Systematic Botany (2006), 31(2): pp. 380–397 䉷 Copyright 2006 by the American Society of Plant Taxonomists

The Mimulus moschatus Alliance (Phrymaceae): Molecular and Morphological Phylogenetics and their Conservation Implications JUSTEN B. WHITTALL,1,4 MATTHEW L. CARLSON,2 PAUL M. BEARDSLEY,3 ROBERT J. MEINKE,1 and AARON LISTON1 Department of Botany and Plant Pathology, 2082 Cordley Hall, Oregon State University, Corvallis, Oregon 97331-2902, U.S.A.; 2 Alaska Natural Heritage Program, Environment and Natural Resources Institute and Department of Biological Sciences, 707 A Street, Anchorage, Alaska 99501, U.S.A.; 3 Department of Biological Sciences, Idaho State University, Pocatello, Idaho 83209, U.S.A.; 4 Author for correspondence, present address: Section of Evolution and Ecology, University of California, One Shields Avenue, Davis, California 95616, U.S.A. ([email protected]) 1

Communicating Editor: Sara B. Hoot ABSTRACT. The Mimulus moschatus alliance consists of 13 morphologically similar species, the majority of which have been considered for conservation protection. Phylogenetic analyses of four rapidly evolving molecular DNA regions (ITS, ETS, trnL-F, and rpl16) and a morphological data set under several optimality criteria reveal that the M. moschatus alliance is composed of three geographically defined clades: the Sierra Nevada Clade (M. floribundus, M. norrisii, and M. dudleyi), the Snake River Clade (M. hymenophyllus, M. ampliatus, and M. patulus), and the Columbia River Clade (M. washingtonensis and M. jungermannioides). The relationships within and among the clades are well resolved. Numerous instances of morphological homoplasy among the clades are inferred, including three independent origins of the autogamous mating system. Although nearly half of the morphological characters are highly homoplasious, the inclusion of morphological data in the combined maximum parsimony and Bayesian analyses improves topological resolution and branch support. The phylogenetic results support the specific recognition of three rare taxa (M. ampliatus, M. patulus, and M. dudleyi), previously synonymized with more widespread species. A key to the species within the M. moschatus alliance is provided. KEYWORDS: autogamy, ETS, homoplasy, ITS, rpl16, trnL-F.

Mimulus L. is a model genus for understanding several fundamental evolutionary processes such as speciation (Schemske and Bradshaw 1999; Fishman et al. 2001; Bradshaw and Schemske 2003; Ramsey et al. 2003) and mating system evolution (Fishman et al. 2002; Ivey et al. 2004; Karron et al. 2004). Intensive studies of a few taxa or closely related species have made significant contributions to our understanding of evolutionary biology, yet Mimulus is a large genus harboring numerous species complexes characterized by phylogenetic uncertainty, morphological ambiguity, and a history of taxonomic confusion (Grant 1924; Beardsley et al. 2004). Determining interspecific relationships for such complexes provides a foundation for an evolutionary-based taxonomy and explicit tests of character evolution. Appropriate taxonomic circumscription is particularly important for species of conservation concern. Several morphologically complex, yet geographically restricted, species complexes are concentrated in Mimulus section Paradanthus (Grant 1924; Beardsley et al. 2004). This section was originally designated as an artificial grouping of several unrelated species complexes (Grant 1924). One such complex, the M. moschatus alliance, is a group of 13 closely related species with uncertain species boundaries and interspecific relationships (Grant 1924; Pennell 1951; Argue 1986; Meinke 1992; Whittall 1999; Carlson 2002). These taxa are unit-

ed by a characteristic viscid pubescence, small calyx teeth of equal length, and acrescent peduncles in fruit (Grant 1924; Argue 1986; Meinke 1992). This alliance exhibits an exceptionally high proportion of rare species centered in western North America (Fig. 1): 10 of the 12 sampled species have been assigned conservation rankings (Appendix 1; Idaho Conservation Data Center 2003; California Natural Diversity Database 2004; Montana Natural Heritage Program 2004; Oregon Natural Heritage Information Center 2004; Washington Natural Heritage Program 2004). The conservation status for several species in the M. moschatus alliance remains tentative due to uncertain taxonomic rankings (Meinke 1992; Thompson 1992; Idaho Conservation Data Center 2003; Washington Natural Heritage Program 2004). For example, M. ampliatus is the rarest species in this alliance, known from only four populations in the Snake River region of western Idaho (Idaho Conservation Data Center 2003). It has been allied with M. washingtonensis based on morphological characteristics (Meinke 1992), but its conservation status has awaited additional phylogenetic information to determine its evolutionary affinities and molecular distinctiveness (Meinke 1992; Idaho Conservation Data Center 2003). The M. moschatus alliance was recently included in a comprehensive molecular phylogenetic study of Mimulus (Beardsley et al. 2004). A maximum parsimony

380

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

381

FIG. 1. Geographic ranges of the species of the M. moschatus alliance in western North America. A. Distributions for the rare species are indicated as follows: M. ampliatus (vertical hatching), M. hymenophyllus (black squares), M. patulus (horizontal hatching), M. jungermannioides (light grey shading), M. washingtonensis (dark grey shading), M. evanescens (stars), M. dudleyi (diagonal hatching), and M. norrisii (open diamonds). B. Range limits for the widespread species are M. floribundus (evenly dashed line), M. moschatus (long-short dashed line), M. breviflorus (dotted line), M. pulsiferae (grey solid line), and M. latidens (black solid line).

analysis of three regions (ITS, ETS, and trnL-F) from a nearly complete sampling of the M. moschatus alliance (except M. arenarius Grant) confirmed the monophyly of this complex, while several other species complexes in section Paradanthus were polyphyletic (Beardsley et al. 2004). This investigation also supported the hypothesis of a hybrid origin for M. evanescens suggested

previously by Meinke (1995). The M. moschatus alliance is comprised of three well supported geographically distinct clades as previously identified by Whittall (1999); Snake River clade ⫽ SR; Columbia River clade ⫽ CR; Sierra Nevada clade ⫽ SN. The phylogenetic backbone of the M. moschatus alliance suggested several independent origins of the autogamous mating

382

[Volume 31

SYSTEMATIC BOTANY

FIG. 2. Phylogenetic hypothesis for the M. moschatus alliance based on a MP analysis of the ITS, ETS, and trnL-F regions (redrawn from Beardsley et al. 2004 and excluding M. evanescens). The M. moschatus alliance (labeled MoA) is monophyletic and consists of three geographic clades: the Snake River clade (SR), the Columbia River clade (CR), and the Sierra Nevada clade (SN). Outgroups are labeled ‘‘OG’’. BS percentages are indicated at the nodes when greater than 70%. Branches with BS ⬍ 70% are collapsed.

system, cliff dwelling habit, and restricted geographic ranges. Further investigations into patterns of character evolution, convergent mating system shifts, and the evolutionary patterns of range size differences have awaited increased phylogenetic resolution within the M. moschatus alliance. The relationships among the three geographically cohesive clades and two widespread taxa, M. pulsiferae and M. moschatus, remained unresolved (Fig. 2; Beardsley et al. 2004). The placement of widespread taxa are critical for testing evolutionary hypotheses such as whether particular characters are associated with the origins of rarity or its converse (Carlson 2002). In addition, the relationships within the SR clade also were unresolved (Fig. 2; Beardsley et al. 2004), obstructing our understanding of putative mating system shifts (Carlson 2002). Additional data and more complex models of DNA sequence evolution may improve our ability to accurately estimate relationships within the M. moschatus alliance. Therefore, we have added the rapidly evolving chloroplast intron from the ribosomal protein gene L16 (rpl16) to supplement the preexisting nuclear rDNA (ITS, ETS) and chloroplast (trnL-F) molecular data of Beardsley et al. (2004). The rpl16 intron has provided interspecific phylogenetic resolution in other closely related species complexes (Jordan et al. 1996; Small et al. 1998). Furthermore, the reduced number of taxa in this data and the development of new analytical tools allow us to more rigorously examine our phylogenetic results under different optimality criteria

not previously possible due to computational limitations (Beardsley et al. 2004). Specifically, the development of increasingly more complex models of sequence evolution implemented in maximum likelihood (ML) and Bayesian MCMC analyses provides powerful new approaches to combined phylogenetic analyses (Nylander et al. 2004). A unique feature only implemented in some Bayesian MCMC software (MrBayes) is the application of data specific models of sequence evolution that can now be independently estimated and partitioned in an efficient and statistically-based combined phylogenetic analysis (Ronquist and Huelsenbeck 2003). The utility of morphological data in phylogenetic reconstruction has been debated partly because of the high levels of homoplasy associated with some morphological characters (Wiens 2004). In order to determine the amount of homoplasy in the morphological data, one needs the ‘‘true’’ phylogeny. Yet, the correct phylogeny is rarely known with certainty. Instead, one can estimate the degree of morphological homoplasy by comparing standard character indices (i.e., consistence index or retention index) when the morphological data are mapped onto a molecular phylogeny (Gernandt et al. 2003). This approach can illuminate the types of characters that are highly homoplasious, provide insights into patterns of morphological evolution, and help identify clade-distinguishing morphological synapomorphies. With an estimate of the degree of morphological homoplasy, the phylogenetic effects of combining morphological and molecular data can be examined. Recent results suggest that even though morphological characters may be less abundant and more homoplasious than the molecular data, they can improve topological resolution and branch support values in combined analyses (Prather 1999; Wiens et al. 2003; Swigonova and Kjer 2004). Therefore, we have examined the phylogenetic effects of combining the morphological and molecular data in analyses of the M. moschatus alliance. This study describes a phylogenetic investigation of the M. moschatus alliance using DNA sequence data from the nuclear rDNA ITS and ETS regions, the chloroplast trnL-F intron and spacer and rpl16 intron, and morphological data. We compare the phylogenetic results from analyses of the individual datasets, the combined molecular data, and the combined molecular and morphological data. The resulting phylogenetic inferences are used to interpret patterns of morphological evolution. Taxonomic implications and conservation recommendations are inferred from the phylogeny. A key to the species within the M. moschatus alliance is provided. MATERIALS

