| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Botany, Washington State University, Pullman, Washington 99164-4238
Received for publication December 15, 1998. Accepted for publication April 27, 1999.
ABSTRACT
Ceanothus comprises ~55 morphologically and ecologically diverse species of woody perennials endemic to North America. Interpretations of the natural history of Ceanothus have served as a general model of evolution for woody perennials with simple entomophilous pollination systems, but these interpretations lacked explicit phylogenetic context. We used cladistic analysis of sequences of the chloroplast-encoded matK and the internal transcribed spacers (ITS) and 5.8S coding region of nuclear ribosomal DNA (nrDNA) to reconstruct the phylogeny of Ceanothus. The nuclear and organellar phylogenies exhibited very low levels of both topological and character congruence. Subgenera Ceanothus and Cerastes are monophyletic sister taxa in both phylogenies, but both data sets suffer from a lack of resolution below the level of subgenus. Lack of taxonomic congruence between the two data sets may be a result of introgression and/or lineage sorting. The ITS tree was accepted as the better estimate of a species phylogeny for Ceanothus, on the assumption that nuclear markers are less prone to introgression. Three of five polytypic species in the ITS data set were paraphyletic, and four of six polytypic species in the matK data set were paraphyletic. This study demonstrates the degree to which matched independent data sets can produce conflicting summaries of evolutionary history.
Key Words: Ceanothus • congruence • ITS • matK • phylogeny • Rhamnaceae
Phylogenetic histories inferred from different sets of characters need not be congruent with each other, nor with the phylogenetic history of the organisms from which the characters were collected. This problem is generally referred to as the "gene tree vs. species tree" problem (e.g., Pamilo and Nei, 1988
); unlinked genetic loci evolve independently and may be transmitted differentially to descendent species. The handling of independent or "process-partitioned" (Bull et al., 1993
; Kluge and Wolf, 1993
; Miyamoto and Fitch, 1995
) data sets is an active debate of systematists. Concerns regarding the relative degree and nature of correlation between separate phylogenetic data sets and the handling of these data sets for inferences of evolutionary history have resulted in three methodological philosophies: total evidence (e.g., Kluge, 1989
), data partitioning (e.g., Miyamoto and Fitch, 1995
), and consensus (e.g., de Queiroz, 1993
)(for review see Huelsenbeck, Bull, and Cunningham, 1996
). The concerted effects of hybridization and introgression are one mechanism whereby cytoplasmic and nuclear elements may gain phylogenetic independence (i.e., become incongruent). Therefore, taxa marked by frequent and widespread hybridization provide an opportunity to assess the potential degree to which separate nuclear and cytological estimates of phylogeny might conflict.
Ceanothus L. (California lilac, New Jersey tea) comprises ~55 morphologically and ecologically diverse species of woody perennials endemic to North America (Table 1). The genus displays its greatest diversity in the California Floristic Province (CFP) (McMinn, 1942
), where it is a distinctive component of the flora. Species of Ceanothus occur mostly in semiarid forests, oak woodland, and chaparral, on a wide variety of soil types, typically in climates with low summer rainfall and wet winters (Mason, 1942
). Ceanothus is characterized by small, bisexual, actinomorphic flowers borne in dense, showy inflorescences and exhibits great diversity in leaf morphology and anatomy, habit, and habitat. Some species of Ceanothus produce root nodules (McMinn, 1942
), and the nitrogen-fixing actinomycete Frankia has been isolated from the nodules of several species (Rose, Daniels, and Trappe, 1979
). The 23 species of Ceanothus that have been examined cytologically are diploid with n = 12 (McMinn, 1942
; Nobs, 1963
).
|
; Brandegee, 1895
; Howell, 1940
; McMinn, 1942
; Nobs, 1963
). Much of the taxonomy of Ceanothus has been complicated by apparent hybridization (Watson, 1875
; Brandegee, 1895
; Howell, 1940
). A biosystematic study of Ceanothus (Nobs, 1963
) supported the proposed subgeneric distinctions; only two of 11 intersubgeneric crosses attempted were partially successful, producing weak hybrids, none of which survived longer than one year. In contrast, 20 of the 24 attempted intrasubgeneric crosses in Cerastes produced vigorous offspring; the products of seven of those crosses were carried into a second generation. All of the interspecific hybrids in subgenus Cerastes exhibited normal seed set and high pollen fertility, and second-generation products showed no evidence of weakness that could be indicative of latent genetic incompatibility between parental genotypes (Nobs, 1963
).
Subgenus Ceanothus occurs throughout much of North America and comprises 32 species (Table 1). The greatest concentration of species occurs in the CFP; four species are restricted to regions of Mexico. Subgenus Ceanothus is characterized by alternate leaves, hornless fruits, small, early-falling stipules, and stomata located on the lower surfaces of the leaves (Watson, 1875
; McMinn, 1942
). Fewer reports of natural interspecific hybridization in subgenus Ceanothus than in Cerastes have led to the conclusion that, in general, isolating barriers among members of subgenus Ceanothus are stronger than those in subgenus Cerastes (McMinn, 1942
). Subgenus Ceanothus is considered to represent the ancestral group of species of the genus, with an origin during the Paleocene or Eocene period (Mason, 1942
). No formal infrasubgeneric groups have been recognized for Ceanothus, but several informal groups have been variously proposed based on morphology, habit, and leaf persistence (Parry, 1888b
; McMinn, 1930
; Mason, 1942
).
Subgenus Cerastes comprises 22 species (Table 1) that are endemic to western North America, with the greatest diversity occurring in the CFP; one species is restricted to Mexico. The subgenus is characterized by opposite, thick, coriaceous leaves; horned fruits; persistent, corky stipules; and stomata located in crypts on the undersides of the leaves (Watson, 1875
; McMinn, 1942
). Based on morphological and distributional evidence, McMinn (1930)
concluded that subgenus Cerastes originated in either southern California or northern Mexico and subsequently migrated northward along the Coast Ranges and Sierra Nevada. Mason (1942)
proposed that subgenus Cerastes arose from within subgenus Ceanothus sometime during the Miocene when a general climatic drying trend promoted an adaptive radiation. One adaptation putatively associated with such a radiation is the xeromorphic leaf anatomy, including stomatal crypts, which now typifies subgenus Cerastes. This adaptation is postulated (Mason, 1942
) to have occurred in a maritime environment in the coastal archipelago that now forms the Coast Ranges of California. The ancestor of subgenus Cerastes is believed to have radiated along the coastal archipelago, sending adaptive radiants into the interior (Mason, 1942
). Nobs (1963)
recognized three types of stomatal crypts within subgenus Cerastes and suggested that crypt types identified three major waves of differentiation within the subgenus, each with a different geographic focus (Nobs, 1963
). No formal (i.e., taxonomic) infrasubgeneric groups have been proposed for subgenus Cerastes, but several informal groups have been suggested based on morphological and distributional data (e.g., McMinn, 1930, 1942
). Furthermore, several species of subgenus Cerastes (i.e., C. sonomensis, C. masonii, C. divergens, C. confusus) have putative hybrid origins, based on morphological, anatomical, and habitat characteristics (Howell, 1940
; McMinn, 1942
; Nobs, 1963
). The hypothesis of hybrid speciation in subgenus Cerastes has become widely accepted (e.g., Raven, Evert, and Eichhorn, 1992
).
Species of Ceanothus are typically a prominent component of dynamic seral communities (e.g., fire-adapted chaparrals and coastal scrubs), are often engaged in intimate and ecologically significant, symbiotic relationships with soil microbes, and demonstrate prominent diversity in several plant characteristics. Stebbins (1950)
and Grant (1981)
both used interpretations of character diversity in Ceanothus in the formulation of a model of diversification for woody perennials with open pollination (e.g., Quercus, Salix, Populus). Explicit phylogenetic hypotheses for Ceanothus would provide both the necessary context for assessing the evolution of a variety of characters and the basis for future comparative studies. No detailed phylogenetic analysis of Ceanothus has been reported. Jeong, Liston, and Myrold (1997)
presented two molecular-based phylogenetic trees, derived from the chloroplast-encoded genes rbcL and ndhF, for a limited subset (ten species) of Ceanothus. Their ndhF phylogeny supported the traditional subgeneric distinctions, and an rbcL-based estimate of divergence time suggested that subgenera of Ceanothus originated between 18 and 39 Mya, a finding consistent with previous estimates based on fossil evidence (Mason, 1942
).
