Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

Phylogeny of the popcorn flowers: Use of genome skimming to evaluate monophyly and interrelationships in subtribe Amsinckiinae (Boraginaceae) Michael G. Simpson,1 C. Matt Guilliams,2 Kristen E. Hasenstab-Lehman,1,3 Makenzie E. Mabry4 & Lee Ripma1 1 Department of Biology, San Diego State University, San Diego, California 92128, U.S.A. 2 Department of Conservation and Research, Santa Barbara Botanic Garden, Santa Barbara, California, U.S.A. 3 Botany Department, National Museum of Natural History, Smithsonian Institution, Washington, D.C., U.S.A. 4 Division of Biological Sciences and Bond Life Sciences Center, University of Missouri, Columbia, U.S.A. Author for correspondence: Michael G. Simpson, [email protected] ORCID MGS, http://orcid.org/0000-0002-6197-2132

DOI https://doi.org/10.12705/666.8

Abstract Subtribe Amsinckiinae, currently containing 13 genera and approximately 287 species, is a species-rich group of the family Boraginaceae. Past studies assessing relationships had a limited sample size and generally weak support. Here we study phylogenetic relationships of Amsinckiinae using a large sample size and considerably more sequence data in order to evaluate the interrelationships of genera and clades within this group. Using high-throughput, genome skimming sequencing of 139 samples of Amsinckiinae and four outgroup taxa, maximum likelihood and Bayesian analyses of separate plastome, cistron, and mitochondrial datasets are presented. In almost all analyses the common ancestor of the Amsinckiinae gives rise to an Andersonglossum or to an Andersonglossum + Adelinia clade. Most genera, including Amsinckia, Eremocarya, Greeneocharis, Harpagonella, Oreocarya, and Pectocarya, are consistently monophyletic with strong support. Plagiobothrys is confirmed to be non-monophyletic, composed of three clades conforming to generic sections. Cryptantha is also non-monophyletic, with most species within a strongly supported Cryptantha s.str. clade, but some nesting within Johnstonella or our Maritimae clade, all with strong support. Although genome skimming verifies the monophyly of many genera and clades of Amsinckiinae, relationships among those clades and along the backbone of the trees remain uncertain, their elucidation possibly a factor of short branch lengths and likely requiring different types of molecular data. Our study may serve as a baseline for future work on the morphology, reproductive biology, and biogeography of the Amsinckiinae.

Keywords Amsinckiinae; Boraginaceae; Cryptanthinae; genome skimming; high-throughput sequencing; phylogenetics

Supplementary Material The Electronic Supplement (Figs. S1–S3) is available in the Supplementary Data section of the on- line version of this article (http://www.ingentaconnect.com/content/iapt/tax); DNA sequence alignments are available from TreeBASE (http://purl.org/phylo/treebase/phylows/study/TB2:S21212)

INTRODUCTION sense, as in Luebert & al. (2016) and Chacón & al. (2016), and our use of the name “Boraginaceae” is with this circumscrip- Boraginales are a group of flowering plants consisting tion for the remainder of this article. of up to 2700 species of trees, shrubs, or (often) herbs, with Boraginaceae contain 1600–1700 species (Chacón & al., a nearly worldwide distribution (Mabberley, 2008; Luebert & 2016) and possess numerous diagnostic morphological char- al., 2016). This plant order has been the focus of several broad acteristics, including a hirsute to hispid vestiture, a usually phylogenetic studies (Långström & Chase, 2002; Nazaire & circinate scorpioid cyme inflorescence, mostly actinomorphic Hufford, 2012; Cohen, 2014; Weigend & al., 2014; Luebert & al., flowers, a strongly four-lobed ovary from a gynobasic style, 2016). The Boraginales have been subject to differing circum- and a fruit that is a schizocarp of nutlets, the last two characters scriptions over the years, having been treated as equivalent to apomorphic for the family (Luebert & al., 2016). The focus of a single, large family Boraginaceae s.l. (e.g., Mabberley, 2008; this study is a group within Boraginaceae, recently delimited APG IV, 2016) with several subfamilies, or split into as many as subtribe Amsinckiinae (of subfamily Cynoglossoideae, tribe as eleven families (e.g., Weigend & al., 2014; Luebert & al., Cynoglosseae, after Chacón & al., 2016), many members of 2016), including the Cordiaceae, Ehretiaceae, Heliotropiaceae, which have been referred to as the “popcorn flowers”. The Hydrophyllaceae, and a more narrowly circumscribed Bora­ Amsinckiinae is currently calculated to contain approximately gin­aceae s.str. We elect to treat the Boraginaceae in this strict 287 species and 330 minimally ranked taxa (see Table 1), a

Received: 29 Nov 2016 | returned for (first) revision: 24 Jan 2014 | (last) revision received: 21 Jun 2017 | accepted: 22 Jun 2017 || publication date(s): online fast track, n/a; in print and online issues, 22 Dec 2017 || © International Association for Plant Taxonomy (IAPT) 2017

1406 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

diversity that will undoubtedly increase as the taxa become recognized tribe Eritricheae (within which Cryptantha had better studied. This is a relatively high richness within the traditionally been placed). Their small sample size prevented Boraginaceae, up to 18% of the species in the family. Here, these authors from making more than cursory observations we aim to infer the phylogenetic relationships within the about affinities of Cryptantha. Nazaire & Hufford (2012), in Amsinckiinae in order to assess the monophyly of genera and a broad study of the Boraginaceae s.l., using 318 samples and clades and to evaluate their interrelationships. DNA sequence data from two cistron regions and four chlo- Past phylogenetic studies. — Prior to molecular phyloge- roplast markers, recovered the same four tribes (within their netic studies, the members of the Amsinckiinae belonged pri- Boraginoideae of Boraginaceae s.l.) as did Långström & Chase marily to tribe Eritrichieae (see, e.g., Riedl, 1997), along with (2002). Within their Cynoglosseae, four species of Cryptantha many other genera now included in other tribes and subtribes and one species of Plagiobothrys constituted a clade sister to of the Boraginaceae. Långström & Chase (2002), in an early another clade of two species of Amsinckia, the two linked molecular study that included 38 taxa of the Boraginaceae (their with mixed support (i.e., strongly supported in one analysis Boraginoideae of Boraginaceae s.l.), using the single chloro- but weakly supported in another). plast marker atpB, argued for the recognition of four, mono- Hasenstab-Lehman & Simpson (2012) focused on the phyletic tribes within the group: Boragineae, Cynoglosseae, interrelationships of Cryptantha in a molecular phylogenetic Lithospermeae, and a new tribe, Echiochileae. These research- analysis of 70 taxa using sequence data from ITS and the ers were the first to include a member of the Amsinckiinae (as chloroplast marker trnLUAA. Although they had limited out- recognized here) in such a study, in this case a single sample group sampling, the authors resolved a mixed to strongly sup- of the genus Cryptantha Lehman ex G.Don, C. virgata (Porter) ported large clade, which they called subtribe Cryptanthinae Payson (= Oreocarya virgata (Porter) Greene), which placed (= Amsinckiinae). (Note that the name “Cryptanthinae” was within their tribe Cynoglosseae, inclusive of the formerly determined to be invalid by the rules of the International Code of Nomenclature for algae, fungi, and plants—McNeil & al., 2012; K. Gandhi, J. McNeill, J. Reveal, J. Strother, pers. comm.; see below.) What had been circumscribed by most taxonomists Table 1. Consensus classification of genera of subtribe Amsinckiinae as the genus Cryptantha (see Simpson & Hasenstab, 2009) with current number of species and minimally ranked taxa (including was determined to be polyphyletic. Species traditionally clas- varieties and subspecies) accepted by the authors. sified in Cryptantha were found to comprise 5–6 separate Minimum- and strongly supported clades, warranting the authors to split Genus Species ranked taxa Cryptantha s.l. into five genera: Cryptantha s.str. and the resur- Adelinia J.I.Cohen 2015 1 1 rected genera Eremocarya, Greeneocharis, Johnstonella, and Amsinckia Lehm. 1831, nom. cons. 15 17 Oreocarya. Moreover, Cryptantha s.str. was resolved as two Andersonglossum J.I.Cohen 2015 2 2 clades, Cryptantha s.str. 1 and Cryptantha s.str. 2, which were either sister taxa (in their parsimony analysis) or well separated Cryptantha Lehm. ex G.Don 1837 103 115 from one another (in their Bayesian and maximum likelihood Dasynotus I.M.Johnst. 1948 1 1 analyses). Other clades with strong support in their analyses Eremocarya Greene 1887 2 3 were the genera Amsinckia and Pectocarya. Finally, the genus Greeneocharis Gürke & Harms 1899 2 3 Plagiobothrys was inferred to be non-monophyletic, composed of a main clade containing members of sect. Allocarya and sect. Harpagonella A.Gray 1876 2 2 Plagiobothrys only (Plagiobothrys s.str.), a clade correspond- Johnstonella Brand 1925 13 15 ing to Plagiobothrys sect. Sonnea (= genus Sonnea) and a clade Oncaglossum Sutorý 2010 1 1 corresponding to Plagiobothrys sect. Amsinckiopsis (this last Oreocarya Greene 1887 63 72 group monophyletic in one analysis, diphyletic in another, both with weak support). The interrelationships of these clades/gen- Pectocarya DC. ex Meisn. 1840 13 14 era varied in their analyses. Although some interrelationships Plagiobothrys Fisch. & C.A.Mey. 1836 among genera remained constant (see below), the so-called sect. Amsinckiopsis I.M.Johnst. 2 3 “backbone” interrelationships of these genera and clades varied sect. Allocarya I.M.Johnst., sect. Echidio­carya by analysis with generally weak support along nodes. I.M.Johnst., sect. Plagiobothrys I.M.Johnst. 65 79 Subsequent molecular phylogenetic analyses of the sect. Sonnea I.M.Johnst. (= Sonnea Greene 2 2 Boraginaceae encompassing members of this plant complex 1887) include the studies of Weigend & al. (2013), Cohen (2014, 2015), Otero & al. (2014), and Chacón & al. (2016), all using vari- Total 287 330 ous combinations of ITS and coding and/or intergenic chloro- Consenus from the phylogenetic analyses of Hasenstab-Lehman & plast markers. These studies have almost entirely converged Simpson (2012), Weigend & al. (2013), Cohen (2014, 2015), Otero on the recognition of a strongly supported clade that includes & al. (2014), and Chacón & al. (2016), Note that, for now, we are not including the monospecific Nesocaryum I.M.Johnst. (Johnston, Cryptantha and relatives, within tribe Cynoglosseae. This 1927), which was accepted by Chacón & al. (2016), as a member of clade is now formally recognized (see Chacón & al., 2016) the subtribe. as subtribe Amsinckiinae Brand, the first available name for

Version of Record 1407 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

this complex of plants that currently includes members of 13 Chacón & al., 2016), these being sister to Greeneocharis in genera: Adelinia, Amsinckia, Andersonglossum, Cryptantha, some studies (Hasenstab-Lehman & Simpson, 2012; Otero Dasynotus, Eremocarya Greeneocharis, Harpagonella, & al., 2014, in part). Eremocarya and Greeneocharis, where Johnstonella, Oncaglossum, Oreocarya, Pectocarya, and included, are always monophyletic, but sample size of these Plagiobothrys (Table 1). two genera (each containing only two species) has often been Although there are differences in the details of these stud- limited to a single specimen. The genus Johnstonella is mono- ies, these likely a function of taxon and DNA sequence sam- phyletic in two studies (Hasenstab-Lehman & Simpson, 2012; pling, several common patterns have emerged. In those studies Otero & al., 2014, in part). Cryptantha s.str. (minus the seg- in which they were included, three North American species of regate genera Eremocarya, Greeneocharis, Johnstonella, and Cynoglossum arise as part of a polytomy (Weigend & al., 2013) Oreocarya) remains consistently polyphyletic in all analyses, as a grade (Cohen, 2015; Chacón & al., 2016) from the com- often with weak support. In these analyses Cryptantha s.str. mon ancestor of what is now recognized as the Amsinckiinae. is often (but not always) classified apart from Eremocarya, This novel placement of these North American species of Greeneocharis, and Oreocarya, but species of Johnstonella are Cynoglossum within the complex was first discovered by sometimes included within Cryptantha s.str. Both Hasenstab- Weigend & al. (2013), with Cohen (2015) summarizing their Lehman & Simpson (2012) and the ITS analysis of Otero & al. comparative morphology and naming them as genera nova, (2014) recovered similar Cryptantha s.str. 1 and s.str. 2 clades. Adelinia and Andersonglossum. Two other North American Finally, interrelationships of genera and major clades of the genera (both monospecific) were also discovered to be mem- Amsinckiinae, especially along the backbone of the tree, gener- bers of the Amsinckiinae. The genus Oncaglossum, which had ally vary significantly between analyses and often with poor also been earlier classified in the genus Cynoglossum, was support, the clades sometimes forming polytomies. shown to be part of the Amsinckiinae by Cohen (2014, 2015), Members of the Amsinckiinae exhibit a great range of vari- in the only studies where the genus was included. Oncaglossum ation in nutlet morphology (size, shape, sculpturing patterns, was resolved as sister either to all other members of the gynobase attachment, heteromorphism, and dimorphism), plant Amsinckiinae (Cohen, 2014) or to all other members of the duration (annuals, biennials, or perennials), and reproductive subtribe except Adelinia and Andersonglossum (when the latter biology (cleistogamous versus chasmogamous, heterostyly ver- two genera were included; Cohen, 2015). The genus Dasynotus sus homostyly; see Simpson & Hasenstab, 2009). In addition, was established by Weigend & al. (2013), Cohen (2014, 2015), many members of the group exhibit a classic American amphi- and Chacón & al. (2016) to belong in the Amsinckiinae clade. tropical distribution, occurring on either side of the American Weigend & al. (2013) resolved Dasynotus in a polytomy tropics (Raven, 1963; Wen & Ickert-Bond, 2009) (see Fig. 1). with all other Amsinckiinae examined except Adelinia and Thus, the Amsinckiinae constitutes a particularly interesting Andersonglossum. Cohen (2014) resolved Dasynotus as sister group of plants with respect to several facets of biology and to three Oreocarya species or to a clade containing Cryptantha, biogeographic history. Greeneocharis, Pectocarya, and Plagiobothrys species. Goals. — The overriding goal of this study is to infer, However, Cohen (2015), using the same molecular markers with greater confidence, the phylogeny of the Amsinckiinae. but a reduced taxon sampling focusing on tribe Cynoglosseae, Using the high-throughput sequencing method of genome resolved it as sister to all other examined Amsinckiinae, minus skimming, we compare different gene trees from analyses of Adelinia, Andersonglossum, and Oncaglossum. Finally, Chacón most or all of the chloroplast (cpDNA) genome, most or all & al., 2016 resolved Dasynotus to be sister to all other exam- of the nuclear ribosomal cistron (nrDNA) genome (including ined Amsinckiinae except Adelinia, Andersonglossum, and a ITS and ETS), plus a significant portion of the mitochondrial Pectocarya-Harpagonella clade. (The placement of the South (mtDNA) genome. From these analyses, we assess the mono- African Cynoglossum obtusicalyx in a polytomy with all other phyly of named genera, infer the interrelationships of the major examined Amsinckiinae in the plastid analysis of Otero & al., clades, and evaluate the need for future taxonomic changes 2014 is likely erroneous and needs to be tested.) in the group. We hope this study will lay the groundwork for Phylogenetic relationships of the other genera of the additional work on the character evolution, divergence timing, Amsinckiinae have also varied in the aforementioned analy- and phylogeographic history of this interesting complex of ses. However, some general trends can be noted. The genera plants, which we plan to address in future studies. Amsinckia and Pectocarya are recovered as monophyletic in all analyses, and (where included) Harpagonella is always sister to Pectocarya. Oreocarya is monophyletic in the analyses of MATERIALS AND METHODS Hasenstab-Lehman & Simpson (2012), Weigend & al. (2013), Otero & al. (2014, in part), and Cohen (2015), and, when so, Taxon sampling. — DNA was extracted from 143 silica- support values for the clade are high. Eremocarya is sister to dried leaf samples collected concurrently with vouchered Oreocarya in Hasenstab-Lehman & Simpson (2012) and in specimens or taken directly from previously collected herbar- Otero & al. (2014, in part). The species of Plagiobothrys in ium specimens, generally within the last 20 years. Voucher sect. Allocarya and sect. Plagiobothrys consistently form a specimens are housed at the following herbaria: ARIZ, well-supported clade (Hasenstab-Lehman & Simpson, 2012; CONC, DUKE, GH, JEPS, MERL, MO, NSW, RSA, SBBG, Weigend & al., 2013; Cohen, 2014, 2015; Otero & al., 2014; SD, SDSU, SGO, SI, UC, UCR, and UTC. Appendix 1 lists

