Content uploaded by Sara B. Hoot
Author content
All content in this area was uploaded by Sara B. Hoot
Content may be subject to copyright.
1409
Int. J. Plant Sci. 162(6):1409–1418. 2001.
䉷2001 by The University of Chicago. All rights reserved.
1058-5893/2001/16206-0022$03.00
PHYLOGENETIC RELATIONSHIPS OF CAPSICUM (SOLANACEAE) USING DNA SEQUENCES
FROM TWO NONCODING REGIONS: THE CHLOROPLAST atpB-rbcL
SPACER REGION AND NUCLEAR waxy INTRONS
Brian M. Walsh and Sara B. Hoot
1
Department of Biological Sciences, University of Wisconsin—Milwaukee, Milwaukee, Wisconsin 53201, U.S.A.
This study focuses on three phylogenetic problems related to Capsicum (Solanaceae): (1) the monophyly of
the genus, (2) species delimitation within the genus, and (3) phylogenetic relationships of species within
Capsicum. The chloroplast atpB-rbcL noncoding spacer region was used to derive a phylogeny for seven
outgroup genera and 11 species of Capsicum. Data derived from five introns within the nuclear gene waxy
were used, both separately and in combination with the atpB-rbcL spacer data, to resolve further questions
of species delimitation and phylogenetic relationships within Capsicum.Capsicum is monophyletic, with
moderate support. Capsicum ciliatum, which is both molecularly and morphologically distinctive, is sister to
a highly supported clade consisting of all other Capsicum species studied. Capsicum cardenasii and C. eximium
are sister species and are, in turn, sisters to a moderately supported clade consisting of C. tovarii,C. pubescens,
C. chacoense,C. baccatum,C. galapagoense,C. chinense,C. frutescens, and C. annuum.Capsicum galapa-
goense, whose taxonomic affinities have been largely unstudied, is included in a weakly supported clade
consisting of C. annuum,C. chinensis, and C. frutescens. Many species of Capsicum have sufficient molecular
markers in the waxy data set (both nucleotide substitutions and insertions/deletions) to be useful in species
delimitation. An informal classification of the genus is proposed.
Keywords: Capsicum,atpB-rbcL spacer, waxy, chilies, peppers, Solanaceae, phylogeny, species delimitation.
Introduction
Capsicum (chilies and other peppers) consists of annual or
perennial herbs or shrubs native to South and Central America
and the Gala´pagos. Because humans have been affecting dis-
persal since prehistoric times, the original geographic distri-
bution of Capsicum is difficult to determine. Of the 20–27
currently recognized species within the genus that appear to
be native to Central and South America, ca. 17 have ranges
overlapping in Bolivia. In the past 50 yr, several Capsicum
species have been identified that were previously unknown to
botanists, including C. tovarii (Eshbaugh et al. 1983), C. car-
denasii,C. praetermissum, and C. galapagoense (Heiser and
Smith 1958). Because of limited study, the affinities of C. gala-
pagoense (Gala´pagos Islands) is of particular biogeographic
and morphological interest.
Capsicum exhibits considerable morphological variation, es-
pecially in fruit shape, color, and size. Pubescence of leaves
and stems range from glabrous to very pubescent. Inflores-
cences vary from solitary to seven flowers at one node. The
calyx may range from long, green sepals to truncate sepals to
spinelike projections. The corolla is rotate or infrequently cam-
panulate, with highly variable coloration between and among
species. Seeds are cream colored, except for C. pubescens,
which has black seeds. Capsicum species, with few exceptions,
are diploid (2np24, infrequently 2np26) and have similar
1
E-mail hoot@uvm.edu.
Manuscript received April 2001; revised manuscript received July 2001.
karyotypes (Lippert et al. 1966; Moscone et al. 1993). Many
species have overlapping morphological character states, po-
tentially leading to unresolved or erroneous species identifi-
cation. A combination of diagnostic characters is usually re-
quired to identify and differentiate Capsicum species.
Archaeological evidence from Mexico indicates that humans
have been using wild chili peppers as a food source possibly
as early as 7200
B
.
C
. (Pickersgill 1966; Heiser 1969). The
oldest evidence of domesticated chilies was found in a cave in
Tehuacan Valley (south-central Mexico) and dates to
5000–6500
B
.
C
. (Davenport 1970), which establishes chilies
as one of the earliest domesticated plants in the New World.
There are four ancient agricultural centers in the New World,
three of which are believed to have domesticated the chili pep-
per independently (Pickersgill 1969, 1977). After several thou-
sand years of domestication, the varieties of chilies, along with
other crops and technologies, were traded between the agri-
cultural centers and dispersed over half of North and South
America. Trading and migration rapidly expanded the ranges
of many Capsicum species into small, fragmented populations
scattered over vast regions, increasing the potential for inter-
breeding between domesticated and wild populations. While
interbreeding is quite common in laboratory situations, it does
not appear to occur frequently in the wild, possibly due to a
strong tendency toward self-pollination in domesticates (Esh-
baugh 1970, 1976).
The first known Europeans to come in contact with chilies
were the crew of Columbus’s initial transatlantic voyage to
the New World. Peter Martyr, a historian who accompanied
Columbus on his voyages, wrote in 1493 that the New World
1410 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Fig. 1 Capsicum dendrogram constructed from standard genetic
distance estimates based on isozyme data compared with a classifi-
cation based on flower color (from McLeod et al. 1982).
has a pepper more pungent than the black and white pepper
from Asia (Piper nigrum; Lippert et al. 1966). The association
between the spicy bite of black pepper and chilies is how the
name “pepper” was inappropriately linked with Capsicum.
Columbus returned to the Old World with several pungent
forms of Capsicum, most of which were members of the species
C. annuum. In Europe, the chili was enthusiastically and rap-
idly incorporated into many cultures. Within 50 yr, chilies
spread from Spain to England (Lippert et al. 1966) and as far
west as India (Deb 1979).
From the time chilies arrived in Europe up to the mid-1900s,
taxonomists have disagreed on the criteria delimiting Capsi-
cum species and varieties. Often the characters used in these
studies were the same morphological features manipulated by
domestication for 3500–7000 yr. In these studies, workers
compared morphological differences between Capsicum vari-
eties and deduced common ancestry based on such shared fea-
tures as fruit shape, color, position, and pungency. These stud-
ies served only to obscure evolutionary relationships. Some
early botanists recognized up to 100 species of Capsicum,
while others recognized only a few (Eshbaugh 1980).
The morphological differences between wild and cultivated
chilies are easily discerned. All wild forms of chilies have small,
red, berry-like fruits with colors and sizes attractive to birds.