The purpose of this study was to infer phylogenetic relationships in Ceanothus using gene sequence data, with the subsequent goal of assessing the degree of congruence between organellar and nuclear gene trees in a group of species considered to hybridize frequently. We conducted phylogenetic analyses using sequences of the internal transcribed spacers (ITS) and 5.8S coding region of the 18S–26S ribosomal DNA (nrDNA) and of the chloroplast-encoded matK and assessed congruence between both the data sets and the trees.
MATERIALS AND METHODS
Plant material
Leaf material for ITS and matK sequence analyses was collected in the field, in botanical gardens, and from herbarium specimens (Table 1). Fresh material was collected, transported on ice to the laboratory, and stored under refrigeration until extraction was conducted. DNA from the same individual was used for both ITS and matK sequencing. Amplification and sequencing of matK from herbarium specimens had limited success, and not all individuals are represented by matK sequences. Some sequences of C. cuneatus var. cuneatus, C. gloriosus var. exaltatus, C. masonii, and C. sonomensis were taken from population samples collected for a separate study of diploid hybridization in Ceanothus (Hardig, Soltis, and Soltis, unpublished data). No material of C. microphyllus was available for DNA extraction. All other missing sequences resulted from failed attempts at sequencing DNA from herbarium specimens.
DNA extraction and sequencing
Extraction of DNA from fresh material followed the procedure of Doyle and Doyle (1987)
as modified by Soltis et al. (1991)
. Extraction of DNA from herbarium and desiccant-dried material followed the procedure of Doyle and Doyle (1987)
as modified by Cullings (1992)
, with the omission of an RNAase treatment.
The ITS regions and intervening 5.8S coding region (~626 basepairs in Ceanothus) of all individuals were amplified using primers N-nc18S10 and C26A (Suh et al., 1993
). Polymerase chain reactions (PCR) were conducted in 100-µL reaction volumes using the following recipe: 10 µL reaction buffer (0.2 mol/L Tris, pH 8.3, 0.5 mol/L KCl, 1.5 mmol/L MgCl2, 0.01% Tween-20), 16 µL dNTPs (10 mmol/L), 10 µL of each primer (5 ng/µL), 5 µL dimethyl sulfoxide (DMSO), 0.5 µL Taq polymerase, sample DNA, and dH2O. The double-stranded products were purified using polyethylene glycol (PEG) precipitation to permit direct sequencing; PCR products were mixed with an equivalent amount of 20% w/v PEG/ 2.5 mol/L NaCl, heated for 15 min at 37°C, and microfuged for 15 min (4°C). The supernatant was removed, and the remaining pellet was cleaned with consecutive washes using 200 µL of 80 and 95% ethanol, respectively, with 7-min centrifugations (4°C) after each. Samples were dried under vacuum, without heat, and resuspended in 20 µL of sterile water. PCR products were sequenced using the forward primers ITS1 and ITS3 and reverse primers ITS2 and ITS4 (White et al., 1990
). Fluorescent dye-labeled sequences were obtained by cycle sequencing using ABI PRISM rhodomine dye terminator (Applied Biosystems, Inc., Foster City, California) following the manufacturer's procedures, except that reaction volumes were 25% those described by ABI. Cycle sequencing conditions were: 30-s denaturation, 96°C; 15-s annealing, 50°C; 4-min extension, 60°C; 25 cycles. Data were collected using an Applied Biosystems, Inc., Model 377 DNA sequencer. The resulting sequences were assembled and edited using the computer program SequencherTM 2.1 (Gene Codes Corp., Inc., 1995). All sequences were aligned manually, alignment gaps were scored as missing data, and missing data were scored as ambiguous. ITS sequences are GenBank accessions GBAN-AF048901 to GBAN-AF048975.
Amplification and sequencing primers used for analysis of the chloroplast-encoded matK are given in Johnson and Soltis (1994, 1995)
. Double-stranded amplification was achieved using primers trnK-3924F and trnK-2R. Complementary single-stranded copies of the gene were amplified using each of the primers individually, with the double-stranded product as template. Single-stranded DNA products were purified by PEG precipitation (Morgan and Soltis, 1993
) to permit direct sequencing. Sequencing primers used were: trnK-710F, matK-1235R, matK-1506F, and matK-1412F. Divergence between some of these sequencing primers and their target sequences in the 3' region of Ceanothus matK necessitated the development of three additional primers (matK-1272F: 5'-CGC TAT TGG GTG AAA GAT GCC-3' matK-1712F: 5'-GTG TGG GCT CAA CCA GGA AGG-3'; matK-1972R: 5'-GAC GAA TCG GYC CAG GTC GAC-3'). S35-labeled sequences were obtained by Sequenase-mediated dideoxynucleotide reactions (U.S. Biochemical Corp., Cleveland, Ohio) and visualized by autoradiography following fragment separation in 4% polyacrylamide gels. We sequenced 1147 bp of matK, beginning 31 bases upstream from the transcription initiation codon. No insertions or deletions were apparent among the 33 species of Ceanothus examined. GenBank accessions GBAN-AF049797 to GBAN-AF049849 are the matK sequences.
Phylogenetic analyses
Data sets were examined for taxa with identical character states using the FILTER TAXA utility of MacClade ver. 3.04 (Maddison and Maddison, 1992
). Such taxa were merged to reduce the computational burden on subsequent phylogenetic analyses.
Zizyphus obtusifolia, Spyridium parvifolium, and Adolphia californica were used as outgroups, based on their close relationship to Ceanothus in phylogenetic trees of Rhamnaceae using sequences of the chloroplast-encoded rbcL and trnL-F genes (Richardson et al., 1997
).
Each data set was analyzed separately with heuristic parsimony searches using PAUP* 4.0d59 (used by permission of D. Swofford). Search strategy consisted of 1000 replicates of random taxon addition with TBR branch-swapping, MULPARS on, and STEEPEST DESCENT on; branches with minimum lengths equal to zero ("amb-") were collapsed to form polytomies. Bootstrap analysis (Felsenstein, 1985
) using 1000 replicates of fast bootstrap (PAUP* 4.0d59) and converse constraint decay analysis (Baum, Sytsma, and Hoch, 1994
) were used to assess the relative support for branches.
Congruence testing
The handling of separate data sets for the same group of taxa continues to be debated. Some have argued that all data should be combined into a single analysis (e.g., Kluge, 1989
; Barrett, Donoghue, and Sober, 1991
; Donoghue and Sanderson, 1992
; Eernisse and Kluge, 1993
; Kluge and Wolf, 1993
; Chippindale and Wiens, 1994
; Baker and DeSalle, 1997
), others have argued for independent analyses and subsequent examination for congruence (e.g., Bledsoe and Raikow, 1990
; Lanyon, 1993
; Miyamoto and Fitch, 1995
), and still others have advocated an ad hoc approach, wherein separate data sets are combined after independent analyses have demonstrated minimal heterogeneity (Bull et al., 1993
; de Queiroz, 1993
; de Queiroz, Donoghue, and Kim, 1995
; Huelsenbeck and Bull, 1996
; Huelsenbeck, Bull, and Cunningham, 1996
; Johnson and Soltis, 1998
). Increasing the number of characters by combining data sets increases the potential number of parsimony-informative characters, and thus the phylogenetic accuracy of the analysis (e.g., Soltis et al., 1998
; Yoder et al., unpublished data). However, if separate data sets are incongruent as a result of evolutionary independence, then analysis of combined data sets may result in reduced or erroneous resolution with respect to the actual organismal phylogeny. The decision to combine data sets cannot be limited to a simple metric of congruence; intrinsic qualities must be considered (e.g., process-partitioning). Incongruence between nuclear and organellar phylogenetic trees is typically attributed to the introgression of a cytoplasmic genome from one species into the nuclear background of another species (e.g., Ferris et al., 1983
; Gyllensten and Wilson, 1987
; Harrison, Rand, and Wheeler, 1987
; Tegelström, 1987
; Soltis et al., 1991
; Rieseberg, 1991
; Rieseberg and Wendel, 1993
; Soltis and Kuzoff, 1995
; Soltis, Johnson, and Looney, 1996
). Given the potential of introgression in the history of Ceanothus, it was decided a priori that the matK and ITS data sets represent different process-partitions and that trees derived from these data sets would be compared for topological and character congruence prior to any combined analysis.
The matK data set comprised sequences from 38 species of Ceanothus. An equivalent ITS data set, using sequences collected from the same individuals represented in the matK data set, was used for congruence testing. Individuals with identical sequences were not excluded from the data sets. For both matK and ITS, 5000 trees were saved from a single tree bisection-reconnection (TBR) search. Topological congruence between trees derived from ITS and matK was assessed with Rolf's modified consensus information index (1982; CI1) and Templeton's (1983)
nonparametric statistical framework, as implemented in PAUP* 4.0d59. Character incongruence was assessed using the incongruence metric of Mickevich and Farris (1981
; IMF). The CI1 and IMF metrics were selected because they provide quantitative measures of congruence, are easy to calculate using PAUP-generated values, and are relatively straightforward to interpret. CI1 values were obtained using the INDICES = YES option of the CONTREE command, and IMF values were calculated using the LENFIT command with the CI option activated in PAUP* 4.0d59 (following Johnson and Soltis, 1998
). The IMF value was tested for significance following Farris et al. (1995)
using the PARTITION HOMOGENEITY TEST in PAUP* 4.0d59. Given the limited amount of phylogenetic information inherent in both data sets, semistrict (combinable component; Bremer, 1990
) consensus trees were used for congruence testing (Johnson and Soltis, 1998
).
RESULTS
ITS/5.8S analysis
Sequence lengths
ITS-1 differed in length between the two subgenera. Members of subgenus Ceanothus possessed an ITS-1 region 252 bp (base pairs) in length, with the exception of C. foliosus var. medius whose ITS-1 region was 248 bp. Members of subgenus Cerastes possessed an ITS-1 region 254 bp in length, with the exception of C. fresnensis, C. roderickii, C. verrocosus, and C. megacarpus, each of which possessed an ITS-1 region of 253 bp. ITS-1 was 251 bp in both Zizyphus obtusifolia and Spyridium parvifolium and 255 bp in Adolphia californica. Aligned ITS-1 sequences were 261 bp. ITS-2 was 185 bp in most species of both subgenera. In C. cordulatus, C. impressus, C. prostratus, and some individuals of C. sonomensis and C. cuneatus, ITS-2 was 184 bp, while C. confusus had an ITS-2 of 186 bp. ITS-2 was 190 bp in Spyridium parvifolium, 194 bp in Zizyphus obtusifolia, and 197 bp in Adolphia californica. Aligned ITS-2 sequences were 201 bp. The 5.8S gene was 164 bp in all taxa sampled.
Ceanothus oliganthus var. sorediatus and C. tomentosus had identical ITS sequences. Ceanothus papillosus var. roweanus, C. griseus, C. integerrimus, and C. hearstiorum formed a second group with identical ITS sequences. Ceanothus masonii, C. ferrisae, and C. sonomensis also had identical ITS sequences. A single species of each group (C. tomentosus, C. griseus, and C. masonii, respectively) was included in the phylogenetic analysis, reducing the number of species in the ingroup to 45.
Intraspecific sequence variation
Sequences from conspecific individuals were evaluated pairwise using Kimura's two-parameter distances to estimate levels of sequence divergence within species (Table 2). ITS sequences were collected from multiple individuals of Ceanothus masonii, C. cuneatus var. cuneatus, C. sonomensis, C. gloriosus var. exaltatus, C. jepsonii var. albiflorus, C. ferrisae, and C. hearstiorum. The smallest range of divergence occurred in Ceanothus masonii (0.00–0.32%, seven individuals examined), and the greatest range occurred in C. gloriosus var. exaltatus (0.00–1.17%, five individuals examined) (Table 2).
|
Search results
ITS-1, ITS-2, and 5.8S sequences were analyzed as a single data set following the search strategy outlined above. The TBR search resulted in 528 trees of length 333 [retention index (RI) = 0.93; consistency index (CI) = 0.63, excluding autapomorphies] on a single island of trees. This tree island was hit in 968 of the TBR replicates. The remaining 32 TBR replicates hit islands of suboptimal trees.
Phylogenetic relationships
The species of Ceanothus fall into two equivalent groups that correspond entirely to subgeneric circumscriptions (Fig. 1). Subgenus Ceanothus forms a large clade (bootstrap 99%; decay value of 8) that comprises a 21-tomy with only three small subclades. Ceanothus foliosus var. medius and C. foliosus var. vineatus appear as sister taxa (bootstrap 59%; decay 1), as do C. palmeri and C. lemmonii (bootstrap 59%; decay 1) and C. parvifolius and C. leucodermis (bootstrap 56%; decay 1). Ceanothus cyaneus is sister to C. parvifolius and C. leucodermis (bootstrap 52%; decay 1). Ceanothus tomentosus and C. griseus, two species representing additional taxa with identical ITS sequences, occur as terminal taxa in the 21-tomy (Fig. 1).
|
Paraphyletic species
Individuals of all varieties of three polytypic species (C. megacarpus, C. gloriosus, and C. jepsonii) were included in the ITS data set. Of the three, only C. megacarpus demonstrated potential monophyly in both the strict and majority-rule consensus trees (Fig. 1). All three varieties of Ceanothus gloriosus were represented in the ITS analysis, and C. gloriosus var. gloriosus and C. gloriosus var. exaltatus form a well-supported sister pair (bootstrap 80%; decay value of 2) and form another strongly supported clade with C. jepsonii var. jepsonii (bootstrap 81%; decay value of 2) (Fig. 1); however, C. gloriosus var. porrectus occurs in the polytomy of the Northwestern clade, thus making C. gloriosus paraphyletic. A subsequent parsimony analysis of the ITS data, with the imposition of monophyly on the three varieties of C. gloriosus, resulted in an increase in tree length of four steps (1.2%). Ceanothus jepsonii var. jepsonii occurs inside the C. gloriosus clade, as previously discussed, while C. jepsonii var. albiflorus occurs in the Coast clade (Fig. 1), also making C. jepsonii paraphyletic. A subsequent parsimony analysis of the ITS data, with the imposition of monophyly on the two varieties of C. jepsonii, resulted in an increase in tree length of five steps (1.5%). Two other polytypic species (C. foliosus and C. cuneatus) were represented by some, but not all, of their recognized varieties in the ITS data set. Ceanothus foliosus demonstrated potential monophyly, with two of the three recognized varieties (medius and vineatus) as sisters in the ITS strict consensus tree (bootstrap 59%; decay value of 1) (Fig. 1), but without an individual of C. foliosus var. foliosus, the monophyly of C. foliosus cannot be addressed further. Ceanothus cuneatus vars. cuneatus and rigidus are both part of the polytomy of the Northwestern clade in the ITS tree, but C. cuneatus var. cuneatus is the sister to C. roderickii (bootstrap 39%; decay value of 1) rather than to C. cuneatus var. rigidus (Fig. 1). Given the ITS tree, a monophyletic C. cuneatus would therefore also include at least C. roderickii and perhaps several other species. Ceanothus cuneatus var. fascicularis was not included in the analysis. A subsequent parsimony analysis of the ITS data, with the imposition of monophyly on the two varieties of C. cuneatus present, resulted in an increase in tree length of two steps (0.6%).
matK analysis
Species possessing identical sequences
Ceanothus hearstiorum and C. lemmonii had identical matK sequences. Ceanothus velutinus var. velutinus, C. foliosus var. vineatus, C. arboreus, C. diversifolius, and C. tomentosus formed a second group with identical matK sequences. Ceanothus pinetorum, C. masonii, C. gloriosus var. porrectus, C. pumilus, C. gloriosus var. exaltatus, and C. cuneatus also had identical matK sequences. A single species of each group (C. lemmonii, C. foliosus var. vineatus, and C. pinetorum, respectively) was included in the phylogenetic analysis, reducing the number of taxa in the ingroup to 43 (27 species).
Intraspecific sequence variation
Sequences from conspecific individuals were evaluated pairwise using Kimura's two-parameter distances to estimate levels of sequence divergence within species (Table 2). Multiple individuals were sampled for C. gloriosus var. exaltatus, C. masonii, C. cuneatus var. cuneatus, and C. sonomensis. Ceanothus cuneatus var. cuneatus possessed the narrowest range of divergence for matK sequences (0.00–0.26%; three individuals examined), and C. sonomensis exhibited the greatest range (0.17–0.52%; three individuals examined) (Table 2).