1408 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

voucher information and National Center for Biotechnology library preparation, three separate runs of high-throughput Information (NCBI) accessions. Sampling included four out- sequencing were performed. Run one utilized the Illumina group taxa, Cynoglossum creticum Mill., Hackelia micrantha HiSeq 2000 genetic analysis system (Illumina, San Diego, (Eastw.) J.L.Gentry, Microula tibetica Benth., and Myosotis California, U.S.A.) at the University of Delaware Sequencing laxa Lehm. Amsinckiinae ingroup taxa (n = 139) encompassed and Genotyping Center for 38 samples in one lane, with an 12 genera (all but the monotypic Oncaglossum), 123 species average of 3,474,074 (1,332,453–7,593,640) reads. A second run (42% of the subtribe), and 127 minimum-ranked taxa (38% of of 53 samples was performed at the University of California the subtribe) (Appendix 1). A map showing the distribution of at Riverside Genomics Core in a single lane, with an aver- all Amsinckiinae sampled in this study and general ranges of age of 2,926,555 (707,000–5,092,337) reads. A third run was the subtribe is seen in Fig. 1. performed on an Illumina HiSeq2500 at Global Biologics DNA isolation and sequencing. — Genomic DNA extrac- for 96 samples in a single lane, with an average of 1,998,447 tion and sequencing follow Ripma & al. (2014). Briefly, ge- (820,347–4,835,141) reads. All three sequencing runs produced nomic DNA was isolated using a modified cetyltrimethylam- single-end 100 bp reads with an average insert size of 250 bp. monium bromide (CTAB) protocol (Doyle & Doyle, 1987; Friar, The last two runs were shared with other researchers to reduce 2005). Whole genomic DNA was then sent to Global Biologics costs. All reads are deposited at the Short Read Archive of the (Columbia, Missouri, U.S.A.) for library preparation. After NCBI (https://www.ncbi.nlm.nih.gov; see Appendix 1).

Fig. 1. Distribution map of the 139 sampled specimens (dots) and native ranges of genera (highlighted areas) in subtribe Amsinckiinae, members of which show an American amphitropi- cal disjunct distribution. Only Andersonglossum virginianum (inclusive of A. boreale (Fernald) J.I.Cohen) is natively distributed Adelinia in eastern U.S.A. (also occur- Andersonglossum o. ring in western Canada), and Amsinckia Cryptantha only species of Plagiobothrys Dasynotus natively occur in Australia. Harpagonella Andersonglossum *Oncaglossum not included in Eremocarya virginianum our study. Distribution data, in Greeneocharis part, from Global Biodiversity Johnstonella Information Facility (http://www. Oncaglossum* Oreocarya gbif.org). Pectocarya Plagiobothrys

Amsinckia Cryptantha Greeneocharis Johnstonella Pectocarya Plagiobothrys

Plagiobothrys spp.

Version of Record 1409 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

DNA quality, assembly, and alignment. — Raw read quality MegaBLAST searched against the mtDNA contig bin using a control and filtering follows the methods in Ripma & al. (2014) query centric alignment output and the settings for LCNG gene using PRINSEQ (Schmieder & Edwards, 2011), with removal searches. The result was an alignment of each N. tabacum exon of reads being exact duplicates, having a mean quality Phred and the corresponding sequence(s) from each sampled taxon. score below 30, or having more than one N. Both the 3′ and 5′ Only alignments with a single copy from each sample were ends were trimmed to a Phred quality score of 30 using a win- retained, and several of these exons were partial. Exons were dow size of 1 (Straub & al., 2013). Any read less than 50 bp in aligned using the MAFFT plugin (Katoh & al., 2002) with length was removed, as well as any remaining sample barcodes. default settings. Sequences were edited with the same methods Post-quality control reads were imported into Geneious v.7.1.5 as those used for the cpDNA plastome (above). (Biomatters, Auckland, New Zealand, http://www.geneious. Phylogenetic analyses. — Only sequence data represented com) in FASTQ format, hereafter referred to as read pools. in all samples for each genome were used in subsequent analy- For assembly of the plastome (cpDNA), we used a ref- ses. All analyses were run with the outgroups Cynoglossum erence-guided assembly to an annotated 124,868 bp partial creticum, Hackelia micrantha, Microula tibetica, and Myosotis plastome sequence of Pectocarya penicillata (Hook. & Arn.) laxa and rooted afterwards with Hackelia micrantha, following A.DC. (from Ripma & al., 2014), which was constructed us- the results of Weigend & al. (2013) and Chacón & al. (2016). ing Geneious v.7.017 with default settings. Reference-guided All partitioning schemes and models of molecular evolution assembly to the P. penicillata plastome was implemented in were determined using PartitionFinder2 v.2.1.1 (Lanfear & Geneious, with default settings and 25 iterations of the read al., 2016). Each genome was analyzed separately, with the cp- pool from each sample. Consensus contigs for each sample DNA partitioned by codon position for the entire plastome, were generated using a 75% similarity threshold, masking areas the nrDNA partitioned by gene and non-coding spacer regions with less than 20 bp sequence depth with gaps, and retain- (ETS, 18S, ITS1, 5.8S, ITS2, 26S), and the mtDNA partitioned ing IUPAC ambiguity codes. Sequences were aligned using by exons. Phylogenetic trees were inferred using maximum the MAFFT plugin v.1.3.3 in Geneious v.7.017 (Katoh & al., likelihood (ML) and Bayesian (BI) frameworks implemented 2002) with default settings. Lastly, final alignments were ex- on the CIPRES portal (Miller & al., 2010). For all three align- amined for misaligned areas by eye, and adjusted accordingly. ments, the GTR + I + G model of evolution was used for both ML Total number of informative characters were calculated using and BI analyses. Searches for the tree topology with the highest GARLI v.2.0 (Zwickl, 2006), implemented in Geneious v.7.017 likelihood score were conducted in RAxML v.8 (Stamatakis, using default settings. 2014), with support assessed using 1000 rapid bootstrap (BS) Ribosomal cistron (nrDNA) assembly followed methods replicates. Bayesian analyses were implemented in MrBayes detailed in Ripma & al. (2014), with some exceptions. The v.3.2 (Ronquist & al., 2012), with the MCMC chain run for 10 assembled cistron of Cryptantha torreyana (A.Gray) Greene million generations. var. torreyana from Ripma & al. (2014) was used for reference- Resulting trees were viewed in FigTree (Rambaut, 2006– guided assembly of the read pools in Geneious using 25 2014), with final graphics prepared with Adobe InDesign CC. iterations, medium-low sensitivity, and default settings. The Bootstrap support was considered strong if ≥ 70% and poste- methods for generating nrDNA consensus sequences, sequenc- rior probabilities (PP) strong if ≥ 0.95; otherwise, either were ing editing, and alignment followed those employed for the considered to be weak in support. A combination of strong and plastome (above). Positions that contained an ambiguity code weak support between the ML and BI analyses was indicated were removed from the alignment, as we were not able to de- as “mixed”. termine whether these ambiguities were due to polymorphic states within the ITS/ETS regions or due to sequencing error. Mitochondrial (mtDNA) exons were obtained using a RESULTS reference-guided assembly of each read pool to a GenBank Nicotiana tabacum L. mitochondrion (GenBank: BA000042), The total alignment lengths, number of parsimony-infor- modified to include only one copy of each annotated repeat mative characters, and number of uninformative characters region. A consensus contig was saved using the same meth- for the three datasets are seen in Table 2 and are available ods presented for the cpDNA plastome (above). A single con- at TreeBase for download (http://purl.org/phylo/treebase/ tig from each sample was made into a custom database in phylows/study/TB2:S21212). The ML and BI analyses result Geneious, henceforth referred to as the mtDNA contig bin. Each N. tabacum exon was separated using the extract anno- tations feature in Geneious. To exclude chloroplast sequences Table 2. Alignment lengths, number of parsimony-informative char- from the mitochondrial assembles, these exons were BLASTN acters (PICs), and number of uninformative characters (UCs) for the searched against the cited plastome of S. lycopersicum, using three types of DNA used in this study. an E-value of 1e-10, a k-mer length of 15, a scoring match- DNA Alignment length PICs UCs mismatch of 2-3, and a 5-2 open extend gap cost (this BLASTN cpDNA 107,155 bp 9,837 6,765 search being more likely to find matches than the MegaBLAST mtDNA 20,511 bp 4,567 2,154 search used elsewhere). Only N. tabacum exons with no match to chloroplast sequences were retained. Each retained exon was nrDNA 5,532 bp 419 190

1410 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

An. virginianum An. occidentale And. Ad. grandis Adel. D. daubenmirei H. arizonica Dasy. H. palmeri Pe. pusilla Pe. setosa Pect./ Pe. linearis var. f. Pe. penicillata Harp. Pl. jonesii Pl. kingii var. h. Am. tessellata var. t. Am. intermedia 1 Amsinckia/ Am. intermedia 2 Am. intermedia 3 E. lepida 1 Amsinckiopsis E. lepida 2 E. micrantha var. m. 1 E. micrantha var. p. Eremocarya E. micrantha var. m. 2 C. martirensis C. clokeyi C. muricata var. m. J. echinosepala C. subamplexicaulis Maritimae clade C. maritima var. m. O. crymophila O. hyposphila O. sobolifera O. suffruticosa var. a. O. virgata O. setosissima O. “ursina” O. nubigena “mam.” O. “howelliana” Oreocarya O. subretusa “type” O. schoolcraftii O. glomerata O. subretusa “mt ed.” O. nubigena “craters” O. hoffmannii O. humilis subsp. h. O. nubigena “granite” O. subretusa “w. mts” Pl. hispidus Pl. glomeratus Pl. fulvus var. c. Sonnea Pl. nothofulvus Pl. collinus var. u. Pl. canescens var. c. Pl. myosotoides NA Pl. tenellus Pl. torreyi Pl. plurisepalus Pl. polycaulis Plagiobothrys Pl. gracilis Pl. greenei Pl. austinae Pl. undulatus s.str. Pl. glyptocarpus Pl. leptocladus Pl. mollis Pl. congestus Pl. linifolius cf. G. similis 1 G. similis 2 G. circumscissa var. c. G. circumscissa var. r. Greeneocharis J. angustifolia J. inaequata J. fastigiata J. racemosa J. holoptera J. costata J. grayi var. g. J. angelica Johnstonella/Albidae clade J. diplotricha C. hispida C. albida C. mexicana C. texana C. recurvata C. crassisepala C. minima C. kelseyana C. fendleri C. nevadensis C. scoparia C. utahensis C. gracilis C. oxygona C. mohavensis C. pterocarya f. p. C. ganderi C. intermedia var. i. C. barbigera var. b. C. microstachys C. corollata C. clevelandii C. wigginsii C. decipens C. alfalfalis C. calycotricha C. aspera C. cynoglossoides C. glomerata subsp. g. C. alyssoides Cryptantha C. kingii C. glomerulifera s.str. C. gnaphalioides J. micromeres C. dumetorum

C. simulans C. watsonii C. torreyana C. ambigua C. crinita C. mariposae cpDNA C. incana C. echinella C. leiocarpa C. hispidissima 1 C. hispidissima 2 C. nemaclada C. juniperensis C. diffusa C. peruviana C. phaceloides C. globulifera 1 C. globulifera 2 C. globulifera 3 Fig. 2. Maximum likelihood phylogram from complete plastome (cpDNA)C. globuliferaanalysis, 4 illustrating major clades of the Amsinckiinae. Graphics and abbreviations as in Fig. 3. 0.003 Version of Record 1411 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

An. virginianum An. occidentale Thick horizontal lines = strong Ad. grandis And. D. daubenmirei Adel. support; medium horizontal lines Pe. setosa Dasy. Pe. pusilla H. arizonica = mixed support; thin horizontal H. palmeri Pect./ Pe. linearis var. f. lines = weak support or terminat- Pe. penicillata Harp. E. lepida 1 ing taxa. Dashed vertical line E. micrantha var. m. 2 E. lepida 2 represents alternative arrangement E. micrantha var. m. 1 Eremocarya E. micrantha var. p. from Bayesian analysis. Major and O. virgata backbone nodes with weak support O. crymophila O. suffruticosa var. a. O. setosissima indicated by thick-lined circles, O. humilis subsp. h. O. sobolifera those with mixed support indicated O. hyposphila O. subretusa “mt ed.” by thin-lined circles. Adel. = Adeli­ O. nubigena “craters” O. “ursina” Oreocarya nia; And. = Andersonglossum; O. nubigena “mam.” O. subretusa “type” O. “howelliana” Cy. cr. = Cynoglossum creticum; O. hoffmannii O. schoolcraftii Dasy. = Dasynotus; Harp. = Har­ O. glomerata O. subretusa “w. mts” pagonella; Pect. = Pectocarya. O. nubigena “granite” Pl. glomeratus Pl. hispidus Sonnea J. diplotricha C. albida C. mexicana C. texana J. racemosa J. holoptera J. fastigiata Johnstonella/ J. angelica C. hispida Albidae clade J. grayi var. g. J. costata J. inaequata J. angustifolia Pl. jonesii Pl. kingii var. h. Am. tessellata var. t. Amsinckia/ Am. intermedia 2 Am. intermedia 3 Am. intermedia 1 Amsinckiopsis Pl. fulvus var. c. Pl. canescens var. c. Pl. myosotoides NA Pl. torreyi Pl. tenellus Pl. nothofulvus Pl. collinus var. u. Pl. plurisepalus Plagio- Pl. linifolius cf. Pl. congestus Pl. gracilis bothrys Pl. polycaulis Pl. undulatus Pl. leptocladus Pl. mollis s.str. Pl. austinae Pl. glyptocarpus Pl. greenei Cy. creticum G. similis 1 Cy. cr. G. circumscissa var. c. G. similis 2 G. circumscissa var. r. Greeneocharis C. martirensis C. muricata var. m. C. clokeyi C. subamplexicaulis C. maritima var. m. Maritimae clade J. echinosepala C. recurvata C. fendleri C. minima C. crassisepala C. kelseyana

C. simulans C. ambigua C. torreyana var. t. C. watsonii C. mariposae C. crinita C. dumetorum C. gnaphalioides C. cynoglossoides C. calycotricha C. aspera C. alyssoides C. kingii C. glomerulifera C. glomerata subsp. g. C. alfalfalis C. nevadensis Cryptantha C. scoparia C. utahensis C. gracilis C. mohavensis s.str. C. pterocarya f. p. C. oxygona J. micromeres C. hispidissima 1 C. leiocarpa C. nemaclada C. juniperensis C. ganderi C. microstachys C. decipens C. intermedia var. i. C. barbigera var. b. nrDNA C. wigginsii C. clevelandii C. corollata C. hispidissima 2 C. echinella C. incana C. peruviana Fig. 3. Maximum likelihood C. phaceloides C. diffusa phylogram from complete cistron C. globulifera 2 C. globulifera 1 (nrDNA) analysis, illustrating C. globulifera 3 C. globulifera 4 major clades of the Amsinckiinae.