Wild chilies have deciduous fruits, which, if not eaten by birds,
fall to the ground while the seeds are still at peak viability.
Domesticated forms exhibit variable fruit and flower colora-
tion (designed to appeal to the human eye); gigantism of the
fruits, seeds, flowers, and leaves (Cochran 1940; Eshbaugh
1976); and retention of the fruit on the peduncle at maturity
(Pickersgill 1969; Eshbaugh 1976). When early taxonomists
compared various Capsicum taxa, they noted that chilies
sorted into two distinct groups: one typified by small, red fruits
and the other by large fruits. This classification effectively sep-
arated the wild and domesticated forms of Capsicum but bore
no relevance to evolutionary relationships.
Capsaicin, a volatile phenolic amine, is a very stable mol-
ecule and is responsible for the pungency commonly associated
with chili peppers (Heiser 1969). When some chilies, such as
the haban˜ero, are ground into a powder, capsaicin can be de-
tected by taste at dilutions up to 1 ppm. Presence of capsaicin
was once thought to be an identifying characteristic found in
all species within the genus (excluding only nonpungent, do-
mesticated varieties). However, C. ciliatum is never pungent
(Eshbaugh 1980), and several wild nonpungent forms of C.
chacoense have been found. However, C. anomalum, despite
its pungent fruit (determined by taste test by B. M. Walsh),
has been removed from Capsicum to the monotypic genus
Tubocapsicum, which is relatively distantly related to Capsi-
cum (Olmstead et al. 1999; this study).
Enzymatic studies of Capsicum (Jensen et al. 1979; McLeod
et al. 1979a, 1979b, 1982, 1983) have demonstrated that spe-
cies could be grouped into taxonomic categories that some-
what agreed with groupings based on flower color (fig. 1).
This system of classification is useful for separating some Cap-
sicum species into subgeneric categories. However, less than
half of the commonly recognized species of Capsicum were
included in these studies. In addition, some of the excluded
species do not fit into this categorization, such as the yellow
flowers of C. ciliatum and C. scolnikianum and the white flow-
ers of C. chacoense, which seem to be more closely related to
the purple-flowered group than to the white-flowered group
(McLeod et al. 1982).
During the past 40 yr, hybrid analyses have been used ex-
tensively to resolve species relationships in Capsicum (Heiser
and Smith 1948, 1953, 1958; Smith and Heiser 1951, 1957;
Emboden 1961; Lippert et al. 1966; Eshbaugh 1970, 1976;
Pickersgill 1971; Eshbaugh et al. 1983). To determine the vi-
ability of hybrids between various species of Capsicum, pollen
staining and F1 seed germination studies were used. The results
of these hybrid analyses are helpful in grouping closely related
species into subgeneric categories but have limited usefulness
in determining evolutionary relationships (fig. 2).
Numerical comparisons of morphological traits (Cochran
1940; Eshbaugh 1970; Jensen et al. 1979; Pickersgill et al.
1979) and cytogenetic analyses (Shopova 1966; Ballard et al.
1970; McLeod et al. 1979a, 1979b, 1982; Moscone et al.
1993) have been used to resolve relationships. The numerical
analyses typically included a limited number of species and
focused primarily on the relationships of cultivated varieties
to their wild progenitors. Cytogenetic analyses allowed greater
resolution of the relationships between species and varieties
but have achieved limited taxonomic resolution between
closely related species, such as the C. annuum/frutescens/chi-
nense and the C. cardenasii/eximium complexes. All of these
studies correlate well with the hybrid analyses.
Species delimitation within two Capsicum species complexes
remain problematic: (1) the C. annuum complex, consisting
of C. annuum,C. frutescens, and C. chinense, and (2) the C.
eximium complex, consisting of C. eximium and C. cardenasii.
Species of the C. annuum complex contain both domesticated
WALSH & HOOT—PHYLOGENETIC RELATIONSHIPS OF CAPSICUM 1411
Fig. 2 Summary of Capsicum–hybrid crossing studies with associated citations indicated by the letters below
and wild varieties, as well as a wide range of intermediates,
which are all similar morphologically and indistinguishable
based on enzyme profiles (Jensen et al. 1979). Some researchers
have argued that C. frutescens and C. chinense should be com-
bined into one species (Pickersgill 1966, 1971; McLeod et al.
1979b) because they interbreed fairly readily (Smith and Heiser
1957; Lippert et al. 1966; Pickersgill 1966) and intergraded
into a morphological continuum. Capsicum frutescens displays
features considered typical of a wild species and is not culti-
vated on a large scale, except relatively recently on the Tabasco
farms of Louisiana (Pickersgill 1971). Capsicum chinense does
not have any true wild form and is cultivated extensively in
South America. Several characteristics, such as nondehiscent
fruit, fruit shape, and gigantism of leaves, fruit, and flower
structure, suggest it has been cultivated for a long time (Pick-
ersgill 1966, 1971).
The C. eximium complex consists of C. eximium and C.
cardenasii, which are morphologically quite distinct. Capsicum
eximium produces the rotate flowers typical of Capsicum,
while C. cardenasii produces vaselike, campanulate flowers.
Furthermore, C. cardenasii is the only species in the genus that
obligately outbreeds (McLeod et al. 1979a). However, the
ranges of these species overlap, and they appear to form nat-
ural hybrids (McLeod et al. 1979a). Hybrid studies indicate a
high level of fertility, with 90%–100% pollen stainability (fig.
2; Lippert et al. 1966; Eshbaugh 1976). Hybrids of C. eximium
and C. cardenasii are more fertile than some crosses between
varieties within a species (Eshbaugh 1976). In addition, based
on allozyme data, these two species are indistinguishable from
each other (Jensen et al. 1979; McLeod et al. 1979a). It has
been suggested that C. eximium and C. cardenasii be consol-
idated into a single, morphologically variable species (Ballard
et al. 1970; Eshbaugh 1976; Jensen et al. 1979; McLeod et al.
1979a).
