Nomenclatural Changes in Lithospermum (Boraginaceae) and Related Taxa Following a Reassessment of Phylogenetic Relationships

Nomenclatural changes in Lithospermum (Boraginaceae) and related taxa following a reassessment of phylogenetic relationships

1,2 1 JAMES I. COHEN AND JERROLD I. DAVIS

1 L. H. Bailey Hortorium, Department of Plant Biology, Cornell University, 412 Mann Library Building, Ithaca, NY 14853, USA 2 Author for Correspondence: [email protected]

Abstract. Lithospermum (Boraginaceae) comprises approximately 40 species in both the Old and New Worlds, with a center of diversity in the southwestern United States and Mexico. Using ten cpDNA regions, a phylogeny of Lithospermum and related taxa was reconstructed. Lithospermum (including New World and Old World species) and related New World members of Lithospermeae form a monophyletic group, with Macromeria, Onosmodium, Nomosa, Lasiarrhenum, and Psilolaemus nested among species of Lithospermum. New World Lithospermeae also is a monophyletic group, with Eurasian species of Lithospermum sister to this group. Because Lithospermum is not monophyletic without the inclusion of the other New World genera, species from these genera are transferred to Lithospermum, and appropriate nomenclatural changes are made. New combinations are Lithospermum album, Lithospermum barbigerum, Lithospermum dodrantale, Lithospermum exsertum, Lithospermum helleri, Lithospemum leonotis, Lithospermum notatum, Lithospermum oaxacanum, Lithospermum pinetorum, Lithospermum rosei, Lithospermum trinverium, and Lithospermum unicum; new names are Lithospermum chiapense, Lithospermum johnstonii, Lithospermum macromeria, Lithospermum onosmodium, Lithosper- mum rzedowskii, and Lithospermum turneri. Key Words: Lasiarrhenum, Lithospermum, Macromeria, Nomosa, Onosmodium, Perittostema, Psilolaemus, Lithospermeae, Boraginaceae.

The genus Lithospermum L. (Boraginaceae) monophyletic, but resolution is lacking within comprises approximately 40 species (Johnston, the tribe and no New World members of 1954a). Although it has a cosmopolitan distri- Lithospermeae were included in the analysis; bution, most species of the genus occur in the therefore, Johnston’s hypothesis that all six New World, from Canada southward to Peru, genera of Lithospermeae endemic to the New with a center of diversity in Mexico and the World (i.e., all New World genera of the tribe southwestern United States. Lithospermum and other than Lithospermum) are closely related to the tribe to which it belongs, Lithospermeae Lithospermum (Johnston, 1954a) has yet to be Dumort., were revised by Ivan Johnston in a tested (but see Thomas et al., 2008). This group series of papers from 1952 to 1954 (1952, includes Macromeria D. Don, Onosmodium 1953a, b, 1954a, b). In these papers, Johnston Michx., Lasiarrhenum I. M. Johnst., Nomosa I. recognized new species within the genus, M. Johnst., Psilolaemus I. M. Johnst., and described new genera, and proposed hypothe- Perittostema I. M. Johnst. Species of these ses concerning evolutionary relationships with- genera occur from Canada to Guatemala, and in the tribe (1952, 1954a, b). A recent as with Lithospermum, the center of diversity phylogeny of Boraginaceae (Langstrom & for the group is in Mexico. Only two of these Chase, 2002) reconstructs Lithospermeae as genera, Macromeria and Onosmodium, include

Brittonia, 61(2), 2009, pp. 101–111 ISSUED: 1 June 2009 © 2009, by The New York Botanical Garden Press, Bronx, NY 10458-5126 U.S.A. 102 BRITTONIA [VOL 61 more than one species; Macromeria includes Although both Lehmann and de Candolle 11 species and Onosmodium includes seven. recognized a large, inclusive Lithospermum, Johnston’s generic demarcations within New they also recognized two endemic New World Lithospermeae are primarily based on World genera, Macromeria and Onosmodium floral characters, such as exserted vs. included (as Purshia Lehm. [Lehmann, 1818], non anthers; flattened vs. cylindrical filaments; and Purshia DC. ex Poir. [Rosacaeae]). Species campanulate, urceolate, or long funnelform from these two genera have only rarely been vs. salverform or funnelform corolla shapes included in Lithospermum (e.g., Muhlenberg, (Johnston, 1954a, b). However, phylogenetic 1813). This treatment of species in the New analyses (Cohen, 2007) suggest that the genera World differs from that of many species in the recognized on the basis of these floral features Old World in which it has been common to are not monophyletic. include species from putatively related genera Lithospermum and other New World mem- of Lithospermeae, such as Buglossoides, bers of Lithospermeae share characteristic Lithodora, and Neatostema, in Lithospermum erect, smooth to slightly dimpled, white, (see Johnston, 1953a [but see Johnston, lustrous nutlets with basal attachment, and 1954a, b for taxonomic changes], Flora all have three-celled glands inside the corol- Iranica [Rechinger, 1967], and Flora of las, except for ca. 12 taxa in which glands, China [Zhu et al., 1995]). presumably, have been lost or modified. At present, the most widely accepted Closely related genera of Lithospermeae that taxonomic system for the genus derives from share these features, such as Buglossoides Johnston’s revisionary efforts on the genus Moench and Lithodora Griseb. (Thomas et and tribe during the 1950s, but that treatment al., 2008), have blue or purple corollas, but also contributed additional complexity to the these corolla colors do not occur in Litho- group’s taxonomy. Johnston placed Arnebia spermum or other New World members of and Echioides Fabr. in Lithospermum in Lithospermeae. 1952, but two years later he removed them, The taxonomic history of Lithospermum is noting differences in stamens and pollen. complex, and the complexity arises, in part, Also, he placed some newly described spe- from treatments of the genus in the 1800s. For cies in Lithospermum in 1953, and transferred example, in Plantae e Familia Asperifoliarum them to Arnebia the following year. In his Nuciferae (1818), Lehmann circumscribed a 1952 revision of Lithospermum, Johnston broadly inclusive Lithospermum, which includ- recognized the difficulty of dividing Litho- ed species currently placed in Lithospermum spermum into smaller groups, stating that (Johnston, 1954a) as well as species currently “The genus has great internal coherence. It included in other genera, such as Mertensia is exasperatingly lacking in lines of cleavage” Roth, Buglossoides, Arnebia Forssk., and Lith- (1952, p. 300). odora. A later taxonomic treatment of Lith- Taxonomic questions regarding generic ospermum in the Prodromus Systematis boundaries of Lithospermum still arise, as Naturalis Regni Vegetabilis by A. P. de Can- seen in the treatment of Boraginaceae in the dolle (1846) was, like Lehmann’s treatment, Flora of China (Zhu et al., 1995), in which quite inclusive; however, unlike Lehmann, de Lithospermum is circumscribed as including Candolle divided the genus into three sections. species often placed in Buglossoides. Treating One section, Eulithospermum DC. included Lithospermum in this broad sense is common only species currently included in Lithosper- in Old World literature; however, Buglos- mum (with one exception, L. chinensis Hook. soides is not monophyletic (cf. Langstrom & and Arn., which is a synonym of Heliotropium Chase, 2002; Cohen, unpublished data) and strigosum Willd. [Zhu et al., 1995]). The other may not be sister to Lithospermum. two sections in the treatment, Rhytispermum A recent taxonomic study of Lithospermum Link and Margarospermum Rchb. f., included (Ralston, 1993) included a phylogenetic anal- species that are often included in other genera, ysis of the genus utilizing a total of twenty such as Buglossoides, Lithodora, Neatostema I. morphological, palynological, karyological, M. Johnst., and Moltkia Lehm. and chemical characters. This analysis provid- 2009] COHEN & DAVIS: LITHOSPERUMUM NOMENCLATURE 103 ed support for the recognition of several species Materials and methods groups, but did not resolve relationships among these groups. Recent work on other taxa of TAXON SAMPLING New World Lithospermeae includes Turner’s Twenty-seven species were included in the synopses of Macromeria (1994a), Onosmo- analyses. Twenty-two belong to the ingroup, dium (1995), and Lasiarrhenum (1994b), and and five are outgroup taxa from related Boyd’s (2003) phylogeny of Macromeria genera in Lithospermeae (Table I). The based on morphological characters. ingroup sampling represents the range of The present study was undertaken to variation within Lithospermum, and species examine the phylogenetic relationships from all but one elusive genus of New World among species of Lithospermum and other Lithospermeae, Perittostema, were included. elements of New World Lithospermeae, Perittostema was not included in the analyses with two of the goals being to determine because it is only known from the type, whether the non-monotypic genera Litho- housed in Paris (P), and the type locality is spermum, Macromeria, and Onosmodium unknown; however, Johnston hypothesized constitute monophyletic groups and to re- that it was collected in Oaxaca, Mexico evaluate the taxonomy of New World (Johnston, 1935, 1954a). The majority of the Lithospermeae. taxa were collected from wild populations.

TABLE I LIST OF TAXA INCLUDED IN ANALYSES ALONG WITH COLLECTION OR SOURCE DATA.

Taxon Collection Location

OUTGROUP SPECIES Alkanna tuberculata Greuter 2002 1260-96 Natl. Bot. Gard. Belgium Buglossoides arvense (L.) I. M. Johnst. 19792083 Natl. Bot. Gard. Belgium Echium vulgare L. J. Cohen 212 Cornell Plantations Lithodora diffusa. (Lag.) I. M. Johnst. M. Chase 6063 Cantabria, Spain Onosma stellulata Waldst. & Kit. 1992 1317-39 Natl. Bot. Gard. Belgium INGROUP SPECIES Lasiarrhenum trinervium (Lehm.) I. M. Johnst J. Cohen 228 Michoacán, Mexico Lithospermum californicum A. Gray J. Cohen 43 Oregon, U.S.A. Lithospermum canescens Lehm. J. Cohen & S. Straub 12 Ohio, U.S.A. Lithospermum caroliniense MacMill. J. Cohen 11 Indiana, U.S.A. Lithospermum cobrense Greene J. Cohen 78 Texas, U.S.A. Lithospermum erythrorhizon Siebold & Zucc. J. Cohen 173 Cornell Plantations Lithospermum multiflorum A. Gray J. Cohen 57 Arizona, U.S.A. Lithospermum nelsonii Greenm. J. Cohen 184 Nuevo León, Mexico Lithospermum obovatum J. F. Macbr. J. Cohen 208 Durango, Mexico Lithospermum officinale L. J. Cohen 171 Cornell Plantations Lithospermum strictum Lehm. J. Cohen 225 Michoacán, Mexico Lithospermum tuberosum DC. J. Cohen 108 Georgia, U.S.A. Macromeria exserta D. Don J. Cohen 224 Michoacán, Mexico Macromeria hispida M. Martens & Galeotti J. Cohen 218 Michoacán, Mexico Macromeria leonotis I. M. Johnst. J. Cohen 195 Nuevo León, Mexico Macromeria longiflora D. Don J. Cohen 226 Michoacán, Mexico Macromeria notata I. M. Johnst. J. Cohen 188 Nuevo León, Mexico Macromeria viridiflora DC. J. Cohen 141 Arizona, U.S.A. Nomosa rosei I. M. Johnst. J. Cohen 207 Durango, Mexico Onosmodium helleri Small J. Cohen 132 Texas, U.S.A. Onosmodium molle (Michx.) Muhl. 2003-0671 Missouri Botanical Garden Psilolaemus revolutus (B. L. Rob.) I. M. Johnst. J. Cohen 199 San Luis Potosí, Mexico 104 BRITTONIA [VOL 61

For these taxa, herbarium specimens were TABLE II collected and deposited at BH, and leaf tissue ANNEALING TEMPERATURES FOR THE CPDNA REGIONS was dried and preserved in silica gel for AMPLIFIED FOR THIS STUDY. subsequent DNA extraction. Taxa not collect- Region Annealing Temperature (°C) ed from natural populations were obtained from gardens as leaves preserved in silica ndhF - rpl32 56 (Cornell Plantations, Missouri Botanical Gar- psbA - trnH 58 psbJ - petA 50 den, and National Botanic Garden of Bel- rpl16 intron 54–59 gium) or as DNA isolations from the DNA trnK - rps16 50 bank of Royal Botanic Gardens, Kew. See trnL - rpl32 50 Table I for specimen list and information. trnL - trnF 50 trnQ - rps16 50–54 ycf6 - psbM 54 MatK 48–50 DNA EXTRACTION, PCR AMPLIFICATION, AND SEQUENCING DNA extraction was performed with dried was performed by the Life Sciences Core plant tissue using a modified CTAB extrac- Laboratory Center at Cornell University with tion method (Doyle & Doyle, 1990) which an Applied Biosystems (ABI) 3730 DNA included 2% PVP-40 and in some cases the Analyzer. Alternatively, sequencing reactions addition of 0.5 M glucose in the CTAB and subsequent steps were performed by the extraction buffer. PCR amplifications of the Life Science Core Laboratory Center at ten chloroplast (cpDNA) regions, ndhF - Cornell University, using Big Dye termina- rpl32, psbA - trnH, psbJ - petA, the rpl16 tors and either an ABI 3700 or an ABI 3730. intron, trnK - rps16, trnL - rpl32, trnL - trnF, Sequence trace files were compiled, exam- trnQ - rps16, ycf6 - psbM (Shaw et al. 2005, ined, and edited with Sequencher ver. 4.6 – 2007), and matK (390F [Cuénoud et al., 4.8 (Gene Codes Corporation, MI U.S.A.), 2002] and 1710R [Barfuss et al., 2005]), and sequences were deposited in Genbank were performed using the published primers (Genbank accessions FJ827256 – FJ827480). of the cited authors. PCR mixtures, 25 μL in volume, consisted of 67 mM Tris-HCl with ALIGNMENT, GAP CODING, AND PHYLOGENETIC 2.1% DMSO and 0.01% TritonX per reaction or 1X Ex Taq Buffer (Takara Bio Inc., Japan), ANALYSIS 2 mM MgCl2, 0.2 mM to 0.25 mM dNTPs , Initial alignments were performed with MUS- 1 μM of primers, 0.125 μL to 1 μL of Taq CLE (Edgar, 2004) as implemented by the polymerase, and 0.1 μL to 2.5 μL DNA European Bioinformatics Institute’s MUSCLE sample, depending on the DNA concentra- server (http://www.ebi.ac.uk/Tools/muscle/ tion. Amplifications were performed in an index.html) using the default settings. Subse- Eppendorf Mastercycler Gradient 5331 ther- quent adjustments were made manually in mocycler with the programs in Appendix 1, Bioedit ver. 7.0.5.3 (Hall, 1999). Gaps were using the annealing temperatures listed in coded using simple indel coding (Simmons & Table II. PCR products were run on a 1% to Ochoterena, 2000). Inversions were coded as 1.5% agarose gel and stained with ethidium present/absent with the inverted sequence bromide to determine if amplification oc- regions excluded from the analyses (Ochoterena, curred. Prior to sequencing, some PCR in press). All characters were weighted equally reactions were purified with the QIAquick and treated as unordered. Regions that aligned PCR purification kit (Qiagen, Germany). ambiguously (ca. 234 bp of the rpl16 intron) Sequencing reactions were performed with were excluded from analyses. BigDye 3.1 (Applied Biosystems, CA U.S. Maximum parsimony phylogenetic analy- A.) terminators and locus-specific amplifica- ses were conducted using TNT (Goloboff et tion primers. Sequencing products were pre- al., 2000), with 10,000 trees held in memory, cipitated using a modification of the ethanol/ and 1000 parsimony ratchet iterations per- EDTA/sodium acetate method (Applied Bio- formed (Nixon, 1999), with 10% probability systems), and automated cycle sequencing of upweighting and 10% probability of down- 2009] COHEN & DAVIS: LITHOSPERUMUM NOMENCLATURE 105 weighting, followed by 100 cycles of tree characters (substitutions, no indels or inver- drifting, 100 rounds of tree fusing, and sions) out of 1140 bp (although 234 aligned random sectorial searches (Goloboff, 1999). nucleotides were difficult to align and, there- This search strategy was repeated for ten fore, not included in analyses), or 5% of the rounds with five replications per round using total sequence length. The least informative the Xmult setting. In total, 50 searches, cpDNA region, in terms of total number of 50,000 ratchets, 5000 cycles of drift, and characters, was trnL - trnF, with four infor- 5000 rounds of tree fusing were performed. mative characters; however, matK, with sev- The commands for the analysis are provided en informative characters, was the least in Appendix 2. Clade support was measured informative region in terms of percent infor- with TNT by conducting 10,000 jackknife mative characters (0.6% vs. 0.7% for trnL - replicates (36% removal probability, tradi- trnF). These percentages are not strictly tional search strategies, and groups below 1% comparable because different numbers of taxa support collapsed) (Farris et al., 1996). were sampled for each region; a more Consistency indices were calculated after re- complete evaluation of variation levels moval of parsimony-uninformative characters. among the cpDNA regions is in preparation.

Results PHYLOGENETIC RESULTS SEQUENCE VARIATION Four most-parsimonious trees of 374 steps A total of 8232 aligned nucleotides from (CI=0.62, RI=0.61) were discovered, the the 27 species were included in the analyses strict consensus tree is provided in Fig. 1. (Table III). The mean number of species Most relationships are well resolved. Litho- sequenced for each cpDNA region was 23, spermum is not monophyletic, as sampled with all 27 species sequenced for ndhF - representatives of other genera of New World rpl32 and the rpl16 intron, and the fewest Lithospermeae are nested within it. The group number of species, 13, sequenced for matK. including Lithospermum and these other The sequence data yielded 207 informative genera is monophyletic with 99% jackknife characters, 176 (85%) from nucleotides, and support, and New World Lithospermeae is a 31 (15%) from gaps and inversions. The monophyletic group with 54% jackknife sequenced cpDNA regions ranged in length support. Most of the larger clades in the from 420 bp to 1144 bp. The most informa- group that includes Lithospermum and other tive region was rpl16, with 45 informative members of New World Lithospermeae are

TABLE III SUMMARY OF CPDNA REGIONS INCLUDED IN ANALYSES.

Informative Structural % Informative Number Aligned Informative Nucleotide Characters Total Informative Nucleotide Site Region of Taxa Length (bp) Site Characters (gaps and inversion) Characters Characters ndhF - rpl32 27 801 27 4 31 3.4% psbA - trnH 23 420 8 5 13 1.9% psbJ - petA 23 821 17 6 23 2.1% rpl16 intron 27 1140a 45 3 48 5.0% trnK - rps16 24 894 23 3 26 2.6% trnL - rpl32 25 838 21 5 26 2.5% trnL - trnF 17 549 4 1 5 0.7% trnQ - rps16 23 497 14 2 16 2.8% ycf6 - psbM 23 1128 10 1 11 0.9% matK 13 1144 7 1 8 0.7% Mean 23 823 18 3 21 2.2%

Total 225 8232 176 31 207 2.3%

a (includes 234 characters deactivated, but these characters not included in number of Informative Characters or % Informative Characters) 106 BRITTONIA [VOL 61

Alkanna tuberculata Echium vulgare Old World Onosma stellulata Buglossoides arvense 99 Lithodora diffusa

99 Lithospermum officinale 99 Lithospermum erythrorhizon Lithospermum canescens 99 Macromeria longiflora Lithospermum nelsonii 54 99 Macromeria leonotis

99 Macromeria hispida Macromeria exserta

55 Psilolaemus revolutus Lithospermum cobrense New World Lithospermum caroliniense Lithospermum obovatum Macromeria notata Lithospermum strictum Lithospermum tuberosum

99 Onosmodium helleri Onosmodium molle Lithospermum californicum Nomosa rosei Lasiarrhenum trinervium

98 Lithospermum multiflorum Macromeria viridiflora

FIG. 1. Strict consensus of 4 MP trees of 374 steps, CI=0.62, RI=0.61. Jackknife values >50% are shown above branches. not well supported; however, each of five Lithospermum officinale and L. erythrorhi- species pairs, L. officinale L. and L. erythro- zon, the two Old World species of Lithosper- rhizon Siebold and Zucc., L. nelsonii mum included in the analysis, form a clade Greenm. and M. leonotis I. M. Johnst., M. that is sister to the rest of Lithospermum and hispida M. Martens & Galeotti and M. other members of New World Lithospermeae. exserta D. Don, O. helleri Small and O. Lithodora diffusa (Lag.) I. M. Johnst. is sister molle Michx., and L. multiflorum A. Gray and to Lithospermum and other members of New M. viridiflora DC has 98% or 99% jackknife World Lithospermeae with 99% jackknife support. support. Although Onosmodium was found to be monophyletic, Macromeria is not a mono- Discussion phyletic group. Psilolaemus is sister to Lithospermum cobrense Greene, and Nomosa PHYLOGENETIC ANALYSES and Lasiarrhenum are sister taxa, but neither The analysis resolves Lithospermum and of these relationships is supported by jack- other New World members of Lithospermeae knife values greater than 50%. as a monophyletic group, and this is also the 2009] COHEN & DAVIS: LITHOSPERUMUM NOMENCLATURE 107 case when more taxa are added to the phyloge- in Lithospermum, or alternatively the clade netic analyses (Cohen, unpublished data). containing Lithospermum and other New However, Lithospermum is not monophyletic World Lithospermeae could be divided along because species of other genera of New World different lines than have previously been Lithospermeae are interdigitated among species drawn. We have not observed diagnostic of this genus (Fig. 1). Although species of character combinations by which a series of Lithospermum differ from each other in habit, readily recognizable monophyletic genera can corolla shape, and corolla color, the other be delimited. In addition, most of the mono- genera of New World Lithospermeae have been phyletic groups that are readily delimited segregated from Lithospermum on the basis of have five or fewer species. Splitting along the same sorts of features, which appear to be these lines would result in several small derived vegetative and floral characters. For genera that would be quite difficult to example, Nomosa and Lasiarrhenum are the distinguish from each other, and this would only two genera of New World Lithospermeae not provide a useful taxonomy. The former with completely white corollas and three- option, an expansion of the genus, is more veined leaves. In addition, both have corolla useful because Lithospermum and the other shapes, urceolate and campanulate respectively, species of New World Lithospermeae share which are not found in any other species in the multiple synapomorphies, including a nutlet group. that is erect, smooth to slightly dimpled, Macromeria is a polyphyletic assemblage white, and lustrous. Although some species of species that share a similar phenotype. of the endemic New World genera of Lith- Species that have been placed in this genus ospermeae have not been included in these are often large in habit (> 0.75 m tall) with phylogenetic analyses, careful examination long corollas (> 4 cm) and exserted anthers has shown that these species also share the and stigmas. This suite of characteristics characteristic nutlet, a synapomorphy for the appears to have arisen multiple times inde- clade, as well as other vegetative and floral pendently among species of New World characteristics similar to those of species that Lithospermeae. are included in the analyses. Therefore, a Onosmodium, on the other hand, is re- broader circumscription of Lithospermum is solved as monophyletic, but only two of the proposed, with the genus recognized by the seven species of the genus were included in characteristic nutlet in association with corol- this study. Although the sampling is limited, las that are yellow, yellow-green, white, and/ the group has 99% jackknife support, and or orange in color. Including all members of with the inclusion of other taxa in the the New World Lithospermeae in Lithosper- analyses Onosmodium remains monophyletic mum is a useful and conservative approach to (Cohen, unpublished data). the taxonomy of the group. Lithodora diffusa is sister to the clade that Several of the species that are assigned to includes Lithospermum and other genera of genera other than Lithospermum have previ- New World Lithospermeae. This sister rela- ously been included in Lithospermum, but tionship differs from the results obtained by most others have never been included in the Thomas et al. (2008) for Lithodora and genus. Most species that are part of the related taxa, in which Buglossoides, rather former group retain their specific epithet than Lithodora, was placed as sister to when included in Lithospermum. Species that Lithospermum and other genera of New have not been classified as part of Lithosper- World Lithospermeae. Differences in sam- mum must be transferred to that genus. As pling of both taxa and loci could contribute to Lithospermum is a Linnaean name that has these dissimilar results. proven to be a catch-all for many species with smooth, white nutlets, several common specific epithets, such as hispida, have been used TAXONOMY AND NOMENCLATURE previously. Therefore, six species require new As currently circumscribed, Lithospermum specific epithets, and those proposed here is not monophyletic. All of the species of refer to the unique characteristics and history New World Lithospermeae could be included of these species. 108 BRITTONIA [VOL 61

Lithospermum album (G. L. Nesom) J. Type: U.S.A. Texas: Kerr Co., along Bear Cohen, comb. nov. Basionym: Macromeria Creek, 1600 – 2000 ft, 30 Apr 1894, A. A. alba G. L. Nesom, Madroño 36: 28. 1989. Heller 1682 (holotype: NY; isotypes: UC- Type: Mexico. Tamaulipas, Mpio. Gomez n.v., US). Farias, 5–7 km, NW of Gomez Farias, just S of Agua del Indio, 30 May 1969, A. Lithospermum johnstonii J. Cohen, nom. Richardson 1763 (holotype: TEX). nov. Basionym: Macromeria hispida M. Martens & Galeotti, Bull. Acad. Roy. Sci. Lithospermum barbigerum (I. M. Johnst.) J. Bruxelles 11: 339. 1844, non Lithosper- Cohen, comb. nov. Basionym: Macromeria mum hispidum Forssk., Fl. Aegypt.-Arab. barbigera I. M. Johnst. J. Arnold Arbor. 16: 39. 1775. Type: Mexico. Michoacán: near 189. 1935. Type: Mexico. Nuevo León, N Morelia, Cerro de Quinzeo, 6500 – 8000 ft, slope of Sierra Tronconal between Canyon 1840–1844, Galeotti 1917 (holotype: BR- de los Charcos and Canyon de San Miguel, n.v., isotypes: GH). ca. 15 mi SW of Galeana, “common in dense oak woods beyond pine-fir belt,” This species is named for Ivan M. Johnston, 6000–9000 ft., 4 Jun 1934, C. H. Mueller Boraginaceae systematist. His work on & M. T. Mueller 741 (holotype: GH; Lithospermum, Lithospermeae, and Boragi- isotypes: F-n.v., MEXU, MICH-n.v., TEX). naceae has provided great insight for many botanists. Lithospermum chiapense J. Cohen, nom. nov. Basionym: Macromeria guatemalen- Lithospermum leonotis (I. M. Johnst.) J. sis I. M. Johnst., J. Arnold. Arbor. 29: 232. Cohen, comb. nov. Basionym: Macrome- 1948, non Lithospermum guatemalense ria leonotis I. M. Johnst. [“leontis”], J. Donn.Sm., Bot. Gaz. 27: 436. 1899. Type: Arnold Arbor. 16: 188. 1935. Type: Mex- Guatemala. San Marcos, south facing slope ico. Nuevo León: “ascent into Taray,” ca. of Volcán Tajumulco, between Las Canojas 15 mi SW of Galeana, ca. 8000 ft, 6 June and top of ridge, 7 mi from San Sebastián, 1934, C. H. Mueller & M. T. Mueller 754 3300–3900 m, 16 Feb 1940, J. A. Steyer- (holotype: GH; isotype: TEX). mark 35898 (holotype: GH; isotype: F). Lithospermum macromeria J. Cohen, nom. This species has a range restricted to nov. Basionym: Macromeria viridiflora Guatemala and Chiapas, Mexico; the specific DC. in A. P. de Candolle & A. L. P. P. de epithet is based on the latter. Candolle, Prodr. 10: 68. 1846, non Lith- ospermum viridiflorum Roxb., in Carey & Lithospermum dodrantale (I. M. Johnst.) J. Wallich, Fl. Ind. 2: 4. 1824. Type: Mexico, Cohen, comb. nov. Basionym: Onosmodium without locality, 1787–1804, Sessé & dodrantale I. M. Johnst., J. Arnold Arbor. 18: Mociño 5131 (holotype: M-n.v.; probable 22. 1937. Type: Mexico. Nuevo León, Mpio. fragment of holotype: F-n.v., GH; photo of de Galeana, Peak of Cerro Potosí, “Scattered holotype: GH, MICH-n.v.). in colonies in the upper pine wood,” 21 July 1935, C. H. Mueller 2259 (holotype: GH-n.v.; This specific epithet is derived from the isotypes: F, GH-n.v., MEXU, MICH-n.v.). genus Macromeria, the largest genus of Lithospermeae endemic to the New World. Lithospermum exsertum (D. Don) J. Cohen, This species was given this epithet because of comb. nov. Basionym: Macromeria exserta D. its broad, international range, from western Don, Edinburgh New Philos. J. 13: 239. 1832. Mexico to the southwestern United States. Type: Mexico. Without locality, 1787–1804, Sessé & Moçiño s.n. (holotype: not found). Lithospermum notatum (I. M. Johnst.) J. Cohen, comb. nov. Basionym: Macrome- Lithospermum helleri (Small) J. Cohen, ria notata I. M. Johnst., J. Arnold Arbor. comb. nov. Basionym: Onosmodium hel- 35: 13. 1954. Type: Mexico. Nuevo León: leri Small, Fl. S.E. U.S. 1000, 1337. 1903. Ascent of Sierra Infernillo, ca. 15 mi. SW 2009] COHEN & DAVIS: LITHOSPERUMUM NOMENCLATURE 109

of Galeana, “Common over small areas just Mexico. Hidalgo, Sierra de Pachuca, below the peak,” 9000–10,000 ft, 16 June 10,000 ft., 4 Aug 1898, C. G. Pringle 6949 1934, C. H. Mueller & M. T. Mueller 830 (holotype: GH; isotypes: F, MEXU, PH-n.v., (holotype: GH; isotypes: F, MICH-n.v., UC-n.v., US). TEX). This species is named for Jerzy Rzedowski. Lithospermum oaxacanum (B. L. Turner) J. Whether by chance or by design, he has Cohen, comb. nov. Basionym: Onosmo- collected a great number of specimens of dium oaxacanum B. L. Turner, Phytologia Lithospermum and other New World Lithosper- 78: 53. 1995. Type: Mexico. Oaxaca, Llano meae, and these collections have been very Udadi, 4 km SE of San Andreas Lagunas, helpful in studying the variation in these groups. Pinus pseudostrobus woodland, 2380 m, 10 July 1981, A. Garcia M. 504 (holotype: Lithospermum trinervium (Lehm.) J. Cohen, MEXU). comb. nov. Basionym: Onosma trinervium Lehm., Pl. Asperif. Nucif. 2: 37. 1818. Lithospermum onosmodium J. Cohen, nom. Lasiarrhenum trinvervium (Lehm.) B. L. nov. Basionym: Onosmodium bejariense Turner Phytologia 77: 39. 1994. Type: DC. in A. P. de Candolle & A. L. P. P. de “Habitat in America Meridionali,” Humboldt Candolle, Prodr. 10: 70. 1846, non Litho- & Bonpland s.n.? (holotype: not located). spermum bejariense DC. in A. P. de Can- dolle & A. L. L. P. de Candolle, Prodr. 10: According to Turner (1994b), the type is 79. 1946. Type: U.S.A. Texas, “Bejar ad Rio from Mexico, with the following information de la Trinidad,” May 1828, J. L. Berlandier provided by Lehmann: “Habitat in America 1681 (holotype: G-DC; isotype: GH; photo Meridionali;” however, Lehmann did not of isotype: F-n.v., GH, MICH-n.v.). name the collector or provide a collection date for the specimen. Turner (1994b) states This speciﬁc epithet is derived from the that the type is possibly based upon a genus Onosmodium, the second largest genus specimen from Humboldt and Bonpland’s of Lithospermeae endemic to the New World. collection from Michoacán in 1803, and that this specimen is possibly at MEL, but no Lithospermum pinetorum (I. M. Johnst.) J. specimen by that name could be found there. Cohen, comb. nov. Basionym: Lasiarrhe- num pinetorum I. M. Johnst., J. Arnold Lithospermum turneri J. Cohen, nom. nov. Arbor. 16: 187. 1935. Perittostema pine- Basionym: Macromeria hintoniorum B. L. torum (I. M. Johnst.) I. M. Johnst. J. Turner, Phytologia 77: 398. 1994, non Arnold Arbor. 35: 30. 1954. Type: Mexico. Lithospermum hintoniorum B. L. Turner, Oaxaca(?), September 1841(?), Ghies- Sida 20: 501. 2002. Type: Mexico. Guer- breght 311 (holotype: P). rero, Teotepec, ca. 12˚ 27′ N, 100˚ 10′ W, pine forest, 2750 m, 16 July 1939, G. B. Lithospermum rosei (I. M. Johnst.) J. Cohen, Hinton et al. 14439 (holotype: NY; iso- comb. nov. Basionym: Nomosa rosei I. M. types: F, NY, US). Johnst., J. Arnold. Arbor. 35: 25. 1954. Type: Mexico, in Sierra Madre near the southern This species is named for Billie L. Turner. border of the state of Durango, Aug 16 1897, His work on Lithospermum and New World J. N. Rose 2360 (holotype: GH). Lithospermeae has been helpful in under- standing these taxa. Lithospermum rzedowskii J. Cohen, nom. nov. Basionym: Macromeria pringlei Lithospermum unicum (J. F. Macbr.) J. Greenm., Proc. Amer. Acad. Arts 34: 570. Cohen, comb. nov. Basionym: Onosmo- 1899, non Lithospermum pringlei I.M.. dium unicum J. F. Macbr., Contr. Gray Johnst., Contr. Gray Herb. 70: 22. 1924. Herb. n.s. 49: 21. 1917. Type: Mexico. San Onosmodium pringlei (Greenm.) J. F. Macbr. Luis Potosí, Alvarez, 13 – 23 July 1904, E. Contr. Gray Herb. n.s. 49: 20. 1917. Type: Palmer 185. (holotype: GH; isotype: F). 110 BRITTONIA [VOL 61

Acknowledgments Edgar, R. C. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. This research was supported by grants from Nucleic Acids Research 32: 1792–1797. the American Society of Plant Taxonomists, Farris, J. S., V. A. Albert, M. Kallersjo, D. Lipscomb the Harold Moore Jr. Funds, