
and even small trees. The flowers are usually white, but spe-
cies with pink, red, yellow, and orange flowers are also known.
It ranges from southern Thailand and the Nicobar Islands,
through peninsular Malaysia, throughout Malaysia and Indo-
nesia and the Philippines, to Papua New Guinea, the New
Hebrides (Gillett, 1974), the Solomon Islands (Gillett, 1975),
and Northern Queensland, Australia. Regions of particularly
high diversity include Sumatra, Borneo, the Philippines (Mer-
rill, 1922), Sulawesi, and (especially) New Guinea. Most of
the species are regionally localized or even endemic to single
mountains (Atkins and Cronk, 2001; Bramley and Cronk,
2003).
Cyrtandra is also extremely common on the oceanic islands
of the Pacific with approximately 58 species in the Hawaiian
Islands (Wagner, 1999), eight species in the Marquesas Islands,
and 12 species in the Society Islands (Gillett, 1973). It also
occurs in Micronesia, Fiji (Smith, 1953, 1991; Gillett, 1967),
Samoa, Tonga, and throughout French Polynesia (Gillett,
1973). Most of the species occur on single islands only, often
being extremely local endemics. For instance, the recently de-
scribed Hawaiian endemic, Cyrtandra paliku, is known only
from about 70 individuals on northern exposed rock faces on
Mount Namahana (Wagner et al., 2001).
Ecologically, Cyrtandra is almost entirely restricted to rain
forest regions, being absent from dry, seasonal areas. It is very
sensitive to disturbance of the canopy and disappears quickly
after logging; hence it is now almost entirely absent from re-
gions of the SE Asian tropics where the forest experienced
extensive human disturbance. This sensitivity, coupled with
the large numbers of local endemics, suggests that Cyrtandra
may have suffered a high rate of extinction in some areas. It
is also vulnerable to extinction on oceanic islands, where it
typically grows in submontane forests and thickets on steep
slopes or gulches on the high islands. Thus the genus has
potential as an indicator of historical and current disturbance
of island forest systems (Kiehn, 2000). Relatively few hybrids
have been reported, and it is only on Hawaii, where different
species grow together that some species have been shown to
hybridize (Wagner et al., 1990; Smith et al., 1996).
The last complete survey of the genus was carried out by
Clarke (1883), who divided the species then known into sev-
eral sections, although these sections have generally been re-
garded by subsequent taxonomists as highly tentative. Gillett
(1973) did not consider any infrageneric classification (formal
or informal) possible, based on the instability of character
combinations. The most recent treatment of the genus for the
Hawaiian islands (Wagner, 1999) divides the species among
six sections. Considering the size and importance of the genus
in SE Asia, little recent taxonomic work has been carried out
on the Asian species. Most of the recent studies have been
local studies describing new species (Burtt, 1970, 1978, 1990;
Atkins and Cronk, 2001; Bramley and Cronk, 2003). However,
recent molecular work suggests the existence of regional
clades within SE Asia (Atkins et al., 2001). Despite consid-
erable morphologic variability, certain geographical trends
have been suggested. Foliar sclereids (Burtt and Bokhari,
1973) are very common in Bornean species, but are less com-
mon in species further east. Foliar sclereids are not found in
Pacific Ocean species. As regards pollen morphology, Hawai-
ian and South Pacific species of Cyrtandra show a high degree
of uniformity, whereas West Malaysian species proved rather
heterogeneous (Luegmayr, 1993a, b; Kiehn, 2001; Schlag-Ed-
ler and Kiehn, 2001). Seed surface morphology studied by
Beaufort-Murphy (1983) and Mu?hlbauer and Kiehn (1997) re-
vealed two types (reticulate and striate) present across almost
the whole distribution area of Cyrtandra except the Hawaiian
Islands, where only the reticulate type occurs. Cytologically,
the large majority of the counted Cyrtandra species revealed
a diploid number of 2n 5 34, indicating a high degree of
conservation of chromosome number in the genus (Mo?ller and
Kiehn, 2003). Differences may only be present at the chro-
mosome structural level (Storey, 1966; Kokubugata and Mad-
ulid, 2000). Fruits of eastern species tend to be tough walled
and green or brown at ripeness, whereas eastern species (and
June 2005] 1019CRONK ET AL.?PACIFIC CYRTANDRA
TABLE 1. Summary of ITS sequence variability for all taxa and for
Pacific and non-Pacific species. Non-Pacific species include infor-
mation for the outgroup.
Pacific species Non-Pacific species Total
Length range 427?464 438?506 427?506
Sequence divergence (%) 0.22?9.62 6.87?27.4 0.22?27.4
Number of indels 19 27 75
Indel size 1?15 1?9 1?15
GC content (%) 55.7 57.2 56.3
Phylogenetically infor-
mative sites (%) 9.0 17.9 28.7
particularly Pacific species) tend to have fleshy fruits, often
ripening white.
Given such a large and widely distributed genus, it would
not be surprising if the oceanic islands of the Pacific had been
colonized several times independently from the species-rich
source region in Asia. Nevertheless, compared to the genus as
a whole, the Pacific island species are relatively uniform, but
this may be due to the ecologically uniform conditions on
oceanic islands. This paper addresses the question of the origin
of the oceanic island species using phylogenetic methods,
which with certain caveats (Emerson, 2002), are powerful
tools for the study of oceanic island biogeography. Given the
extremely wide dispersion of Cyrtandra among the oceanic
islands of the Pacific, knowledge of the origin of these colo-
nization events from the various continental progenitors might
be expected to shed light on the interesting question of the
evolution of dispersability.
MATERIALS AND METHODS
Sampling and plant material: outgroup and ingroup taxa?Twenty-three
Pacific island species of Cyrtandra, including 14 species representing all six
Hawaiian sections, along with 13 species from mainland Asia, Indonesia,
Philippines, Australia, and Taiwan, representing all sections of Clarke (1883),
and clades recovered from recent phylogenetic analyses (Atkins et al., 2001),
were sampled for a total of 36 species (Appendix, see Supplemental Data
accompanying online version of this article). A single species of Aeschyn-
anthus was used as the outgroup for the analysis, following Atkins et al.