This study focused on three phylogenetic problems associ-
ated with Capsicum: (1) monophyly of the genus Capsicum,
(2) species delimitation, and (3) phylogenetic relationships of
the species within Capsicum. Recent work by Olmstead and
Palmer (1997), Bohs and Olmstead (1997), and Olmstead et
al. (1999) using both chloroplast sequences and restriction site
data indicates that the genus Capsicum is derived from Ly-
cianthes, making Lycianthes paraphyletic. Because of this close
1412 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Table 1
List of Species, Source of Plant Material or DNA, Voucher Information, and GenBank Numbers (atpB-rbcL spacer/waxy)
Species Source/type of material Voucher information GenBank number
Aureliana fasciculata (Sendt.) Barb. & A. Hunz. R. Olmstead/DNA K. Brown s.n., UEC AF397083
Capsicum annuum:
var. annuum* L. USDA-ARS/seeds B. Walsh 1, UWM AF397108/AF397129
var. annuum* USDA-ARS/seeds B. Walsh 24, UWM AF397110/AF397131
var. annuum cv. Early CalWonder* Green Valley Seed/seeds B. Walsh 14, UWM AF397109/AF397130
var. aviculare (Dierb) D’Arcy & Eshbaugh USDA-ARS/seeds B. Walsh 5, UWM AF397106/AF397127
var. aviculare CATIE 8191 B. Walsh 12, UWM AF397107/AF397128
C. baccatum:
var. baccatum L. USDA-ARS/seeds B. Walsh 6, UWM AF397100/AF397120
var. pendulum* (Willd.) Eshbaugh USDA-ARS/seeds B. Walsh 9, UWM AF397101/AF397121
C. cardenasii Heiser & Smith USDA-ARS/seeds B. Walsh 26, UWM AF397095/AF397116
C. chacoense Hunz. USDA-ARS/seeds B. Walsh 7, UWM AF397099/AF397122
C. chinense Jacq. USDA-ARS/seeds B. Walsh 3, UWM AF397102/AF397123
C. ciliatum (H.B.K.) O. Kuntze R. Olmstead/DNA C. Heiser 7518, IND AF397094/AF397115
C. eximium Hunz. B. Pickersgill/seeds B. Walsh 35, UWM AF397096/AF397117
C. frutescens L. USDA-ARS/seeds B. Walsh 20, UWM AF397104/AF397124
cv. Tabasco* Shepherds Garden Seeds/seeds B. Walsh 15, UWM AF397105/AF397125
C. galapagoense Hunz. USDA-ARS/seeds B. Walsh 18, UWM AF397103/AF397126
C. pubescens* Ruiz. & Pav. USDA-ARS/seeds B. Walsh 17, UWM AF397098/AF397119
C. tovarii Eshbaugh, Smith & Nickrent B. Pickersgill/seeds B. Walsh 34, UWM AF397097/AF397118
Datura stramonium L. S. Hoot/leaves B. Walsh 29, UWM AF397076
Jaltomata auriculata (Miers) Mione R. Olmstead/DNA BIRM S1596/76 AF397081
Lycianthes ciliolata (Mart. & Gal.) Bitter R. Olmstead/DNA BIRM S0607/70 AF397085
L. cuchumatanensis J. L. Gentry R. Olmstead/DNA R. Olmstead 94-06, WTU AF397092
L. glandulosa Bitter R. Olmstead/DNA BIRM S1616/75 AF397089/AF397114
L. heteroclita (Sendtn.) Bitter L. Bohs/DNA Bohs 2376, UT AF397091/AF397113
L. lenta Bitter R. Olmstead/DNA R. Olmstead 96-92, WTU AF397093
L. lycioides Hassl. R. Olmstead/DNA R. Olmstead S-87, WTU AF397087/AF397111
L. magdalenae Bitter R. Olmstead/DNA Det. D. Symon s.n. AF397090
L. rantonnei (Carrie`re) & Bitter R. Olmstead/DNA R. Olmstead S-96, WTU AF397086/AF397112
Solanum aviculare Forst. f. USDA-ARS/seeds B. Walsh 33, UWM AF397077
S. lycopersicum L. R. Olmstead/DNA No voucher
a
AF397080
S. pimpinellifolium (L.) P. Miller S. Hoot/leaves B. Walsh 13, UWM AF397079
S. pseudocapsicum L. USDA-ARS/seeds B. Walsh 32, UWM AF397078
S. shanesii F. Muell. (pLycianthes sp.)
b
R. Olmstead/DNA Clarkson 6674, AD AF397088
Tubocapsicum anomalum (Franchet & Savat.) Makino USDA-ARS/seeds B. Walsh 27, UWM AF397082
Withania coagulans (Stocks) Dun. R. Olmstead/DNA BIRM S0678/69 AF397084
Note. An asterisk denotes species and varieties of Capsicum commonly found in cultivation. BIRM pUniversity of Birmingham Solanaceae seed collection;
CATIE pCentro Agrono´ mico de Investigacio´n y Ensen˜ anza, Costa Rica.
a
Same DNA accession used in Olmstead and Palmer (1992, 1997) and Bohs and Olmstead (1997).
b
Solanum shanesii is now considered a Lycianthes species, but a formal recombination has not yet been made (L. Bohs, personal communication).
relationship between Capsicum and Lycianthes and the limited
sampling of both genera in the above studies, a goal of this
research was to verify the monophyly of Capsicum and, with
broader sampling (18 species and eight genera), determine the
placement of Capsicum within Solanaceae. To accomplish this
goal, the noncoding chloroplast DNA region between atpB
and rbcL was sequenced. This spacer region is ca. 800 bp long
in Capsicum and is suitable for taxonomic studies at the ge-
neric and family level (Golenberg et al. 1993; Savolainen et
al. 1994; Manen and Natali 1995; Natali et al. 1995; Hoot
and Douglas 1998).
To test species delimitations and phylogenetic relationships
within Capsicum, sequence data from both the chloroplast
atpB-rbcL spacer region and a 1200-bp segment from the nu-
clear gene waxy, encoding an essential enzyme in granule-
bound starch synthesis (GBSS), were used. The waxy gene
contains 12 introns and is ca. 3 kb long in Solanum tuberosum
(van der Leij et al. 1991). The waxy region used in this study
includes introns 2–6 and is ca. 900 bases long. The waxy gene
has had limited use for phylogenetic work to date but is be-
coming increasingly popular. Unlike the ribosomal internal
transcribed spacer (ITS) regions, which contain at least two,
nonidentical paralogues in Capsicum,waxy appears to be sin-
gle copy (van der Leij et al. 1991; Miller et al. 1999) and is
most useful at the generic level (Mason-Gamer et al. 1998;
Miller et al. 1999).
Material and Methods
Eleven of the 27 most commonly recognized species of Cap-
sicum (Eshbaugh 1980), several varieties of some species, and
seven outgroup genera were sampled in this study (table 1).