1412 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

Ad. grandis Adel. An. virginianum And. An. occidentale D. daubenmirei H. arizonica Dasy. H. palmeri Pe. pusilla Pe. setosa Pect./ Pe. penicillata Pe. linearis var. f. Harp. Pl. jonesii Pl. kingii var. h.Am. tessellata var. t. Am. intermedia 1 Amsinckia/ Am. intermedia 3 Am. intermedia 2 Amsinckiopsis Pl. fulvus var. c. Pl. collinus var. u. Pl. nothofulvus Pl. canescens var. c. Pl. myosotoides NA Pl. torreyi Pl. tenellus Pl. plurisepalus Pl. polycaulis Pl. gracilis Plagiobothrys Pl. greenei Pl. austinae Pl. glyptocarpus Pl. leptocladus s.str. Pl. undulatus Pl. mollis Pl. congestus Pl. linifolius cf. E. lepida 1 E. lepida 2 E. micrantha var. m. 2 E. micrantha var. m. 1 Eremocarya E. micrantha var. p. C. subamplexicaulis J. echinosepala C. maritima var. m. C. martirensis C. clokeyi Maritimae clade C. muricata var. m. O. crymophila O. hyposphila O. sobolifera O. suffruticosa var. a. O. setosissima O. virgata O. subretusa “mt ed.” O. schoolcraftii O. glomerata O. subretusa “w. mts” Oreocarya O. nubigena “granite” O. nubigena “craters” O. “ursina” O. humilis subsp. h. O. hoffmannii O. nubigena “mam.” O. subretusa “type” O. “howelliana” Pl. glomeratus Pl. hispidus Sonnea G. similis 1 G. similis 2 G. circumscissa var. r. G. circumscissa var. c. Greeneocharis J. costata J. holoptera J. racemosa J. fastigiata J. inaequata J. angustifolia J. grayi var. g. J. angelica Johnstonella/Albidae clade C. hispida J. diplotricha C. albida C. mexicana C. texana C. recurvata C. minima C. crassisepala C. fendleri C. kelseyana C. scoparia C. nevadensis C. utahensis C. gracilis C. oxygona C. pterocarya f. p. C. mohavensis C. ganderi C. intermedia var. i. C. barbigera var. b. C. microstachys C. corollata C. clevelandii C. decipens C. aspera C. wigginsii C. calycotricha C. alfalfalis C. gnaphalioides C. glomerata subsp. g. C. alyssoides C. glomerulifera C. kingii C. cynoglossoides J. micromeres C. dumetorum C. simulans C. watsonii C. ambigua C. torreyana Cryptantha s.str. C. crinita C. mariposae C. echinella mtDNA C. incana C. leiocarpa C. hispidissima 1 C. hispidissima 2 C. nemaclada C. juniperensis C. globulifera 1 C. phaceloides C. peruviana C. diffusa C. globulifera 2 C. globulifera 4 C. globulifera 3 Fig. 4. Maximum likelihood phylogram from mitochondrial (mtDNA) analysis, illustrating major clades of the Amsinckiinae. Graphics and abbreviations as in Fig. 3.

Version of Record 1413 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

in the same or almost the same topology for a given genome Within this group, Eremocarya and the Maritimae clade are dataset (ML trees illustrated in Figs. 2–4). For all analyses, sister taxa with mixed to weak support, with Oreocarya sister the Andersonglossum clade alone (cpDNA-ML, nrDNA) or an to these two with strong to mixed support (Figs. 2, 4). In the Andersonglossum-Adelinia clade (cpDNA-BI, mtDNA) is sister nrDNA analyses Eremocarya and Oreocarya are sister taxa to all other examined members of the Amsinckiinae (Figs. with mixed support, and the Maritimae clade is distantly related 2–4). Adelinia is sister to all other subtribe members except to these (Fig. 3). Andersonglossum in the cpDNA-ML and nrDNA analyses Cryptantha s.str., the largest clade in our study, is sister to (Figs. 2, 3). the Johnstonella/Albidae clade with strong to mixed support in In all analyses the examined species of Harpagonella the cpDNA and mtDNA analyses (Figs. 2, 4), but is sister to the and Pectocarya comprise a monophyletic group with strong Maritimae clade with mixed support in the nrDNA analyses support, termed here the Pectocarya/Harpagonella clade. In (Fig. 3). Plagiobothrys s.str. is sister either to the Amsinckia/ all cpDNA and mtDNA analyses, Harpagonella is sister to a Amsinckiopsis clade with weak to strong support (nrDNA and monophyletic Pectocarya with strong support (Figs. 2, 4), but mtDNA-BI analyses), to Sonnea with strong support (cpDNA), in the nrDNA analyses Harpagonella is sister to two of the four or to a group of seven major clades with weak support (mtDNA- sampled species of Pectocarya with mixed support (Fig. 3). ML analysis) (Figs. 2–4). The relationships of Greeneocharis The monospecific Dasynotus is sister to the Pectocarya/ and Sonnea differ among analyses. Finally, Cynoglossum cre­ Harpagonella clade with strong support in the cpDNA and ticum, which is sister to Microula tibetica among the outgroups mtDNA analyses, and in these same analyses this Dasynotus- in the cpDNA and mtDNA analyses (Figs. 2, 4), is nested within Pectocarya/Harpagonella clade is sister to the rest of the the Amsinckiinae in the nrDNA analyses, but with weak to Amsinckiinae (minus Adelinia and Andersonglossum) with mixed support (Fig. 3). strong support (Figs. 2, 4). However, in the nrDNA analyses In each of the individual genome phylograms (cpDNA, Dasynotus is sister to all other Amsinckiinae (minus Adelinia nrDNA, and mtDNA; only ML illustrated) relative branch and Andersonglossum) with mixed support (Fig. 3). lengths along the backbone of the Amsinckiinae are generally A clade containing all Amsinckia spp. plus Plagiobothrys short (Figs. 2–4). Relative branch lengths of Oreocarya taxa jonesii, and P. kingii (the last two of Plagiobothrys sect. in the cpDNA (Fig. 2) and nrDNA (Fig. 3) analyses are quite Amsinckiopsis), termed here the Amsinckia/Amsinckiopsis short, whereas those in the mtDNA analysis (Fig. 4) are long, clade, is recovered in all analyses with strong (cpDNA, mtDNA) longer than for most other taxa. Also in the mtDNA-ML analy- to weak (nrDNA) support. Within this clade Plagiobothrys sis, five species of Cryptantha s.str.—C. ambigua (A.Gray) jonesii is sister to P. kingii + Amsinckia, and P. kingii is sis- Greene, C. crinita Greene, C. mariposae I.M.Johnst., C. tor­ ter to Amsinckia, with strong support in all analyses. This reyana (A.Gray) Greene, and C. watsonii (A.Gray) Greene— Amsinckia/Amsinckiopsis clade is sister to the remainder of the show particularly long branches relative to the rest of the genus Amsinckiinae (minus Adelinia, Andersonglossum, Dasynotus, (Fig. 4). These taxa do not show long branches in either the and the Pectocarya/Harpagonella clade) with strong support cpDNA or nrDNA analyses (Figs. 2, 3). A comparison of the in the cpDNA and the mtDNA-ML analyses, whereas, in the three ML cladograms from the cpDNA, nrDNA, and mtDNA mtDNA-BI analysis the Amsinckia/Amsinckiopsis clade is sister analyses illustrates the differences in the placement of major to Plagiobothrys s.str. with strong support (Figs. 2, 4). In the clades (Electr. Suppl.: Figs. S1–S3). nrDNA analyses the Amsinckia/Amsinckiopsis clade is part of a clade with Plagiobothrys s.str. with mixed support (Fig. 3). Johnstonella, as circumscribed by Hasenstab-Lehman & DISCUSSION Simpson (2012), is not monophyletic in our study. However, in all analyses, 10 of the 12 included species of Johnstonella plus Phylogenetic relationships and taxonomic considerations. 4 species of Cryptantha—C. albida (Kunth) I.M.Johnst., C. his­ — The analyses presented here of the “popcorn flowers”, pida (Phil.) Reiche, C. mexicana I.M.Johnst., and C. texana subtribe Amsinckiinae (Boraginaceae), have a significantly Greene—together form a strongly supported monophyletic increased sample size over past studies, and utilized high- group, termed here the Johnstonella/Albidae clade (Figs. 2–4). throughput genome skimming sequencing, yielding much more The remaining genera/clades—Cryptantha s.str., Eremo­ data. For these analyses, we were able to obtain virtually all carya, Greeneocharis, Maritimae clade, Oreocarya, Plagio­ of the cpDNA (plastome), virtually all of the nrDNA (cistron), bothrys s.str. (sect. Allocarya, sect. Echidiocarya, and sect. and many mtDNA markers. Nonetheless these analyses yielded Plagio­bothrys only), and Sonnea (= Plagiobothrys sect. trees of somewhat different topologies (Figs. 2–4). Bootstrap Sonnea)—are recovered in all analyses with strong support, and posterior probability support values for genera or major except for what we term the Maritimae clade (see Discussion), clades are generally strong, but those for nodes along the “back- recovered with mixed support in the nrDNA analyses (Figs. bone” of the Amsinckiinae are mostly weak to mixed, except 2–4). The interrelationships of these seven genera/clades vary for the cpDNA (Fig. 2) analysis, in which backbone nodes are among analyses. However, some common patterns are evident. mostly strongly supported. (See summary cladograms for the Three clades—Eremocarya, Oreocarya, and the Maritimae analyses in Electr. Suppl.: Figs. S1–S3.) Thus, we feel that our clade—together form a monophyletic group in the cpDNA and studies are insufficient to confidently infer evolutionary rela- mtDNA analyses, with strong to mixed support (Figs. 2, 4). tionships among many of the major clades of the Amsinckiinae.

1414 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

The reasons for the conflicts among datasets could be for a although studies with a greater sample size are needed to evalu- variety of reasons. The only biparententally inherited marker, ate more detailed relationships in the complex. the nuclear cistron, is present within the organisms in thou- In two of our analyses (Figs. 2, 4; see Electr. Suppl.: Figs. sands of copies, but may not have homogenized in all cases S1–S3), Dasynotus is sister to the Pectocarya /Harpagonella (Alvarez & Wendel, 2003). Sequences from the plastome and clade with strong support. This novel relationship was not ob- mitochondrion, though showing strongly supported relation- tained by Weigend & al. (2013), Cohen (2014, 2015), or Chacón ships, are almost always uniparentally (usually maternally) & al. (2016), which have Dasynotus placed in polytomies or inherited. Such organelle evolution may not necessarily cor- weakly supported relationships, but none of which are in- respond with species evolution. Moreover, the relatively short compatible with our results based on support values. Aside branch lengths along the backbone of the Amsinckiinae (Figs. from its relationships to Dasynotus, the general placement of 2–4) may be indicative of relatively rapid divergence of the the Pectocarya /Harpagonella clade in the subtribe generally major clades of the subtribe. This may explain both the vary- agrees with or is compatible with that of several earlier studies ing results we obtained with different datasets and analyses (Hasenstab-Lehman & Simpson, 2012; Weigend & al., 2013; (Figs. 2–4), and those across different studies (Hasenstab- Chacón & al., 2016). Lehman & Simpson, 2012; Cohen, 2014; Otero & al., 2014; Aside from Adelinia, Andersonglossum, Dasynotus, and Weigend & al., 2014, Chacón & al., 2016). These relatively Pectocarya /Harpagonella, all but our nrDNA analyses recover short branches along the major nodes of the backbone may a monophyletic grouping within the Amsinckiinae comprised confound future attempts to obtain a robust phylogeny of the of nine clades: Amsinckia /Amsinckiopsis, Cryptantha s.str., subtribe (see Future Work). Eremocarya, Greeneocharis, Johnstonella /Albidae clade, Despite these conflicts among datasets, we have made Maritimae clade, Oreocarya, Plagiobothrys s.str., and Sonnea significant progress in our understanding of the group. The (Figs. 2, 4). Relationships within this clade of nine groups vary Amsinckiinae as a whole is strongly supported as monophyletic among the different analyses, but some patterns are consistent. in all but one (nrDNA) of the analyses, in which Cynoglossum Plagiobothrys s.l., as traditionally treated, is not mono- creticum is nested within the subtribe, and its placement gener- phyletic. However, three segregate clades are monophyletic ally with low support values. Chacón & al. (2016) placed C. cre­ with strong support: Sonnea (Plagiobothrys sect. Sonnea), ticum within the more distantly related subtribe Cynoglossineae, the Amsinckia /Amsinckiopsis clade (including two spe- one of the four subtribes of tribe Cynoglosseae, to which the cies of Plagiobothrys sect. Amsinckiopsis), and our Plagio­ Amsinckiinae belongs. Thus, we suspect that the placement bothrys s.str. clade (including sect. Allocarya, sect. Echidio­ of C. creticum in our nrDNA analysis is in error. Although we carya, and sect. Plagiobothrys only). Sonnea is sister to recognize that our outgroup sampling (n = 4) was limited, our either Plagiobothrys s.str. (with strong to mixed support), to study largely continues to support the recognition of subtribe Greeneocharis (with mixed support), or to a large compo- Amsinckiinae, having the circumscription cited in Chacón & nent of the Amsinckiinae (with weak support; Figs. 2–4). In al. (2016), summarized in Table 1. their study, Hasenstab-Lehman & Simpson (2012) sampled Our analyses show generally strong support for the mono- only Plagiobothrys hispidus of Sonnea, which consistently phyly of most of the (non-monospecific) genera: Amsinckia, occurred separate from other Plagiobothrys species and was Andersonglossum, Eremocarya, Greeneocharis, Harpagonella, sister to their Eremocarya + Oreocarya clade. Other studies Oreocarya, and Pectocarya, although Pectocarya is paraphy- cited did not include samples of Plagiobothrys sect. Sonnea. letic in one analysis (nrDNA), this with weak support (below). Given its strongly supported monophyly, but ambiguous re- The phylogenetic positions of Andersonglossum and lationship to other clades in the subtribe, we propose that the Adelinia generally correspond to or are compatible with the species of Plagiobothrys sect. Sonnea (P. glomeratus A.Gray results of Weigend & al. (2013), Cohen (2015), and Chacón and P. hispidus (Greene) I.M.Johnst.) should be treated in & al. (2016). One discrepancy is the recovery of a paraphy- genus Sonnea, as S. glomerata (A. Gray) Greene and S. his­ letic Andersonglossum by Chacón & al. (2016), who found pida (A.Gray) Greene. Andersonglossum virginianum to be sister to Adelinia grandis. Within the Amsinckia /Amsinckiopsis clade, Plagiobothrys However, all of our analyses strongly support the two species of jonesii and P. kingii (both of sect. Amsinckiopsis) consistently Andersonglossum as monophyletic (Figs. 2–4), agreeing with form a grade with the monophyletic genus Amsinckia in all Weigend & al. (2013) and Cohen (2015). We were, however, of our analyses, generally with strong support. The morpho- limited to a single sample per species for both Andersonglossum logical resemblance of these two Plagiobothrys species to and Adelinia. An increased sampling of these two genera, and Amsinckia has been recognized for some time (Johnston, of outgroup taxa, may be needed to verify their monophyly 1923), reflected in the sectional name, Amsinckiopsis. This and interrelationships. close relationship between Plagiobothrys sect. Amsinckiopsis Harpagonella is sister to Pectocarya in all except our and the genus Amsinckia was first confirmed phylogeneti- nrDNA analysis, in which the former is nested within the lat- cally in one of the analyses of Hasenstab-Lehman & Simpson ter, but with weak support. This sister relationship of the two (2012), although in that study the two Amsinckiopsis species genera corresponds with that of Weigend & al. (2013), Otero & formed a clade (rather than a grade) sister to Amsinckia. A sim- al., 2014, and Chacón & al. (2016). Our study confirms others ilar pattern to our study was recovered by Guilliams (2015), in recognizing the close relationship between these two genera, who is revising the classification of Plagiobothrys, including

Version of Record 1415 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

sect. Amsinckiopsis (work in prep.), which will likely entail In the genus Oreocarya, the relatively short branch lengths generic transfers of these two species. for taxa in the cpDNA (Fig. 2) and nrDNA (Fig. 3) analyses The placement of our Plagiobothrys s.str. clade also varied might be expected for this clade of perennials, which have a by analysis, being sister either to Sonnea (with strong support), to longer generation time correlated with relatively slow sequence the Amsinckia /Amsinckiopsis clade (with weak to mixed support), divergence (Andreasen & Baldwin, 2001; Cenci & al., 2013). or to a large component of the Amsinckiinae (with weak sup- However, in our mtDNA analysis (Fig. 4), the branch lengths of port). Given the strongly supported monophyly of Plagiobothrys Oreocarya are actually longer than most other taxa, an unex- s.str., but its variable placement in the Amsinckiinae, we propose pected result. Why the mitochondrial data would show this re- that the genus Plagiobothrys be recognized only as inclusive of verse trend is a mystery, one we hope to elucidate in the future. sect. Allocarya, sect. Echidiocarya, and sect. Plagiobothrys. The Johnstonella /Albidae clade largely corresponds to the What we term the Maritimae clade, based on Cryptantha Johnstonella clade of Hasenstab-Lehman & Simpson (2012) ser. Maritimae of Johnston (1925) (but inclusive of only but with the addition of the North American Cryptantha al­ Cryptantha maritima of that classification), is composed of bida and C. mexicana (both of ser. Albidae of Johnston, 1925, Cryptantha clokeyi I.M.Johnst., C. maritima (Greene) Greene, 1961), the North American C. texana (ser. Texanae of Johnston, C. martirensis M.G.Simpson & Rebman, C. muricata (Hook. 1925), and the South American C. hispida (ser. Phaceloides & Arn.) A.Nelson & J.F.Macbr., C. subamplexicaulis (Phil.) of Johnston, 1927). These four taxa were not included in the Reiche, and Johnstonella echinosepala (J.F.Macbr.) Hasenstab study by Hasenstab-Lehman & Simpson (2012). The place- & M.G.Simpson. This Maritimae clade is roughly equiva- ment within Johnstonella of the two species of Cryptantha ser. lent to the “Cryptantha s.str. 2” group of Hasenstab-Lehman Albidae, C. albida and C. mexicana, is not surprising based & Simpson (2012), which included two (North and South on nutlet morphology. Members of the Albidae group have American) samples of C. maritima, plus the South American white nutlet tubercles typical of Johnstonella and resemble the species C. chaetocalyx (Phil.) I.M.Johnst. and C. granulosa shape of certain Johnstonella species (e.g., J. angustifolia; see (Ruiz & Pav.) I.M.Johnst. Unfortunately, we were unable to ob- Hasenstab-Lehman & Simpson, 2012). The discovery of the tain quality sequences of the last two species. However, Otero South American Cryptantha hispida (of Johnston’s 1927 ser. & al. (2014) sequenced both C. martima and C. granulosa Phaceloides) nested within the Johnstonella /Albidae clade is also and found them to be sister taxa, agreeing with Hasenstab- supported by morphology, as the ovate, acutely margined nutlets Lehman & Simpson (2012) that at least the latter taxon should of this taxon are similar to several other Johnstonella species, be placed in the Maritimae clade. However, the inclusion in differing only in having a smooth sculpturing (see Simpson & our Maritimae clade of the South American C. subamplexicau­ al., 2014). However, the fourth Cryptantha species nesting within lis, the North American Johnstonella echinosepala, and three Johnstonella, C. texana of Johnston’s (1925) ser. Texanae, seems species corresponding to Johnston’s 1925 North American aberrant within Johnstonella. The nutlet and calyx morphology ser. Muricatae (C. muricata, C. clokeyi, and C. martirensis) of C. texana do not resemble members of other Johnstonella is novel. No obvious morphological features unite all of the species. Other members of Johnston’s (1925) ser. Texanae in- taxa of the Maritimae clade, and the placement of Johnstonella clude C. crassisepala (Torr. & A.Gray) Greene and C. kelseyana echinosepala outside the Johnstonella /Albidae clade was un- Greene, both of which nested firmly within Cryptantha s.str. in expected. (The identification of this specimen was confirmed our analyses. We believe that this discrepancy between phylog- by us in review; see Appendix 1 for voucher information.) eny and morphology warrants additional sampling of C. texana Despite this, our Maritimae clade is strongly supported in all in future analyses before nomenclatural changes are considered. of our analyses and is well-separated from Cryptantha s.str. in The Johnstonella /Albidae clade is sister to Cryptantha all analyses (with strong support) except in the nrDNA analy- s.str. in two (cpDNA, mtDNA) of our gene analyses, with mixed sis, in which it is sister to Cryptantha s.str. (but with mixed to strong support. However, the stem lineage of this sister- support). More samples are needed to establish the validity taxon grouping is very short (see Figs. 2, 4). The Johnstonella / and full membership of our Maritimae clade, but it seems Albidae clade is sister to clades other than to Cryptantha to represent an intriguing offshoot within the Amsinckiinae. s.str. in the nrDNA analyses, with mixed to weak support. If Given the strong molecular evidence for its phylogenetic dis- the Johnstonella /Albidae clade were unequivocally sister to tinctiveness, the group may require elevation to the rank of Cryptantha s.str., one choice of classification would be to com- genus in the future. bine the two groups, transferring all Johnstonella species back A clade consisting of Eremocarya, Oreocarya, and the to the genus Cryptantha. However, given that the relationship Maritimae clade, with Oreocarya sister to the other two, was of the Johnstonella /Albidae clade relative to Cryptantha s.str. recovered with strong support in most of our analyses. This is still ambiguous, and given the short branch length linking grouping was not detected in previously published studies. them in analyses in which they are sister taxa, a second choice However, an Eremocarya + Oreocarya clade was detected by is to retain Johnstonella and to transfer those Cryptantha spe- Hasenstab-Lehman & Simpson (2012), who recovered this rela- cies nested within the Johnstonella /Albidae clade to the genus tionship in both of their analyses (one with strong support), and Johnstonella. We are convinced that the Johnstonella /Albidae was also recovered in the ITS tree of Otero & al. (2015). No clade as presented here is a monophyletic group (with the morphological apomorphy is evident for either of these group- possible exception of C. texana) and should be recognized as ings, but this complex appears to represent a natural group. Johnstonella, either at the generic or subgeneric rank.

1416 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

As mentioned earlier, in all analyses, two previously clas- contains a number of taxa /clades with American amphitropical sified Johnstonella species occurred outside the Johnstonella / disjunct distributions (Guilliams & al., 2016). The number and Albidae clade, namely J. echinosepala, placing in the Maritimae timing of putative dispersal events and their correlation with clade, and J. micromeres, occurring in the Cryptantha s.str. current vegetation types, geologic and climatic history, and clade. We hope to obtain additional samples of these two spe- possible selective pressures leading to the above morphological cies in order to test these phylogenetic relationships before features are of great biological interest and a work in progress. considering formal changes in classification. The techniques we used in this study may continue to be The Cryptantha s.str. clade corresponds largely to “Crypt­ useful in studies of some genera or clades of the Amsinckiinae, antha s.str. 1” of Hasenstab-Lehman & Simpson (2012). Other especially those with relatively longer branch lengths. However, analyses, e.g., Weigend & al. (2013), Cohen (2014, 2015), Chacón one of the conclusions from our work is that, despite the acqui- & al. (2016), recovered a non-monophyletic Cryptantha, with sition of virtually the entire plastome, the entire cistron, and quite variable and weakly supported relationships. However, numerous mitochondrial markers, we have still been unable to our analyses firmly place the bulk of this genus (50 of the obtain strong support for the interrelationships of many genera 59 species sequenced in our study) within the strongly sup- and major clades, especially along the backbone of the tree. ported Cryptantha s.str. clade. As noted earlier, however, four This may be a function of the fact that divergences of major Cryptantha species place within the Johnstonella /Albidae clade clades in the Amsinckiinae occurred rapidly. Thus, these lin- and five within the Maritimae clade. Even if these nine species eages diverged not only relatively long ago, but also over a were ultimately transferred to other genera, Cryptantha still brief period of time, confounding our efforts to obtain an ac- would be the largest genus of the Amsinckiinae. In the mtDNA curate phylogeny. Different molecular techniques will likely be analysis, the five species of Cryptantha (C. ambigua, C. crinita, needed for refining the backbone relationships of the subtribe. C. mariposae, C. torreyana, C. watsonii) that have particularly Targeted sequences of numerous nuclear genes using methods long branches relative to the rest of the genus (Fig. 4) show such as Hyb-Seq (Weitemier & al., 2014) would enable a more no remarkable rate increases in either the cpDNA or nrDNA rigorous application of species coalescent methods. In addi- analyses. We are puzzled as to an explanation for this pattern tion, there are many research problems that could be addressed in the mitochondrial analyses. Mitochondrial DNA in plants is on various species complexes within the subtribe. These lat- known to have a high degree of plasticity in terms of genomic ter studies may be tractable using methods such as RAD-Seq rearrangements, insertion or transfer of DNA to the chloro- (Eaton & Ree, 2013; Eaton, 2014) for elucidating relationships plast or nucleus, disruption of intron/exon gene continuity, and within a recently evolved group. Still, we hope that the present evolutionary changes in gene expression (Knoop, 2004) and investigation has enhanced our understanding of phylogenetic is generally less useful in phylogenetic studies. We will be relationships in Amsinckiinae and can serve as a framework interested to see if this pattern is maintained in future analyses for additional studies in the future. with greater taxon sampling. Future work. — Our first recourse in future work will be to acquire sequence data for additional samples in the ACKNOWLEDGEMENTS Amsinckiinae. These should ideally include Oncaglossum and more samples of Adelinia, Andersonglossum, and Dasynotus, to The authors thank the staff of the herbaria that provided mate- better evaluate their monophyly and placement in the subtribe. rial for this study: University of Arizona (ARIZ), Universidad de We hope to acquire sequence data for additional species not yet Concepción (CONC), Duke University (DUKE), Gray Herbarium sampled, particularly the larger genera Cryptantha (59 of ca. 103 (GH), Instituto Argentino de Investigaciones de las Zonas Áridas spp. = ca. 57% sampled in this study) and Plagiobothrys (22 of (MERL), Missouri Botanical Garden (MO), National Herbarium of ca. 69 spp.(including members of Sonnea sect. Amsinckiopsis) New South Wales, Australia (NSW), Rancho Santa Ana Botanical = ca. 32% sampled in this study). We wish to acquire several Garden (RSA), Santa Barbara Botanic Garden (SBBG), San Diego additional samples of taxa that had an unexpected placement Natural History Museum (SD), San Diego State University (SDSU), in our analyses, particularly Cryptantha texana, Johnstonella Museo Nacional de Historia Natural (SGO), Institutio de Botánica echinosepala, and J. micromeres. The latter in particular merits Darwinion, Argentina (SI), University of California and Jepson a population-level study or more nuclear data to understand its Herbaria (UC, JEPS), and University of Riverside (UCR), and Utah evolutionary history, as different samples of this species have State University (UTC). We thank the South American botanists placed in different areas of the subtribe with traditional Sanger Gina Arancio, Victor Ardilles, Roberto Kiesling, Alicia Marticorena, sequence data (Hasenstab, unpub. data). Melica Muñoz, Gloria Rojas, and Rosita Scherson who supported Following the acquisition of sequences of some additional this project by providing logistical support during our trips to Chile taxa, future studies in subtribe Amsinckiinae will include broad and Argentina. And, we thank Ron B. Kelley and Jon P. Rebman for surveys of character evolution, divergence timing, and phylo- contributing North American field collections and giving valuable geographic history. Of particular interest in character assess- insight into the taxonomy of the group. Lastly, we thank our funding ment are the evolutionary transitions in cleistogamy, heterostyly, sources for this study: the American Society of Plant Taxonomists, plant duration, ploidy levels, and details of nutlet morphology California Native Plant Society, California Botanical Society, Joshua such as sculpturing, number per fruit, heteromorphism, and Tree National Park, National Geographic Society grant 9533-14 to the dimorphism (after Guilliams & al., 2013). The Amsinckiinae first author, and San Diego State University Travel Grants.