(2001). Aeschynanthus is a genus unambiguously distinct from Cyrtandra,
and available evidence (Mayer et al., 2003) suggests that Aeschynanthus is
part of the sister clade to Cyrtandra.
Molecular methods?Fresh, frozen, or silica-gel-dried leaf material of one
plant of each species was used for total DNA extraction using a modification
of the CTAB procedure of Doyle and Doyle (1987) with no further purifi-
cation or with Qiagen DNeasy extraction kits (Qiagen Ltd, Dorking, Surrey,
UK) following the manufacturer?s instructions. The complete ITS region was
amplified with the polymerase chain reaction (PCR) using the primers and
following the protocols of Mo?ller and Cronk (1997) or those of Smith (2000).
The PCR products were purified prior to sequencing using either the QIAquick
PCR purification kit (Qiagen) or Wizard purification kits (Promega, Madison,
Wisconsin, USA). Purified PCR products were sequenced using either a dye
terminator cycle-sequencing ready-reaction kit on an ABI 377 Prism Auto-
matic DNA sequencer (Perkin Elmer, Applied Biosystems, Foster City, Cal-
ifornia, USA) as described in Atkins et al. (2001), manually using silver
sequencing (Promega) and the methods of Smith et al. (1997) or on a Li-Cor
LongreadIR 4200 automated sequencer (Li-Cor, Lincoln, Nebraska, USA). For
each taxon, forward and reverse sequencing reactions were performed for
sequence confirmation. Both ITS regions were aligned using the CLUSTAL
option in the multiple alignment program Sequence Navigator Version 1.0.1
software package (Perkin Elmer, Applied Biosystems), with minor manual
adjustments. The GC content was determined by inspection. Sequence diver-
gence among taxa was calculated using the distance matrix option in PAUP*
4.0b10 (Swofford, 2000), based on unambiguously alignable regions. All se-
quences have been submitted to Genbank (Appendix).
Phylogenetic analysis?The data were analyzed using PAUP* 4.0b10
(Swofford, 2000). Maximum parsimony (MP) analyses were conducted for
each data set using 1000 random stepwise addition replicates with tree-bisec-
tion-reconnection (TBR) branch swapping algorithm and MulTrees on, saving
all shortest trees. The shortest trees from all searches were used to generate
a strict consensus. Bootstrap support (BS) for nodes (Felsenstein, 1985) was
estimated with 100 heuristic simple taxon addition searches using TBR and
Multrees on. Decay indices (DI; Bremer, 1988) for individual clades were
obtained by comparing the strict consensus of all equal-length trees up to six
steps longer than the shortest tree, using simple addition sequence and TBR
in PAUP*. Clades that persisted in the strict consensus for six steps longer
than the most parsimonious trees were examined using the constraint option
of PAUP*. Descriptive statistics reflecting the amount of phylogenetic signal
in the parsimony analyses were given by the consistency index (CI; Kluge
and Farris, 1969), retention index (RI; Farris, 1989), and the resulting rescaled
consistency index (RC). Only combined ITS-1 and ITS-2 sequence data
(boundaries determined by alignment with known sequences) were subjected
to phylogenetic analyses. Indels were scored as a separate presence/absence
character and added to the sequence data matrix. To investigate the effect of
these additional data, separate analyses with and without a gap matrix were
undertaken.
Bayesian analyses were conducted using MRBAYES version 3.0B4 (Huel-
senbeck and Ronquist, 2001). Initial analyses used 100 000 generations, four
chains, and temperature of 0.2 to determine the burn-in period. These results
indicated that a burn-in of 25 000 generations was more than sufficient to
reach a plateau, and subsequent analyses used two million generations and
the general time reversible (GTR) model with a gamma distribution and in-
variant characters as suggested by the Aikake information criterion (AIC) of
MODELTEST (Posada and Crandall, 1998).
RESULTS
Sequence analysis?Sequences for combined ITS-1 and
ITS-2 (excluding the 5.8S gene) ranged from 427 to 506 bp
with a mean of 465 bp. In general, sequences for the non-
Pacific species were longer than those of the Pacific species,
ranging from 438 to 506 and 427 to 464, respectively. Pair-
wise sequence variability ranged from 0.22 to 27.4%, and
again, the greater variability was among the non-Pacific spe-
cies, 6.87?27.4% compared with 0.22?9.62%. Aligned se-
quences had a length of 564 bp that included a total of 75
indels (based on opening a gap in the sequence, regardless of
length), ranging from one to 15 bp. Of these indels, 31 were
a single bp and 16 were autapomorphic (often for the outgroup
Aeschynanthus). A summary of sequence variability for both
the Pacific and non-Pacific clades is presented in Table 1.
Phylogenetic analysis?Maximum parsimony analysis of
the aligned matrix of 37 taxa resulted in 5905 most parsimo-
nious trees of length 693 with indels excluded from the anal-
ysis (CI 5 0.56, RI 5 0.71). A strict consensus tree based on
the results of analyses that included the gap matrix did not
differ from the strict consensus that excluded indels, therefore
only the data that excluded indels will be presented herein. In
addition, a Bayesian analysis was performed using a GTR
model with gamma distribution and invariant characters as the
best fit determined using the AIC of MODELTEST (Posada
and Crandall, 1998). The MP strict consensus and the Bayes-
ian analysis resulted in identical topologies, although the
Bayesian consensus was better resolved. The MP strict con-
1020 [Vol. 92AMERICAN JOURNAL OF BOTANY
Fig. 1. Maximum parsimony strict consensus cladogram of 5905 trees.
Numbers above branches are bootstrap values before the slash (only values
greater than 50% are shown), Bayesian inference values after the slash. Decay
indices are below branches. Triangles represent insertions relative to other
taxa, and inverted triangles represent deletions. Only indels of 3 bp or greater
that could be plotted onto the strict consensus tree are shown. The size of the
indels is indicated with each triangle. Solid triangles are unique, open ones
indicate homoplasy elsewhere in the tree. The consistency index is 0.56, re-
tention index 0.71, and rescaled consistency index 0.50.