AND

METHODS

Taxonomic Sampling. Taxon sampling followed a broad interpretation of the M. moschatus alliance consisting of 13 taxa (Ap-

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

pendix 1; Grant 1924; Argue 1986; Meinke 1992; Carlson 2002). Multiple outgroups were selected to verify the monophyly of the M. moschatus alliance. These included the hypothesized closest relatives from section Paradanthus (M. latidens, M. bodinieri, M. tenellus, M. nepalensis, M. dentatus, and M. alsinoides) and other morphologically similar taxa from sections Simiolus (M. guttatus) and Erythranthe/Paradanthus Clade H (M. lewisii, M. primuloides, M. bicolor; Beardsley et al. 2004). Mimulus arenarius was the only species in the alliance not included due to failure to extract amplifiable DNA from herbarium specimens and difficulty in relocating historical populations. DNA extraction and PCR conditions of the ITS, ETS, and trnL-F regions were as described previously (Whittall 1999; Beardsley et al. 2004). Identical DNA samples were used for sequencing the rpl16 intron (Appendix 1). Mimulus evanescens, a member of the M. moschatus alliance, has been identified as a hybrid taxon between M. latidens and M. breviflorus (Meinke 1995; Beardsley et al. 2004). The rpl16 intron was sequenced from this taxon to verify its chloroplast haplotype, and was excluded from all analyses thereafter. Morphological Data. Fifty three morphological characters were scored from fresh material and herbarium specimens (OSC, UCSB; Appendix 2). Data were supplemented using taxonomic descriptions when necessary (Grant 1924; Heckard and Shevock 1985; Meinke 1992; Thompson 1992; Beardsley et al. 2004). The same outgroup taxa included in the molecular data were scored for the 53 morphological characters. Characters were divided into three categories: general (N⫽6), vegetative (N⫽21), and reproductive (N⫽26). Only parsimony informative characters were scored and taxa with multiple character states were considered polymorphic. All characters were considered unordered and all but nine characters were binary. Rpl16 Sequencing. Amplification and sequencing of the rpl16 intron followed the ITS protocol described in Whittall et al. (1999) except for the following: each PCR reaction (50 ␮L) contained a final concentration of 1.5 mM MgCl2, 200 ␮M each dNTP, 1⫻ Taq polymerase buffer, 1⫻ MasterAmp PCR Enhancer (Epicentre Technologies), 50 pmol of primers F71 and R1661 (Jordan et al. 1996; Kelchner and Clark 1997), and 1 U Taq polymerase (Epicentre Technologies). Thermocycling conditions followed Jordan et al. (1996). Approximately 30 ng purified rpl16 PCR products were cycle sequenced following the BigDye Terminator protocol (Applied Biosystems, Inc.) and separated on an ABI Prism 377 (Applied Biosystems). Sequencing was completed with the forward PCR primer and an internal reverse primer, R1516 (Kelchner and Clark 1997). Sequences were compiled using the GAP 4.0 editor (Bonfield et al. 1995) and manually aligned in GCG10 (Genetics Computer Group 1999). Gaps were coded according to a simple model (Simmons and Ochoterena 2000) in which all gaps that have different 5⬘ and/or 3⬘ ends were scored as separate characters. Although gaps often provide reliable characters for phylogenetic analyses, we used a conservative approach by mapping synapomorphic gaps onto the phylogeny. Phylogenetic Analysis of Individual Data. DNA sequence alignments from the ITS, ETS, and trnL-F regions are those of Beardsley et al. (2004). Alignments of all four DNA sequence regions are available from TreeBASE (study accession no. S1326). Alignment lengths are: ETS ⫽ 446; ITS ⫽ 633; trnL-F ⫽ 932; and rpl16 ⫽ 936bp. There are no missing accessions for the ITS and ETS data. Missing outgroup accessions for the trnL-F include M. dentatus, M. tenellus, and M. bodinieri. For the rpl16 data, M. floribundus CO, M. tenellus, M. bodinieri, M. nepalensis, and M. guttatus CA were not included. Missing cells total 282 for ETS (2.63%), 213 cells for ITS (1.40%), 258 ⫹ three missing accessions for trnL-F (total missing cells ⫽ 13.65%), and five missing accessions for rpl16 (20.83%). The morphological data of 53 characters contained 18 polymorphic cells (1.89%) and 93 missing cells (7.31%), but there were only 2.7% missing cells among the 12 members of the M. moschatus alliance used in the phylogenetic analyses (Appendix 3). The maximum parsimony (MP) analyses consisted of a branchand-bound search with the initial MAXTREES set to 1100, then

383

allowed to automatically increase until completed. For maximum likelihood, models of sequence evolution were determined using the likelihood ratio tests provided in Modeltest v.3.06 (Posada and Crandall 1998) after removing taxa with large blocks of missing data since these sites are excluded for all samples during the model-fitting process. All four sequence regions were analyzed individually with ML using the appropriate model in PAUP*4.0b10 (Swofford 1998). Heuristic searches were conducted with ADDSEQ ⫽ AS-IS and MAXTREES ⫽ 100, but allowed to automatically increase as needed. Bootstrap support (BS) was determined from 1000 replicates implemented in PAUP*4.0b10 for the MP and ML analyses with the same search options as described for ML. The morphological data was analyzed under MP as described above with a similar BS methodology. Phylogenetic Analysis of Combined Molecular Data. For the combined molecular data, there were 8209 missing cells (11.42%). The only missing accession in the combined molecular analysis for the M. moschatus alliance was a duplicate accession, M. floribundus CO, for the rpl16 region. MP and ML analyses of the combined data were conducted using PAUP*4.0b10 (Swofford 1998) following the methodologies described for the analysis of the separate regions (see above). ML analysis on the entire four gene data (unpartitioned) was conducted using the best model determined from a series of hierarchical likelihood ratio tests as implemented in Modeltest v.3.06 (Posada and Crandall 1998). Bayesian analysis provides an alternative approach to inferring phylogenies, yet it is not an uncontroversial method (Buckley 2002; Suzuki et al. 2002; Wilcox et al. 2002; Alfaro et al. 2003; Douady et al. 2003; Rannala and Yang 2003; Nylander et al. 2004; Lewis et al. 2005; Pickett and Randle 2005; Randle et al. 2005). For example, the Bayesian posterior probability (PP) has been used as a measure of phylogenetic confidence, but is often much higher than nonparametric BS values. Based on simulations where the true phylogeny is known, the PP has been determined to provide a more accurate measure of phylogenetic confidence than the non-parametric BS when the assumptions of the method are satisfied (Wilcox et al. 2002; Alfaro et al. 2003). But when the assumptions are violated, which may be quite common (e.g., using overly simplified models of evolution), PP are often inflated, suggesting higher confidence in a particular branch than truly exists (Huelsenbeck and Rannala 2004). Given these caveats, we have conducted a Bayesian analysis of the combined molecular data with the data partitioned by region using MrBayes v.3.0B4 (Huelsenbeck and Ronquist 2001) under the GTR⫹⌫ model with rates allowed to vary among partitions (Ronquist and Huelsenbeck 2003). All model parameters were unlinked between partitions, except the branch lengths and the topology. Four chains were run for one million generations, sampling every 1000 generations. This analysis was run three times to estimate the degree of convergence in the resulting topologies. Burn-in was determined by stabilization of the likelihood score, and these trees were removed for subsequent analyses. A majority rule consensus tree based on the combination of trees resulting from all three runs was generated in MrBayes v.3.0B4. Branch lengths are averaged from the posterior distribution of trees and the PP for each branch reported. To estimate the relative degree of parameter mixing, several parameter values were examined for stabilization in the three separate runs (Nylander et al. 2004). Phylogenetic Analysis of Combined Morphological and Molecular Data. Concerns over the use of morphological data in phylogenetic inference have recently focused on the limited number of unambiguously scorable markers, difficulty in accurately determining homologies, and concerns over appropriate character coding (Scotland et al. 2003; Wiens 2004; Smith and Turner 2005). Recent studies suggest that the addition of morphological data, although often relatively small and sometimes highly homoplasious, can bolster support for a molecular phylogeny (CabreroSan˜udo and Zardoya 2004; Nylander et al. 2004; but see Funk et al. 1995 and McCracken et al. 1999 for exceptions, and De Queiroz et al. 1995 for a historical review). Therefore, molecular and morphological data were combined to determine the phylogenetic effects of morphological data in resolving the M. moschatus alliance

384

SYSTEMATIC BOTANY

phylogeny. MP analysis was conducted as described above with DATATYPE ⫽ STANDARD to accommodate the morphological characters. MP analyses were compared with and without polymorphic character states (rescored as missing data) to determine their effect on the topology, since polymorphic morphological characters are considered missing data in the subsequent Bayesian analysis (Ronquist and Huelsenbeck 2003). The combined molecular and morphological data were analyzed using the Bayesian methods implemented in MrBayes v.3.0B4 (Huelsenbeck and Ronquist 2001). The Bayesian protocol for the combined molecular and morphological data was similar to that described above for the combined molecular data, except that the morphological data partition was analyzed using the Mk⫹⌫ likelihood model of Lewis (2001) with a correction for scoring only variable characters (CODING ⫽ VARIABLE; Nylander et al. 2004). Although this model allows for different evolutionary rates between characters, it should be used with caution since it assumes all transitions within a character are equally likely. The dirichlet prior was fixed so that all character states have equal frequency priors [SYMDIRIHYPERPR ⫽ FIXED(INFINITY)] (Nylander et al. 2004). Morphological Homoplasy. The degree of morphological homoplasy was estimated from the retention index (RI), an unbiased measure of homoplasy (Farris 1989). RI estimates homoplasy by comparing the observed number of steps for a character on a given phylogeny to the minimum number of steps for a character irrespective of the phylogeny. The observed number of steps is determined using maximum parsimony mapping of the character on the given topology. The minimum number of steps is the number of states in the data minus one. For a two state character the minimum number of steps is one and for a three state character the minimum number of states is two. RI is considered ‘‘unbiased’’ since it removes the inflating effects of autapomorphies by subtracting both the observed number of steps and the minimum number of steps from the maximum number of steps. The maximum number of steps is determined by invoking the highest amount of homoplasy (i.e., the number of steps required given a star phylogeny). Therefore, the RI ⫽ (Maximum-Observed)/(Maximum-Minimum). RI values for each morphological character were calculated in PAUP*4.0b10 based on the combined molecular phylogeny (Fig. 4). Characters were mapped onto the phylogenetic tree resulting from the Bayesian analysis with the highest maximum likelihood value pruned to include only the M. moschatus alliance plus M. latidens. RI values were compared between vegetative and reproductive characters and significance of differences was tested with a Mann-Whitney test because of the non-normal distribution and the discrete nature of the RI values for characters with few maximum steps on a given phylogeny. Non-homoplasious morphological synapomorphies were identified for clades within the M. moschatus alliance using the same pruned phylogeny.