Version of Record 1417 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

LITERATURE CITED Johnston, I.M. 1925. Studies in the Boraginaceae IV. The North American species of Cryptantha. Contr. Gray Herb. 74: 1–114. Alvarez, I. & Wendel, J.F. 2003. Ribosomal ITS sequences and plant Johnston, I.M. 1927. Studies in the Boraginaceae VI. A revision of phylogenetic inference. Molec. Phylogen. Evol. 29: 417–434. the South American Boraginoideae. Contr. Gray Herb. 78: 1–118. https://doi.org/10.1016/S1055-7903(03)00208-2 Johnston, I.M. 1961. Notes on some Texan borages. Wrightia 2: Andreasen, K. & Baldwin, B.G. 2001. Unequal evolutionary rates be- 158–162. tween annual and perennial lineages of checker mallows (Sidalcea, Katoh, K., Misawa, K., Kuma, K. & Miyata, T. 2002. MAFFT: Malvaceae): Evidence from 18s-26s rDNA internal and external A novel method for rapid multiple sequence alignment based on transcribed spacers. Molec. Biol. Evol. 18: 936–944. fast Fourier transform. Nucl. Acids Res. 30: 3059–3066. https://doi.org/10.1093/oxfordjournals.molbev.a003894 https://doi.org/10.1093/nar/gkf436 APG IV 2016. An update of the Angiosperm Phylogeny Group clas- Knoop, V. 2004. The mitochondrial DNA of land plants: Peculiarities sification for the orders and families of flowering plants: APG IV. in phylogenetic perspective. Curr. Genet. 46: 123–139. Bot. J. Linn. Soc. 181: 1–20. https://doi.org/10.1007/s00294-004-0522-8 https://doi.org/10.1111/boj.12385 Lanfear, R., Frandsen, P.B., Wright, A.M., Senfeld, T. & Calcott, B. Cenci, A., Combes, M. & Lashermes, P. 2013. Differences in evolution 2016. PartitionFinder 2: New methods for selecting partitioned rates among Eudicotyledon species observed by analysis of protein models of evolution for molecular and morphological phylogenetic divergence. J. Heredity 104: 459–464. analyses. Molec. Biol. Evol. 34: 772–773. https://doi.org/10.1093/jhered/est025 https://doi.org/10.1093/molbev/msw260 Chacón, J., Luebert, F., Hilger, H.H., Ovcinnikova, S., Selvi, F., Långström, E. & Chase, M.W. 2002. Tribes of Boraginoideae Cecchi, L., Guilliams, C.M., Hasenstab-Lehman, K., Sutorý, (Boraginaceae) and placement of Antiphytum, Echiochilon, K., Simpson, M.G. & Weigend, M. 2016. The borage family Ogastemma and Sericostoma: A phylogenetic analysis based on (Boraginaceae s.str.): A revised infrafamilial classification based atpB plastid DNA sequence data. Pl. Syst. Evol. 234: 137–153. on new phylogenetic evidence, with emphasis on the placement of https://doi.org/10.1007/s00606-002-0195-z some enigmatic genera. Taxon 65: 523–546. Luebert, F., Cecchi, L., Frohlich, M.W., Gottschling, M., Guilliams, https://doi.org/10.12705/653.6 C.M., Hasenstab-Lehman, K.E., Hilger, H.H., Miller, J.S., Cohen, J.I. 2014. A phylogenetic analysis of morphological and mo- Mittelbach, M., Nazaire, M., Nepi, M., Nocentini, D., Ober, D., lecular characters of Boraginaceae: Evolutionary relationships, tax- Olmstead, R.G., Selvi, F., Simpson, M.G., Sutorý, K., Valdés, onomy, and patterns of character evolution. Cladistics 30: 139–169. B., Walden, G.K. & Weigend, M. 2016. Familial classification https://doi.org/10.1111/cla.12036 of the Boraginales. Taxon 65: 502–522. Cohen, J.I. 2015. Adelinia and Andersonglossum (Boraginaceae), two https://doi.org/10.12705/653.5 new genera from New World species of Cynoglossum. Syst. Bot. Mabberley, D.J. 2008. Mabberley’s plant-book: A portable diction­ 40: 611–619. http://dx.doi.org/10.1600/036364415X688385 ary of the higher plants; Their classification and uses, ed. 3. Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for Cambridge: Cambridge University Press. small quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. McNeill, J., Barrie, F.R., Buck, W.R., Demoulin, V., Greuter, W., Amer. 19: 11–15. Hawksworth, D.L., Herendeen, P.S., Snapp, S., Marhold, Eaton, D.A.R. 2014. PyRAD: Assembly of de novo RADseq loci for K., Prado, J., Prud’homme von Reine, W.F., Smith, G.F., phylogenetic analyses. Bioinformatics 30: 1844–1849. Wiersema, J.H. & Turland, N.J. 2012. International Code of https://doi.org/10.1093/bioinformatics/btu121 Nomenclature for algae, fungi, and plants (Melbourne Code). Eaton, D.A.R. & Ree, R.H. 2013. Inferring phylogeny and introgres- Regnum Vegetabile 154. Königstein: Koeltz Scientific Books. sion using RADseq data: An example from flowering plants Miller, M.A., Pfeiffer, W. & Schwartz, T. 2010. Creating the CIPRES (Pedicularis: Orobanchaceae). Syst. Biol. 62: 689–706. Science Gateway for inference of large phylogenetic trees. Pp. https://doi.org/10.1093/sysbio/syt032 45–52 in: Proceedings of the Gateway Computing Environments Friar, E.A. 2005. Isolation of DNA from plants with large amounts of Workshop (GCE), New Orleans, Louisiana, 14 Nov 2010. secondary metabolites. Pp. 3–14 in: Zimmer, E.A. & Roalson, E.H. Piscataway: IEEE. https://doi.org/10.1109/GCE.2010.5676129 (eds.), Methods in enzymology, vol. 395, Producing the biochemical Nazaire, M. & Hufford, L. 2012. A broad phylogenetic analysis of data, part B. San Diego: Academic Press. Boraginaceae: Implications for the relationships of Mertensia. https://doi.org/10.1016/S0076-6879(05)95001-5 Syst. Bot. 37: 758–783. https://doi.org/10.1600/036364412X648715 Guilliams, C.M. 2015. Diversification, biogeography, and classifica­ Otero, A., Jiménez-Mejías, P., Valcárcel, V. & Vargas, P. 2014. tion of the Amsinckiinae (Boraginaceae), with an emphasis on the Molecular phylogenetics and morphology support two new gen- popcorn flowers (Plagiobothrys). Ph.D. Dissertation, University era (Memoremea and Nihon) of Boraginaceae s.str. Phytotaxa 173: of California, Berkeley, California, U.S.A. 241–277. http://dx.doi.org/10.11646/phytotaxa.173.4.1 Guilliams, C.M., Veno, B.A., Simpson, M.G. & Kelley, R.B. 2013. Rambaut, A. 2006–2014. FigTree, version 1.4.0. http://tree.bio.ed.ac. Pectocarya anisocarpa, a new species of Boraginaceae, and a uk/software/figtree revised key for the genus in western North America. Aliso 31: 1–13. Raven, P.H. 1963. Amphitropical relationships in the floras of North https://doi.org/10.5642/aliso.20133101.02 and South America. Quart. Rev. Biol. 38: 151–177. Guilliams, M.C., Mabry, M., Hasenstab-Lehman, K., Baldwin, https://doi.org/10.1086/403797 B.G. & Simpson, M.G. 2016. Exploring patterns and mecha- Riedl, H. 1997. Boraginaceae. Pp. 43–144 in: Kalkman, C., Kirkup, nisms of American amphitropical disjunction in the Amsinckiinae D.W., Nooteboom, H.P., Stevens, P.F. & De Wilde, W.J.J.O. (Boraginaceae). Botany 2016, Savannah, Georgia, Abstract 1006. (eds.), Flora Malesiana, ser. 1, Seed plants, vol. 13. Leiden: http://2016.botanyconference.org/engine/search/index. Rijksherbarium/Hortus Botanicus. php?func=detail&aid=1006 Ripma, L.A., Simpson, M.G. & Hasenstab-Lehman, K. 2014. Hasenstab-Lehman, K.E. & Simpson, M.G. 2012. Cat’s eyes and Geneious! Simplified genome skimming methods for phylogenetic popcorn flowers: Phylogenetic systematics of the genus Cryptantha systematic studies: A case study in Oreocarya (Boraginaceae). s.l. (Boraginaceae). Syst. Bot. 37: 738–757. Applic. Pl. Sci. 2: 1400062. https://doi.org/10.3732/apps.1400062 https://doi.org/10.1600/036364412X648706 Ronquist, F., Teslenko, M., Van der Mark, P., Ayres, D.L., Darling, Johnston, I.M. 1923. Studies in the Boraginaceae. A synopsis and A., Höhna, S., Larget, B., Liu, L., Suchard, M.A. & Huelsenbeck, redefinition of Plagiobothrys. Contr. Gray Herb. 68: 57–80. J.P. 2012. MrBayes 3.2: Efficient Bayesian phylogenetic inference

1418 Version of Record TAXON 66 (6) • December 2017: 1406–1420 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae)

and model choice across a large model space. Syst. Biol. 61: 539– Navel seeds (Omphalodes Mill.) – The phylogeny of the borage 542. http://dx.doi.org/10.1093/sysbio/sys029 family (Boraginaceae s.str.). Molec. Phylogen. Evol. 68: 604–618. Schmieder, R. & Edwards, R. 2011. Quality control and preprocessing https://doi.org/10.1016/j.ympev.2013.04.009 of metagenomic datasets. Bioinformatics 27: 863–864. Weigend, M., Luebert, F., Gottschling, M., Couvreur, T.L.P., Hilger, https://doi.org/10.1093/bioinformatics/btr026 H.H. & Millerd, J.S. 2014. From capsules to nutlets—Phylogenetic Simpson, M.G. & Hasenstab, K.E. 2009. Cryptantha of Southern relationships in the Boraginales. Cladistics 30: 508–518. California. Crossosoma 35: 1–59. https://doi.org/10.1111/cla.12061 Simpson, M.G., Hasenstab-Lehman, K. & Kelley, R.B. 2014. Wen, J. & Ickert-Bond, S.M. 2009. Evolution of the Madrean–Tethyan Typification of the name Johnstonella (Boraginaceae). Taxon 63: disjunctions and the North and South American amphitropical 930–931. https://doi.org/10.12705/634.8 disjunctions in plants. J. Syst. Evol. 47: 331–348. Stamatakis, A. 2014. RAxML Version 8: A tool for phylogenetic anal- https://doi.org/10.1111/j.1759-6831.2009.00054.x ysis and post-analysis of large phylogenies. Bioinformatics 30: Weitemier, K., Straub, S.C.K., Cronn, R.C., Fishbein, M., Schmickl, 1312–1313. https://doi.org/10.1093/bioinformatics/btu033 R., McDonnell, A. & Liston, A. 2014. Hyb-Seq: Combining tar- Straub, S.C.K., Cronn, R.C., Edwards, C., Fishbein, M. & Liston, get enrichment and genome skimming for plant phylogenomics. A. 2013. Horizontal transfer of DNA from the mitochondrial to Applic. Pl. Sci. 2: apps.1400042. the plastid genome and its subsequent evolution in milkweeds http://dx.doi.org/10.3732/apps.1400042 (Apocynaceae). Genome Biol. Evol. 5: 1872–1885. Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic https://doi.org/10.1093/gbe/evt140 analysis of large biological sequence datasets under the maximum Weigend, M., Luebert, F., Selvi, F., Brokamp, G. & Hilger, H.H. likelihood criterion. Ph.D. dissertation, The University of Texas 2013. Multiple origins for Hound’s tongues (Cynoglossum L.) and at Austin, U.S.A.