Character status
There were 1006 constant characters, 71 variable but parsimony-uninformative characters, and 70 parsimony-informative characters in the matK data set. Within the ingroup, there were 1111 constant characters, 20 variable but parsimony-uninformative characters, and only 16 parsimony-informative characters.
Search results
The initial 1000 TBR swapping replicates generated 899 most-parsimonious trees (length = 152, RI = 0.97, CI = 0.92, excluding autapomorphies), from a single island of trees. This tree island was hit in 998 of the TBR replicates. The remaining two TBR replicates hit islands of suboptimal trees. The strict consensus tree is given in Fig. 2. Approximately 75% (113 steps) of the overall tree length of a single most-parsimonious tree recovered in the analysis occurred between the ingroup and the outgroup (results not shown). This pattern of character distribution is due to the high degree of sequence divergence between Ceanothus and members of the outgroup and results in a high CI value (0.92), despite the limited number of parsimony-informative characters relative to the overall length.
|
Paraphyletic species
Individuals of all varieties of three polytypic species (C. oliganthus, C. gloriosus, and C. megacarpus) were included in the matK data set. Ceanothus oliganthus vars. oliganthus and sorediatus could potentially be resolved as sisters in the polytomy of subgenus Ceanothus (Fig. 2). Ceanothus gloriosus var. porrectus and an individual of C. gloriosus var. exaltatus possessed an identical matK sequence that was also found in four other species of Ceanothus. Ceanothus gloriosus var. gloriosus occurs as a terminal taxon in a large polytomy that also contains the other varieties of C. gloriosus (Fig. 2); this polytomy could potentially be resolved to create a monophyletic C. gloriosus, however, it must be noted that the second individual of Ceanothus gloriosus var. exaltatus occurred as sister to an individual of C. sonomensis in 95% of the recovered trees (Hardig, unpublished data). A subsequent parsimony analysis of the matK data, with the imposition of monophyly on the three varieties of C. gloriosus, resulted in an increase in tree length of only one step (0.6%). Ceanothus megacarpus var. megacarpus and C. megacarpus var. insularis both occur as terminal taxa in the large subgenus Cerastes polytomy (Fig. 2) and could potentially be resolved as sister species, however, C. megacarpus var. insularis did occur as a sister species to C. ferrisae in 90% of the recovered trees (Hardig, unpublished data). A constrained parsimony analysis resulted in an increase in tree length of two steps (1.3%). Three other polytypic species (C. cuneatus, C. foliosus, and C. velutinus) were represented by some, but not all, of their recognized varieties in the matK data set. Ceanothus foliosus var. medius and C. foliosus var. vineatus occur as terminal taxa in the subgenus Ceanothus polytomy and could potentially be resolved as sister taxa in the strict consensus tree, yet C. foliosus var. medius occurred as a sister taxon to C. impressus in 93% of the recovered trees (Hardig, unpublished data). Ceanothus velutinus var. velutinus shared an identical matK sequence with C. lemmonii, which in turn was found to occur as a sister taxon to C. cyaneus in 82% of the recovered trees (Hardig, unpublished data); C. velutinus var. hookeri occurs as a terminal taxon in the basal polytomy and could not be resolved as a sister taxon to C. velutinus var. velutinus without including C. lemmonii (and possibly C. cyaneus) as well as the other species possessing the same matK sequence as C. lemmonii. One individual of C. cuneatus var. cuneatus shares an identical matK sequence with individuals of C. pinetorum, C. masonii, C. pumilus, and C. gloriosus vars. exaltatus and porrectus and occurs as a single terminal taxon in a large polytomy (bootstrap 44%; decay value of 1), and a second individual of C. cuneatus var. cuneatus occurs as a terminal taxon in the same large polytomy as the first individual, and a third individual of C. cuneatus var. cuneatus occurs in the well-supported North Coast clade, with individuals of C. sonomensis, C. masonii, and C. purpureus (bootstrap 79%; decay value of 2) (Fig. 2). Two individuals of C. cuneatus var. rigidus occur as terminal taxa in the subgenus Cerastes polytomy. It is not possible to resolve the polytomies in subgenus Cerastes to create a monophyletic C. cuneatus.
Congruence testing
Search results
Trees derived for congruence testing from ITS data had a length of 249 steps (RI = 0.95, CI = 0.68, excluding autapomorphies) and those derived from matK data had a length of 118 steps (RI = 0.98, CI = 0.85, excluding autapomorphies) (Fig. 3).
|
Templeton's (1983)
nonparametric test compares the fit of alternative phylogenies with respect to a "best" phylogeny. This test assesses support within a data set for an alternative topology that is not obvious owing to the greater support present for the "best" topology. We tested both phylogenies as the potential "best" phylogeny. Both phylogenies were significantly different with respect to each other (ITS: P < 0.0001; matK: P = 0.0077) (Table 3).
|
) measures character incongruence between data sets in terms of the extra homoplasy required to explain a specific alternative topology [the combined topology(ies)]. IMF ranges from zero, when the independent topologies are identical, to 1, when all homoplasy exists only between data sets and there is no homoplasy within data sets. The high IMF value we observed (0.703; P < 0.01) indicates that substantial homoplasy arises when the data are combined. DISCUSSION
Congruence
When appropriate, combining data sets increases the number of parsimony-informative characters, which may increase support for the "correct" phylogenetic tree. However, combining incongruent data sets may yield estimates of relationships that are erroneous. We have used a conditional combination approach (e.g., de Queiroz, 1993
; Bull et al., 1993
; Lutzoni and Vilgalys, 1995
; Huelsenbeck, Bull, and Cunningham, 1996
) in assessing the combinability of the ITS and matK data sets. The conditional combination approach attempts to reduce mistakes in estimating phylogeny by preventing the combination of heterogeneous data sets. Two issues are central to the conditional combination approach: (1) identification of sources of conflict, and (2) determination of what constitutes significant incongruence. We examined potential topological and character conflict between the two data sets using the CI1 and IMF metrics, respectively, and assessed the statistical significance of topological incongruence between the data sets with Templeton's (1983)
nonparametric statistical test of fit.
No taxonomic congruence between the ITS and matK trees, below the level of subgenus, was evident with visual examination (Fig. 3). Of the 29 clades present in the two semistrict consensus trees derived from the ITS and matK data sets, only those corresponding to the subgenera are identical (Fig. 3). CI1 quantifies resolved clades on consensus trees, that is, clades that are shared between trees. CI1 values for consensus trees derived from ITS and matK data indicate little topological congruence among trees within either data set (0.20 and 0.16, respectively) and considerably less among trees between data sets (0.08). Both data sets are plagued by limited numbers of parsimony-informative characters and moderate degrees of homoplasy, as is evident in the limited resolutions and low CI1 values. When the semistrict consensus trees from both the ITS and matK data sets were combined as either a strict or semistrict consensus tree, little congruence among the limited number of resolved clades present in either was evident.
IMF measures incongruence in terms of the extra homoplasy required to explain the shortest topologies recovered from analysis of the combined data set. The calculated IMF value of 70.3% indicates that conflict between the data sets is much greater than any conflict within either data set.
Templeton's (1983)
nonparametric statistical test of the ITS and matK trees demonstrated that they were significantly different with respect to the quality and quantity of substitutions inferred for both. The CI1 and IMF values, as well as the nonparametric test of fit, suggest that the ITS and matK data sets result in trees that contain conflicting information about the phylogenetic history of Ceanothus. To investigate potential sources of incongruence between the two data sets, we will examine several notable instances of incongruence evident between ITS and matK trees (Figs. 1, 2). In some instances, where there is insufficient resolution between the taxa being considered, we will use relationships summarized in a 50% majority rule tree (Hardig, unpublished data)
Ceanothus verrocosus and C. crassifolius (subgenus Cerastes) both have narrow ranges that overlap in southern California and northern Baja California. In the ITS analysis, C. verrocosus appears in the Coast clade, while C. crassifolius appears in a trichotomy near the base of subgenus Cerastes. In contrast, C. verrocosus and C. crassifolius occur as sister species in 93% of the matK trees (Hardig, unpublished data). The current sympatry and degree of phylogenetic divergence of C. crassifolius and C. verrocosus suggest that these species may share a similar matK lineage as a result of introgression, rather than as a retained character from a common ancestor; C. verrocosus apparently acquired the chloroplast genome of C. crassifolius.
Another potential example of introgression involves the species found in the North Coast clade of the matK tree (Fig. 2). This clade comprises individuals of four species, three of which (C. sonomensis, C. masonii, and C. purpureus) are endemic to the northern San Francisco Bay area and southern North Coast Ranges. The fourth member of this clade, C. cuneatus var. cuneatus, occurs throughout the CFP; the individual represented in the North Coast clade was collected in the southern region of the Outer North Coast Ranges, within 50 km of the other three individuals. In the ITS tree, C. cuneatus var. cuneatus and C. purpureus occur in separate subclades, while C. sonomensis and C. masonii share an identical sequence and occur in the basal polytomy of the Northwestern clade. Ceanothus masonii, C. sonomensis, and C. purpureus are all allopatric to each other within this region, but all are sympatric with the widespread C. cuneatus var. cuneatus. ITS and matK sequences were collected from two other individuals of C. cuneatus, one from southern Oregon and the other from the northern Sierra Nevada foothills. The ITS sequences from these three individuals are identical with respect to derived characters; however, the latter two individuals possess a matK lineage distinct from the one defining the North Coast clade (Fig. 2). Other individuals of C. sonomensis and C. masonii also lacked the North Coast clade matK synapomorphies. The matK lineage found in species of the North Coast clade has a limited geographic distribution and does not occur in all individuals of the species in the North Coast clade. The geographic proximity of the individuals represented in the North Coast clade suggests that the distribution of this matK lineage may represent the consequences of introgression. It is impossible to determine the ultimate origin of this lineage, or to reconstruct the path of its spread.
Ceanothus impressus occurs along the Central Coast of California, and C. foliosus var. medius occurs in the San Francisco Bay area and South Coast Ranges. These taxa are potentially parapatric along the adjoining margins of the Central Coast and South Coast Ranges. In the ITS tree, C. foliosus var. medius occurs as sister to C. foliosus var. vineatus (Fig. 1); C. impressus occurs in a trichotomy with C. tomentosus and C. oliganthus var. sorediatus in 90% of the ITS trees (Hardig, unpublished data) Ceanothus foliosus var. medius and C. impressus appear as sister taxa in 93% of the matK trees, which suggests the result of hybridization and chloroplast introgression. If introgression has occurred, it appears that C. foliosus var. medius has captured the chloroplast DNA (cpDNA) of C. impressus, forming a C. impressus/foliosus var. medius clade, with the exclusion of C. foliosus var. vineatus.
Ceanothus gloriosus var. exaltatus grows in the Outer North Coast Ranges and northern San Francisco Bay area. Ceanothus sonomensis, as previously discussed, is a narrow endemic from the southern portion of the Outer North Coast Ranges. Ceanothus gloriosus var. exaltatus forms the most strongly supported clade in the ITS tree and is well removed from C. sonomensis. Individuals of Ceanothus gloriosus var. exaltatus and C. sonomensis form a sister pair in 95% of the matK trees (Hardig, unpublished data), despite their divergent placements in the ITS tree, again suggesting that the similarity of their matK sequence is due to introgression; C. sonomensis has apparently acquired the chloroplast genome of C. gloriosus var. exaltatus.
Individuals of C. hearstiorum and C. lemmonii possess identical matK sequences and occur as sisters to C. cyaneus in 82% of the matK trees (Hardig, unpublished data). Ceanothus hearstiorum is a narrow endemic from near Arroyo de la Cruz in San Luis Obispo County, C. lemmonii grows on open, wooded slopes in the Inner North Coast Ranges, Cascade Range foothills, and northern and southern Sierra Nevada Foothills, and C. cyaneus grows on dry, shrubby slopes in the southern Peninsular Ranges. In the ITS tree, C. lemmonii occurs as a sister species to C. palmeri, while C. cyaneus is sister to C. parvifolius-leucodermis. In their current ranges, C. hearstiorum and C. lemmonii are allopatric by a minimum distance of 400 km, C. hearstiorum and C. cyaneus are allopatric by a minimum distance of 440 km, and C. cyaneus and C. lemmonii are allopatric by a minimum distance of 840 km. This degree of separation would seem to preclude any modern introgression of chloroplast genomes. The common matK sequence found in C. hearstiorum and C. lemmonii may be either retained from a common ancestor, as the polytomy containing both species in the ITS tree could potentially be resolved to make C. hearstiorum a sister taxon to the C. lemmonii-palmeri pair, or may have been introgressed at some earlier time when the ranges of both species might have been parapatric or sympatric.
Individuals of C. tomentosus, C. diversifolius, C. arboreus, C. foliosus var. vineatus, and C. velutinus var. velutinus also possess identical matK sequences. Ceanothus tomentosus, C. diversifolius, and C. velutinus var. velutinus are all wide-ranging species that are sympatric within the Sierra Nevada portion of their respective ranges. Ceanothus arboreus is endemic to the Channel Islands of California, and C. foliosus var. vineatus is endemic to the Outer North Coast Ranges of California. Ceanothus foliosus var. vineatus and C. diversifolius are potentially parapatric along the margin of the Outer and High North Coast Ranges of California, although they occur at different elevations (<300 m and 900–1800 m, respectively). Ceanothus tomentosus is sister to C. impressus in 90% of the ITS trees (Hardig, unpublished data), and C. foliosus var. vineatus is sister to C. foliosus var. medius (Fig. 1). The geographic and phylogenetic relationships among these species are too complex to allow a ready interpretation of matK sequence distribution.
Anecdotal tales of hybridization in Ceanothus are numerous, and the biosystematic study of Nobs (1963)
provides convincing evidence that there are no postzygotic barriers (e.g., infertility) to preclude continued hybridization and potential backcrossing between species within the subgenera. A lack of congruence between nuclear and organellar data sets has often been attributed to hybridization, particularly in plants (e.g., Neigel and Avise, 1986
; Pamilo and Nei, 1988
; Harrison, 1989
; Soltis et al., 1991
; Rieseberg, 1991
; Rieseberg and Wendel, 1993
; Soltis and Kuzoff, 1995
; Soltis, Johnson, and Looney, 1996
). Phylogenies based on gene sequences from a maternally inherited, clonal lineage (i.e., mitochondrial or chloroplast genes) reflect the evolutionary history of that lineage and will only indicate the organismal phylogeny in the absence of introgression. Reports of hybridization and introgression in many plant groups have increased significantly since the inception of cpDNA analysis, based primarily on the occurrence of unexpected or incongruent phylogenetic results between cpDNA and nuclear data (reviewed in Wendel and Doyle, 1998
). Numerous instances of chloroplast transfer between species via introgression have been documented in a wide variety of angiosperms (e.g., Rieseberg and Soltis, 1991
; Wendel and Doyle, 1998
). Thus, the degree of incongruence between trees from nuclear ITS and organellar matK phylogenies for Ceanothus is entirely concordant with a priori expectations based on the frequency of anecdotal reports of hybridization among species of Ceanothus (e.g., Parry, 1888a, b
; Brandegee, 1895
; Howell, 1940
; McMinn, 1942
; Nobs, 1963
). Introgression appears to be the most probable explanation for many of the instances of incongruence discussed above.
However, hybridization is not the only process that could produce such incongruence. If species of Ceanothus developed rapidly from a common ancestor that was polymorphic for a molecular character (e.g., matK or ITS sequences), then incipient species may possess different subsets of molecular diversity (Neigel and Avise, 1986
; Harrison, 1991
). With time and subsequent extinction of gene lineages, descendant populations will have randomly sorted out separate nuclear and organellar sequences, irrespective of the actual organismal phylogeny (i.e., stochastic lineage sorting). Although comparison of nuclear and cytoplasmic-based data has been proposed as one method of detecting introgression (e.g., Rieseberg and Brunsfeld, 1992
), it is impossible to distinguish unequivocally between introgression and lineage sorting based solely upon geographic and phylogenetic relationships. This problem is perhaps best illustrated by the example of identical matK sequences found in individuals of the allopatric species C. lemmonii and C. hearstiorum. The current degree of allopatry between theses species readily precludes any modern hybridization between them, although the possibility that they were historically parapatric or sympatric cannot be dismissed. In fact, the second alternative (lineage sorting) necessitates that both species were at one time parapatric, during the process of speciation and the apportioning of the ancestral matK sequences. Therefore, we conclude that the identical matK sequences found in C. lemmonii and C. hearstiorum indicate historical changes in the range of either or both species, but we cannot determine the historical event that is responsible for the modern distribution of those sequences. The problem of distinguishing between two alternative evolutionary processes (e.g., introgression and lineage sorting) that may give rise to similar phylogenetic patterns is a significant challenge.