RESULTS Morphological Data. The morphological data of 53 characters (Appendix 3) consists of 35 parsimony informative characters for the M. moschatus alliance. Four characters are categorized as general, 12 characters are vegetative, and 19 characters are reproductive. Rpl16 Sequencing. The rpl16 intron provides additional variation and phylogenetically informative data to the previously existing three gene data (Beardsley et al. 2004). The rpl16 data consisted of 6.1% variable characters for the M. moschatus alliance accessions plus M. latidens. This level of variation is similar to the other cpDNA region, trnL-F (5.6%), but is lower than that for the ETS (14.6%) and ITS (8.8%) regions. There are 18 parsimony informative sites in the rpl16 intron for the

[Volume 31

M. moschatus alliance and M. latidens. This is higher than that for the trnL-F region (12 sites), yet less than the ETS (30 sites) and ITS (29 sites) regions. The rpl16 genotype of the hybrid species, M. evanescens, differs from that of M. breviflorus by only one substitution, consistent with the ITS and trnL-F results of Beardsley et al. (2004). This taxon was removed from further analyses due to its putative hybrid origin. Phylogenetic Analysis of Individual Data. MP and ML analyses of the individual DNA regions yield mostly congruent topologies for the M. moschatus alliance, yet resolve only a few internal nodes per region with BS ⬎ 80% (MP trees available on TreeBASE). In the MP analysis, each individual region resolves between three and seven of the 10 interspecific nodes in the majority rule consensus topology for the M. moschatus alliance. The likelihood models chosen for these regions were HKY⫹G for the ETS region, trn⫹G for the ITS region, and k81uf⫹G for the trnL-F region and rpl16 intron. Likelihood analysis of each individual region (Fig. 3) provides similar resolution, and equivalent or higher ML BS values compared to the MP analysis. The three geographically distinct clades (SN, SR, and CR) are completely resolved in the ETS and ITS likelihood analyses, yet sometimes with only moderate to low BS (Fig. 3A, B). The trnL-F and rpl16 regions support one and two of these clades, respectively (Fig. 3C, D). Several parsimony informative indels map to previously identified clades within the M. moschatus alliance (Fig. 3; Whittall 1999; Beardsley et al. 2004). Although no single region resolves every node within the M. moschatus alliance, each region does provide support for different portions of the alliance. There are no instances of conflicting relationships supported by BS ⬎ 80%, but the placement of M. pulsiferae and M. moschatus remains uncertain. The ETS allies M. moschatus with M. pulsiferae (BS ⫽ 73%) as sister to the SN clade, whereas the trnL-F places M. pulsiferae as sister to the SN clade (BS ⫽ 94%). MP analysis of the morphological data generated a single most parsimonious tree (not shown), which is considerably different and only weakly supported compared to any of the individual molecular analyses (CI ⫽ 0.3351; RI ⫽ 0.5208; tree available from TreeBASE). Two of the three geographically distinct clades are recovered, the SN clade (BS ⫽ 86%) and the CR clade (BS ⬍ 50%). Mimulus moschatus is sister to the SN clade, but without significant BS. There is no support for the monophyly of the M. moschatus alliance and only one additional interspecific clade is supported with BS ⫽ 85% (the Eurasian outgroup clade consisting of M. bodinieri, M. tenellus and M. nepalensis). When the distant outgroups are removed leaving only the two closest outgroup lineages, the six resulting equally parsimonious trees consistently recover a monophyletic M. moschatus alliance with BS ⫽ 73%.

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

385

FIG. 3. ML analyses of the separate DNA sequence regions with ML BS values indicated at the nodes. A. ETS. B. ITS. C. trnL-F. D. rpl16. The M. moschatus alliance (MoA) is monophyletic and the acronyms are follow those described in Fig. 2. Branch percentages ⬎ 70% are indicated at the nodes. Parsimony informative ingroup indels (not included in the analysis) are indicated with black bars.

Only one branch within the M. moschatus alliance has BS ⬎ 70% (CR clade plus M. ampliatus and M. hymenophyllus). The monophyly of the SN and CR clades previously resolved are supported with BS ⫽ 65% and 66%, respectively. Combined Molecular Data. The combined molecular analysis including the rpl16 region substantially

improves topological resolution and branch support for relationships within the M. moschatus alliance when compared to the analyses of the individual molecular data. MP generates a single most parsimonious tree with seven of the 10 interspecific nodes for the M. moschatus alliance resolved with BS ⱖ 70% (CI ⫽ 0.7918; RI ⫽ 0.7913; Fig. 4). MP analysis provides strong sup-

386

SYSTEMATIC BOTANY

FIG. 4. The Bayesian phylogram with the highest likelihood score resulting from the analysis of the combined molecular data. Parsimony bootstrap values are shown above the nodes, followed by Bayesian PP values. Values of 100% are indicated with an asterisk and values below 70% are indicated with a hyphen. Clade acronyms are as in Fig. 2.

port for the three biogeographically defined clades of the M. moschatus alliance. Mimulus moschatus and M. pulsiferae are sister taxa allied with the SN clade, but with BS ⬍ 50% (tree available from TreeBASE). An unpartitioned ML analysis using the trn⫹G best fit model provides a similar level of resolution compared to the combined MP analysis (trees available from TreeBASE). Six of 10 interspecific nodes for the M. moschatus alliance are resolved with BS ⬎ 70% (BS for the sister relationship between M. ampliatus and M. hymenophyllus drops from 72% in the MP analysis to 62% in the ML analysis). Mimulus moschatus and M. pulsiferae are allied with the SN clade (identical to that presented in Fig. 4), but with BS ⬍ 50%. In general, the MP and ML combined molecular analyses provide improved branch support compared to previous combined molecular analyses (Whittall 1999; Beardsley et al. 2004). One branch with increased support is the sister relationship of M. breviflorus and the CR clade which improves from 90% in Beardsley et al. (2004) to 99% and 100% in the MP and ML combined molecular analyses, respectively. A similar improvement is found in the increased BS for the mono-

[Volume 31

phyly of the Mimulus moschatus alliance from 94% in Beardsley et al. (2004) to 99% in both MP and ML analyses. A Bayesian analysis can potentially increase the accuracy of a combined phylogenetic analysis by allowing mixed substitution models to be simultaneously applied across data partitions during phylogenetic analysis. The first 100,000 generations were removed from each of three replicate runs as burn-in. The remaining trees from these three runs were combined to form a majority-rule consensus (Fig. 4). Branch lengths reflect the posterior distribution from the three runs (Fig. 4). The majority rule consensus tree from the Bayesian analyses is largely congruent with the MP and ML analyses. Within the M. moschatus alliance, there are only two branches with PP ⬍ 0.95 (Fig. 4). Mimulus pulsiferae is sister to the SN clade with PP ⫽ 0.60; M. moschatus is unresolved. The tree with the highest ML value across all three Bayesian runs places M. moschatus as sister to the SN/ pulsiferae clade (Fig. 4). Combined Molecular and Morphological Data. Combining the molecular and morphological data substantially improves the phylogenetic resolution and branch support within the M. moschatus alliance. A MP analysis including polymorphic sites for the combined molecular and morphological data generated a single most parsimonious tree with nine interspecific nodes supported with BS ⬎ 70% within the M. moschatus alliance. This topology is largely congruent with the combined molecular analyses except for the placement of two taxa; M. pulsiferae and M. moschatus. There is moderate support for the novel placement of M. pulsiferae as sister to a large clade including the SR clade, the CR clade, and M. breviflorus (BS ⫽ 78%; tree not shown). In addition, the sister group relationship of M. moschatus to the rest of the M. moschatus alliance is also novel, but not supported (BS ⬍ 50%). All three geographically distinct clades are resolved, as are the relationships within them. Coding polymorphic sites in the morphological data as missing data did not significantly affect the topology or associated BS values in the MP analysis (trees available from TreeBASE). The Bayesian analysis of the combined molecular and morphological data improves the branch support with all nodes in the M. moschatus alliance now supported by PP ⫽ 1.0 except one (M. moschatus; Fig. 5B). Three nodes that increased in BS in the MP analysis also show increased PP in the combined molecular and morphological analysis (Fig. 5). MP and Bayesian analyses of the combined molecular and morphological data are congruent except in the placement of the phylogenetically labile taxon, M. pulsiferae. As described above, M. pulsiferae occupies a novel relationship as sister to a large clade of ingroup taxa (including the SN clade) in the MP analysis (BS ⫽ 78%; tree not shown).

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

387

FIG. 5. A comparison of the Bayesian majority rule consensus trees from the combined molecular data (A) and the combined molecular and morphological data (B) shows the positive effect of the morphological data on branch support for the M. moschatus alliance. Parsimony bootstrap values are followed by the Bayesian PP values on the combined molecular phylogeny (A). Branches with increased support in the combined molecular and morphological phylogeny are circled. The number of morphological changes attributed to each branch under the MP criterion is shown below the branches.