Appendix 1. Taxa included in the phylogenetic analyses with their corresponding country of origin, collector/collection number, herbarium accessions, and NCBI Short Read Archive accession numbers. OUTGROUP TAXA: Cynoglossum creticum Mill., Chile, Hasenstab s.n. (SDSU 21418), SRR5713409; Hackelia micrantha (Eastw.) J.L.Gentry, U.S.A., Guilliams 2606 (SBBG 132390), SRR5713385; Microula tibetica Benth., China, Boufford 31295 (GH 00466293), SRR5713384; Myosotis laxa Lehm., U.S.A., Guilliams 1246 (SBBG 132391), SRR5713383. INGROUP TAXA: Adelinia grandis (Douglas ex Lehm.) J.I.Cohen, U.S.A., Simpson 3007 (SDSU 19197), SRR5713435; Amsinckia intermedia 1 Fisch. & C.A.Mey., U.S.A., Guilliams 1466 (SBBG 132389), SRR5713381; Amsinckia intermedia 2 Fisch. & C.A.Mey., U.S.A., Mabry 65 (SDSU 20756), SRR5713430; Amsinckia intermedia 3 Fisch. & C.A. Mey., U.S.A., Guilliams 1447 (SBBG 132388), SRR5713433; Amsinckia tessellata A.Gray var. t., U.S.A., Mabry 29 (SDSU 20350), SRR5713421; Andersonglossum occidentale (A.Gray) J.I.Cohen, U.S.A., Bell 3822 (RSA 793153), SRR5713450; Andersonglossum virginianum (L.) J.I.Cohen, U.S.A., Wilbur 73075 (DUKE 379623), SRR5713449; Cryptantha affinis (A.Gray) Greene, U.S.A., Rebman 18116 (SD 199070), SRR5713395; Cryptantha albida (Kunth) I.M.Johnst., U.S.A., Kelley 1426 (SDSU 20612), SRR5713394; Cryptantha alfalfalis (Phil.) I.M.Johnst., Chile, Arroyo 995313 (CONC 163659), SRR5713397; Cryptantha alyssoides (DC.) Reiche, Chile, Teiller 5210 (CONC 156553), SRR5713396; Cryptantha ambigua (A.Gray) Greene, U.S.A., Benet-Pierce 524 (SDSU 20524), SRR5713391; Cryptantha aspera (Phil.) J.Grau, Chile, Munoz & al. 2634 (MO 4317599), SRR5713390; Cryptantha barbigera (A.Gray) Greene var. b., U.S.A., Mabry 27 (SDSU 20349), SRR5713393; Cryptantha calyco- tricha I.M.Johnst., Chile, Luebert 3023 (CONC 150898), SRR5713392; Cryptantha capituliflora (Clos) Reiche, Chile, Arroyo 991122 (CONC 166914), SRR5713389; Cryptantha clevelandii Greene, U.S.A., Simpson 3733 (SDSU 20782), SRR5713388; Cryptantha clokeyi I.M.Johnst., U.S.A., Andre 4153 (UCR 164170), SRR5713360; Cryptantha corollata (I.M.Johnst.) I.M.Johnst., U.S.A., Mabry 83 (SDSU 20775), SRR5713361; Cryptantha crassisepala (Torr. & A.Gray) Greene, U.S.A., Kelley 1997 (SDSU 20623), SRR5713362; Cryptantha crinita Greene, U.S.A., Lepley s.n. (SDSU 20823), SRR5713363; Cryptantha cynoglossoides (Phil.) I.M.Johnst., Argentina, Kiesling 8083 (SI 87776), SRR5713364; Cryptantha decipens (M.E.Jones) A.Heller, U.S.A., Simpson 3661 (SDSU 20014), SRR5713365; Cryptantha diffusa (Phil.) I.M.Johnst., Argentina, Mendez 9862 (MERL 56799), SRR5713366; Cryptantha dumetorum (Greene ex A.Gray) Greene, U.S.A., Hasenstab 57 (SDSU 18694), SRR5713367; Cryptantha echinella Greene, U.S.A., Simpson 3319 (SDSU 19611), SRR5713335; Cryptantha fendleri (A.Gray) Greene, U.S.A., Ripma 372 (SDSU 20114), SRR5713334; Cryptantha flaccida (Douglas ex Lehm.) Greene, U.S.A., Simpson 3619 (SDSU 19846), SRR5713333; Cryptantha ganderi I.M.Johnst., U.S.A., Hasenstab 40 (SDSU 20345), SRR5713332; Cryptantha globulifera 1 (Clos) Reiche, Chile, Teillier 3845 (SGO 147985), SRR5713328; Cryptantha globulifera 2 (Clos) Reiche, Chile, Arroyo 995294 (SGO 146942), SRR5713329; Cryptantha globulifera 3 (Clos) Reiche, Chile, Arroyo 993602 (CONC 163475), SRR5713331; Cryptantha globulifera 4 (Clos) Reiche, Chile, Arroyo 993602 (SGO 147688), SRR5713330; Cryptantha glomerata Lehmann ex G. Don subsp. g., Chile, Arroyo 995177 (SGO 146941), SRR5713337; Cryptantha glomerulifera (Phil.) I.M.Johnst., Chile, Teiller 5579 (CONC 166867), SRR5713336; Cryptantha gnaphalioides (Phil.) Reiche, Chile, Eggli 2983 (SGO 146002), SRR5713453; Cryptantha gracilis Osterh., U.S.A., Andre 12644 (UCR 217631), SRR5713454; Cryptantha hispida (Phil.) Reiche, Chile, Teillier 4754 (CONC 150914), SRR5713451; Cryptantha hispidissima 1 Greene, U.S.A., Helmkamp 8471 (RSA 710334), SRR5713358; Cryptantha hispidissima 2 Greene, U.S.A., Hasenstab 30 (SDSU 18342), SRR5713359; Cryptantha incana Greene, U.S.A., Myers 1032 (UCR 227031), SRR5713452; Cryptantha intermedia (A.Gray) Greene var. i., U.S.A., Simpson 3686 (SDSU 20037), SRR5713457; Cryptantha junipereneis R.B.Kelley & M.G.Simpson, U.S.A., Mabry 75 (SDSU 20766), SRR5713448; Cryptantha kelseyana Greene, U.S.A., Kelley 2254 (SDSU 20630), SRR5713458; Cryptantha kingii (Phil.) Reiche, Chile, Muñoz 2580 (SGO 123832), SRR5713455; Cryptantha leiocarpa (Fisch. & C.A.Mey.) Greene, U.S.A., Mabry 68 (SDSU 20759), SRR5713456; Cryptantha mariposae I.M.Johnst., U.S.A., Helmkamp 15796 (SDSU 20826), SRR5713459; Cryptantha maritima (Greene) Greene var. m., U.S.A., Simpson 3665 (SDSU 20050), SRR5713460; Cryptantha martirensis M.G.Simpson & Rebman, Mexico, Rebman 15973 (SDSU 18625), SRR5713440; Cryptantha mexicana (Brandegee) I.M.Johnst., U.S.A., Kelley 1230 (SDSU 20610), SRR5713439; Cryptantha microstachys (Greene ex A.Gray) Greene, U.S.A., Rebman 21420B (SD 216851), SRR5713442; Cryptantha minima Rydb., U.S.A., Kelley 2248 (SDSU 20629), SRR5713441; Cryptantha mohavensis (Greene) Greene, U.S.A., Ripma 348 (SDSU 20877), SRR5713444; Cryptantha muricata (Hook. & Arn.) A.Nelson & J.F.Macbr. var. m., U.S.A., Simpson 3818 (SDSU 20749), SRR5713443; Cryptantha nemaclada Greene, U.S.A., Mabry 82 (SDSU 20774), SRR5713446; Cryptantha nevadensis A.Nelson & P.B.Kenn., U.S.A., Barth 913 (SDSU 20393), SRR5713445; Cryptantha oxygona (A.Gray) Greene, U.S.A., Honer 811 (RSA 685321), SRR5713447; Cryptantha peruviana I.M.Johnst., Chile, Teillier 4100 (SGO 140959), SRR5713426; Cryptantha phaceloides (Clos) Reiche, Chile, Ackerman 211 (SGO 146206), SRR5713427; Cryptantha pterocarya (Torr.) Greene f. p., U.S.A., Mabry 33 (SDSU 20355), SRR5713428; Cryptantha recurvata Coville, U.S.A., Sanders 39404 (UCR 225245), SRR5713429; Cryptantha scoparia A.Nelson, U.S.A., Andre 10360 (UCR 211150), SRR5713422; Cryptantha simulans Greene, U.S.A., Hains 258 (SDSU 20390), SRR5713423; Cryptantha sparsiflora (Greene) Greene, U.S.A., Sanders 34146 (UCR 184326), SRR5713424; Cryptantha subamplexicaulis (Phil.) Reiche, Chile, Teillier 2620 (SGO 129437), SRR5713425; Cryptantha texana (A.DC.) Greene, U.S.A., Kelley 1415 (SDSU 20611), SRR5713431; Cryptantha torreyana (A.Gray) Greene var. t., U.S.A., Ripma 377 (SDSU 20124), SRR5713432; Cryptantha utahensis (A.Gray) Greene, U.S.A., Mabry 28 (SDSU 20348), SRR5713416; Cryptantha watsonii (A.Gray) Greene, U.S.A., Andre 15116 (UCR 226737), SRR5713415; Cryptantha wigginsii I.M.Johnst., U.S.A., Clonessy s.n. (SDSU 20082), SRR5713414; Dasynotus daubenmirei I.M.Johnst., U.S.A., Kelley 1951 (SDSU 20343), SRR5713413; Eremocarya lepida 1 (A.Gray) Greene, U.S.A., Simpson 3851 (SDSU 21209), SRR5713419; Eremocarya lepida 2 (A.Gray) Greene, U.S.A., Simpson 3184 (SDSU 19533), SRR5713420; Eremocarya micrantha (Torrey) Greene var. m. 1, U.S.A., Guilliams 602 (SDSU

Version of Record 1419 Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae) TAXON 66 (6) • December 2017: 1406–1420

Appendix 1. Continued. 18956), SRR5713418; Eremocarya micrantha (Torrey) Greene var. m. 2, U.S.A., Hendrickson 2640 (SD 203297), SRR5713417; Eremocarya micrantha var. pseudolepida M.G.Simpson, L.M.Simpson & Rebman, Mexico, Simpson 3847 (SDSU 21205), SRR5713412; Greeneocharis circumscissa (Hook. & Arn.) Rydb. var. c., U.S.A., Simpson 3875 (SDSU 21417), SRR5713411; Greeneocharis circumscissa var. rosulata (J.T.Howell) Hasenstab & M.G.Simpson, U.S.A., Kelley 1625 (SDSU 20663), SRR5713374; Greeneocharis similis (K.Mathew & P.H.Raven) Hasenstab & M.G.Simpson, U.S.A., Kelley 1015 (SDSU 20605), SRR5713375; Greeneocharis similis (K.Mathew & P.H.Raven) Hasenstab & M.G.Simpson, U.S.A., Kelley 1015 (SDSU 20605), SRR5713410; Harpagonella arizonica (I.M.Johnst.) Guilliams & B.G.Baldwin, U.S.A., Tedford 599 (ARIZ 388168), SRR5713372; Harpagonella palmeri A.Gray, U.S.A., Guilliams 1414 (SBBG 132412), SRR5713373; Johnstonella angelica (I.M.Johnst.) Hasenstab & M.G.Simpson, Mexico, Rebman 18550 (SDSU 19425), SRR5713370; Johnstonella angustifolia (Torr.) Hasenstab & M.G.Simpson, U.S.A., Boyd 11841 (RSA 731212), SRR5713371; Johnstonella costata (Brandegee) Hasenstab & M.G.Simpson, U.S.A., Guilliams 538 (SDSU 18964), SRR5713368; Johnstonella diplotricha (Phil.) Hasenstab & M.G.Simpson, Argentina, Mabry 89 (SDSU 21232), SRR5713369; Johnstonella echinosepala (J.F.Macbride) Hasenstab & M.G.Simpson, Mexico, Rebman 25402 (SD 228804), SRR5713376; Johnstonella fastigiata (I.M.Johnst.) Hasenstab & M.G.Simpson, Mexico, West 99-23 (SDSU 15588), SRR5713377; Johnstonella grayi (Vasey & Rose) Hasenstab & M.G.Simpson var. g., Mexico, Fritsch 1264 (RSA 544081), SRR5713343; Johnstonella holoptera (A.Gray) Hasenstab & M.G.Simpson, U.S.A., Simpson 8III98H (SDSU 13036), SRR5713342; Johnstonella inaequata (I.M.Johnst.) Brand, U.S.A., Andre 8132 (RSA 732141), SRR5713345; Johnstonella micromeres (A.Gray) Hasenstab & M.G.Simpson, U.S.A., Mabry 71 (SDSU 20762), SRR5713344; Johnstonella parviflora (Phil.) Hasenstab & M.G.Simpson, Chile, Teillier 4616 (CONC 150821), SRR5713339; Johnstonella racemosa Brand, U.S.A., Hasenstab 68 (SDSU 18710), SRR5713338; Oreocarya crymophila (I.M.Johnst.) Jeps. & Hoover, U.S.A., Ripma 390 (SDSU 20116), SRR5713341; Oreocarya flavoculata A.Nelson, U.S.A., Ripma 307 (SDSU 20030), SRR5713340; Oreocarya glomerata (Pursh) Greene, U.S.A., Ripma 379 (SDSU 20113), SRR5713347; Oreocarya hoffmannii (I.M.Johnst.) Abrams, U.S.A., Ripma 306 (SDSU 20036), SRR5713346; Oreocarya “howelliana” (R.B. Kelley, in prep.), U.S.A., Ripma 312 (SDSU 20004), SRR5713324; Oreocarya humilis (A.Gray) Greene subsp. h., U.S.A., Ripma 303 (SDSU 20029), SRR5713325; Oreocarya hyposphila (I.M.Johnst.) Hasenstab & M.G.Simpson, U.S.A., Ripma 374 (SDSU 20086), SRR5713326; Oreocarya nubigena “craters” Greene, U.S.A., Ripma 399 (SDSU 20094), SRR5713327; Oreocarya nubigena “granite” Greene, U.S.A., Ripma 363 (SDSU 20079), SRR5713320; Oreocarya nubigena “mammoth” Greene, U.S.A., Ripma 301 (SDSU 20055), SRR5713321; Oreocarya schoolcraftii (Tiehm) R.B.Kelley, U.S.A., Ripma 370 (SDSU 20123), SRR5713322; Oreocarya setosissima (A.Gray) Greene, U.S.A., Kelley 1466 (SDSU 20242), SRR5713323; Oreocarya sobolifera (Payson) R.B.Kelley, U.S.A., Kelley 1173 (SDSU 20211), SRR5713318; Oreocarya subretusa “mt eddy” (I.M.Johnst.) Abrams, U.S.A., Kelley 928 (SDSU 20232), SRR5713319; Oreocarya subretusa “type” (I.M.Johnst.) Abrams, U.S.A., Ripma 384 (SDSU 20107), SRR5713353; Oreocarya subretusa “warner mts” (I.M.Johnst.) Abrams, U.S.A., Ripma 389 (SDSU 20110), SRR5713352; Oreocarya suffruticosa var. abortiva (Greene) J.F.Macbr., U.S.A., Ripma 308 (SDSU 20024), SRR5713351; Oreocarya “ursina” (R.B. Kelley, in prep.), U.S.A., Ripma 395 (SDSU 20098), SRR5713350; Oreocarya virgata (Porter) Greene, U.S.A., Ripma 371 (SDSU 20117), SRR5713357; Pectocarya linearis var. ferocula I.M.Johnst., U.S.A., Kelley 1962 (SBBG 132392), SRR5713356; Pectocarya penicillata A.DC., U.S.A., Kelley 1967 (SBBG 132393), SRR5713355; Pectocarya pusilla A.Gray, U.S.A., Guilliams 995 (SBBG 132394), SRR5713354; Pectocarya setosa A.Gray, U.S.A., Gowen 306 (JEPS 108272), SRR5713349; Plagiobothrys austinae (Greene) I.M.Johnst., U.S.A., Guilliams 1010B (SBBG 132398), SRR5713348; Plagiobothrys canescens Benth. var. c., U.S.A., Guilliams 929 (SBBG 132399), SRR5713437; Plagiobothrys collinus var. ursinus (A.Gray) Higgins, U.S.A., Guilliams 1067 (SBBG 132400), SRR5713438; Plagiobothrys congestus (Wedd.) I.M.Johnst., Bolivia, Villavicencio 323 (MO 5203201), SRR5713382; Plagiobothrys fulvus var. campestris (Greene) I.M.Johnst., U.S.A., Guilliams 1105 (SBBG 132401), SRR5713436; Plagiobothrys glomeratus A.Gray, U.S.A., Tiehm 12542 (UTC 230182), SRR5713380; Plagiobothrys glyptocarpus (Piper) I.M.Johnst., U.S.A., Guilliams 1142 (SBBG 132402), SRR5713434; Plagiobothrys gracilis (Ruiz & Pav.) I.M.Johnst., Chile, Guilliams 1688 (SBBG 132396), SRR5713378; Plagiobothrys greenei (A.Gray) I.M.Johnst., U.S.A., Forrestal 4-15-09 (SBBG 132403), SRR5713379; Plagiobothrys hispidus A.Gray, U.S.A., Oswald & Ahart 5655 (JEPS 87508), SRR5713386; Plagiobothrys jonesii A.Gray, U.S.A., André & Clifton 10750 (UCR 215416), SRR5713387; Plagiobothrys kingii var. harknessii (Greene) Jeps., U.S.A., Taylor 15044 (UC 1876874), SRR5713403; Plagiobothrys leptocladus (Greene) I.M.Johnst., Mexico, Guilliams 1772 (SBBG 132404), SRR5713402; Plagiobothrys linifolius cf. (Willd. ex Lehm.) I.M.Johnst., Peru, Weigend & Schwarzer 8073 (MO 6145366), SRR5713405; Plagiobothrys mollis (A.Gray) I.M.Johnst., U.S.A., Guilliams 1347 (SBBG 132404), SRR5713404; Plagiobothrys myosotoides NA (Lehm.) Brand, U.S.A., Gowen 1029 (SBBG 132405), SRR5713399; Plagiobothrys nothofulvus (A.Gray) A.Gray, U.S.A., Guilliams 1481 (SBBG 132406), SRR5713398; Plagiobothrys plurisepalus I.M.Johnst., Australia, Weber 5653 (NSW 647734), SRR5713401; Plagiobothrys polycaulis (Phil.) I.M.Johnst., Chile, Guilliams 1687 (SBBG 132395), SRR5713400; Plagiobothrys tenellus (C.A.Mey. ex Ledeb.) A.Gray, U.S.A., Guilliams 1183 (SBBG 132407), SRR5713407; Plagiobothrys torreyi (A.Gray) A.Gray, U.S.A., Guilliams 1888 (SBBG 132408), SRR5713406; Plagiobothrys undulatus (Piper) I.M.Johnst., U.S.A., Guilliams 1138 (SBBG 132409), SRR5713408.

1420 Version of Record Vol. 66 (6) • December 2017

International Journal of Taxonomy, Phylogeny and Evolution

Electronic Supplement to

Phylogeny of the popcorn flowers: Use of genome skimming to evaluate monophyly and interrelationships in subtribe Amsinckiinae (Boraginaceae)

Michael G. Simpson, C. Matt Guilliams, Kristen E. Hasenstab-Lehman, Makenzie E. Mabry & Lee Ripma

Taxon 66: 1406–1420 (TAXON 66 (6) • December 2017 Electr. Suppl. to: Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae

Myosotis_spHac. micrantha -SBBG 132390 Hackelia_micranthaMy. laxa -SBBG 132391 87 Cyno_creticumCy. creticum -SDSU 21418 .91 87 Microula_tibetica_GH00466293Mi. tibetica -GH 00466293 100 An. virginianum -DUKE 379623 NR 100 Cyno_virginianum_DUKE379623 100 1.0 Cyno_occidentale_RSA793153An. occidentale And. -RSA 793153 1.0 100 63 Cyno_gran_MGSAd. grandis -SDSUAdel. 19197 Dasynotus_daub_SDSU20343D. daubenmirei -SDSU 20343 84 100 H. arizonica Dasynotus-ARIZ 388168 100 84 100 Harpagonella_arizonica_ARIZ388168 100 .69 100 1.0 Harpagonella_palmeriH. palmeri -SBBG 132412 1.0 100 1.0 97 Pectocarya_pusillaPe. pusilla Pectocarya/-SBBG 132394 97100 Pectocarya_setosa_JEPS108272Pe. setosa -JEPS 108272 1.0 100100 67 1.0 Pectocarya_linearis_JEPS108272Pe. linearis var. f. Harpagonella-SBBG 132392 67 1.0 100 Pec_penicillataPe. penicillata -SBBG 132393 NR Pec_penicillata 93 Plagio_jonesii_UCR215416Pl. jonesii -UCR 215416 93100 Plagio_kingii_UC1876874Pl. kingii var. h. -UC 1876874 1.0 100 Am. tessellata var. t. Amsinckia/-SDSU 20350 1.0 100 Amsinckia_tessellata_SDSU20350 100100 Amsinckia_spAm. intermedia 1 -SBBG 132389 1.0 100100 Am. intermedia 2 -SDSU 20756 1.0 100 Amsinckia_menziesii Amsinckiopsis 1.0 Amsinckia_intermedia_SDSU20756Am. intermedia 3 -SBBG 132388 100 100 E. lepida 1 -SDSU 21209 100 100 Eremo_lepida_SDSU21209 1.0 100 1.0 Eremo_lepida_SDSU19533E. lepida 2 -SDSU 19533 100 1.0 100 Eremo_micrantha_SDSU18956E. micrantha var. m. 1 Eremocarya-SDSU 18956 100100 Eremo_micrantha_pseudolepida_SDSU21205E. micrantha var. p. -SDSU 21205 68 1.0 100 E. micrantha var. m. 2 -SD 203297 68 1.0 Eremo_micrantha_micrantha_SD203297 .64 100 C_martirensis_SDSU18625C. martirensis -SDSU 18625 100100 C_clokeyi_UCR164170C. clokeyi -UCR 164170 100 1.0 100 C_clokeyi_UCR164170 100 1.0 C_muricata_SDSU20749C. muricata var. m. Maritimae-SDSU 20749 1.0 100 J_echinosepala_SD228804J. echinosepala -SD 228804 10099 C. subamplexicaulis -SGO 129437 96 1.0 99 C_subamplexicaule_SGO129437 clade 1.0 96 1.0 C_maritima_SDSU20050C. maritima var. m. -SDSU 20050 100 68 OrecryL37O. crymophila -SDSU 20116 100 6898 1.0 1.0 OrehypL36O. hyposphila -SDSU 20086 1.0 98 OresobL31O. sobolifera -SDSU 20211 100 50 OrciabLR6O. suffruticosa var. a. -SDSU 20024 100 5052 OreflaLR5 -SDSU 20030 1.0 .99 5285 O. virgata -SDSU 20117 1.0 85 O_virgata_SDSU20117 1.0 OresetL54O. setosissima -SDSU 20242 95 95 86 OreursL40O. “ursina” -SDSU 20098 1.0 86100 OrnummLR1O. nubigena “mam.” -SDSU 20055 .99 100 1.0 100 OrehowLR9O. “howelliana” Oreocarya-SDSU 20004 91 100 OrsutyL65O. subretusa “type” -SDSU 20107 91 1.0 OrsutyL65 1.0 52 OreschL41O. schoolcraftii -SDSU 20123 5243 OrecelL69O. glomerata -SDSU 20113 89 1.0 43 89 .98 OrsumeL52O. subretusa “mt ed.” -SDSU 20232 1.0 44 OrnucrL61O. nubigena “craters” -SDSU 20094 44 OrehofLR4O. hoffmannii -SDSU 20036 1.0 61 O. humilis subsp. h. -SDSU 20029 1.0 6148 OrehumL59 84 48100 O. nubigena “granite” -SDSU 20079 84 1.0 100 OrnugrL38 1.0 1.0 OrsuwmL39O. subretusa “w. mts” -SDSU 20110 100 Plagio_hispidus_JEPS87508Pl. hispidus -JEPS 87508 1.0 100 Plagio_glomeratus_UTC230182Pl. glomeratus Sonnea-UTC 230182 100 100 Plagio_fulvusPl. fulvus var. c. -SBBG 132401 100 100100 Pl. nothofulvus -SBBG 132406 1.0 1.0 100 Plagio_nothofulvus 100 1.0 Plagio_collinusPl. collinus var. u. -SBBG 132400 1.0 100 Plagio_canescensPl. canescens var. c. -SBBG 132399 100 100 Plagio_myosotoidesNAPl. myosotoides NA -SBBG 132405 100 100100 Plagio_tenellusPl. tenellus -SBBG 132407 1.0 1.0 100 Plagio_tenellus 99 1.0 Plagio_torreyiPl. torreyi -SBBG 132408 99 Pl. plurisepalus -NSW 647734 1.0 Plagio_plurisepalus_NSW647734 100 100 Plagio_nonpricklyChilePl. polycaulis Plagiobothrys-SBBG 132395 100 1.0 100 1.0 99 Plagio_pricklyChilePl. gracilis -SBBG 132396 99 1.0 99 Plagio_greeneiPl. greenei -SBBGs.str. 132403 1.0 99 Plagio_austiniaePl. austinae -SBBG 132398 91 100 Plagio_undulatusPl. undulatus -SBBG 132409 91 10070 Plagio_glyptocarpusPl. glyptocarpus -SBBG 132402 1.0 100 1.0 70 Plagio_glyptocarpus 92 100 1.0 Plagio_leptocladusPl. leptocladus -SBBG 132404 92 1.0 100 Plagio_mollisPl. mollis -SBBG 132404 1.0 10099 Plagio_congestus_MO5203201Pl. congestus -MO 5203201 1.0 99 Plagio_congestus_MO5203201 1.0 Plagio_linifolius_MO6145366Pl. linifolius cf. -MO 6145366 100 G_simulis_SDSU20605G. similis 1 -SDSU 20605 1.0 100100 G_similis_SDSU20605G. similis 2 -SDSU 20605 1.0 100100 G_circumscissa_c_SDSU21417G. circumscissa var. c. Greeneocharis-SDSU 21417 1.0 100 G_circumscissa_rosulata_SDSU20663G. circumscissa var. r. -SDSU 20663 100 J_angustifolia_RSA731212J. angustifolia -RSA 731212 100 1.0 100 J_inaequata_RSA732141J. inaequata -RSA 732141 1.0 100100 J_fastigata_SDSU15588J. fastigiata -SDSU 15588 10066 J_racemosa_SDSU18710J. racemosa -SDSU 18710 100 1.0 6692 J_holoptera_SDSU13036J. holoptera -SDSU 13036 1.0 92 J_holoptera_SDSU13036 1.0 100 1.0 J_costata_SDSU18964J. costata Johnstonella/-SDSU 18964 J_grayi_RSA544081J. grayi var. g. -RSA 544081 100 J_grayi_RSA544081 100 J_parviflora_CONC150821 -CONC 150821 Amsinckiinae 100 100 100 J_parviflora_CONC150821 Albidae 1.0 1.0 100 J_angelica_SDSU19425J. angelica -SDSU 19425 100 J_diplotrichia_SDSU21418J. diplotricha -SDSU 21232 1.0 100 100 C_hispida_CONC150914C. hispida clade-CONC 150914 1.0 100100 C_albida_SDSU20612C. albida -SDSU 20612 1.0 100100 C_mexicana_SDSU20610C. mexicana -SDSU 20610 1.0 100 C_texana_SDSU20611C. texana -SDSU 20611 C_recurvata_UCR225245C. recurvata -UCR 225245 100 C_recurvata_UCR225245 100 100 C_crassisepala_SDSU20623C. crassisepala -SDSU 20623 1.0 100 1.0 100 C_minima_SDSU20629C. minima -SDSU 20629 1.0 100100 C_kelseyana_SDSU20630C. kelseyana -SDSU 20630 79 100 C_kelseyana_SDSU20630 .97 79 1.0 C_fendleri_SDSU20114C. fendleri -SDSU 20114 cpDNA 100 C_nevadensis_SDSU20393C. nevadensis -SDSU 20393 100 100 C. scoparia -UCR 211150 1.0 100 100 1.0 C_scoparia_UCR211150 1.0 100100 C_utahensis_SDSU20348C. utahensis -SDSU 20348 1.0 100100 C_gracilis_UCR217631C. gracilis -UCR 217631 1.0 100100 C_oxygona_RSA685321C. oxygona -RSA 685321 97 1.0 100100 C_mohavensis_SDSUC. mohavensis -SDSU 20877 1.0 97 1.0 100 C_pterocarya_SDSU20355C. pterocarya f. p. -SDSU 20355 100 C_ganderi_SDSU20345C. ganderi -SDSU 20345 100 1.0 100 C_intermedia_SDSU20037C. intermedia var. i. -SDSU 20037 100 1.0 100 C. barbigera var. b. -SDSU 20349 100 C_barbigera_SDSU20349 1.0 97 C_microstachys_SD16851C. microstachys -SD 216851 C_corollata_SDSU20775C. corollata -SDSU 20775 100 1.0 97100 100 -SD 199070 1.0 100 100 C_affinis_SD199070 1.0 78 1.0 100 C_clevelandii_SDSU20782C. clevelandii -SDSU 20782 NR 78100 C_wigginsii_SDSU20082C. wigginsii -SDSU 20082 100 C_wigginsii_SDSU20082 1.0 C_decipens_SDSU20014C. decipens -SDSU 20014 C_alfalfalis_CONC163659C. alfalfalis -CONC 163659 100 100 C_calycotricha_CONC150898C. calycotricha -CONC 150898 Fig. S1. Maximum likelihood analysis 100 100 C_calycotricha_CONC150898 1.0 84 1.0 C_aspera_MO4317599C. aspera -MO 4317599 84 C_cynoglossoides_SI87776C. cynoglossoides Cryptantha-SI 87776 from complete plastome (cpDNA) data. 1.0 100 100 C_glomerata_SGO146941C. glomerata subsp. g. -SGO 146941 100 100 C_glomerata_SGO146941 1.0 100 1.0 C_alyssoides_CONC156553C. alyssoides s.str.-CONC 156553 Cladogram, showing outgroups and 100 1.0 100 84 C_capituliflora_CONC166914 -CONC 166914 100 99 1.0 84 C_kingii_SGO123832C. kingii -SGO 123832 major clades within the Amsinckiinae, 1.0 1.0 9980 C_glomerulifera_CONC166867C. glomerulifera -CONC 166867 1.0 80 C_gnaphalioides_SGO146002C. gnaphalioides -SGO 146002 J_micromeres_SDSU20762J. micromeres -SDSU 20762 with bootstrap support values (above) C_dumetorum_SDSU18694C. dumetorum -SDSU 18694 100 100 C_flaccida_SDSU19846 -SDSU 19846 100 C_flaccida_SDSU19846 and Bayesian inference posterior prob- 1.0 1.0 100100 C_sparsiflora_UCR184326 -UCR 184326 100 1.0 100 C_simulans_SDSU20390C. simulans -SDSU 20390 100 100 C. watsonii -UCR 226737 abilities (below; NR = not recovered in 1.0 100 1.0 100 C_watsonii_UCR226737 100 100 C_torreyanaC. torreyana -SDSU 20124 Bayesian inference analysis). Dashed 1.0 100 1.0 100 C_ambigua_SDSU20524C. ambigua -SDSU 20524 1.0 100100 C_crinita_SDSU2082C. crinita -SDSU 20823 100 100 C_crinita_SDSU2082 1.0 100 1.0 C_mariposae_SDSU20826C. mariposae -SDSU 20826 line represents alternative arrange- 100 C_incana_UCR227031C. incana -UCR 227031 1.0 100 C_echinella_SDSU19611C. echinella -SDSU 19611 ment from Bayesian analysis. Major 100 C_leiocarpa_SDSU20759C. leiocarpa -SDSU 20759 100 100 C_leiocarpa_SDSU20759 100 100 1.0 C_clevelandii_RSA710334C. hispidissima 1 -RSA 710334 1.0 1.0 100100 C_clevelandii_SDSU18342C. hispidissima 2 -SDSU 18342 and backbone nodes with weak support 100 C_clevelandii_SDSU18342 100 1.0 100 C_nemaclada_SDSU20774C. nemaclada -SDSU 20774 indicated by thick-lined circles, those 100 1.0 100 C_nevadensisR_SDSU20766C. juniperensis -SDSU 20766 1.0 C_diffusa_MERL56799C. diffusa -MERL 56799 100 100 100 C_peruviana_SGO140959C. peruviana -SGO 140959 with mixed support indicated by thin- 1.0 42 1.0 100 C_phaceloides_SGO146206C. phaceloides -SGO 146206 NR 4251 C_globulifera_SGO147985C. globulifera 1 -SGO 147985 lined circles. Adel. = Adelinia; And. = NR 5158 C_involucrata_SGO146942C. globulifera 2 -SGO 146942 NR 5871 C_globulifera_CONC163475C. globulifera 3 -CONC 163475 Andersonglossum. NR 71 C_linearis_SGO147688C. globulifera 4 -SGO 147688

0.003 S1 (TAXON 66 (6) • December 2017 Electr. Suppl. to: Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae

Hac. micrantha My. laxa Mi. tibetica 98 An. virginianum 1.0 An. occidentale And. Ad. grandis 59 D. daubenmirei Adel. 1.0 Pe. setosa Dasynotus 66 Pe. pusilla 48 1.0 54 100 H. arizonica Pectocarya/ .99 .89 58 1.0 H. palmeri .98 100 Pe. linearis var. f. Harpagonella 39 1.0 Pe. penicillata 100 E. lepida 1 .99 99 1.0 E. lepida 2 1.0 100 E. micrantha var. m. 2 96 Eremocarya 30 1.0 E. micrantha var. m. 1 1.0 E. micrantha var. p. NR O. virgata 100 28 1.0 O. crymophila .81 36 .55 50 O. suffruticosa var. a. 13 31 .74 O. setosissima NR .68 51 O. humilis subsp. h. .99 98 O. sobolifera 3 1.0 O. hyposphila NR O. subretusa “mt ed.” Oreocarya NR 35 O. nubigena “craters” .50 42 O. “ursina” .11 NR 12 NR 55 O. nubigena “mam.” 23 .77 88 O. subretusa “type” NR .51 O. “howelliana” 18 O. hoffmannii .83 30 NR 68 O. schoolcraftii .99 100 O. glomerata 1.0 31 O. subretusa “w. mts” NR O. nubigena “granite” 100 Pl. glomeratus 1.0 Pl. hispidus Sonnea 54 J. diplotricha 68 1.0 C. albida 1.0 100 C. mexicana NR C. texana 100 94 J. racemosa 1.0 44 J. holoptera .97 J. fastigiata Johnstonella/ 87 1.0 9 1.0 94 Albidae NR 1.0 44 J. angelica 10 NR C. hispida clade .90 22 J. grayi var. g. .92 100 J. costata 1.0 98 J. inaequata 1.0 J. angustifolia Pl. jonesii 56 Pl. kingii var. h. .51 78 Amsinckia/ 100 Am. tessellata var. t. 47 1.0 Am. intermedia 2 1.0 100 Amsinckiopsis 1.0 46 1.0 76 Am. intermedia 3 .90 NR Am. intermedia 1 58 Pl. fulvus var. c. .98 Pl. canescens var. c. 100 100 Pl. myosotoides NA 1.0 1.0 92 Pl. torreyi NR 71 1.0 Pl. tenellus 1.0 99 Pl. nothofulvus .67 1.0 Pl. collinus var. u. 76 Pl. plurisepalus 1.0 100 Pl. linifolius cf. Plagiobothrys NR 1.0 66 .56 Pl. congestus 15 .95 58 52 Pl. gracilis s.str. NR 1.0 NR Pl. polycaulis 57 78 Pl. undulatus NR 44 1.0 Pl. leptocladus 1.0 36 Pl. mollis .97 75 Pl. austinae 1.0 63 Pl. glyptocarpus .99 Pl. greenei Cy. creticum G. similis 1 Cy. cr. 100 G. circumscissa var. c. Amsinckiinae 1.0 100 .99 32 G. similis 2 Greeneocharis 8 .86 G. circumscissa var. r. NR 99 C. martirensis 1.0 98 C. muricata var. m. 58 1.0 C. clokeyi Maritimae 25 .98 100 C. subamplexicaulis NR 1.0 98 C. maritima var. m. clade 1.0 J. echinosepala C. recurvata 97 C. fendleri 1.0 100 1.0 100 C. minima 1.0 75 C. crassisepala 22 14 1.0 C. kelseyana .95 .94 100 1.0 50 50 1.0 C. simulans 1.0 100 C. ambigua 100 1.0 C. torreyana var. t. 1.0 79 C. watsonii 1.0 100 C. mariposae 1.0 C. crinita 88 C. dumetorum 1.0 25 C. gnaphalioides NR 58 C. cynoglossoides 1.0 79 C. calycotricha .98 59 35 .95 NR 48 C. aspera 66 .62 C. alyssoides nrDNA 10 C. kingii .98 5 C. glomerulifera Cryptantha 11 NR 88 C. glomerata subsp. g. NR 1.0 C. alfalfalis s.str. 96 C. nevadensis 1.0 C. scoparia 42 C. utahensis 1.0 82 C. gracilis 1.0 21 C. mohavensis NR 40 1.0 45 C. pterocarya f. p. 8 1.0 C. oxygona NR J. micromeres 75 C. hispidissima1 91 .98 C. leiocarpa 3 1.0 100 C. nemaclada Fig. S2. Maximum likelihood analysis .86 1.0 C. juniperensis 92 C. ganderi from complete cistron (nrDNA) data. 59 C. microstachys 15 1.0 C. decipens NR 98 1.0 Cladogram, showing outgroups and 1.0 100 C. intermedia var. i. 88 1.0 C. barbigera var. b. major clades within the Amsinckiinae. 1.0 87 C. wigginsii 7 91 .96 32 C. clevelandii Support values and abbreviations as in NR 1.0 C. corollata NR C. hispidissima 2 Fig. S1. 36 100 C. echinella .57 1.0 C. incana 70 99 C. peruviana 1.0 .1.0 C. phaceloides 95 C. diffusa 1.0 21 C. globulifera 2 S2 NR 41 C. globulifera 1 NR 65 C. globulifera 3 1.0 71 NR C. globulifera 4 (TAXON 66 (6) • December 2017 Electr. Suppl. to: Simpson & al. • Phylogeny of Amsinckiinae (Boraginaceae

100 MHac.yosotis_sp micrantha 1.0 HaMckelia_miy. laxacrantha 62 CCyno_cyr.eti ccreticumum MiMi.croula_ ttibeticaibetica_GH00466293 91 Cyno_gran_MAd. grandisGS Adel. 1.0 100 Cyno_virginianum_DUAn. virginianumKE379623 1.0 CAn.yno_occ ideoccidentalentale_RSA793153 And. 100 DaD.syno tus_daub_daubenmireiSDSU20343 1.0 100 100 Harpagonella_ariH. arizonicazonica_ARIZ388168 Dasynotus 1.0 100 1.0 Harpagonella_palmerH. palmerii .96 86 PPe.ectocarya_pusill pusillaa Pectocarya/ 100 1.0 100 PPe.ectocarya_ setosasetosa_JEPS108272 1.0 1.0 100 PPe.ec_peni cillapenicillatata Harpagonella 1.0 PPe.ectocarya_lineari lineariss_JEPS10827 va2 r. f. PPl.lagio_jone jonesiisii_UCR215416 95 PPl.lagio_kingii_UC187687 kingii va4 r. h. .99 83 Amsinckia/ 1.0 100 AAm.msinckia_ testessellatasellata_SDSU20350 var. t. 100 AAm.msinckia_ spintermedia 1 NR 1.0 100 1.0 98 AAm.msinckia_menzie intermediasii 3 Amsinckiopsis 1.0 .93 .99 AAm.msinckia_in intermediatermedia_SDSU20756 2 PPl.lagio_f ulvufulvuss var. c. 100 100 PPl.lagio_collinu collinuss var. u. 1.0 1.0 PPl.lagio_no nothofulvusthofulvus 79 PPl.lagio_cane canescensscens var. c. .92 100 PPl.lagio_myoso myosotoidestoidesNA NA 81 1.0 100 PPl.lagio_ torreytorreyii 1.0 1.0 PPl.lagio_t etenellusnellus 82 31 PPl.lagio_plurisepalus_N plurisepalusSW647734 Plagiobothrys 100 1.0 NR 39 PPl.lagio_nonpri polycauliscklyChile NR 59 PPl.lagio_pric gracilisklyChile 1.0 100 NR PPl.lagio_greene greeneii s.str. 1.0 PPl.lagio_au austinaestiniae 56 100 PPl.lagio_gl glyptocarpusyptocarpus 1.0 1.0 76 PPl.lagio_lep leptocladustocladus 94 .99 PPl.lagio_undula undulatustus 1.0 81 PPl.lagio_molli molliss .79 89 PPl.lagio_conges congestustus_MO5203201 1.0 PPl.lagio_lini linifoliusfolius_MO614536 6cf. 100 EE.remo_lepida_ lepidaSDSU2120 19 100 1.0 EE.remo_lepida_ lepidaSDSU1953 23 1.0 100 EE.remo_micran micranthatha_micrantha_SD20329 var7. m. 2 Eremocarya 36 EE.remo_micran micranthatha_SDSU18956 var. m. 1 48 1.0 44 .99 EE.remo_micran micranthatha_pseudolepida_ SvaDSU2120r. p.5 .94 C_C.subample subamplexicaulisxicaule_SGO129437 NR 100 J_eJ.chino echinosepalasepala_SD228804 100 1.0 C_mariC. tima_maritimaSDSU20050 var. m. Maritimae 1.0 100 C_marC.t irenmartirensissis_SDSU18625 100 67 1.0 C_C.cloke yclokeyii_UCR164170 clade 1.0 1.0 C_muriC. camuricatata_SDSU20749 var. m. 100 OO.recryL3 7crymophila 1.0 100 OO.rehypL3 hyposphila6 1.0 OO.resobL3 sobolifera1 94 OO.rciabLR 6suffruticosa var. a. 100 100 1.0 OO.resetL5 setosissima4 1.0 1.0 93 OO._virga ta_virgataSDSU20117 1.0 OreflaLR5 100 100 OO.rsumeL5 subretusa2 “mt ed.” 1.0 1.0 100 OO.reschL4 schoolcraftii1 100 OO.recelL6 9glomerata Oreocarya 1.0 100 100 1.0 OO.rsuwmL3 subretusa9 “w. mts” 1.0 OO.rnugrL3 nubigena8 “granite” 1.0 45 OO.rnucrL6 nubigena1 “craters” .99 30 OO.reursL4 “ursina”0 38 .85 OO.rehumL5 humilis9 subsp. h. 57 1.0 75 OO.rehofLR hoffmannii4 .97 1.0 100 OO.rnummLR nubigena1 “mam.” 1.0 100 OO.rsutyL6 subretusa5 “type” 1.0 OO.rehowLR “howelliana”9 100 PPl.lagio_glomera glomeratustus_UTC230182 61 1.0 PPl.lagio_hi shispiduspidus_JEPS87508 Sonnea 1.0 100 GG._simuli ssimilis_SDSU20605 1 100 GG._simili s_similisSDSU20605 2 1.0 G_cG.ircu mscircumscissacissa_rosulata_SDSU20663 var. r. Greeneocharis 1.0 GG._circumsci circumscissassa_c_SDSU21417 var. c. J_cosJ.t acostatata_SDSU18964 100 77 J_holopJ. holopteratera_SDSU13036 1.0 68 1.0 J_raJ.cemo racemosasa_SDSU18710 1.0 59 J_J.fast igafastigiatata_SDSU15588 98 100 1.0 J_inaequaJ. inaequatata_RSA732141 1.0 J_anguJ. angustifoliastifolia_RSA731212 Johnstonella/ 28 1.0 41 J_graJ.y i_RgrayiSA54408 1var. g.

100 Amsinckiinae .96 .84 J_angelica_J. angelicaSDSU19425 100 1.0 J_parviflora_CONC150821 Albidae 1.0 99 C_hispida_CC. hispidaONC150914 100 1.0 J_diploJ. tdiplotricharichia_SDSU21418 clade 1.0 100 C_albida_C. albidaSDSU20612 1.0 100 C_mexicana_C. mexicanaSDSU20610 1.0 C_C.texana_ texanaSDSU20611 C_reC.curva recurvatata_UCR225245 100 100 C_minima_C. minimaSDSU20629 1.0 100 1.0 C_C.crassi scrassisepalaepala_SDSU20623 99 47 1.0 C_C.fendleri_ fendleriSDSU20114 1.0 C. kelseyana .96 C_kelseyana_SDSU20630 74 100 C_C.scoparia_UCR21115 scoparia0 1.0 100 1.0 C_neC.vaden nevadensissis_SDSU20393 1.0 100 C_uC.tahen utahensissis_SDSU20348 1.0 100 C_graC.c ilis_UCR21763gracilis1 1.0 100 C_oxygona_RC. oxygonaSA685321 76 1.0 100 C_pC.tero cpterocaryaarya_SDSU20355 f. p. 1.0 1.0 C_mohaC. vmohavensisensis_SDSU 100 C_ganderi_C. ganderiSDSU20345 1.0 78 C_inC.termedia_ intermediaSDSU20037 var. i. 100 .92 C_barbigera_C. barbigeraSDSU20349 var. b. 1.0 C_miC.cro stmicrostachysachys_SD16851 84 C_C.corolla corollatata_SDSU20775 100 1.0 100 100 C_C.clevelandii_ clevelandiiSDSU20782 1.0 1.0 43 1.0 C_affinis_SD199070 1.0 66 C_decipens_C. decipensSDSU20014 .99 C_wigginsii_C. wigginsiiSDSU20082 89 C_aC.spera_ asperaMO4317599 100 1.0 C_C.calyco calycotrichatricha_CONC150898 C_alC.falf alialfalfaliss_CONC163659 1.0 28 C_gnaphalioideC. gnaphalioidess_SGO146002 Cryptantha mtDNA .57 59 98 C_glomeraC. glomeratata_SGO146941 subsp. g. .96 72 1.0 C_alysC.s oidealyssoidess_CONC156553 s.str. C_glomeruliC. glomeruliferafera_CONC166867 100 NR 63 C_capituliflora_CONC166914 1.0 NR 25 NR 77 C_C.kingii_ kingiiSGO123832 NR C_C.cynoglossoides cynoglossoides_SI87776 J_miJ.cromere micromeress_SDSU20762 C_dumetorum_SC. dumetorumDSU18694 94 100 C_flaccida_SDSU19846 1.0 1.0 87 C_C.simulans_ simulansSDSU20390 95 1.0 C_sparsiflora_UCR184326 1.0 C_waC.tsonii_UCR2 watsonii26737 100 C_ambigua_C. ambiguaSDSU20524 1.0 100 C_torreyanC. torreyanaa 100 1.0 100 CC._crinit a_SDSU2crinita082 1.0 1.0 C_maripoC. mariposaesae_SDSU20826 100 C_echinella_C. echinellaSDSU19611 1.0 C_incana_UCR22703C. incana1 Fig. S3. Maximum likelihood analysis 70 100 C_leioC.c arpa_leiocarpaSDSU20759 .96 1.0 C_C.clevelandii_R hispidissimaSA710334 1 from mitochondrial (mtDNA) data. 100 99 C_C.clevelandii_ hispidissimaSDSU18342 2 1.0 1.0 100 C_nemaclada_C. nemacladaSDSU20774 51 1.0 C_neC.vaden juniperensissisR_SDSU20766 Cladogram, showing outgroups and .74 C_globuliC. globuliferafera_SGO147985 1 100 C_phaC.c eloidephaceloidess_SGO146206 major clades within the Amsinckiinae. 1.0 65 CC._peru viperuvianaana_SGO140959 1.0 38 C_diC.ffusa_M diffusaERL56799 Support values and abbreviations as in 1.0 65 C_inC.volu cglobuliferarata_SGO146942 2 1.0 70 C_linearis_C. globuliferaSGO147688 4 Fig. S1. 1.0 C_globuliC. globuliferafera_CONC163475 3

0.02 S3