WALSH & HOOT—PHYLOGENETIC RELATIONSHIPS OF CAPSICUM 1413
Included in the sampling are the five domesticated and some
of the wild progenitor species (table 1). Because many species
are difficult to attain and little systematic work has been done
on the genus as a whole, it is difficult to know how represen-
tative this sampling is of the overall variation found in Cap-
sicum. All taxa in this study were positively identified (see
tables 2 and 3 in Walsh 1999 for a list of diagnostic characters
used in species verification and botanical descriptions of Cap-
sicum species). Sequencing, accession, voucher, and GenBank
information are included in table 1. All seeds were germinated
and grown in the greenhouses of the Department of Biological
Sciences, University of Wisconsin—Milwaukee.
Total cellular DNA was isolated from fresh leaf material
according to the miniprep method of Doyle and Doyle (1987).
In most cases, DNA was further purified and concentrated after
extraction using Wizard Column PCR Preps (Promega) or eth-
anol precipitation (Sambrook et al. 1989).
For amplification (PCR) of the atpB-rbcL spacer, four dif-
ferent 25-mer amplification primers were used (Hoot et al.
1995). The primer rbcL1 was used in all reactions. This primer
complements the 5
end of the rbcL gene but with opposite
orientation, allowing amplification through the spacer region
toward the atpB gene. The primer S385R within atpB was
used in conjunction with rbcL1 for all samples, except Ly-
cianthes cuchumatanensis and L. lenta, which were amplified
with the primers S2R and S766R. The protocol for the am-
plification of the spacer region used a reaction mixture con-
taining the final concentrations: 10 mM Tris-HCl, pH 8.3, 50
mM KCl, 1.5 mM MgCl2, 0.2 mM of each dNTP, 0.5 mMof
each amplification primer, 2.5 U Taq polymerase, and 0.3–2.0
mL template DNA per 100 mL reaction (depending on concen-
tration). The sample was then amplified using the cycling pa-
rameters described in Hoot et al. (1995).
The regions of waxy were amplified using primers waxy 5
and waxy 3
based on the van der Leij et al. (1991) Solanum
tuberosum sequence (kindly provided by D. Spooner) or two
20-mer primers located ca. 100 bases in from the two previous
primers (primer sequences available by request from S. Hoot).
Samples were amplified using ready-to-go PCR Beads (Phar-
macia Biotech), adding 0.3 mM of each primer, 0.5 mLof
template DNA, and 30 mL of mineral oil for a 25-mL reaction.
The thermocycler was programmed with the following param-
eters: premelt at 94⬚C for 3 min, 40 cycles each consisting of
a denaturation step at 94⬚C for 30 s, annealing step at 45⬚C
for 30 s, and an extension step at 72⬚C for 2 min, followed
by a final extension step of 72⬚C for 5 min. In some cases (13
of the 21 samples), 2% low-melting agarose gel plugs of the
PCR product were diluted with 75–125 mLofddH
2
O, de-
pending on the size of the plug, and used as a template for
further PCR.
One sample, Capsicum annuum var. annuum cv. Early
CalWonder, which did not amplify for waxy in large enough
quantities for sequencing with the protocol described above,
was cloned from the PCR product using a T-overhang vector
kit (T-Easy, Promega). The vector was used to transform Pro-
mega ultracompetent cells. DNA-Pure Plasmid Mini-Prep Kit
(CPG) was used to isolate, purify, and concentrate the vector
DNA from cloudy cultures for use in automated sequencing.
PCR products were purified for automated sequencing by
electrophoresis on a 2% low-melting agarose gel (Fisher Bio-
tech) with 1% TAE buffer. Ethidium-bromide-stained bands
were visualized over UV illumination and then removed as gel
plugs. Either Wizard PCR Preps (Promega) or QIAquick PCR
Purification columns (Qiagen) were used to remove agarose
and concentrate the PCR product.
Both 5
and 3
strands of DNA were sequenced for the atpB-
rbcL spacer and waxy with 100% overlap for the spacer region
and at least 80% for waxy. In the case of waxy, the sequences
amplified using waxy 5
and waxy 3
were pruned to the same
length as those using the more internal amplification primers.
In spite of the numerous indels, alignments could be accom-
plished to a rough approximation using Sequencher 3.0 (Gene
Codes Corporation) with subsequent manual corrections.
The following alignment criteria and methodology were
used: (1) Alignments maximized two elements—matching nu-
cleotides at a sequence position and consideration of gap or
indel type. Recognition of two types of indels (type Ia indels,
runs of the same nucleotide of any length, and type Ib indels,
regions of two or more base pairs with more complex repeated
nucleotide motifs) are often helpful in assessing gap homology
and reliability (Golenberg et al. 1993; Hoot and Douglas
1998). (2) Gaps were scored using simple indel coding (Sim-
mons and Ochoterena 2000). (3) Phylogenetically informative
indels (variable in two or more taxa) were scored as one event
at the end of the data set. (4) Regions of the alignment that
consisted of gaps in at least 50% of the taxa were removed
from the data set before analysis (these regions were almost
universally phylogenetically uninformative).
Phylogenetic analyses were performed using PAUP* 4.0b4a
(Swofford 1998) with the branch-and-bound search option.
PAUP* was also used to perform bootstrap analyses with2000
replications using the branch-and-bound search option (Fel-
senstein 1985). Before combining the data sets, several meth-
ods of assessing congruence among the two data sets were
implemented: visual comparison of the various clades found
in the minimal trees, their bootstrap support, and implemen-
tation of the incongruence length difference test (Farris et al.
1995; implemented in PAUP*), which tests whether the pre-
defined partitions in the data differ significantly from random
partitions of the combined data set. The analysis was con-
ducted with 1000 replications, heuristic search with simple
addition, TBR (tree bisection/reconnection) branch swapping,
and saving up to 2000 trees for each replicate.
Outgroup taxa for both analyses were selected based on the
results of several previous phylogenetic analyses of the Sola-
naceae (Olmstead and Palmer 1992, 1997; Bohs and Olmstead
1997; Olmstead et al. 1999). The studies cited above indicate
that, of the outgroup taxa sampled in this study, Datura oc-
cupies the most basal position within the Solanaceae. For this
reason, Datura was selected as the outgroup in the analysis at
the family level. We also rooted the family-level analysis with
the immediate sister group (Tubocapsicum/Areliana/Wither-
ingia)toCapsicum to test for changes in tree topology. Several
studies have shown Lycianthes to be paraphyletic (Olmstead
and Palmer 1992, 1997; Bohs and Olmstead 1997; Olmstead
et al. 1999) or as a monophyletic group sister to Capsicum
(this study). To be consistent with the results from our family-
level analysis, we assumed monophyly of both Lycianthes and
Capsicum and rooted the generic-level tree accordingly.