In the Bayesian analysis, it is sister to the SN clade, with PP ⫽ 1.0 (Fig. 5B). Morphological Homoplasy. We estimated the degree of morphological homoplasy in the M. moschatus alliance by comparing RI values for individual morphological characters (and subsets of morphological characters, e.g., floral vs. vegetative) on the phylogeny with the highest ML value from the combined molecular Bayesian analysis (Table 1). After all outgroup taxa except M. latidens were removed (and retaining the observed ingroup root), nine morphological characters were constant and another nine characters were autapomorphic for the M. moschatus alliance. Seventeen of the remaining 35 characters have RI values ⱕ 0.5, suggesting a significant amount of homoplasy in the

morphological data (Table 1). The eight characters exhibiting no homoplasy on the tree are all reproductive characters (interior corolla palate ridges, corolla aperture, lateral corolla lobe orientation, pollen type, stigma color, style pubescence, seed color, and seed shape). The average RI value for reproductive characters (mean ⫽ 0.65) is significantly higher than that of vegetative characters (mean ⫽ 0.30; Mann-Whitney test, p ⫽ 0.0128), indicating that reproductive characters exhibit less homoplasy than vegetative characters in the M. moschatus alliance. Non-homoplasious synapomorphies identified from mapping the morphological data onto the molecular phylogeny help define several clades within the M. moschatus alliance. The taxa within the SN clade pos-

388

[Volume 31

SYSTEMATIC BOTANY

TABLE 1. Homoplasy found in the morphological characters by mapping them onto the most likely Bayesian tree from the combined molecular analysis. Characters are sorted by retention index (RI) for the M. moschatus alliance and M. latidens. Character numbers refer to those in Appendix 2. Nine autapomorphic characters and nine constant characters were removed. Characters are categorized as general (G), vegetative (V), or reproductive (R). No. Steps Char. No.

Description

Category

RI

Minimum

Tree

Maximum

30 31 39 40 43 48 51 52 17 18 35 41 42 46 47 14 19 23 4 34 49 32 15 1 2 6 10 11 12 21 24 26 31 36 50

Seed color Seed shape Interior palate ridges Corolla aperture Lateral corolla lobe orient. Pollen type Stigma color Style pubescence Fine-glandular herbage Pubescence cell number Corolla shape Corolla lobe margins Upper corolla lobe orient. Anther opening Anther sacs Leaf base shape Calyx pubescence distr. Calyx pubescence pres. Seed dormancy Corolla palate prominence Stigma lobe shape Corolla length Leaf margin Longevity Habit Substrate Rosette Leaf venation Upper leaf attachment Calyx lobes in fruit Calyx shape in fruit Calyx lobe length Capsule shape Corolla throat floor Stigma lobe symmetry

R R R R R R R R V V R R R R R V V V G R R R V G G G V V V V V V R R R

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.80 0.75 0.67 0.67 0.67 0.67 0.67 0.60 0.60 0.60 0.50 0.50 0.33 0.25 0.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 1 1 1 1 1 1 1 1 1

2 1 1 1 1 1 2 1 2 2 2 2 2 2 2 3 3 3 3 2 4 4 5 2 2 6 2 2 2 3 3 2 2 2 2

5 4 4 2 2 4 5 2 6 6 4 4 4 4 4 6 6 6 5 3 5 5 6 2 2 6 2 2 2 3 3 2 2 2 2

sess similar seed shape, seed color, and glabrous interior palate ridges. Taxa in the CR clade share three traits: constricted corolla apertures, downward angled lateral petals, and pubescent styles. There are no cladedefining characters for the SR clade. In searching for non-homoplasious synapomorphies for the placement of the phylogenetically labile M. pulsiferae and M. moschatus, we found the non-homoplasious character, pollen type IIb, which unites M. pulsiferae with the SN clade. Two traits, multicellular pubescence and wide to slightly reflexed anther opening, ally M. moschatus to the SN clade (though, not including M. pulsiferae). Stigma coloration may also be shared between M. moschatus and the SN clade, but missing data for some M. moschatus alliance taxa makes this result only tentative. DISCUSSION Phylogenetics. With the addition of the rpl16 intron sequences, morphological data, and DNA sequence re-

gion-specific models of sequence evolution, we have bolstered support for several previously identified clades and resolved some formerly ambiguous relationships in the M. moschatus alliance. All four DNA sequence regions confirm the monophyly of the M. moschatus alliance to varying degrees. The closest outgroup to the M. moschatus alliance is often M. latidens, but the Asian outgroup clade of M. tenellus, M. bodinieri, and M. nepalensis is also closely related, as previously reported by Beardsley et al. (2004). Combined analyses of molecular and molecular plus morphological data consistently recover the three geographically distinct clades of the M. moschatus alliance that have been previously described (Whittall 1999; Beardsley et al. 2004). This study also resolves, for the first time, the relationships within the SR clade and the relationships between the SN clade and its closest relative, M. pulsiferae. The basal position of M. patulus in the SR clade

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

suggests its early differentiation and, if the autogamous mating system characteristic of this taxon (Carlson 2002) evolved during this speciation event, this increases the time that it has persisted as an autogamous lineage. The affinity of M. pulsiferae with the SN clade was suggested Beardsley et al. (2004), but with very weak MP bootstrap support (60%). The strongly supported sister relationship of M. pulsiferae with the SN clade in the combined molecular and morphological Bayesian analysis highlights the significance of similar geographic distributions and pollen type IIb, a synapomorphy of this clade (Argue 1980). Although nearly half of the morphological data shows high levels of homoplasy (Table 1), in combination with the molecular data, MP and Bayesian analyses resulted in improved topological resolution and higher support values for several branches. Although two branches outside of the M. moschatus alliance show decreases in BS and PP following the addition of the morphological data, five nodes within the M. moschatus alliance show increases in BS and/or PP (Fig. 5B). The most significant improvement with the addition of the morphological data is in the placement of M. pulsiferae as the sister to the SN clade with BS ⫽ 78% and PP ⫽1.0 compared to BS ⬍ 50% and PP ⫽ 0.60 in the combined molecular analyses (Fig. 5). Such improvement has been described in several studies combining molecular and morphological characters in phylogenetic inference (Cabrero-San˜udo and Zardoya 2004; Nylander et al. 2004, and see Sanderson and Donoghue 1989 and Scotland et al. 2003 for comparisons of homoplasy in molecular versus morphological data). At least two explanations are possible for this pattern. First, by using an overly simplified model of evolution (i.e., forward and reverse transitions within a character are considered equal), analyses may be biased toward inflated PP values the same way that overly simplified models of sequence evolution can artificially increase PP values (Huelsenbeck and Rannala 2004). Alternatively, the improved support for several branches of the M. moschatus alliance following the inclusion of the morphological data may reflect the strong influence of a few additional characters with low homoplasy at nodes with little supporting molecular data. For example, in the M. moschatus alliance, a single non-homoplasious character, pollen type IIb (Argue 1980, 1986), maps to the branch with the most improved support values. The addition of the rpl16 intron and morphological data has not completely resolved the placement of M. moschatus with significant support. Each individual DNA sequence region places this taxon differently within the M. moschatus alliance, but never with BS ⬎ 70% (Fig. 3). Its relationship in the combined molecular Bayesian analyses is unresolved (PP ⬍ 0.5). When the morphological data is added, the Bayesian com-

389

bined molecular and morphological analysis place M. moschatus as sister to the SN clade plus M. pulsiferae, but still with low support (PP ⫽ 0.74). The phylogenetic affinity of M. moschatus to the SN clade and M. pulsiferae is similar to the results from the combined molecular MP and ML analyses. The presence of two non-homoplasious morphological characters supports the close relationship of M. moschatus and the SN clade. To clarify the phylogenetic placement of M. moschatus among the early radiation of the M. moschatus alliance lineages will require more rapidly evolving DNA sequences. Morphological Evolution. A well-resolved phylogeny provides the historical foundation for interpreting morphological evolution in the M. moschatus alliance. Although the M. moschatus alliance has been previously considered a natural grouping based on the presence of viscid pubescent herbage (Meinke 1992), this trait also arises in the distantly related species, M. alsinoides. Subsequent modifications of pubescence in the M. moschatus alliance have occurred in two clades; the predominantly multi-cellular pubescence group (SN clade and M. moschatus, but not M. pulsiferae) and the unicellular pubescence group (the remaining taxa of the M. moschatus alliance). There are three non-homoplasious characters defining the SN clade (seed color, seed shape, and glabrous internal palate ridges) and three non-homoplasious characters defining the CR clade (corolla aperture, lower petals down-curved, and style pubescence). These clade defining traits emphasize the congruence of seed and floral characters with the molecular phylogeny for the M. moschatus alliance. There is substantially more homoplasy in the morphological data than in the molecular data. For example, on the combined molecular and morphological Bayesian phylogeny, the RI for the molecular data is much higher than that of the morphological data (RI ⫽ 0.79 and 0.51, respectively). Similar disparity in homoplasy of molecular and morphological characters exists when only examining the M. moschatus alliance plus M. latidens (molecular data RI ⫽ 0.76, morphological data RI ⫽ 0.54). Some level of homoplasy is expected in most data and it is these homoplasious characters that often provide the majority of information useful in phylogenetic reconstruction (Sanderson and Donoghue 1989; Kallersjo et al. 1999). The causes of the elevated levels of homoplasy in the morphological data may illuminate patterns of morphological evolution if we can first exclude sampling errors. Two sources of sampling error include sampling characters with a bias towards choosing those with high homoplasy (Hillis and Wiens 2000) and mistaken assignments of homology (Weins 2004). Our morphological dataset contains 35 parsimony informative characters within the M. moschatus alliance. The characters have been drawn from three broadly defined categories (vegeta-

390

SYSTEMATIC BOTANY

[Volume 31

FIG. 6. Floral evolution in the M. moschatus alliance based on the Bayesian combined molecular and morphological phylogeny. Species with a highly autogamous mating system are indicated by shaded branches. Flowers are drawn to scale.