Fig. 2. Area cladogram showing the geographic structuring of the anal-
ysis. Numbers above branches are bootstrap values before the slash (only
values greater than 50% are shown), Bayesian inference values after the slash.
The solid triangle marks a 15-bp deletion. The size of the triangles used as
terminals is relative to the number of species sampled from that geographic
region.
sensus is shown in Fig. 1, with bootstrap values greater than
50% and Bayesian inference (BI) values shown. Cyrtandra
cupulata, a species from peninsular Malaysia, is sister to the
rest of the sample (BS 5 85, BI 5 92). Next, the phylogeny
divides into two groups, although with weak (,50%) boot-
strap support. One includes all the Pacific species, the Phil-
ippine species, and a species from the Bataan Islands (Taiwan).
The other clade includes species from Malaysia, Indonesia,
Borneo, Australia, and New Guinea. Within the former clade,
two large subgroups have high support. These are the ??Pacific
clade?? (BS 5 100, BI 5 100), sister to C. umbellifera of the
Bataan Islands (Taiwan), and the ??Hawaiian clade?? (BS 5 95,
BI 5 100). A 15-bp deletion supports the monophyly of the
Pacific species, and a 12-bp deletion supports the monophyly
of the Hawaiian clade. The species from northern Queensland
(C. baileyii) groups closely with C. monticola (BS 5 100, BI
5 100), which is from adjacent parts of Papua New Guinea,
implying that the only species in Australia is the result of
colonization from the hugely species-rich New Guinea Cyr-
tandra flora.
Relationships of Pacific Cyrtandra?The Pacific clade is
divided into two distinct clades, each receiving high support.
The first clade, with BS of 95 and BI of 100 (Fig. 1) is com-
prised of the Hawaiian endemic species. This clade is clearly
monophyletic based on the data presented here. Further sup-
port comes from a 12-bp indel unique to the Hawaiian species.
Current sampling places C. paliku as the sister to all remaining
species, with moderate to strong support from these data (BS
5 74, BI 5 100). Cyrtandra paliku is a recently described
species from Kaua?i, and although it is tempting to speculate
on a Kaua?i origin for the Hawaiian clade, this can only be
speculative based on the current sampling (14 of 58 species).
Three other species?C. longifolia, C. kauaiensis, and C. wa-
wrae?are also endemic to Kaua?i, but do not show a closer
relationship to C. paliku than to other species, nor do they
form a monophyletic group. Beyond the placement of C. pa-
liku, no relationship within the Hawaiian clade receives strong
support from BS, although one clade of species is supported
with BI of 99 (C. propinqua, C. menziesii, C. paludosa, C.
macraei, C. sandwicensis, and C. grayana).
The second clade with BS of 85 and BI of 99 (Figs. 1, 2)
comprises the non-Hawaiian Pacific species. As with the Ha-
waiian clade, different island groups of species receive mod-
erate to strong support, but relationships among these clades
are less strongly supported. The MP analysis does not resolve
relationships among the different island clades, although a BS
consensus tree weakly (BS 5 38, data not shown) places C.
samoensis from Tonga as sister to the remaining sampled spe-
cies. In contrast, the Bayesian analysis tree (data not shown)
places C. tahuatensis from the Marquesas Islands as sister to
the remaining species, although again with much less support
than is seen in other parts of the tree (BI 5 68), and places
C. samoensis from Tonga as sister to the Samoan species (BI
5 84). Two clades for the Tongan/Samoan species, one com-
prising C. samoensis and the other the ??inland species,?? were
revealed by AFLP analysis (M. Kiehn, unpublished results).
DISCUSSION
Phylogenetic relationships between Pacific and continental
Cyrtandra?The area cladogram (Fig. 2) shows that the sam-
ples of Cyrtandra analyzed here (36 of ;600 species) fall into
June 2005] 1021CRONK ET AL.?PACIFIC CYRTANDRA
a series of geographically structured clades, as also noted by
Atkins et al. (2001). All current evidence indicates that Cyr-
tandra as a whole is monophyletic. Although it is possible that
other genera will be found to be phylogenetically nested within
Cyrtandra, no morphological evidence supports this, and ge-
neric-level sequencing work in the Cyrtandroideae has not re-
vealed any likely candidates. The geographical structuring of
clades within an apparently monophyletic Cyrtandra, there-
fore, represents an interaction between the evolution of the
genus and geotectonic, climatological, and dispersal events.
The Pacific species (Polynesian and Micronesian) form a
highly supported monophyletic group (BS 5 100, BI 5 100).
Because most of the islands are oceanic in origin, tectonically
driven vicariance or land connection events can be ruled out,
and a single dispersal into the Pacific followed by numerous
subsequent interisland dispersals is the most likely explana-
tion. Although the Pacific Cyrtandra species have been con-
sidered highly variable (and have been placed in several sec-
tions), they are remarkably uniform relative to the diversity
across all species. The overwhelming majority are medium-
sized shrubs with white, somewhat campanulate flowers and
fleshy leaves. The pollen of these species is uniform (Lueg-
mayr, 1993a, b; Schlag-Edler and Kiehn, 2001), and foliar
sclereids are absent (Burtt and Bokhari, 1973). Even in the
Hawaiian Islands where the morphological variation is great-
est, Carlquist (1970, p. 155.) was of the opinion that ??[v]ery
likely, the Hawaiian species represent a single stock.?? The
Pacific clade in this analysis is sister to the Taiwanese species
C. umbellifera, and together with this species they are sister
to a group of primarily eastern SE Asian species that share the
shrubby habit, fleshy leaves, and white, somewhat campanu-
late corollas. The apparent relationship with C. umbellifera is
interesting because this species is restricted to the Bataan Is-
lands (between Taiwan and the Philippines) and so is an in-
sular species itself. It therefore appears that the major barrier
to dispersal has been leaving the SE Asian rainforests; once
the Pacific Ocean was colonized, frequent long-distance dis-
persal between islands across the Pacific subsequently oc-
curred.
Radiation within the Pacific clade and the origin of the
Hawaiian species?Within the Pacific Cyrtandra clade are
two relatively well-supported clades (Figs. 1, 2). One of these
represents the Hawaiian clade, the second all other Pacific is-
land species (Fig. 2). The sister group relationship between the
Hawaiian clade and all other Pacific islands implies an early
dispersal into the Hawaiian Islands. The pattern is suggestive
of a stepping stone model with the Hawaiian Islands being the
first stop into the Pacific. However, from these data it is clear
that the origin of the non-Hawaiian Pacific species is not from
within the extant Hawaiian species of Cyrtandra. The Hawai-
ian clade is well-supported as sister to the non-Hawaiian clade,
rather than being a paraphyletic assemblage that includes a
monophyletic non-Hawaiian group, as would be expected if
the Hawaiian Islands were serving as a stepping stone as seen
with Metrosideros (Wright et al., 2000). One alternative ex-
planation may be sampling. The Hawaiian Islands are home
to 58 endemic species of Cyrtandra (Wagner, 1999), and with
only 14 species sampled here, it is possible that greater sam-
pling may reveal the non-Hawaiian clade to be imbedded with-
in the group of Hawaiian species. However, ITS sequences
from 32 additional species of Hawaiian Cyrtandra align more
completely with the Hawaiian species sampled here and share
a 12-bp deletion unique to the Hawaiian Cyrtandra species
sampled here (J. Smith, unpublished results). A better expla-
nation for the sister group relationship among the Pacific spe-
cies may be a missing ancestor that either has not been sam-
pled (e.g., many Fijian species have not been sampled) or is
extinct. The current high islands of the Hawaiian chain are of
recent origin; however, the Hawaiian hot spot is known to have
existed in the same general region for 84 million years
(Clague, 1997), and it is possible that one of these earlier
islands served as the first stepping stone for Cyrtandra into
the Pacific. This missing ancestor may have given rise to two
separate radiations, one in the Hawaiian Islands, the other in
the remaining Pacific islands.
The ITS data presented here also imply that Cyrtandra col-
onized the Hawaiian Islands once (Figs. 1, 2). This is contrary
to previous hypotheses based on morphological diversity (St.
John, 1966; Gillett, 1973; Wagner, 1990), but in accordance
with earlier molecular studies (Samuel et al., 1997; Kiehn,
2001) and with the proposal of a single (or two) origin(s) for
the Hawaiian Cyrtandra by Carlquist (1970). The 58 Hawaiian
species of Cyrtandra are morphologically disparate and have
been classified into six distinct sections, each thought to rep-
resent a separate introduction (except section Apertae), with
possibly one additional introduction within section Crotono-
calyces for C. kealiae. The separate colonizations have been
proposed because of the shared morphological characters with-
in each of these groups of species and morphological similar-
ities to non-Hawaiian species. Despite the high degree of mor-
phological diversity among the Hawaiian species and shared
morphological characteristics with some non-Hawaiian, espe-
cially Fijian, species, the ITS data presented here imply a sin-
gle colonization of the Hawaiian Islands (Figs. 1?3). Again, a
limitation may be in our sampling, both from within the Ha-
waiian species and from the non-Hawaiian clades. However,
we have sampled from many of the other Pacific islands that
have been presumed to be the ancestral home of the Hawaiian
species (Samoa, Fiji, Marquesas, New Guinea), and we have
sampled from each of the sections (Appendix) of the Hawaiian
species. (ITS sequence variability in C. kealiae also implies it
is monophyletic with the Hawaiian species sampled here and
is not more closely related to any of the non-Hawaiian species
[J. Smith, unpublished results].) Although a slight chance re-
mains that some of the Hawaiian species may be from an
independent colonization event, the evidence currently avail-
able strongly suggests a single event. Single colonization
events for Hawaiian island plants have become the most com-
mon scenario since they were analyzed phylogenetically and
is now the current hypothesis for Psychotria (Nepokroeff et
al., 2003), the lobelioids (Givnish et al. 1996), and some mem-
bers of Araliaceae (Costello and Motley, 2001).
Biogeography of Cyrtandra in the Pacific: the supertramp
clade?The monophyly of the Pacific oceanic species raises
two questions: Why did Cyrtandra colonize the Pacific only
once? How was Cyrtandra subsequently able to disperse to all
the larger Micronesian and Polynesian islands? It is possible
that the successful colonization of the Pacific required pre-
adaptations for long-distance dispersal, both in dispersal mech-
anism itself and in the ecological potential for establishment
on oceanic islands, which are not widely present in the species
of the SE Asian rainforests. One possible factor is fruit type:
the Pacific species have rather fleshy fruits (usually ripening
to white, but also to orange or green), whereas many (but not
1022 [Vol. 92AMERICAN JOURNAL OF BOTANY
Fig. 3. The topology of the cladogram superimposed onto a map of the Pacific showing the relationships of the species from the various areas.
all) of the western species, for instance from Borneo and Su-
matra, have hard fruits that ripen to green or brown. Colum-
biform birds, highly dispersant frugivores, are characteristic
features of the Pacific islands (and were more so before an-
thropically caused extinctions; Steadman, 1997). However, al-
though columbiforms would be good candidates for interisland
dispersal of the fleshy fruits, no observations have yet been
made of fruit dispersal in Cyrtandra. Palatability tests of Cyr-
tandra fruits by various vertebrate candidates would be inter-
esting in this regard. The seeds of Cyrtandra are numerous,
minute, and covered by sticky fruit flesh, making external (as
well as internal) transport by birds a further possibility, as
pointed out by Carlquist (1970). Short-distance transport of
seeds could also be through water because some of the fleshy
fruits (e.g., Hawaiian C. paludosa) open with slits and release
the seeds as dispersal units (Kiehn, 2001). Sakai et al. (1995)
and Price and Wagner (2004) have hypothesized external bird
dispersal of the minute seeds by shore birds, such as the Pa-
cific Golden Plover.
The other important possible pre-adaptation is the ecological
potential for colonization. One aspect of this is the potential
for pollination by pollen vectors present on oceanic islands.
Yet again, however, there are no data on pollination in Cyr-
tandra. However, floral shape and color differs greatly among
SE Asian species, suggesting possible adaptation to specialist
pollinators. The rather open, white flowers of the Pacific spe-
cies suggest possible pollination by generalist insects, which
may be an advantage for colonists. Autogamy is another po-
tential mechanism to promote vagility, but the showy flowers,
the tendency to gynodioecy in C. longifolia (Carlquist, 1970;
Wagner et al., 1990; Kiehn, personal observation) and inter-
specific hybrids (Wagner et al., 1990; Smith et al., 1996; Kiehn
et al., 2001) suggest out-crossing. However, studies should be
conducted to determine if autogamy occurs in both island and
nonoceanic species. Habitat versatility has been suggested as
another important factor in vagility (Wilson, 1959; Williams,
1969; Carlquist, 1970), based on animal studies, which indi-
cate that habitat generalists, particularly those capable of sur-
viving in dry sites, are responsible for colonization, followed
by evolutionary specialization into wet sites. However, there
is no evidence of this in Pacific Cyrtandra, a genus of eco-
logical specialists of wet shady sites, typically in wet montane
gullies.
In gaining the extraordinary dispersability that has made it
ubiquitous in the Pacific, the Pacific Cyrtandra clade has ap-
parently become a supertramp clade (following the terminol-
ogy of Diamond, 1974). Tramp species have high dispersabil-
ity, whereas supertramp species are adapted to conditions on
small and remote islands, which they can reach because of
their extremely high vagility. Evidence for the supertramp hy-
pothesis in birds comes from disharmonic patterns of distri-
bution of bird species on islands (Diamond, 1974, 1975,
1982). Pacific Cyrtandra species are all island endemics be-
cause they have evolved into endemic species wherever inter-
island colonization events have occurred. It is therefore nec-
essary to think in terms of a supertramp clade with high in-
terisland dispersability, but where there is no evidence of back-
colonization to the rainforest habitats of SE Asia. The single
colonization of the Pacific followed by wide dispersion within
Micronesia and Polynesia suggests that the Pacific Cyrtandra
clade, by gaining adaptations for dispersal as well as specialist
adaptations to island ecology, has become a supertramp clade
with only limited evolutionary potential to recolonize the trop-
ical rainforest from whence it came. To test the supertramp
hypothesis, and hence absence of back-colonization of rain-
forest (which the dispersability of Pacific Cyrtandra would
certainly make possible), extensive further sampling of Cyr-
tandra species along the eastern margin of the SE Asian plate,
particularly the Philippines, New Guinea, New Hebrides, and
the Bismark Archipelago would be necessary. This further
sampling would also help to refine the origins of the Pacific
clade.
LITERATURE CITED
ALICE, L. A., AND C. S. CAMPBELL. 1999. Phylogeny of Rubus (Rosaceae)
based on nuclear ribosomal DNA internal transcribed spacer region se-
quences. American Journal of Botany 86: 81?97.
June 2005] 1023CRONK ET AL.?PACIFIC CYRTANDRA
ATKINS, H., AND Q. C. B. CRONK. 2001. The genus Cyrtandra (Gesneriaceae)
in Palawan, Philippines. Edinburgh Journal of Botany 58(3): 443?458.
ATKINS, H. J., J. PRESTON, AND Q. C. B. CRONK. 2001. A molecular test of
Huxley?s line: Cyrtandra (Gesneriaceae) in Borneo and the Philippines.
Biological Journal of the Linnean Society 72: 143?159.
BALDWIN, B. G., D. W. KYHOS, J. DVORAK, AND G. D. CARR. 1991. Chlo-
roplast DNA evidence for a North American origin of the Hawaiian
silversword alliance (Asteraceae). Proceedings of the National Academy
of Sciences, USA 88: 1840?1843.
BEAUFORT-MURPHY, H. T. 1983. The seed surface morphology of the Ges-
neriaceae utilizing the scanning electron microscope and a new system
for diagnosing seed morphology. Selbyana 6: 220?422.
BRAMLEY, G. L. C., AND Q. C. B. CRONK. 2003. The Cyrtandra (Gesneri-
aceae) species of Mount Kerinci, Sumatra. Harvard Papers in Botany 7:
407?421.
BREMER, K. 1988. The limits of amino acid sequence data in angiosperm
phylogenetic reconstruction. Evolution 42: 795?803.
BURTT, B. L. 1970. Studies in the Gesneriaceae of the Old World. XXXIII.
Some species of Cyrtandra, chiefly Bornean. Notes from the Royal Bo-
tanic Garden Edinburgh 30: 23?42.
BURTT, B. L. 1978. Studies in the Gesneriaceae of the Old World. XLIV.
New and little-known species of Cyrtandra, chiefly from Sarawak. Notes
from the Royal Botanic Garden Edinburgh 36: 157?179.
BURTT, B. L. 1990. Gesneriaceae of the Old World, I. New and little-known
species of Cyrtandra from Malesia. Edinburgh Journal of Botany 47:
201?233.
BURTT, B. L., AND M. H. BOKHARI. 1973. Studies in the Gesneriaceae of the
Old World. XXXVI. Foliar sclereids in New Guinea and Pacific Cyrtan-
dra. Notes from the Royal Botanic Garden Edinburgh 32: 397?402.
CARLQUIST, S. 1970. Island biology. Columbia University Press, New York,
New York, USA.
CLAGUE, D. A. 1997. The growth and subsidence of the Hawaiian-Emperor
volcanic chain. In A. Keast and S. E. Miller [eds.], The origin and evo-
lution of Pacific Island biotas, New Guinea to Eastern Polynesia: patterns
and processes, 35?50. SPB Academic Publishing, Amsterdam, Nether-
lands.
CLARKE, C. B. 1883. Cyrtandreae. In A. de Candolle and C. de Candolle
[eds.], Monographia phanerogamarum, vol. 5, 1?303, 32 plates. G. Mas-
son, Paris, France.
COSTELLO, A., AND T. J. MOTLEY. 2001. Molecular systematics of Tetraple-
sandra, Munroidendron, and Reynoldsia sandwicensis (Araliaceae) and
the evolution of superior ovaries in Tetraplesandra. Edinburgh Journal
of Botany 58: 229?242.
DEJOODE, D. R., AND J. F. WENDEL. 1992. Genetic diversity and origin of
the Hawaiian Islands cotton, Gossypium tomentosum. American Journal
of Botany 79: 1311?1319.
DIAMOND, J. M. 1974. Colonization of exploded volcanic islands by birds:
the supertramp strategy. Science 184: 803?806.
DIAMOND, J. M. 1975. Assembly of species communities. In M. L. Cody and
J. M. Diamond [eds.], Ecology and evolution of communities, 342?444.
Harvard University Press, Cambridge, Massachusetts, USA.
DIAMOND, J. M. 1982. The biogeography of the Pacific Basin. Nature 298:
604?5.
DOYLE, J. J., AND J. L. DOYLE. 1987. A rapid DNA isolation procedure for
small quantities of fresh leaf tissue. Phytochemistry Bulletin 19: 11?15.
EMERSON, B. C. 2002. Evolution on oceanic islands: molecular phylogenetic
approaches to understanding pattern and process. Molecular Ecology 11:
951?966.
FARRIS, J. S. 1989. The retention index and homoplasy excess. Systematic
Zoology 38: 406?407.
FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an approach using
the bootstrap. Evolution 39: 783?791.
GANDERS, F. R., M. BERBEE, AND M. PIRSEYEDI. 2000. ITS base sequence
phylogeny in Bidens (Asteraceae): evidence for the continental relatives
of Hawaiian and Marquesan Bidens. Systematic Botany 25: 122?133.
GILLETT, G. W. 1967. The genus Cyrtandra in Fiji. Contributions of the U.S.
National Herbarium 37: 107?159.
GILLETT, G. W. 1970. The taxonomic status of Protocyrtandra (Gesneri-
aceae). Journal of the Arnold Arboretum 51: 241?246.
GILLETT, G. W. 1973. The genus Cyrtandra (Gesneriaceae) in the South Pa-
cific. University of California Publications in Botany 66: 1?59.
GILLETT, G. W. 1974. New records for Cyrtandra (Gesneriaceae) in the New
Hebrides. Kew Bulletin 29: 699?709.
GILLETT, G. W. 1975. Cyrtandra (Gesneriaceae) in the Bismarck Archipelago
and Solomon Islands. Kew Bulletin 30: 371?412.
GIVNISH, T. J., E. KNOX, T. B. PATTERSON, J. R. HAPEMAN, J. D. PALMER,
AND K. J. SYTSMA. 1996. The Hawaiian lobelioids are monophyletic
and underwent a rapid initial radiation roughly 15 million years ago.
American Journal of Botany 83: 159 (Abstract).
GRESSITT, J. L. 1982. Pacific-Asian biogeography with examples from the
Coleoptera. Entomologia Generalis 8: 1?11.
HOWARTH, D. G., D. E. GARDNER, AND C. W. MORDEN. 1997. Phylogeny
of Rubus subgenus Ideobatus (Rosaceae) and its implication toward col-
onization of the Hawaiian Islands. Systematic Botany 22: 433?441.
HOWARTH, D. G., M. H. G. GUSTAFSSON, D. A. BAUM, AND T. J. MOTLEY.
2003. Phylogenetics of the genus Scaevola (Goodeniaceae): implication
for dispersal patterns across the Pacific Basin and colonization of the
Hawaiian Islands. American Journal of Botany 90: 915?923.
HUELSENBECK, J. P., AND F. RONQUIST. 2001. MRBAYES: Bayesian infer-
ence of phylogeny. Bioinformatics 17: 754?755.
JARAMILLO, M. A., AND P. S. MANOS. 2001. Phylogeny and patterns of floral
diversity in the genus Piper (Piperaceae). American Journal of Botany
89: 1709?1724.
KIEHN, M. 2000. The genus Cyrtandra (Gesneriaceae): its indicative potential
for assessing historical and actual disturbance of island forest systems.
In Program Committee and Secretariat of IAVS 2000 [eds.], 43rd Sym-
posium of the International Association for Vegetation Science, 79 (Ab-
stract). Institute of Environmental Science and Technology, Yokohama
National University,Yokohama, Japan.
KIEHN, M. 2001. South Pacific and Hawaiian Cyrtandra: molecular and mi-
cromorphological studies. Malayan Nature Journal 55: 21?27.
KIEHN, M., M. PFOSSER, AND C. G. JANG. 2001. Implications of infra- and
interspecific AFLP variation in Hawaiian Cyrtandra. In M. Koch and K.-
G. Bernhardt [eds.], Evolution and plasticity in plant populations, 15.
Institute of Botany, University of Agricultural Sciences,Vienna, Austria.
KLUGE, A. G., AND J. S. FARRIS. 1969. Quantitative phyletics and the evo-
lution of anurans. Systematic Zoology 18: 1?32.
KNOX, E. B., S. R. DOWNIE, AND J. D. PALMER. 1993. Chloroplast genome
rearrangements and the evolution of giant lobelias from herbaceous an-
cestors. Molecular Biology and Evolution 10: 414?430.
KOKUBUGATA, G., AND D. A. MADULID. 2000. Comparative study of chro-
mosome number and karyotype of Cyrtandra cumingii var. yaeyamana
(Gesneriaceae) from Batan Island, the Philippines, and Iriomote Island,
Japan. Japanese Biological Magazine, Okinawa 38: 27?31.
LINDQVIST, C., AND V. A. ALBERT. 2002. Origin of the Hawaiian endemic
mints within North American Stachys (Lamiaceae). American Journal of
Botany 89: 1709?1724.
LUEGMAYR, E. 1993a. Pollen characters of Old World Gesneriaceae (Cyrtan-
droideae) with special reference to SE Asian taxa. Grana 32: 221?232.
LUEGMAYR, E. 1993b. Pollen of Hawaiian Cyrtandra (Gesneriaceae) includ-
ing notes on Southeast Asian taxa. Blumea 38: 25?38.
MAYER, V., M. MO? LLER, M. PERRET, AND A. WEBER. 2003. Phylogenetic
position and generic differentiation of Epithemateae (Gesneriaceae) in-
ferred from plastid DNA sequence data. American Journal of Botany 90:
321?329.
MCNAB, B. K. 2002. Minimizing energy expenditure facilitates vertebrate
persistence on oceanic islands. Ecology Letters 5: 693.
MERRILL, E. D. 1922. Gesneriaceae. New or noteworthy Philippines plants.
XVII. Philippine Journal of Science 20: 439?448.
MO? LLER, M., AND Q. C. B. CRONK. 1997. Origin and relationships of Saint-
paulia (Gesneriaceae) based on ribosomal DNA internal transcribed
spacer (ITS) sequences. American Journal of Botany 84: 956?965.
MO? LLER, M., AND M. KIEHN. 2003. Synopsis of cytological studies in Ges-
neriaceae. Edinburgh Journal of Botany 60: 425?447.
MU? HLBAUER, I., AND M. KIEHN. 1997. Seed morphology of Cyrtandra (Ges-
neriaceae). Scripta Botanica Belgica 15: 117.
NEPOKROEFF, M., K. J. SYTSMA, W. L. WAGNER, AND E. A. ZIMMER. 2003.
Reconstructing ancestral patterns of colonization and dispersal in the Ha-
waiian understory tree genus Psychotria (Rubiaceae): a comparison of
parsimony and likelihood approaches. Systematic Biology 52: 820?838.
POSADA, D., AND K. A. CRANDALL. 1998. MODELTEST: testing the model
of DNA substitution. Bioinformatics 14: 817?818.
PRICE, J. P., AND W. L. WAGNER. 2004. Speciation in Hawaiian angiosperm
lineages: cause, consequence, and mode. Evolution 58: 2185?2200.
SAKAI, A. K., W. L. WAGNER, D. M. FERGUSON, AND D. R. HERBST. 1995.
1024 [Vol. 92AMERICAN JOURNAL OF BOTANY
Biogeographical and ecological correlates of dioecy in the Hawaiian flo-
ra. Ecology 76: 2530?2543.
SAMUEL, R., W. PINSKER, AND M. KIEHN. 1997. Phylogeny of some species
of Cyrtandra (Gesneriaceae) inferred from the atpB/rbcL cpDNA inter-
gene region. Botanica Acta 110: 503?510.
SCHLAG-EDLER, B., AND M. KIEHN. 2001. Palynology of South Pacific Cyr-
tandra (Gesneriaceae) with notes on some Hawaiian taxa. Grana 40:
192?196.
SMITH, A. C. 1953. Studies of Pacific island plants. XIV. Notes on the Fijian
species of Cyrtandra. Journal of the Arnold Arboretum 34: 37?50.
SMITH, A. C. 1991. Gesneriaceae. In A. C. Smith [ed.], Flora vitiensis nova,
a new flora of Fiji (spermatophytes only), vol. 5, 88?115. National Trop-
ical Botanical Garden, Lawai, Hawaii, USA.
SMITH, J. F. 2000. Phylogenetic resolution within the tribe Episcieae (Ges-
neriaceae): congruence of ITS and ndhF sequences from parsimony and
maximum-likelihood analyses. American Journal of Botany 87: 883?897.
SMITH, J. F., C. C. BURKE, AND W. L. WAGNER. 1996. Interspecific hybrid-
ization in natural populations of Cyrtandra (Gesneriaceae) on the Ha-
waiian Islands: evidence from RAPD markers. Plant Systematics and
Evolution 200: 61?77.
SMITH, J. F., J. C. WOLFRAM, K. D. BROWN, C. L. CARROLL, AND D. S.
DENTON. 1997. Tribal relationships in the Gesneriaceae: evidence from
DNA sequences of the chloroplast gene ndhF. Annals of the Missouri
Botanical Garden 84: 50?66.
ST. JOHN, H. [ED.]. 1966. Monograph of Cyrtandra (Gesneriaceae) on Oahu,
Hawaiian Islands. Bernice P. Bishop Museum Bulletin 229: 1?465.
STEADMAN, D. W. 1997. The historic biogeography and community ecology
of Polynesian pigeons and doves. Journal of Biogeography 24: 737?753.
STOREY, W. B. 1966. Chromosome numbers in Cyrtandra. In H. St. John
[ed.], Monograph of Cyrtandra (Gesneriaceae) on Oahu, Hawaiian Is-
lands. Bernice P. Bishop Museum Bulletin 229: 31?33.
SWOFFORD, D. L. 2000. PAUP*: phylogenetic analysis using parsimony (*
and other methods), version 4. Sinauer Associates, Sunderland, Massa-
chusetts, USA.
WAGNER, W. L. 1990. Cyrtandra. In W. L. Wagner, D. R. Herbst, and S. H.
Sohmer [eds.], Manual of the flowering plants of Hawai?i, 735?781. Uni-
versity of Hawaii Press, Honolulu, Hawaii, USA.
WAGNER, W. L. 1999. Cyrtandra. In W. L. Wagner, D. R. Herbst, and S. H.
Sohmer [eds.], Manual of the flowering plants of Hawai?i, rev. ed., 735?
781. University of Hawaii Press, Honolulu, Hawaii, USA.
WAGNER, W. L., S. G. WELLER, AND A. K. SAKAI. 2004. Monograph of
Schiedea (Caryophyllaceae subfam. Alsinoideae). Systematic Botany
Monographs 72: 1?169.
WAGNER, W. L., K. R. WOOD, AND D. H. LORENCE. 2001. A new species
of Cyrtandra (Gesneriaceae) from Kaua?i, Hawaiian Islands. Novon 11:
146?152.
WILLIAMS, E. 1969. The ecology of colonization as seen in the zoogeography
of anoline lizards on small islands. Quarterly Review of Biology 44: 345?
389.
WILSON, E. O. 1959. Adaptive shift and dispersal in a tropical ant fauna.
Evolution 13: 122?144.
WRIGHT, S. D., C. G. YONG, J. W. DAWSON, D. J. WHITTAKER, AND R. C.
GARDNER. 2000. Riding the ice age El Nino? Pacific biogeography and
evolution of Metrosideros subg. Metrosideros (Myrtaceae) inferred from
nuclear ribosomal DNA. Proceedings of the National Academy of Sci-
ences, USA 97: 4118?4123.