1414 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Fig. 3 One of the 18 shortest trees based on atpB-rbcL spacer data
for Capsicum and outgroups. Numerals above lines are number of
substitutions supporting branches; numerals below lines are bootstrap
values. Dotted lines indicate where branches collapse in strict consen-
sus tree.
Fig. 4 The shortest tree resulting from the combined atpB-rbcL
spacer and waxy data. Numerals are as in fig. 3; bacc.pvar. baccatum,
pend.pvar. pendulum,Tob .pcv. Tobasco, avic.pvar. aviculare,
ann.pvar. annuum, and CW pcv. Early CalWonder.
Results
The family-level analysis of the atpB-rbcL spacer data re-
sulted in 18 equally parsimonious trees based on 67 variable
sites and a consistency index excluding autapomorphies
and a retention index . One of the 18(CI) p0.77 (RI) p0.90
shortest trees is presented in figure 3. Tree topology within
Capsicum and Lycianthes remained the same whether rooted
with Datura or the more immediate sister clade of Tubocap-
sicum,Aureliana, and Withania. While the tree lacks support
for many of the more basal nodes, several more derived clades
are well supported. The genera Tubocapsicum,Aureliana, and
Withania form a well-supported ( %) trichot-bootstrap p100
omy. Lycianthes and Capsicum together form a highly sup-
ported clade ( %; fig. 3). Within Lycianthes,bootstrap p100
two clades are recognized: a strongly supported clade
( %), consisting of Solanum (pLycianthes)bootstrap p96
shanesii,L. glandulosa,L. magdalenae,L. heteroclita,L. cu-
chumatanensis, and L. lenta, and a smaller, weakly supported
clade ( %), consisting of L. rantonnei,L. ly-bootstrap p67
cioides, and L. ciliolata. All species of Capsicum form a largely
unresolved monophyletic group weakly supported with one
base substitution ( %) with C. ciliatum as sisterbootstrap p66
to all remaining Capsicum. The remaining Capsicum species
(excluding C. ciliatum) form a strongly supported mono-
phyletic group ( %). Several clades are formedbootstrap p96
within this core Capsicum group, which correspond to
known species complexes: the C. annuum/chinense/frutescens
complex, now also including C. galapagoense (bootstrap p
%); the well-supported C. eximium/cardenasii complex61
( %); and the C. baccatum/chacoense groupbootstrap p89
(%).bootstrap p69
The generic-level analysis of four species of Lycianthes and
17 Capsicum taxa using waxy data consisted of 200 variable
characters (including 20 gaps) and 113 parsimony informative
characters. These data resulted in one most parsimonious tree
( , ). The atpB-rbcL spacer data (60 var-CI p0.87 RI p0.94
iable characters including gaps, 20 informative characters) for
the same taxa also resulted in one shortest tree ( ,CI p0.87
). The tree resulting from the atpB-rbcL spacer dataRI p0.93
was similar to the waxy tree but with considerably less reso-
lution. Both visual inspection of the tree topologies and the
partition homogeneity test indicated that the two data sets
were highly congruent (P). For this reason, onlyvalue p1.00
the tree resulting from the combined atpB-rbcL spacer and
waxy data is presented here (fig. 4).
WALSH & HOOT—PHYLOGENETIC RELATIONSHIPS OF CAPSICUM 1415
Table 2
Molecular Markers
Taxon
atpB-rbcL spacer waxy
Substitutions Indels Substitutions Indels
Capsicum annuum (5) 0 0 2 1
C. baccatum (2) 0 0 4 0
C. cardenasii 10 10
C. chacoense 00 41
C. chinense 00 10
C. ciliatum 21204
C. eximium 10 81
C. frutescens (2) 0 0 0 1
C. galapagoense 10 10
C. pubescens 20 30
C. tovarii 20100
Note. Unique substitutions and insertion/deletions not found in
any other Capsicum or Lycianthes taxa, potentially useful in species
delimitation within Capsicum. Numerals in parentheses indicate num-
ber of taxa sequenced within a species.
The combination of the atpB-rbcL spacer and waxy data
(including gaps) resulted in one most parsimonious tree (fig.
4) derived from 133 informative characters ( ,CI p0.87
). Reanalysis excluding gap data resulted in two treesRI p0.94
with identical topology except for the collapse of the mono-
phyletic C. frutescens clade (trees not presented). The division
of Capsicum and Lycianthes is moderately well supported,
with 11 nucleotide changes and 74% bootstrap. The mono-
phyly of the core Capsicum group (excluding C. ciliatum)is
extremely well supported, with 48 characters and 100% boot-
strap. Within the core Capsicum, the following received mod-
erate to strong bootstrap support: a clade consisting of all
Capsicum, excluding C. cilatum,C. cardenasii, and C. exi-
mium (73%); a C. cardenasii and C. eximium clade (97%);
and clades consisting of two varieties of C. baccatum (98%),
two taxa of C. frutescens (76%), and five taxa of C. annuum
(98%). The C. annuum complex (McLeod et al. 1982), con-
sisting of C. annuum,C. chinense,C. frutescens, and C. ga-
lapagoense, form a weakly supported clade (bootstrap p
).58%
Both waxy and spacer data provided markers (nucleotide
substitutions and indels) at the species level, which may be
useful in species delimitation (table 2).
Discussion
Family-Level Analyses
The trees derived from the atpB-rbcL spacer data support
with high bootstrap values (100%) a clade consisting of Tu-
bocapsicum,Aureliana, and Withania and a large clade con-
sisting of Lycianthes and Capsicum (fig. 3). Despite the pres-
ence of a possible capsaicin-like compound in Tubocapsicum
(as detected by a taste test), this genus is not closely related
to Capsicum. Similar results were found in earlier molecular
studies of Tubocapsicum (Olmstead and Palmer 1997; Olm-
stead et al. 1999). These studies and our data indicate that
capsaicin-like compounds may have arisen at least twice in the
evolution of Solanaceae. This possibility is currently being ex-
plored in an evolutionary study of the functions of capsaicin
(J. Tewksbury, personal communication).
A strict consensus tree of the 18 shortest trees obtained from
the atpB-rbcL spacer data indicates that Capsicum is mono-
phyletic but with relatively weak support ( %;bootstrap p66
fig. 3). While the analyses of the combined waxy and spacer
data rooted between Lycianthes and Capsicum cannot confirm
the monophyly of either genus, it does indicate that if C. cil-
iatum is excluded, the monophyly of the remaining Capsicum
species is strongly supported ( %). It is clearbootstrap p100
that further work needs to be done to confirm the monophyly
of both Capsicum s.lat. and Lycianthes.