Debate regarding the handling of data sets is ongoing. For example, DeSalle and Brower (1997)
, using restriction site and sequence data sets to analyze the relationships among four closely related species in the picture-winged species group of Hawaiian Drosophila, found significant incongruence between trees derived from the mitochondrial DNA restriction sites and nuclear alcohol dehydrogenase DNA sequence data; a combined analysis resulted in a tree identical to the tree derived from the mitochondrial data. DeSalle and Brower (1997)
state that they had no reason to dismiss the mitochondrial data based on empirical evidence of ancestral polymorphisms and lineage sorting, and they accept the results of the combined analysis as their best hypothesis of relationship, even though other nuclear gene regions supported a single alternative hypothesis. To critique their application of total evidence, DeSalle and Brower (1997)
suggest that opponents of combined analyses must " ... demonstrate that partitioned analyses are logically and practically superior to the simultaneous analysis approach, that something is gained by dividing data into process partitions for phylogenetic analysis." However, proponents of combined analyses must also demonstrate that nothing is being lost by combining process-partitioned data. For example, DeSalle and Brower's (1997)
statement regarding the absence of empirical evidence for ancestral polymorphisms and lineage sorting seems overly assertive, given that they used specimens of a single genetic line for each of the four species in the study, a sampling scheme that was unlikely to discover polymorphisms present in the wild types. DeSalle and Brower (1997)
also criticize the concept of process partitioning, based on lack of any objective defining criteria, suggesting that since partitioning could ostensibly be carried to a nearly infinitesimal degree it is imprudent to partition at all. We have found evidence of contemporary polymorphic cytoplasmic lineages, and artificial and natural hybrids in Ceanothus are extremely numerous. Both of these conditions allow for future episodes of introgression and/or lineage sorting. Despite problems associated with its application at other levels of organization, process partitioning of independently inherited, nonrecombining nucleus and cytoplasmic genomes seems natural. We cannot demonstrate that our partitioned analysis is logically and practically superior to a simultaneous analysis, but we can be certain that we are not compounding error by combining the phylogenetic data for two separate evolutionary histories.
Phylogenetic relationships
Selection of ITS phylogeny
Comparisons of inferred interspecific nuclear and organellar gene flow in Quercus, Helianthus, Salix, Heuchera, Gossypium, Populus, and Zea have uncovered evidence of cytoplasmic gene flow, without apparent nuclear gene flow (see Soltis and Soltis, 1995
, for review and references). Likewise, empirical evidence suggests that nuclear genes move across hybrid boundaries less freely than organellar genes (reviewed in Rieseberg and Wendel, 1993
). Furthermore, many of the incongruities between the ITS and matK trees for Ceanothus appear to reflect the interspecific transfer of chloroplasts following hybridization. Thus, we will use the ITS tree rather than the matK tree as our hypothesis of the phylogeny of Ceanothus.
Subgeneric relationships
Subgenera Ceanothus and Cerastes appear as monophyletic sister groups in the ITS-derived phylogeny (Fig. 1). The traditional taxonomic distinctions between subgenera are unequivocally and strongly supported by the ITS data. However, the ITS data set suffers from a lack of resolution below the level of subgenus. Sixty parsimony-informative characters, as present in the ITS data set, could resolve 61 taxa into a fully bifurcating tree (with low bootstrap values), given even distribution of characters across taxa and no homoplasy in the data set. However, uneven distribution of parsimony-informative characters across taxa and a moderate degree of homoplasy in the ITS data result in weakly supported clades with short internal branches and limited resolution. Subgenus Cerastes shows greater resolution than subgenus Ceanothus.
Paraphyletic species
The paraphyly of polytypic species (i.e., C. cuneatus, C. gloriosus, and C. jepsonii), based on ITS sequence data, may be attributable to a variety of causes. Due to the inferred sister relationship of C. cuneatus var. cuneatus and C. roderickii (Fig. 1), it is impossible to resolve a monophyletic C. cuneatus. Ceanothus roderickii is a narrow endemic reported only from the western portion of El Dorado County, California, growing on serpentine soils. Examination of morphological characters from herbarium individuals suggests a great deal of similarity in leaf morphology between C. roderickii and C. cuneatus var. cuneatus. If C. roderickii were recognized as a variety of C. cuneatus, then it could be possible to resolve the Northwestern clade polytomy to form a monophyletic C. cuneatus. We do not propose any such realignment at this time; a formal realignment must wait for a more detailed examination of genetic differences between the taxa. Ceanothus gloriosusvar. porrectus is an extremely narrow endemic growing on sandy soils along Inverness Ridge on Pt. Reyes Peninsula, California. This region is very isolated, separated from the remaining varieties of C. gloriosus by Tomales Bay and the San Andreas Rift Zone. Two possibilities may be advanced to explain the paraphyly of C. gloriosus: either C. gloriosus var. porrectus has acquired a foreign ITS sequence via hybridization and introgression, or its inferred relationship with the remainder of C. gloriosus has been in error. Ceanothus jepsonii var. jepsonii occurs inside the C. gloriosus clade, while C. jepsonii var. albiflorus occurs in the Coast clade. The C. jepsonii var. jepsonii sequence appears to be a recombinant sequence possessing two of the four synapomorphies that unite C. gloriosus var. exaltatus and C. gloriosus var. gloriosus (Table 4). Ceanothus jepsonii var. jepsonii and C. gloriosus var. exaltatus both grow in the Outer North Coast Ranges and northern San Francisco Bay Area of California; the individual of C. jepsonii var. jepsonii sampled was collected from a serpentine road bank near Alpine Lake, Marin County, California, a region known for many Ceanothus hybrids (R. Raiche, U. C. Berkeley Botanical Garden, personal communication). Although this individual of C. jepsonii var. jepsonii is morphologically similar to other individuals of C. jepsonii var. jepsonii, it is apparently a descendant of a hybridization event, and its placement within the ITS tree should be considered with caution. It is also possible that these instances of paraphyly are artifacts created by incongruence between the evolution of genes and the evolution of species.
|
; Brooks and McLennan, 1991
). Speciation via peripheral isolation can produce an apomorphic derivative species and a paraphyletic progenitor species (Frey, 1993
). Hedin (1997)
uncovered several instances of paraphyly in both nuclear and mitochondrial gene phylogenies of the North American cave spider Nesticus. The paraphyletic species of Nesticus appear to be the result of peripatric speciation, and derivation of daughter species via microvicariance or habitat colonization resulted in a phylogenetic subsampling of parental genetic variation, leaving a progenitor species that is paraphyletic with respect to the derivative species (Hedin, 1997
). Numerous species of Ceanothus demonstrate narrow habitat ranges and patchy distributions, thus allowing for the possibility of speciation via peripheral isolation and the consequent formation of paraphyletic progenitor species.
Conclusion
The data support equivalent evolutionary ranks for the two subgenera, in contradiction to Mason's (1942)
proposed origin of subgenus Cerastes from within subgenus Ceanothus.
Limited phylogenetic resolution in both the ITS and matK data sets is due, in large part, to a limited number of parsimony-informative characters, relative to the number of taxa, and suggests a relatively rapid radiation of Ceanothus within North America. Combining data sets to increase resolution by increasing the number of parsimony-informative characters available was not attempted because of the high levels of incongruence between the separate data sets. Some instances of incongruence appear to be the result of introgression (e.g., C. verrocosus and C. crassifolius), while other instances are less convincing (e.g., C. lemmonii and C. hearstiorum). Other phenomena, such as the sorting of ancestral polymorphic lineages in descendant species, may produce similar phylogenetic patterns and may be responsible for some of the observed incongruities. Several modern species of Ceanothus are polymorphic for matK sequences (i.e., C. cuneatus var. cuneatus, C. sonomensis, and C. masonii); therefore, these species hold the potential for future episode(s) of speciation, with subsequent lineage sorting among the descendants. Despite the limited resolution and weak statistical support for clades evident in the ITS phylogeny, many clades appear to correspond to specific geographic areas and share suites of morphological characters (Hardig, Soltis, and Soltis, unpublished data).
Results of the phylogenetic analyses presented here have provided a historical context to the diversity found in Ceanothus, and have identified potentially monophyletic subgroups (e.g., the species forming the Northwestern clade in the ITS consensus tree) for future studies of adaptive (e.g., morphological and physiological character-trait evolution) and coevolutionary (e.g., actinomycete symbiosis) processes.
|
1 The authors thank Clifford Schmidt, Holly Forbes, and Roger Raiche for assistance in this project; the University of California at Berkeley Botanical Garden, the Rancho Santa Ana Botanic Garden, the Santa Barbara Botanic Garden, and the Quail Botanical Garden for providing plant samples; Steven J. Brunsfeld, John N. Thompson, and Leigh A. Johnson for helpful comments on earlier drafts; and Dave Swofford for providing a test version of PAUP* 4.0. This work was supported in part by NSF grant BIR-9512890 to PSS, DES, et al., and by grants from Sigma Xi, ASPT, and the Betty W. Higinbotham Trust of Washington State University. 
2 Author for correspondence, current address: Department of Biology and Chemistry, University of Montevallo, Station 6480, Montevallo, Alabama 35115 USA.
LITERATURE CITED
Baker, R. H., and R. DeSalle. 1997 Multiple sources of character information and the phylogeny of Hawaiian Drosophilids. Systematic Biology 46: 654–673. [CrossRef][Web of Science][Medline]
Barrett, M., M. J. Donoghue, and E. Sober. 1991 Against consensus. Systematic Zoology 40: 486–493. [CrossRef]
Baum, D. A., K. J. Sytsma, and P. C. Hoch. 1994 A phylogenetic analysis of Epilobium (Onagraceae) based on nuclear ribosomal DNA sequences. Systematic Botany 19: 363–388. [CrossRef][Web of Science]
Bledsoe, A. H., and R. J. Raikow. 1990 A quantitative assessment of congruence between molecular and nonmolecular estimates of phylogeny. Journal of Molecular Evolution 30: 247–259.
Brandegee, K. 1895 Studies in Ceanothus. Proceedings of the California Academy of Sciences 4: 173–222.
Bremer, K. 1990 Combinable component consensus. Cladistics 6: 369–372. [CrossRef][Web of Science]
Brooks, D. R., and D. A. McLennan. 1991 Phylogeny, ecology, and behavior: a research program in comparative biology. University of Chicago Press, Chicago, Ilinois, USA.
Bull, J. J., J. P. Huelsenbeck, C. W. Cunningham, D. L., Swofford, and P. J. Waddell. 1993 Partitioning and combining data in phylogenetic analysis. Systematic Biology 42: 384–397. [CrossRef][Web of Science]
Chippindale, P. T., and J. J. Wiens. 1994 Weighting, partitioning, and combining characters in phylogenetic analysis. Systematic Biology 43: 278–287. [CrossRef][Web of Science]
Cullings, K. W. 1992 Design and testing of a plant-specific PCR primer for ecological and evolutionary studies. Molecular Ecology 1: 233–240.
de Queiroz, A. 1993 For consensus (sometimes). Systematic Biology 42: 368–372. [CrossRef][Web of Science]
———, M. J. Donoghue, and J. Kim. 1995 Separate versus combined analysis of phylogenetic evidence. Annual Review of Ecology and Systematics 26: 657–681. [Web of Science]
DeSalle, R., and A. V. Z. Brower. 1997 Process partitions, congruence, and the independence of characters: inferring relationships among closely related Hawaiian Drosophila from multiple gene regions. Systematic Biology 46: 751–764. [CrossRef][Web of Science][Medline]
Donoghue, M. J., and M. J. Sanderson. 1992 The suitability of molecular and morphological evidence in reconstructing plant phylogeny. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants. Chapman and Hall, New York, New York, USA.
Doyle, J. J., and J. L. Doyle. 1987 A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.
Eernisse, D. J., and A. G. Kluge. 1993 Taxonomic congruence versus total evidence, and the phylogeny of amniotes inferred from fossils, molecules, and morphology. Molecular Biology and Evolution 10: 1170–1195. [Abstract]
Farris, J. S., M. Källersjö, A. G. Kluge, and C. Bult. 1995 Testing significance of incongruence. Cladistics 10: 315–319. [CrossRef][Web of Science]
Felsenstein, J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. [CrossRef][Web of Science]
Ferris, S. D., R. D. Sage, C. M. Huang, J. T. Nielson, U. Ritte, and A. C. Wilson. 1983 Flow of mitochondrial DNA across a species boundary. Proceedings of the National Academy of Sciences, USA 80: 2290–2294.
Frey, J. K. 1993 Modes of peripheral isolate formation and speciation. Systematic Botany 42: 373–381.
Grant, V. 1981 Plant speciation. Columbia University Press, New York, New York, USA.
Gyllensten, U., and A. C. Wilson. 1987 Interspecific mitochondrial DNA transfer and the colonization of Scandinavia by mice. Genetic Research 49: 25–29.
Harrison, R. G. 1989 Animal mitochondrial DNA as a genetic marker in population and evolutionary biology. Trends in Ecology and Evolution 4: 6–11.
———. 1991 Molecular changes at speciation. Annual Review of Ecology and Systematics 22: 281–308.
———, D. M. Rand, and W. C. Wheeler. 1987 Mitochondrial DNA variation in field crickets across a narrow hybrid zone. Molecular Biology and Evolution 4: 144–158. [Web of Science]
Hedin, M. C. 1997 Speciational history in a diverse clade of habitat-specialized spiders (Araneae: Nesticidae: Nesticus): inferences from geographic-based sampling. Evolution 51: 1929–1945. [CrossRef][Web of Science]
Howell, J. T. 1940 Studies in Ceanothus IV. Leaflets of Western Botany 2(15): 259–262.
Huelsenbeck, J. P., and J. J. Bull. 1996 A likelihood ratio test for detection of conflicting phylogenetic signal. Systematic Biology 45: 92–98. [CrossRef][Web of Science]
———, ———, and C. W. Cunningham. 1996 Combining data in phylogenetic analysis. Trends in Ecology and Evolution 11: 152–158. [CrossRef]
Jeong, S.-C., A. Liston, and D. D. Myrold. 1997 Molecular phylogeny of the genus Ceanothus (Rhamnaceae) using rbcL and ndhF sequences. Theoretical and Applied Genetics 94: 852–857. [CrossRef][Web of Science]
Johnson, L. A., and D. E. Soltis. 1994 matK DNA sequences and phylogenetic reconstruction in Saxifragaceae s.s. Systematic Botany 19: 143–156.
———, and ———. 1995 Phylogenetic inference in Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK sequences. Annals of the Missouri Botanical Garden 82: 149–171. [CrossRef][Web of Science]
———, and ———. 1998 Assessing congruence: empirical examples from molecular data. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.] Molecular systematics of plants. Chapman and Hall. New York, New York, USA.
Kluge, A. G. 1989 A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Systematic Zoology 38: 7–25.
———, and A. J. Wolf. 1993 Cladistics: what's in a word? Cladistics 9: 183–199. [CrossRef][Web of Science]
Lanyon, S. M. 1993 Phylogenetic frameworks: towards a firmer foundation for the comparative approach. Biological Journal of the Linnean Society 49: 45–61.
Lutzoni, F., and R. Vilgalys. 1995 Integration of morphological and molecular data sets in estimating fungal phylogenies. Canadian Journal of Botany 73 (Supplement 1): S649–659.
Maddison, W. P., and D. R. Maddison. 1992 MacClade: analysis of phylogeny and character evolution, version 3.0. Sinauer, Sunderland, Massachusetts, USA.
Mason, H. L. 1942 Distributional history and fossil record of Ceanothus. In Ceanothus. Santa Barbara Botanical Gardens, Santa Barbara, California, USA.
McMinn, H. E. 1930 A geographic and taxonomic study of the California species of the genus Ceanothus. Contributions from the Dudley Herbarium of Stanford University 1: 121–147.
———. 1942 A systematic study of the genus Ceanothus. In Ceanothus. Santa Barbara Botanical Gardens, Santa Barbara, California, USA.
Mickevich, M. F., and J. S. Farris. 1981 The implications of congruence in Menidia. Systematic Zoology 30: 351–370.
Miyamoto, M. M., and W. M. Fitch. 1995 Testing species phylogenies and phylogenetic methods with congruence. Systematic Biology 44: 64–76.
Morgan, D. R., and D. S. Soltis. 1993 Phylogenetic relationships among members of Saxifragaceae s.l. based on rbcL sequence data. Annals of the Missouri Botanical Garden 80: 631–660. [CrossRef][Web of Science]
Neigel, J. E., and J. C. Avise. 1986 Phylogenetic relationships of mitochondrial DNA under various demographic models of speciation. In S. Karlin and E. Nevo [eds.], Evolutionary processes and theory. Academic Press, New York, New York, USA.
Nobs, M. A. 1963 Experimental studies on species relationships in Ceanothus. Carnegie Institute of Washington. Washington, DC.
Pamilo, P., and M. Nei. 1988 Relationships between gene trees and species trees. Molecular Biology and Evolution 5: 568–583. [Abstract]
Parry, C. C. 1888a The North-American genus, Ceanothus, with an enumerated list, and notes and descriptions of several Pacific Coast species. Davenport Academy of Natural Sciences 5: 162–174.
———. 1888b Ceanothus, L. recent field notes, with a partial revision of species. Davenport Academy of Natural Sciences 5: 162–174.
Raven, P. H., R. F. Evert, and S. E. Eichhorn. 1992 Biology of plants, 5th. ed. Worth, New York, New York, USA.
Richardson, J. E., M. F. Fay, C. B. Quentin, Q. C. B. Cronk, D. Bowman, and M. W. Chase. 1997 A molecular analysis of Rhamnaceae using rbcL and trnL-F plastid DNA sequences. American Journal of Botany. 84: 226.
Rieseberg, L. H. 1991 Homoploid reticulate evolution in Helianthus: evidence from ribosomal genes. American Journal of Botany 78: 1218–1237. [CrossRef][Web of Science]
———, and S. J. Brunsfeld. 1992 Molecular evidence and plant introgression. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants. Chapman and Hall, New York, New York, USA.
———, and D. E. Soltis. 1991 Phylogenetic consequences of cytoplasmic gene flow in plants. Evolutionary Trends in Plants 5: 65–84. [Web of Science]
———, and J. F. Wendel. 1993 Introgression and its consequences in plants. In R. G. Harrison [ed.], Hybrid zones and the evolutionary process. Oxford University Press, New York, New York, USA.
Rohlf, F. J. 1982 Consensus indices for comparing classifications. Mathematical Biosciences: 59: 131–144.
Rose, S., B. A. Daniels, and J. M. Trappe. 1979 Glumus gerdemannii sp. nov. Mycotaxon 8: 297–301. [Web of Science]
Soltis, D. E., L. A. Johnson, and C. Looney. 1996 Discordance between ITS and chloroplast topologies in the Boykinia group (Saxifragaceae). Systematic Botany 21: 169–176. [CrossRef][Web of Science]
———, and R. K. Kuzoff. 1995 Discordance between molecular and chloroplast phylogenies in the Heuchera group (Saxifragaceae). Evolution 49: 727–742. [CrossRef][Web of Science]
———, P. S. Soltis, T. G. Collier, and M. L. Edgerton. 1991 variation within and among genera of the Heuchera group: evidence of chloroplast capture and paraphyly. American Journal of Botany 78: 1091–1112. [CrossRef][Web of Science]
———, M. E. Mort, M. W. Chase, V. Savolainen, S. B. Hoot, and C. M. Morton. 1998 Inferring complex phylogenies using parsimony: an empirical approach using three large DNA data sets for angiosperms. Systematic Biology 47: 32–42. [CrossRef][Web of Science][Medline]
Soltis, P. S., and D. E. Soltis. 1995 Plant molecular systematics: inferences of phylogeny and evolutionary processes. In M. K. Hecht [ed.], Evolutionary Biology, vol. 28, 139–194, Plenum Press, New York, New York, USA.
Stebbins, G. L. 1950 Variation and evolution in plants. Columbia University Press, New York, New York, USA.
Suh, Y., L. B. Tien, H. E. Reeve, and E. A. Zimmer. 1993 Molecular evolution and phylogenetic implications of internal transcribed spacer sequences of ribosomal DNA in Winteraceae. American Journal of Botany 80: 1042–1055. [CrossRef][Web of Science]
Tegelström, H. 1987 Transfer of mitochondrial DNA from the Northern Red-Backed Vole (Clethrionomys rutilus) to the Bank Vole (C. glareolus). Journal of Molecular Evolution 24: 218–227. [CrossRef][Web of Science][Medline]
Templeton, A. R. 1983 Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37: 221–244. [CrossRef][Web of Science]
Watson, S. 1875 Revision of the genus Ceanothus, and descriptions of new plants, with a synopsis of the western species of Silene. Proceedings of the American Academy of Arts and Sciences 10:333–339.
Wendel, J. F., and J. J. Doyle. 1998 Phylogenetic incongruence: window into genome history and speciation. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants. Chapman and Hall. New York, New York, USA.
White, T. J., T. Bruns, S. Lee, and J. Taylor. 1990 Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In M. Innis, D. Gelfand, J. Sninsky, and T. White [eds.], PCR protocols: a guide to methods and applications. Academic Press, San Diego, California, USA.
Wiley, E. O. 1981 Phylogenetics: the theory and practice of phylogenetic systematics. John Wiley & Sons, New York, New York, USA.
This article has been cited by other articles:
|
|
 |
|
  Y. JI, P. W. FRITSCH, H. LI, T. XIAO, and Z. ZHOU Phylogeny and Classification of Paris (Melanthiaceae) Inferred from DNA Sequence Data Ann. Bot., July 1, 2006; 98(1): 245 - 256. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. W. Schwilk and D. D. Ackerly Is there a cost to resprouting? Seedling growth rate and drought tolerance in sprouting and nonsprouting Ceanothus (Rhamnaceae) Am. J. Botany, March 1, 2005; 92(3): 404 - 410. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Shaw, E. B. Lickey, J. T. Beck, S. B. Farmer, W. Liu, J. Miller, K. C. Siripun, C. T. Winder, E. E. Schilling, and R. L. Small The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis Am. J. Botany, January 1, 2005; 92(1): 142 - 166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. H. Degnan, A. B. Lazarus, C. D. Brock, and J. J. Wernegreen Host-Symbiont Stability and Fast Evolutionary Rates in an Ant-Bacterium Association: Cospeciation of Camponotus Species and Their Endosymbionts, Candidatus Blochmannia Syst Biol, February 1, 2004; 53(1): 95 - 110. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. A. Preston and D. D. Ackerly Hydraulic architecture and the evolution of shoot allometry in contrasting climates Am. J. Botany, October 1, 2003; 90(10): 1502 - 1512. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Grant Incongruence between cladistic and taxonomic systems Am. J. Botany, September 1, 2003; 90(9): 1263 - 1270. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. C. Muschner, A. P. Lorenz, A. C. Cervi, S. L. Bonatto, T. T. Souza-Chies, F. M. Salzano, and L. B. Freitas A first molecular phylogenetic analysis of Passiflora (Passifloraceae) Am. J. Botany, August 1, 2003; 90(8): 1229 - 1238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Perret, A. Chautems, R. Spichiger, G. Kite, and V. Savolainen Systematics and evolution of tribe Sinningieae (Gesneriaceae): evidence from phylogenetic analyses of six plastid DNA regions and nuclear ncpGS Am. J. Botany, March 1, 2003; 90(3): 445 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Ferguson and R. K. Jansen A chloroplast DNA phylogeny of eastern Phlox (Polemoniaceae): implications of congruence and incongruence with the ITS phylogeny Am. J. Botany, August 1, 2002; 89(8): 1324 - 1335. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.-O. Yoo, P. P. Lowry II, and J. Wen Discordance of chloroplast and nuclear ribosomal DNA data in Osmorhiza (Apiaceae) Am. J. Botany, June 1, 2002; 89(6): 966 - 971. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Gravendeel, M. W. Chase, E. F. de Vogel, M. C. Roos, T. H. M. Mes, and K. Bachmann Molecular phylogeny of Coelogyne (Epidendroideae; Orchidaceae) based on plastid RFLPS, matK, and nuclear ribosomal ITS sequences: evidence for polyphyly Am. J. Botany, October 1, 2001; 88(10): 1915 - 1927. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Knight and D. D. Ackerly Correlated evolution of chloroplast heat shock protein expression in closely related plant species Am. J. Botany, March 1, 2001; 88(3): 411 - 418. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