WRIGHT, S. D., C. G. YONG, J. W. DAWSON, D. J. WHITTAKER, AND R. C.
GARDNER. 2001. Stepping stones to Hawaii: a trans-equatorial dispersal
pathway for Metrosideros (Myrtaceae) inferred from nrDNA (ITS plus
ETS). Journal of Biogeography 28: 769?774.
CRONK ET AL. ? American Journal of Botany 92(6): 1017?1024. 2005. ? Data Supplement ? Page 1 
 
CRONK, QUENTIN C. B., MICHAEL KIEHN, WARREN L. WAGNER, AND JAMES F. SMITH.  2005. Evolution of 
Cyrtandra (Gesneriaceae) in the Pacific Ocean: the origin of a supertramp clade. American Journal of Botany 
92(6): 1017?1024. 
 
APPENDIX. Species used in the phylogenetic analysis, voucher information, and GenBank numbers for ITS. 
 
Species Distribution Section Voucher GenBank number 
Aeschynanthus tricolor Malaysia: Sabah, 
Sarawak (Sabah)  
Aeschynanthus 19812968 (E) AF349170/AF349251 
Cyrtandra baileyi F. Muell. Australia (North 
Queensland) 
 Cronk T118 (E) AF349151/AF349232 
C. calyptribracteata Bakh. f. Java  Van Balgooy 6193 
(WU) 
AY818830/AY818865 
C. compressa C.B. Clarke Samoa  Kiehn et al. 
940823-2/1 (WU) 
AY818840/ AY818875 
C. cordifolia Gaud. Hawai?i, Oahu, Manoa Crotonocalyces Smith 2848 (BISH) AY818854/ AY818889 
C. cumingii C.B. Clarke Philippines  Mendum 29034 (E) AY818833/ AY818868 
C. cupulata Ridley Malaysia  Weber s.n. (WU) AY818826/ AY818861 
C. falcifolia C.B. Clarke Samoa  Kiehn et al. 
940823-3/1 (WU) 
AY818841/ AY818876 
C. ferruginea Merrill Philippines  Mendum 29182 (E) AY818834/ AY818869 
C. grandiflora Gaud. Hawai?i, Oahu, Manoa Verticillatae Smith 2891 (BISH) AY818844/ AY818879 
C. grayana Hillebr. Hawai?i Macrosepalae Wagner 6472 (US) AY818851/ AY818886 
C. hawaiiensis C.B. Clarke Hawai?i, Oahu, 
Waiahole 
Verticillatae Smith 2853 (BISH) AY818855/ AY818890 
C. hirtigera H.J. Atkins and 
Cronk 
Philippines  Cronk et al. 25433 
(E) 
AY818832/ AY818867 
C. induta A. Gray Tahiti  Cronk T97 (E) AY818858/ AY818893 
C. kauaiensis Wawra Hawai?i, Kaua?i Macrosepalae Kiehn 920825-3/4 
(WU) 
AY818847/ AY818882 
C. kusaimontana Hosokawa Kosrae  T. Flynn et al. 5995 
(PTBG) 
AY818857/ AY818892 
C. laxiflora H. Mann Hawai?i, Oahu, 
Waiahole 
Apertae Smith 2858 (BISH) AY818845/ AY818880 
C. longifolia (Wawra) 
Hillebr. ex C.B. Clarke 
Hawai?i, Kaua?i Cylindrocalyces Kiehn 920825-2/1 
(WU) 
AY818846/ AY818881 
C. macraei A. Gray Hawai?i, Oahu, 
Wailupe 
Chaetocalyces Smith 2902 (BISH) AY818853/ AY818888 
C. menziesii Hook. & Arnott Hawai?i Chaetocalyces Wagner 6757 (US) AY818852/ AY818887 
C. cf. milnei Seemann ex A. 
Gray 
Fiji  Vodonaivalu et al. 
940807-1/1 (WU) 
AY818837/ AY818872 
C. monticola Gillespie New Guinea  Takeuchi 6002 (E) AY818860/ AY818895 
CRONK ET AL. ? American Journal of Botany 92(6): 1017?1024. 2005. ? Data Supplement ? Page 2 
 
C. paliku W.L. Wagner, K.R. 
Wood & D.H. Lorence 
Hawai?i, Kaua?i, Mt. 
Namahana 
 Wood 7307 (PTBG) AY818842/ AY818877 
C. paludosa Gaud. Hawai?i, Oahu, Palolo Cylindrocalyces Smith 2884 (BISH) AY818848/ AY818883 
C. patula Ridley Malaysia  Weber s.n. (WU) AY818829/ AY818864 
C. pendula Blume Malaysia  Weber & 
Antonysamy 
860730-1/2 (WU) 
AY818828/ AY818863 
C. propinqua C. Forbes Hawai?i, Oahu, 
Waiahole 
Crotonocalyces Smith 2870 (BISH) AY818849/ AY818884 
C. samoensis A. Gray Tonga  Wood 6873 (PTBG) AY818839/ AY818874 
C. sandwicensis (H. L?v.) St. 
John & Storey 
Hawai?i, Oahu, Manoa Apertae Smith 2850 (BISH) AY818850/ AY818885 
C. splendens C.B. Clarke Sarawak  cult. in Univ. of 
Vienna Botanical 
Garden GS-89-04 
(WU) 
AY818827/ AY818862 
C. tahuatensis F.R. Fosberg 
& M.-H. Sachet 
Marquesas  Wood 6563 (PTBG) AY818838/ AY818873 
C. tohiveaensis G.W. Gillett Moorea  Cronk T28 (E) AY818859/ AY818894 
C. umbellifera Merrill Taiwan  Wagner 6701 (US) AY818836/ AY818871 
C. urvillei C.B. Clarke Kosrae  D. Lorence et al. 
7838 (PTBG) 
AY818856/ AY818891 
C. villosissima Merrill Philippines  Mendum 29053 (E) AY818835/ AY818870 
C. wallichii (C.B. Clarke) 
Burtt 
Malaysia  Weber s.n. (WU) AY818831/ AY818866 
C. wawrae C.B. Clarke Hawai?i, Kaua?i Crotonocalyces Kiehn 920901-1/1 AY818843/ AY818878