Species Delimitation
Most of the variation occurs in the earliest diverging
branches on the tree resulting from the combined waxy and
atpB-rbcL spacer data, with C. ciliatum,C. eximium, and C.
tovarii the most divergent species within Capsicum (fig. 4; table
2). All remaining Capsicum species appear to have diverged
more recently and therefore are not so clearly delimited from
each other using molecular data. The following paragraphs
discuss the potential for using waxy and atpB-rbcL spacer data
to delimit species. The waxy introns are especially useful at
the species level, providing more variation at this level than is
commonly found with sequence data. However, the efficacy of
these markers needs to be tested with increased sampling at
the population level.
Capsicum ciliatum is the only Capsicum species with an
insertion (4 bases in length) in the atpB-rbcL spacer data (table
2). In addition, it has 20 unique substitutions (not found in
any other Capsicum or Lycianthes taxa) and four unique gaps
in the waxy data: three deletions (including one 12 bases long)
and one insertion. D’Arcy and Eshbaugh (1974) speculated
that C. ciliatum may belong in its own genus. The weak-to-
moderate sequence support for its inclusion in Capsicum (figs.
3, 4) and the sequence divergence found in C. ciliatum (24
unique characters; table 2), combined with the unique char-
acters of yellow flower color and the complete absence of cap-
saicin (D’Arcy and Eshbaugh 1974), lend some support to
D’Arcy and Eshbaugh’s argument.
Capsicum annuum, from five different sources and including
two subspecies, is the best test of species delimitation with our
data. All accessions are clearly supported as a monophyletic
group on the combined phylogeny (bootstrap p98%). In
addition, all accessions share three unique characters found in
no other taxon (two substitutions and one 1-base deletion;
table 2). The multiple varieties or cultivars of C. baccatum
and C. frutescens are moderately to hightly supported as
monophyletic groups with four unique molecular markers for
C. baccatum and one marker (a 1-base insertion) for C. fru-
tescens (table 2). Other species of Capsicum are well defined
by both unique substitutions and gaps, which may be useful
in future delimitation (table 2): C. chacoense (four substitu-
tions and one 1-base deletion), C. eximium (nine substitutions
and one 4-base deletion), Capsicum pubescens (five substitu-
tions), and C. tovarii (12 substitutions). Capsicum cardenasii
and C. eximium, sometimes hypothesized as a single species
(see “Introduction”), are well supported as separate species,
1416 INTERNATIONAL JOURNAL OF PLANT SCIENCES
with at least 12 differences between them in the molecular
data.
Species Relationships and Informal
Classification of Capsicum
When comparing the results of this study to the enzyme
studies of McLeod et al. (1979a, 1979b, 1982, 1983; fig. 1),
the species of Capsicum in common between the two studies
assume largely identical patterns of relationship. Similarly, the
molecular data are somewhat congruent with hybridization
studies (fig. 2). Combining all three sources of information,
an informal classification was developed (see below). This is
meant to be a ground plan for future studies in Capsicum,
which will hopefully include some additional species that have
been difficult to obtain. After each species name, geographic
ranges and, when extensively domesticated, proposed places
of origin (PPO, after Mcleod et al. 1982, 1983) are given.
Ciliatum group. Capsicum ciliatum: southern Mexico to
northern Peru.
Eximium group. Capsicum eximium: Bolivia and northern
Argentina; Capsicum cardenasii: Dept. of La Paz, Bolivia.
Baccatum group. Capsicum baccatum: northwestern
South America to northern Argentina (PPO: subtropical Bo-
livia); Capsicum chacoense: Bolivia, Argentina, Paraguay.
Annuum group. Capsicum annuum: southern United
States to northern Peru, Bolivia, and West Indies (PPO: Meso-
america); Capsicum chinense: Central America, Caribbean,
and central South America (PPO: Amazon Basin); Capsicum
frutescens: Mexico, Central America, Carribean, and northern
South America (PPO: Amazon Basin); Capsicum galapagoense:
Gala´pagos Islands (Ecuador).
Unassigned to group. Capsicum tovarii: Dept. of Aya-
cucho, Peru; Capsicum pubescens: Andean highlands to Mex-
ico (PPO: midelevation Bolivia).
The monotypic Ciliatum group is sister to all remaining
Capsicum and, as mentioned above, is genetically distinct from
the core Capsicum species. There is moderate phylogenetic
support for its inclusion within Capsicum, so we retain it
within the genus. This species is characterized by yellow co-
rollas and the complete absence of capsaicin. Capsicum cil-
iatum develops small (under 1.0 cm), red, spherical fruit.
The Eximium group, consisting of C. eximium and C. car-
denasii, is strongly supported as sister species (bootstrap p
%; fig. 4) and moderately supported as sister to all other97
core Capsicum species ( %). These two species,bootstrap p73
when crossed, are the only hybrids within Capsicum that pro-
duce highly fertile progeny (fig. 2). They are characterized by
a “viny” habit (Eshbaugh 1976) and purple corollas with yel-
low-green throats. Like C. ciliatum,C. eximium and C. car-
denasii have small (under 1.0 cm), red, spherical fruit. They
inhabit low montane, xerophytic regions (Eshbaugh 1976)
The placement of C. pubescens and C. tovarii is the most
problematic of all the taxa investigated in this study. Here we
treat these two species as unassigned because their position
within Capsicum is not well resolved (fig. 4). Capsicum tovarii
has cream corollas (sometimes with purple petal margins) with
a pair of yellowish spots at the base of each petal and produces
small (under 1.0 cm), red, spherical fruit. Capsicum pubescens
had previously been considered a member of the C. eximium/
cardenasii complex based on hybridization studies (Heiser and
Smith 1958; Lippert et al. 1966) and enzyme profile studies
(Jensen et al. 1979; McLeod et al. 1979a, 1979b, 1982, 1983).
Capsicum pubescens has purple corollas (sometimes cream
with purple margins) and forms large (over 2.0 cm), globose
fruit with a variety of colors. The unusual fruit size and color
of C. pubescens is probably the result of cultivation. Capsicum
tovarii and C. pubescens are both found in xerophytic regions
at low to midelevations in the Andes (McLeod et al. 1982;
Eshbaugh et al. 1983).
The Baccatum group, consisting of C. baccatum and C. cha-
coense, is morphologically diverse. There are no unique mor-
phological characters that unite these species, and the molec-
ular support for the group is weak (one substitution;
%). However, isozyme data (fig. 1; McLeod etbootstrap p60
al. 1982, 1983) provide additional support for this group.
Capsicum baccatum and its varieties have white to cream co-
rollas (except var. praetermissum, which may have violet co-
rolla margins) with a pair of yellowish spots at the base of
each petal. The varieties of C. baccatum each have distinct
fruit shapes. Capsicum chacoense has a dull white corolla and
develops red, oblong, globose fruit under 1.5 cm long. It had
been previously categorized as a species nearly equidistant be-
tween the C. annuum complex, C. baccatum, and the C. ex-
imium complex (McLeod et al. 1979b; Moscone et al. 1993).
The Baccatum group is believed to have originated in south-
central Bolivia in drier lowland habitats (McLeod et al. 1982).
The Annuum group consists of C. annuum (including nu-
merous varieties and cultivars), C. chinense,C. frutescens, and
C. galapagoense. As with the Baccatum group, there are no
morphological characters that unite this group but strong sup-
port from isozymes (fig. 1; McLeod et al. 1982, 1983) and
crossing studies (fig. 2). Capsicum annuum typically has a
white corolla but may be greenish or purple. Capsicum an-
nuum var. annuum is the most widely cultivated chili and can
develop fruit with a variety of different colors, shapes, and
sizes, which are often fleshy. Some cultivars develop fruit over
20 cm in length. Capsicum annuum var. aviculare, the wild
progenitor of the cultivated C. annuum var. annuum, has small
(rarely exceeding 1.0 cm), red, globose or ovoid fruit. Cap-
sicum chinense is a cultivated species that has a dull white
corolla (rarely greenish white) that forms large (often over 1.5
cm wide and long), fleshy, variously colored and shaped fruit.
Capsicum frutescens has a greenish white corolla and forms
red (rarely orange), fleshy, globose or subconical fruit under
2.5 cm in length. Capsicum galapagoense has a white corolla
with a faint yellow tint and develops small (under 1.0 cm),
red, spherical fruit. The original geographic distribution of the
Annuum group is believed to be moister habitats of lowland
tropical South and Central America (Heiser 1976; Pickersgill
et al. 1979).
The placement of C. galapagoense has never previously been
studied using either morphological or molecular data. All data
sets presented in this study, separate and combined, include
C. galapagoense in the weakly supported Annuum group.
Comparing the patterns of relationship in this study to the
body of hybridization data, all species except C. galapagoense
are capable of producing viable hybrids when crossed with
their closest sister groups (fig. 2). The same evolutionary pro-
cesses that allowed C. galapagoense to become morphologi-
WALSH & HOOT—PHYLOGENETIC RELATIONSHIPS OF CAPSICUM 1417
cally distinct (extreme pubescence, dwarfed fruit, generally dis-
tinct flowers and leaves) from other Capsicum species may now
inhibit successful crossing with other species.
Because geographic distributions within Capsicum are over-
lapping and manipulated by man, it is difficult to test their
correlation with our molecular tree. However, judging from
the earliest branching species on our combined tree (C. cilia-
tum,C. cardenasii, and C. eximium; fig. 4), it appears that the
ancestors of Capsicum may have evolved in the drier regions
of the present-day Andes (Peru and Bolivia) with subsequent
migration north or east into tropical lowland regions. Simi-
larly, McLeod et al. (1982) hypothesized that Capsicum arose
in south-central Bolivia. However, they based this hypothesis
on the belief that C. chacoense or its progenitor were ancestral
in the genus. Further molecular work with more complete sam-
pling (including many of the species that are difficult to obtain
and seldom studied) is needed to test this hypothesis.
Acknowledgments
We are grateful to R. Olmstead and M. Whitson for their
helpful comments in review and L. Bohs for help with no-
menclatural issues. We also thank R. Olmstead, B. Pickersgill,
J. Villand, Centro Agrono´ mico de Investigacio´ n y Ensen˜ anza,
and the USDA-ARS for furnishing seeds or DNA used in this
study. This work was partially funded by a Joseph G. Baier
Memorial Scholarship, Department of Biological Sciences, Uni-
versity of Wisconsin—Milwaukee to B. M. Walsh.
Literature Cited
Ballard RE, JW McClure, WH Eshbaugh, KG Wilson 1970 A che-
mosystematic study of selected taxa of Capsicum. Am J Bot 57:
225–233.
Bohs L, RG Olmstead 1997 Phylogenetic relationships in Solanum
(Solanaceae) based on ndhF sequences. Syst Bot 22:5–17.
Cochran HL 1940 Characters for the classification and identification
of varieties of Capsicum. Bull Torrey Bot Club 67:710–717.
D’Arcy WG, WH Eshbaugh 1974 New World peppers (Capsicum—
Solanaceae) north of Colombia: a resume. Baileya 19:93–105.
Davenport WA 1970 Progress report on the domestication of Cap-
sicum (chili peppers). Proc Assoc Am Geogr 2:46–47.
Deb DB 1979 Solanaceae in India. Pages 87–112 in JG Hawkes, RN
Lester, AD Skelding, eds. The biology and taxonomy of the Solan-
aceae. Academic Press, London.
Doyle JJ, JL Doyle 1987 A rapid DNA isolation procedure for small
quantities of fresh tissue. Phytochem Bull 19:11–15.
Emboden WA Jr 1961 A preliminary study of the crossing relation-
ships of Capsicum baccatum. Butler Univ Bot Stud 14:1–5.
Eshbaugh WH 1970 A biosystematic and evolutionary study of Cap-
sicum baccatum (Solanaceae). Brittonia 22:31–43.
——— 1976 Genetic and biochemical systematic studies of chili pep-
pers (Capsicum—Solanaceae). Bull Torrey Bot Club 102:396–403.
——— 1980 The taxonomy of the genus Capsicum (Solanaceae).
Phytologia 47:153–166.
Eshbaugh WH, PG Smith, DL Nickrent 1983 Capsicum tovarii (So-
lanaceae), a new species of pepper from Peru. Brittonia 35:55–60.
Farris JS, M Ka¨ llersjo¨ , AG Kluge, C Bult 1995 Testing significance
of incongruence. Cladistics 10:315–319.
Felsenstein J 1985 Confidence limits on phylogenies: an approach us-
ing the bootstrap. Evolution 39:783–791.
Golenberg EM, MT Clegg, ML Durbin, J Doebley, DP Ma 1993
Evolution of a noncoding region of the chloroplast genome. Mol
Phylogenet Evol 2:52–64.
Heiser CB Jr 1969 Nightshades, the paradoxical plants. WH Free-
man, San Francisco.