tive, floral and general characters) in which we have emphasized scoring independent characters. The diversity and independence of characters should minimize the effects of choosing characters with a bias towards those with high homoplasy. Homology assignment was optimized in the determination of characters and their states by choosing characters with discrete states (e.g., presence/absence characters). Alternatively, the homoplasy associated with the morphological data could be caused by underlying evolutionary processes such as convergent evolution (Brower 1994; Bernhardt 2000; Wiens et al. 2003; Blackledge and Gillespie 2004; Hu et al. 2005). Some of the observed morphological homoplasy is likely due to parallel natural selection driving convergent evolution. All three geographically distinct clades have independently exploited similar habitats (as defined by soil substrate). The adaptive traits associated with these convergent habitat shifts likely contribute to morphological homoplasy. For example, the separate origins of cliff dwelling in M. hymenophyllus and M. jungermannioides highlight the convergent evolution as-

sociated with this habitat: a prostrate growth form. However, each of these cliff dwelling species also has unique adaptations for reproduction in a vertical habitat: M. hymenophyllus (SR clade) has reflexed peduncles in fruit that increase seed dispersal back onto the vertical wall (Meinke 1983); M. jungermannioides (CR clade) reproduces vegetatively by the growth of long stolons tipped with overwintering buds (turions; Meinke 1992). Similarly, three lineages have independently evolved highly autogamous mating systems (Fig. 6). Parallel shifts to autogamy have also been described for two closely related species in Mimulus section Simiolus (Fenster and Ritland 1994). Independent shifts to autogamy and the homoplasious changes in the associated floral traits are not uncommon within a genus (e.g., Grant 1954; Lewis 1973; Wyatt 1992; Goodwillie and Stiller 2001). In the M. moschatus alliance, there are several homoplasious floral characters associated with this shift to autogamy including decreased bilateral symmetry, less reflexed upper corolla lobes, and reduced overall corolla length. Quantitative genetic stud-

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

ies in some species of Mimulus have identified several quantitative trait loci of relatively small effect conferring the traits associated with the autogamous mating system (Karron et al. 1997; Fishman et al. 2002) compared to the effect size of quantitative trait loci responsible for shifts from bee to hummingbird pollination (Bradshaw et al. 1998). The quantitative genetic and molecular tools necessary to examine the developmental, genetic, and molecular basis of evolutionarily interesting morphological traits, such as the shift to the autogamous mating system, are currently being developed in Mimulus (J. Willis, pers. comm.). Deciphering the underlying genetic architecture responsible for the three independent shifts to autogamy in the M. moschatus alliance could be used to determine the degree of convergence during parallel mating system adaptations within a closely related species complex. Taxonomic Conclusions and Conservation Implications. The morphological homoplasy detected here no doubt contributed to the history of taxonomic confusion surrounding the M. moschatus alliance (Grant 1924; Pennell 1951; Hitchcock and Cronquist 1969; Thompson 1992; Meinke 1995). Although our limited intraspecific sampling does not allow us to test species boundaries, the molecular phylogeny does provide evidence, in combination with morphological and ecological data, for taxonomic revision. Previously, KEY

TO THE

SPECIES

OF THE

391

based on morphological similarities of herbarium specimens, M. ampliatus and M. patulus had been synonymized with M. washingtonensis (Pennell 1951; Hitchcock and Cronquist 1969; Meinke 1992). Given M. patulus’ unique sequences at four DNA sequence regions, well supported placement within the SR clade distinct from M. washingtonensis (CR clade), and morphological uniqueness (Carlson 2002), this species should be maintained as distinct from M. washingtonensis. Recently, M. dudleyi has been synonymized with M. floribundus (Thompson 1992), while M. norrisii has been considered a unique species (Heckard and Shevock 1985; Thompson 1992). Based on the unique morphological features of M. dudleyi and its more distant relationship to M. floribundus, we reiterate the recommendation for specific ranking of this species suggested in Beardsley et al. (2004). A key to the species of the M. moschatus alliance integrating these taxonomic distinctions is provided below. Specific recognition of M. patulus, M. ampliatus, and M. dudleyi calls for a reassessment of their conservation status. These results were partly integrated in a status report on M. patulus (Carlson and Meinke 1994). Meanwhile, the specific status of M. ampliatus associated with its imperiled conservation status calls for surveys of additional potential localities to ensure populations are protected across the species’ limited geographic range.

MIMULUS MOSCHATUS ALLIANCE

AND

M. LATIDENS

1. Foliage mostly viscid-villous, trichomes multicellular . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Plants perennial from rhizomes, stolons, or turions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Calyx lobes 2–5 mm, acute or acuminate, corolla rotate, anthers pubescent, plants arising from relatively thick rhizomes or stolons, widespread western species of montane streams and wetlands . . . . . . . . . . . . . . . . . . . . M. moschatus 3. Calyx lobes 1–2 mm, rounded to mucronate, corolla distinctly bilabiate, anthers not pubescent, plants from very thin stolons that form overwintering turions on or near cliff faces in Oregon’s Columbia Plateau Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. jungermannioides 2. Plants annual, without rhizomes, stolons, or turions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Corollas small, less than 11 mm wide and 13 mm long, red spots of corolla inconspicuous, plants generally erect, leaves abruptly tapered, widespread plants of varied habitats east of the Cascades to Rocky Mountains and south through California . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. floribundus 4. Corollas large, greater than 12 mm wide and 13 mm long, red spots conspicuous, plants often decumbent to short erect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Leaves sessile, lower leaves occasionally short-petioled, margins entire or sparingly dentate, basal rosettes persistent, more or less erect plants of moist sandy substrates in Sierra Nevada Mountains . . . . . . . . . . . . . . M. arenarius 5. Leaves petiolate (often diminished on upper stem), margins entire to dentate, basal rosette absent, plants weakly erect to decumbent and branched . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Calyces with coriaceous ribs, the lobes convergent in fruit, corollas with evident white patches near the corolla throat, plant mostly of marble rock faces of the Sierra Nevada foothills . . . . . . . . . . . . . . . . . . M. norrisii 6. Calyces with chartaceous-membranaceous ribs, the lobes not convergent in fruit, corollas without evident white patches near the corolla throat, plants typically arising from fissures in granitic boulders of the southern Sierra Nevada foothills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. dudleyi 1. Foliage viscid-puberulent, trichomes unicellular, or nearly glabrous (as in M. latidens and M. pulsiferae) . . . . . . . . . . . . . . . . 7. Mature calyces strongly inflated at maturity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Capsules stipitate, basal rosettes persistent, corollas greater than 9 mm long, white or yellowish to rose-pink, leaves sessile or very short petioled. Plants of vernal and ephemeral pools, southeastern Oregon and California . . . . . . M. latidens 8. Capsules sessile, basal rosettes lacking, corollas less than 8 mm long, pale yellow. Plants of the Great Basin . . . . . . . . 9. Leaves sessile, calyces strongly inflated and plicate in fruit, rare species of wetland margins southeastern Oregon and northeastern California . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. evanescens 9. Leaves petioled, calyces moderately inflated in fruit and not strongly plicate, widespread species of the Great Basin (southern B. C., Idaho, Washington, Oregon, Nevada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. breviflorus

392

SYSTEMATIC BOTANY

[Volume 31

7. Mature calyces weakly or not inflated at maturity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. Decumbent, highly branched, plants of moist cliff faces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Petioles shorter than leaf blade, leaves thick, capsules elliptical, more than 2⫻ longer than wide, calyces half the length of the corollas, perennials from very thin stolons that form overwintering turions on or near cliff faces in the Columbia Plateau region of Oregon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. jungermannioides 11. Petioles longer than leaf blade, leaves thin-membranous, capsules short-ovate, less than 2⫻ longer than wide, calyces one third the length of the corolla, annuals of Snake River Canyon . . . . . . . . . . . . . . . . . . M. hymenophyllus 10. Erect, diffusely branched species not growing on cliff faces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. Basal rosettes persistent, capsules long-elliptic to slightly oblong, 3⫻ longer than wide, leaves short petioled, corolla 7–11 mm long, widespread montane species, Washington to California . . . . . . . . . . . . . . . . . . M. pulsiferae 12. Basal rosettes lacking, capsules ovate to elliptic, less than 3⫻ longer than wide at maturity, leaves distinctly petioled, corollas 5–14 mm long . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13. Corollas small, less than 8 mm long, leaves long-petiolate, the petioles usually exceeding the blade in length, blades sharply tapered basally, rare species of moist seeps in Snake River Canyon and surrounding area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. patulus 13. Corollas large, greater than 12 mm long, petioles rarely exceeding blade length, blades weakly tapered basally . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14. Stems four angled in cross section, styles glabrous, corollas more or less open, rare species of seeps on the eastern side of the Snake River Canyon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. ampliatus 14. Stems round in cross section, styles moderately pubescent under 10⫻, corollas constricted at the mouth, uncommon species from northeastern Oregon across the Columbia Plateau . . . . . M. washingtonensis

ACKNOWLEDGEMENTS. We thank Steve Gisler, Kelly Amsberry, Francisco Camacho, Barbara Wilson, Joseph Spatafora, John Willis, and William Robinson for stimulating conversations in the early stages of this study. More recently, Todd Oakley and Jeanne Serb have shared valuable advice on phylogenetic methodology. The authors thank Dave Gernandt, Bruce Baldwin, Sara Hoot, Patrick Herendeen, and an anonymous reviewer for helpful comments. We acknowledge generous funding from the Templeton Foundation (OSU), Hardman Award (OSU), Portland Garden Club, Oregon Department of Agriculture’s Plant Conservation Program, Oregon State Department of Botany and Plant Pathology, and the Converse Fellowship (UCSB) to JBW and an NSF predoctoral award to MLC. The Confederated Tribes of the Warm Springs Reservation and Nez Perce Tribe graciously allowed collection from their lands.