——— 1976 Peppers: Capsicum (Solanaceae). Pages 265–268 in NW
Simmonds, ed. Evolution of crop plants. Longman, London.
Heiser CB Jr, PG Smith 1948 Observations on another species of cul-
tivated pepper, Capsicum pubescens R&P. Proc Am Soc Hortic Sci
52:331–335.
——— 1953 The cultivated Capsicum peppers. Econ Bot 7:214–227.
——— 1958 New species of Capsicum from South America. Brittonia
10:194–201.
Hoot SB, A Culham, PR Crane 1995 The utility of atpB gene se-
quences in resolving phylogenetic relationships: comparison with
rbcL and 18S ribosomal DNA sequences in the Lardizabalaceae.
Ann Mo Bot Gard 82:194–207.
Hoot SB, AW Douglas 1998 Phylogeny of the Proteaceae based on
atpB and atpB-rbcL intergenic spacer region sequences. Aust Syst
Bot 11:301–320.
Jensen RJ, MJ McLeod, WH Eshbaugh, SI Guttman 1979 Numerical
taxonomic analyses of allozymic variation in Capsicum (Solana-
ceae). Taxon 28:315–327.
Lippert LF, PG Smith, BO Bergh 1966 Cytogenetics of the vegetable
crops: garden pepper, Capsicum sp. Bot Rev 32:25–55.
Manen JF, A Natali 1995 Comparison of the evolution of ribulose-
1, 5-biphosphate carboxylase (rbcL) and atpB-rbcL noncoding
spacer sequences in a recent plant group, the tribe Rubieae (Rubi-
aceae). J Mol Evol 41:920–927.
Mason-Gamer RJ, CF Weil, EA Kellogg 1998 Granule-bound starch
synthase: structure, function, and phylogenetic utility. Mol BiolEvol
15:1658–1673.
McLeod MJ, WH Eshbaugh, SI Guttman 1979aAn electrophoretic
study of Capsicum (Solanaceae): the purple flowered taxa. Bull Tor-
rey Bot Club 106:326–333.
——— 1979bA preliminary biochemical systematic study of the ge-
nus Capsicum—Solanaceae. Pages 701–713 in JG Hawkes, RN Les-
ter, AD Skelding, eds. The biology and taxonomy of the Solanaceae.
Academic Press, London.
McLeod MJ, SI Guttman, WH Eshbaugh 1982 Early evolution of
chili peppers (Capsicum). Econ Bot 36:361–368.
McLeod MJ, SI Guttman, WH Eshbaugh, RE Rayle 1983 An elec-
trophoretic study of evolution in Capsicum (Solanaceae). Evolution
37:562–574.
Miller RE, MD Rausher, PS Manos 1999 Phylogenetic systematics of
Ipomoea (Convolvulaceae) based on ITS and Waxy sequences. Syst
Bot 24:209–227.
Moscone EA, M Lambrou, AT Hunziker, F Ehrendorfer 1993 Giemsa
C-banded karyotypes in Capsicum (Solanaceae). Plant Syst Evol 186:
213–229.
Natali A, JF Manen, F Ehrendorfer 1995 Phylogeny of the Rubiaceae-
Rubioideae, in particular the tribe Rubieae: evidence from a non-
coding chloroplast DNA sequence. Ann Mo Bot Gard 82:428–439.
Olmstead RG, JD Palmer 1992 A chloroplast DNA phylogeny of the
Solanaceae: subfamilial relationships and character evolution. Ann
Mo Bot Gard 79:346–360.
——— 1997 Implications for the phylogeny, classification, and bio-
geography of Solanum from cpDNA restriction site variation. Syst
Bot 22:19–29.
Olmstead RG, JA Sweere, RE Spangler, L Bohs, JD Palmer 1999
Phylogeny and provisional classification of the Solanaceae based on
chloroplast DNA. Pages 111–137 in M Nee, DE Symon, JP Jessup,
JG Hawkes, eds. Solanaceae. Vol 4. Advances in biology and util-
ization. Royal Botanic Gardens, Kew.
1418 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Pickersgill B 1966 The variability and relationships of Capsicum chi-
nense Jacq. PhD diss. Indiana University, Bloomington.
——— 1969 The archaeological record of chili peppers (Capsicum
spp.) and the sequence of plant domestication in Peru. Am Antiquity
34:54–61.
——— 1971 Relationships between weedy and cultivated forms in
some species of chili peppers (genus Capsicum). Evolution 25:
683–691.
——— 1977 Taxonomy and the origin and evolution of cultivated
plants in the New World. Nature 268:591–595.
Pickersgill B, CB Heiser, J McNeill 1979 Numerical taxonomic stud-
ies on variation and domestication in some species of Capsicum.
Pages 679–700 in JG Hawkes, RN Lester, AD Skelding, eds. The
biology and taxonomy of the Solanaceae. Academic Press, London.
Sambrook J, EF Fritsch, T Maniatis 1989 Molecular cloning: a lab-
oratory manual. Cold Spring Harbor Laboratory, Plainview, N.Y.
Savolainen V, JF Manen, E Douzery, R Spichiger 1994 Molecular
phylogeny of families related to Celastrales based on rbcL 5
flanking
sequences. Mol Phylogenet Evol 3:27–37.
Shopova M 1966 Studies in the genus Capsicum. I. Species differ-
entiation. Chromosoma 19:340–348.
Simmons MP, H Ochoterena 2000 Gaps as characters in sequence-
based phylogenetic analyses. Syst Biol 49:369–381.
Smith PG, CB Heiser Jr 1951 Taxonomic and genetic studies on the
cultivated peppers, Capsicum annuum L. and C. frutescens L. Am
J Bot 38:362–368.
——— 1957 Taxonomy of Capsicum sinense Jacq. and the geo-
graphic distribution of the cultivated Capsicum species. Bull Torrey
Bot Club 84:413–420.
Swofford DL 1998 PAUP*: phylogenetic analysis using parsimony
(* and other methods), version 4. Sinauer, Sunderland, Mass.
van der Leij FR, RGF Visser, AS Ponstein, E Jacobsen, WJ Feen-
stra 1991 Sequence of the structural gene for granule-bound starch
synthase of potato (Solanum tuberosum L.) and evidence for a single
point deletion in the amf allele. Mol Gen Genet 228:240–248.
Walsh BM 1999 A phylogenetic analysis of the species relationships
within Capsicum (Solanaceae) using sequences from two DNA non-
coding regions: the chloroplast atpB-rbcL intergenic region and in-
trons of the nuclear gene waxy. MS thesis. University of
Wisconsin—Milwaukee.