LITERATURE CITED ALFARO, M. E., S. ZOLLER, and F. M. LUTZONI. 2003. Bayesian Markov Chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. Molecular Biology and Evolution 20: 255–266. ARGUE, C. L. 1980. Pollen morphology in the genus Mimulus (Scrophulariaceae) and its taxonomic significance. American Journal of Botany 67: 68–87. ———. 1986. Some taxonomic implications of pollen and seed morphology in Mimulus hymenophyllus and M. jungermannioides and comparisons with other putative members of the M. moschatus alliance (Scrophulariaceae). Canadian Journal of Botany 64: 1331–1337. BEARDSLEY, P. M., S. E. SHOENIG, J. B. WHITTALL, and R. G. OLMSTEAD. 2004. Patterns of evolution in western North American Mimulus (Phrymaceae). American Journal of Botany 91: 474– 489. BERNHARDT, P. 2000. Convergent evolution and adaptive radiation of beetle-pollinated angiosperms. Plant Systematics and Evolution 222: 293–320. BLACKLEDGE, T. A. and R. G. GILLESPIE. 2004. Convergent evolution of behavior in an adaptive radiation of Hawaiian web-building spiders. Proceeding of the National Academy of Sciences USA 101: 16228–16233. BONFIELD, J. K., K. F. SMITH, and R. A. STADEN. 1995. A new DNA sequence assembly program. Nucleic Acids Research 24: 4992– 4999. BRADSHAW, H. D., K. G. OTTO, B. E. FREWEN, J. K. MCKAY, and

D. W. SCHEMSKE. 1998. Quantitative trait loci affecting differences in floral morphology between two species of monkeyflower (Mimulus). Genetics 149: 367–382. ——— and D. W. SCHEMSKE. 2003. Allele substitution at a flower colour locus produces a pollinator shift in monkeyflowers. Nature 426: 176–178. BROWER, A. V. Z. 1994. Rapid morphological radiation and convergence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proceedings of the National Academy of Sciences of the United States of America 91: 6491–6495. BUCKLEY, T. R. 2002. Model misspecification and probabilitstic tests of topology: evidence from empirical data sets. Systematic Biology 51: 509–523. CABRERO-SAN˜UDO, F.-J. and R. ZARDOYA. 2004. Phylogenetic relationships of Iberian Aphodiini (Coleoptera: Scarabaeidae) based on morphological and molecular data. Molecular Phylogenetics and Evolution 31: 1084–1100. CALIFORNIA NATURAL DIVERSITY DATABASE. 2004. State and federally listed endangered, threatened, and rare plants of California. Department of Fish and Game. Sacramento, California. CARLSON, M. L. 2002. Evolution of mating system and inbreeding depression in the Mimulus moschatus alliance. Ph.D. Thesis. University of Alaska Fairbanks, Fairbanks, Alaska. ——— and R. J. MEINKE. 1994. Status report for Mimulus patulus. U.S. Fish and Wildlife Service, Region 1, Boise, Idaho, and Oregon Natural Heritage Data Base, Portland, Oregon. DE QUEIROZ, A., M. J. DONOGHUE, and J. KIM. 1995. Separate versus combined analysis of phylogenetic evidence. Annual Review of Ecology and Systematics 26: 657–681. DOUADY, C., F. DELSUC, Y. BOUCHER, W. DOOLITTLE, and E. DOUZERY. 2003. Comparison of Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Molecular Biology and Evolution 20: 248–254. FARRIS, J. S. 1989. The retention index and the rescaled consistency index. Cladistics 5: 417–419. FENSTER, C. B. and K. RITLAND. 1994. The quantitative genetics of mating system divergence in the yellow monkeyflower species complex. Heredity 73: 422–435. FISHMAN, L., A. J. KELLY, E. MORGAN, and J. H. WILLIS. 2001. A genetic map in the Mimulus guttatus species complex reveals transmission ratio distortion due to heterospecific interactions. Genetics 159: 1701–1716.

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

———, and J. H. WILLIS. 2002. Minor quantitative trait loci underlie floral traits associated with mating system divergence in Mimulus. Evolution 56: 2138–2155. FUNK, D. J., D. J. FUTUYMA, G. ORTI, and A. MEYER. 1995. Mitochondrial DNA sequences and multiple data sets: a phylogenetic study of phytophagous beetles (Chrysomelidae: Ophraella). Molecular Biology and Evolution 12: 627–640. GENETICS COMPUTER GROUP. 1999. Wisconsin Package 10.0. Madison: Wisconsin. GOODWILLIE, C. and J. W. STILLER. 2001. Evidence for polyphyly in a species of Linanthus (Polemoniaceae): convergent evolution of self-fertilizing taxa. Systematic Botany 26: 273–282. GRANT, A. L. 1924. A monograph of the genus Mimulus. Annals of the Missouri Botanical Garden 11: 99–388. GRANT, V. 1954. Genetic and taxonomic studies in Gilia. IV. Gilia achilleaefolia. Aliso 3: 1–18. HECKARD, L. R. and J. R. SHEVOCK. 1985. Mimulus norrisii (Scrophulariaceae), a new species from the southern Sierra Nevada. Madron˜o 32: 179–185. HILLIS, D. M. and J. J. WIENS. 2000. Molecules versus morphology in systematics: conflicts, artifacts, and misconceptions. Pp. 1– 20 in Phylogenetic analysis of morphological data, ed. J. Weins. Washington D.C.: Smithsonian Institution Press. HITCHCOCK, C. L. and A. CRONQUIST. 1969. Mimulus. Pp. 337–350 in Vascular plants of the Pacific Northwest, eds. C. L. Hitchcock and A. Cronquist. Seattle: University of Washington Press. HU, F. Y., D. Y. TAO, E. SACKS, B. Y. FU, P. XU, J. LI, Y. YANG, K. MCNALLY, G. S. KHUSH, A. H. PATERSON, and Z. K. LI. 2005. Convergent evolution of perenniality in rice and sorghum. Proceeding of the National Academy of Sciences of the United States of America 100: 4050–4054. HUELSENBECK, J. P. and B. RANNALA. 2004. Frequentist properties of Bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Systematic Biology 53: 904–913. IDAHO CONSERVATION DATA CENTER. 2003. Idaho special status of vascular plants. Boise: Idaho Fish and Game Department. IVEY, C. T., D. E. CARR, and M. D. EUBANKS. 2004. Effects of inbreeding in Mimulus guttatus on tolerance to herbivory in natural environments. Ecology 85: 567–574. JORDAN, W. C., M. W. COURTNEY, and J. E. NEIGEL. 1996. Low levels of intraspecific genetic variation at a rapidly evolving chloroplast DNA locus in North American duckweeds (Lemnaceae). American Journal of Botany 83: 430–439. KALLERSJO, M., V. A. ALBERT, and J. S. FARRIS. 1999. Homoplasy increases phylogenetic structure. Cladistics 15: 91–93. KARRON, J. D., R. J. MITCHELL, K. G. HOLMQUIST, J. M. BELL, and B. FUNK. 2004. The influence of floral display size on selfing rates in Mimulus ringens. Heredity 92: 2242–248. KELCHNER, S. A. and L. G. CLARK. 1997. Molecular evolution and phylogenetic utility of the chloroplast intron in Chusquea and the Bambusideae (Poaceae). Molecular Phylogenetics and Evolution 8: 385–397. LEWIS, H. 1973. The origin of diploid neospecies. American Naturalist 107: 161–170. LEWIS, P. O. 2001. A likelihood approach to estimating phylogeny from discrete morphological character data. Systematic Biology 50: 913–925. ———, M. T. HOLDER, and K. E. HOLSINGER. 2005. Polytomies and Bayesian phylogenetic inference. Systematic Biology 54: 241– 253. MCCRACKEN, K. G., J. HARSHMAN, D. A. MCCLELLAN, and A. D. AFTON. 1999. Data set incongruence and correlated character evolution: an example of functional convergence in the hindlimbs of stifftail diving ducks. Systematic Biology 48: 683–714. MEINKE, R. J. 1983. Mimulus hymenophyllus (Scrophulariaceae): a new species from the Snake River Canyon area of eastern Oregon. Madron˜o 30: 147–152.

393

———. 1992. Systematic and reproductive studies of Mimulus (Scrophulariaceae) in the Pacific Northwest: implications for conservation biology. Ph.D. Thesis. Oregon State University, Corvallis, Oregon. ———. 1995. Mimulus evanescens (Scrophulariaceae): a new annual species from the northern Great Basin. Great Basin Naturalist 55: 249–257. MONTANA NATURAL HERITAGE PROGRAM. 2004. Plant species of concern. Helena, Montana. NYLANDER, J. A. A., F. RONQUIST, J. P. HUELSENBECK, and J. L. NIEVES-ALDREY. 2004. Bayesian phylogenetic analysis of combined data. Systematic Biology 53: 47–67. OREGON NATURAL HERITAGE INFORMATION CENTER. 2004. Rare, threatened and endangered species of Oregon. Portland, Oregon. PENNELL, F. W. 1951. Scrophulariaceae. Pp. 686–859 in An illustrated flora of the Pacific states, ed. L. Abrams. Palo Alto: Stanford University Press. PICKETT, K. M. and C. P. RANDLE. 2005. Strange Bayes indeed: uniform topological priors imply non-uniform clade priors. Molecular Phylogenetics and Evolution 34: 203–211. POSADA, D. and K. A. CRANDALL. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818. PRATHER, L. A. 1999. The relative lability of floral vs. non-floral characters and a morphological phylogenetic analysis of Cobaea (Polemoniaceae). Botanical Journal of the Linnean Society 131: 433–450. RAMSEY, J., H. D. BRADSHAW, and D. W. SCHEMSKE. 2003. Components of reproductive isolation between the monkeyflowers Mimulus lewisii and M. cardinalis (Phrymaceae). Evolution 57: 1520–1534. RANDLE, C. P., M. E. MORT, and D. J. CRAWFORD. 2005. Bayesian inference of phylogenetics revisited: developments and concerns. Taxon 54: 9–15. RANNALA, B. and Z. YANG. 2003. Bayes estimation of species divergence times and ancestral population sizes using DNA sequences from multiple loci. Genetics 164: 1645–1656. RONQUIST, F. and J. P. HUELSENBECK. 2003. MrBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. SANDERSON, M. J. and M. J. DONOGHUE. 1989. Patterns of variation in levels of homoplasy. Evolution 43: 1781–1795. SCHEMSKE, D. W. and H. D. BRADSHAW. 1999. Pollinator preference and the evolution of floral traits in monkeyflowers (Mimulus). Proceedings of the National Academy of Sciences USA 96: 11910– 11915. SCOTLAND, R. W., R. G. OLMSTEAD, and J. R. BENNETT. 2003. Phylogeny reconstruction: the role of morphology. Systematic Biology 52: 539–548. SIMMONS, M. P. and H. OCHOTERENA. 2000. Gaps as characters in sequence-based phylogenetic analyses. Systematic Biology 49: 369–381. SMALL, R. L., J. A. RYBURN, R. C. CRONN, T. S. SEELANAN, and J. F. WENDEL. 1998. The tortoise and the hare: choosing between noncoding plastome and nuclear ADH sequences for phylogeny reconstruction in a recently diverged plant group. American Journal of Botany 85: 1301–1315. SMITH, N. D. and A. H. TURNER. 2005. Morphology’s role in phylogeny reconstruction: perspectives from paleontology. Systematic Biology 54: 166–173. SUZUKI, Y., G. GLAZKO, and M. NEI. 2002. Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. Proceeding of the National Academy of Sciences USA 99: 16138– 16143. SWIGONOVA, Z. and K. M. KJER. 2004. Phylogeny and host-plant association in the leaf beetle genus Trirhabda LeConte (Coleoptera: Chrysomelidae). Molecular Phylogenetics and Evolution 32: 358–374. SWOFFORD, D. L. 1998. PAUP*: Phylogenetic analysis using parsi-

394

[Volume 31

SYSTEMATIC BOTANY

mony (*and other methods). 4.0b10. Sunderland: Sinauer Associates. THOMPSON, D. 1992. Mimulus. Pp. 1037–1046 in The Jepson manual of higher plants of California, eds. J. Hickman. Berkeley: University of California Press. WASHINGTON NATURAL HERTIAGE PROGRAM. 2004. Endangered, threatened, and sensitive plants of Washington. Washington State Department of Natural Resources. Olympia, Washington. WHITTALL, J. B. 1999. A molecular phylogeny for the Mimulus moschatus alliance (Scrophulariaceae) and its conservation implications. M. S. Thesis. Oregon State University, Corvallis, Oregon. ———, A. LISTON, S. GISLER, and R. J. MEINKE. 2000. Detecting nucleotide additivity from direct sequences is a SNAP: an example from Sidalcea (Malvaceae). Plant Biology 2: 211–217. WIENS, J. J. 2004. The role of morphological data in phylogeny reconstruction. Systematic Biology 53: 653–661. ———, P. T. CHIPPINDALE, and D. HILLIS. 2003. When are phylogenetic analyses misled by convergence? A case study in Texas cave salamanders. Systematic Biology 52: 501–514. WILCOX, T. P., D. J. ZWICKL, T. A. HEATH, and D. HILLIS. 2002. Phylogenetic relationships of the dwarf boas and a comparison of Bayesian and bootstrap measures of phylogenetic support. Molecular Phylogenetics and Evolution 25: 361–371. WYATT, R., E. A. EVANS, and J. C. SORENSON. 1992. The evolution of self-pollination in granite outcrop species of Arenaria (Caryophyllaceae). VI. Electrophoretically detectable genetic variation. Systematic Botany 17: 201–209.

APPENDIX 1 Species sampled from the M. moschatus alliance and outgroups, location information, conservation status and Genbank accession numbers for rpl16 sequences. Global natural heritage rankings are defined as follows: G1 ⫽ less than five populations remaining; G2 ⫽ 6–20 populations remaining; G4 ⬎ 100 populations, rare in parts of its range, especially at the periphery; NR ⫽ not currently ranked. No rpl16 sequences were obtained from three of the outgroup taxa (M. bodinieri, M. nepalensis, and M. tenellus). M. ampliatus A.L.Grant, ID, Lewis Co., 4 mi N of Ferdinand, E side of old Hwy 95., Whittall 41 (OSC), G1-ID, DQ090895; M. breviflorus Piper, OR, Lake Co., Youkum Valley., Rittenhouse 485 (OSC), G4-MT, DQ090899; M. dudleyi A.L.Grant, CA, Tuolumne Co., Red Hills Rd., 1.7 mi S of Hwy 120., McNeal 1132 (OSC), G4-CA, DQ090906; M. evanescens Meinke, CA, Lassen Co., Moll Reservoir, ca. 13 mi E of Adin., Carlson pers. coll. (OSC), G2-OR, CA, DQ104404; M. floribundus Lindley, OR, Wallowa Co., 2 mi S of Troy along Old Troy Rd., Whittall 43 (OSC), NR, DQ090904; M. hymenophyllus Meinke, OR, Wallowa Co., Horse Creek, Site #3., Whittall 44 (OSC), G1-OR, DQ090896; M. jungermannioides Suksd., OR, Sherman Co., Sherman, S of Highway 84, 3 mi E of county line., Whittall 45 (OSC), G2-OR, DQ090900; M. moschatus Lindley, OR, Benton Co., Oak Creek, 0.5 mi S of MacDonald For., Whittall 46 (OSC), NR, DQ090902; M. norrisii Heckard & J.R.Shevock, CA, Tulare Co., Sequoia Natlional Park, near Potwisha campground., Beardsley 98-013 (UW), G2-CA, DQ090905; M. patulus Pennell, OR, Wallowa Co., Imnaha River Rd., 4 mi S of Imnaha., Whittall 30 (OSC), G2-ID, DQ090897; M. pulsiferae A.Gray, OR, Jefferson Co., Camp Sherman Rd., 6 mi N of Hwy 20., Whittall 48 (OSC), G4WA, DQ090898; M. washingtonensis Gand., OR, Wheeler Co., Spray, Hwy 19, 3 mi E of Horseshoe Cr, Whittall 49 (OSC), G4-WA, DQ090901 Outgroup taxa: M. alsinoides Benth., OR, Douglas Co., Roseburg, West Bank Rd. 6 mi E of I5., Whittall 40 (OSC), NR, DQ090909; M. bicolor Hartw., CA, Madera Co., E shore Bass Lake; Willow Creek & Beasore Rd., Taylor 8594 (UCB), NR, DQ090892; M. dentatus Nutt., OR, Benton Co., Prarie Peak, Alsea., Halse 1977 (OSC), NR, DQ090907; M. guttatus DC., OR, Marion Co., Mount Jefferson, Pacific Coast Trail., Halse 3617 (OSC), NR, DQ090908; M. latidens

(A.Gray) E.Greene, CA, Yolo Co., beside Road 29, 5 mi W of Hwy 99W., Ehlig 37 (OSC), G4-OR, DQ090903; M. lewisii Pursh., WA, University of Washington, greenhouse culture., Christy 760 (ASU), NR, DQ090893; M. primuloides Benth., OR, Klamath Co., Mud Spring, 2 mi S of Hwy 66 on W Branch Rd., Whittall 47 (OSC), G4-MT, DQ090894

APPENDIX 2 Morphological characters scored for the M. moschatus alliance and outgroups. General Characters. 1. Longevity: 0 ⫽ annual, 1 ⫽ perennial. 2. Habit: 0 ⫽ prostrate, 1 ⫽ erect. 3. Runners or stolons: 0 ⫽ absent, 1 ⫽ present. 4. Seed dormancy: 0 ⫽ absent, 1 ⫽ present. This was determined by germination success following an 80d cold stratification (20–100 seeds on moist paper at 4⬚C in the dark), and then 21d at room temperature with 16h day length. 5. Ploidy level: 1 ⫽ diploid (n ⫽ 8), 2 ⫽ aneuploid (n ⫽ 14, 15, 24, 28), 3 ⫽ tetraploid (n ⫽ 16). Mimulus lewisii possess the ancestral chromosome number of eight, the ingroup are ancient tetraploids (Beardsley et al. 2004). Mimulus guttatus has many chromosomal races. 6. Soil substrate: 0 ⫽ inorganic, 1 ⫽ organic, 2 ⫽ cliff. Vegetative Characters. 7. Stem thickness: 0 ⫽ thin, 1 ⫽ wide and succulent. 8. Stem shape in cross-section: 0 ⫽ round, 1 ⫽ four-angled. 9. Stem length: 0 ⫽ much greater than pedicle length, 1 ⫽ much less than pedicle length. 10. Persistent rosette: 0 ⫽ absent, 1 ⫽ present. 11. Mature leaf venation: 0 ⫽ five, 1 ⫽ three. Small or undeveloped leaves of all species often have three veins, but large, well-developed leaves of many species have five primary veins. 12. Upper leaf attachment: 0 ⫽ petiolate, 1 ⫽ sessile. 13. Shape of leaf attachment: 0 ⫽ round or elliptical in cross-section, 1 ⫽ winged in cross section. 14. Leaf base shape: 0 ⫽ abruptly tapered, 1 ⫽ gradually tapered. 15. Leaf margin: 0 ⫽ entire to mildly toothed, 1 ⫽ strongly toothed. 16. Herbage secretions: 0 ⫽ not viscid, 1 ⫽ viscid. 17. Fine-glandular herbage: 0 ⫽ absent, 1 ⫽ present. This character identifies glandular trichomes that are small (i.e., not producing a great amount of glandular secretions individually) and closely spaced. 18. Pubescence cell number: 0 ⫽ single-celled, 1 ⫽ multi-cellular, 2 ⫽ none. 19. Distribution of calyx pubescence: 0 ⫽ uniform,1 ⫽ not uniform. 20. Upper calyx tooth: 0 ⫽ uniform in length with other teeth, 1 ⫽ longer than other teeth. 21. Calyx lobes in fruit: 0 ⫽ not converging, 1 ⫽ converging. 22. Lower calyx tooth: 0 ⫽ uniform in length with other teeth, 1 ⫽ longer than other teeth. 23. Calyx pubescence: 0 ⫽ absent, 1 ⫽ present. 24. Calyx shape in fruit: 0 ⫽ not swollen, 1 ⫽ swollen. 25. Calyx texture: 0 ⫽ not membranous, 1 ⫽ membranous. 26. Calyx lobe length: 0 ⫽ less than 2mm, 1 ⫽ greater than 2mm. 27. Stem architecture: 0 ⫽ rarely or irregularly branched, 1 ⫽ regular branching, progressively bifurcated. Reproductive Characters. 28. Capsule insertion: 0 ⫽ stalked, 1 ⫽ sessile. 29. Capsule shape: 0 ⫽ apex abruptly tapered, 1 ⫽ apex mildly tapered. 30. Seed color: 1 ⫽ light tan, 2 ⫽ brown, 3 ⫽ black. 31. Seed shape: 1 ⫽ oblong to long elliptical, 2 ⫽ ovoid to elliptical. 32. Corolla length: 0 ⫽ corolla less than twice as long as the calyx, 1 ⫽ corolla more than twice as long as the calyx. 33. Corolla color: 1 ⫽ lavender to purple, 2 ⫽ yellow, 3 ⫽ pale pink to white. 34. Corolla palate prominence: 0 ⫽ no or very small white patch, 1 ⫽ prominent white patch. 35. Corolla shape: 0 ⫽ not obviously bilabiate, 1 ⫽ strongly bilabiate. 36. Corolla throat floor: 0 ⫽ not deeply grooved, 1 ⫽ deeply grooved. 37. Exterior palate pubescence: 1 ⫽ tapered, 2 ⫽ clavate. This refers to the shape of the hairs found directly outside the corolla tube, largely concentrated on the palate. 38. Interior palate pubescence: 1 ⫽ mammillate, 2 ⫽ thick-mammillate, 3 ⫽ clavate, 4 ⫽ fine-mammillate. This characterizes the shape of pubescence found within the corolla tube. 39. Interior palate ridges: 0 ⫽ posterior pubescent, 1 ⫽ posterior glabrous. While all species examined had interior palate pubescence, in some species the posterior portion of the palate ridges were glabrous. 40. Corolla aperture: 0 ⫽ con-

2006]

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

stricted, 1 ⫽ open. This character is relative to flower-size: a largeflowered (15–25 mm in width) species would have a constricted aperture if the space at the corolla opening was 0–4.0 mm, while a small-flowered species (5–15 mm) would have a constricted aperture if the space at the corolla opening was 0–2.0 mm. 41. Corolla lobe margins: 0 ⫽ notched, 1 ⫽ entire. 42. Upper corolla lobe orientation: 0 ⫽ not reflexed, 1 ⫽ reflexed. 43. Lateral corolla lobe orientation: 0 ⫽ angled outwardly, 1 ⫽ obviously angled downwardly. 44. Lower-center corolla lobe size: 0 ⫽ similar in size to other lobes, 1 ⫽ obviously larger than all other corolla lobes. 45. Anther pubescence: 0 ⫽ absent, 1 ⫽ present. 46. Anther

395

opening: 1 ⫽ opening widely to the point of being slightly reflexed, 2 ⫽ opening widely but incompletely, 3 ⫽ opened only at the apex. 47. Anther sacs: 1 ⫽ theca equal, 2 ⫽ theca sub-equal. 48. Pollen type: 1 ⫽ IIb, 2 ⫽ I, 3 ⫽ IIc. For a detailed description of the pollen morphological types, see Argue 1980, 1986). 49. Stigma lobe shape: 1 ⫽ obovate, 2 ⫽ lanceolate, 3 ⫽ ovate. 50. Stigma lobe symmetry: 0 ⫽ upper and lower lobes unequal, 1 ⫽ upper and lower lobes equal. 51. Stigma color: 1 ⫽ white, 2 ⫽ cream, 3 ⫽ yellow. 52. Style pubescence: 0 ⫽ absent, 1 ⫽ present. 53. Seed reticulation: 0 ⫽ not visible at 20 ⫻ magnification, 1 ⫽ visible at 20 ⫻ magnification.

396

[Volume 31

SYSTEMATIC BOTANY

APPENDIX 3. Morphological data matrix for the M. moschatus alliance and outgroups. Character numbers refer to those listed in Appendix 2. ? ⫽ unknown character state, parentheses indicate a polymorphic character state. Character No.

M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M.

bicolor lewisii primuloides ampliatus hymenophyllus patulus pulsiferae evanescens breviflorus jungermannioides washingtonensis moschatus latidens floribundus OR norrisii dudleyi dentatus guttatus OR alsinoides floribundus CO tenellus nepalensis guttatus CA bodinieri

1

2

3

4

7

8

9

10 11 12 13 14 15 16 17

0 1 1 0 0 0 0 0 0 1 0 1 0 0 0 0 1 (0,1) 0 0 1 1 (0,1) 1

1 1 0 1 0 1 1 1 1 0 1 (0,1) 1 (0,1) 1 1 (0,1) 1 1 (0,1) (0,1) (0,1) 1 0

0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

? 1 0 0 ? 1 1 1 1 2 1 0 1 3 0 0 1 3 2 0 1 3 0 0 1 3 0 0 1 3 0 0 1 3 0 0 0 3 2 0 0 3 0 0 0 3 1 0 0 3 1 0 0 3 0 0 0 3 2 0 0 3 2 0 0 ? 1 0 0 (2,3) (0,1,2) 1 0 ? 2 0 0 3 0 0 ? ? 1 0 ? 3 1 0 0 (2,3) (0,1,2) 1 ? ? 1 0

5

6

0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 1 1

0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0

? 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 0 0 0 0 1 0 0 0 1 1 0 0 0 1 1 0 0 1 1 1 1

? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 1

1 1 1 0 0 0 1 1 1 0 0 1 1 0 1 1 1 0 0 0 1 1 0 0

0 1 0 0 1 0 0 0 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 0

0 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 1 1 0 0 0 0

18

19

20

21 22 23 24 25 26 27

1 0 1 0 1 0 0 1 1 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 1 0 1 0 1 0 1 0 0 0 1 (0,1) 0 0 1 0 1 0 1 0 0 0 0 1 0 0 0 0 0 2 1 0 1 0 0 1 0 0 0 0 0 1 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 ? 1 1 0 1 0 1 0 0 1 0 1 0 0 0 0 0 2 1 0 1 0 0 2 1 (0,1) 1 0 0 0 ? 1 1 0 0 2 0 0 ? 0

0 1 0 0 0 0 0 0 0 1 0 1 0 1 1 1 1 0 0 1 0 0 0 0

1 0 0 0 0 0 0 1 1 1 0 0 1 0 0 0 0 1 0 0 1 1 1 1

? 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 ? ? 1 ?

1 1 0 0 0 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 1 1 1 1

1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 0 0

2006]

397

WHITTALL ET AL.: MIMULUS MOSCHATUS ALLIANCE PHYLOGENY

APPENDIX 3. Extended.

Character No.

M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M. M.

bicolor lewisii primuloides ampliatus hymenophyllus patulus pulsiferae evanescens breviflorus jungermannioides washingtonensis moschatus latidens floribundus OR norrisii dudleyi dentatus guttatus OR alsinoides floribundus CO tenellus nepalensis guttatus CA bodinieri

28

29

30

31

? 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 ? ? 0 ?

1 0 1 1 0 1 1 0 1 1 1 1 0 1 ? 1 1 0 1 1 1 1 0 1

? 2 2 2 3 2 2 2 2 2 2 2 2 1 1 1 2 2 2 1 ? ? 2 ?

2 1 (2,3) 1 1 1 1 1 2 1 1 2 1 1 2 1 0 2 1 (0,1) 2 1 0 2 1 0 2 1 1 2 1 1 2 1 1 2 1 0 3 2 0 2 2 1 2 2 1 2 1 1 2 1 (0,1) 2 1 1 2 2 0 2 2 0 2 2 0 2 1 (0,1) 2 2 0 2

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

0 0 1 0 0 0 0 0 0 1 1 0 ? 0 1 0 0 1 1 0 0 0 1 0

1 1 0 1 1 0 0 0 0 1 1 0 0 0 0 0 1 1 1 0 0 0 1 0

1 1 0 0 0 0 0 0 0 0 0 1 0 0 ? 1 1 0 0 0 ? ? 0 ?

? 1 2 2 2 2 2 2 ? 2 2 1 2 2 2 2 2 1 1 2 ? ? 1 ?

? 1 2 1 1 1 1 ? ? 1 4 1 1 1 1 1 3 1 4 1 ? ? 1 ?

? 0 0 0 0 0 0 ? ? 0 0 0 0 1 1 1 1 0 0 1 ? ? 0 ?

1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 0 1 1 1 1 0 1

0 0 0 0 0 1 1 1 ? 0 0 1 1 1 1 1 0 0 1 1 1 1 0 1

0 1 0 1 1 0 0 0 0 1 1 0 0 0 0 0 1 1 0 0 0 0 1 0

0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1 1 1 0 0 0 1 0

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0

1 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 ?

? 1 2 2 2 2 2 ? ? 2 2 1 2 1 ? 1 1 3 2 1 ? ? 3 ?

? 1 ? 2 2 2 2 ? ? 2 2 1 ? 1 ? 1 1 1 1 1 ? ? 1 ?

? 1 2 ? 3 3 1 ? 3 3 3 3 3 1 1 ? 3 2 1 1 ? ? 2 ?

1 2 2 3 2 2 1 ? 1 2 1 ? 3 1 1 1 2 1 2 1 ? ? 1 ?

0 0 1 1 1 1 1 1 ? 0 1 ? 0 1 1 1 1 1 0 1 1 1 1 1

? ? ? 2 2 2 ? 1 ? 2 3 1 ? 1 ? 1 1 ? 1 1 ? ? ? ?

0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 0 0 0 0 1 0

1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0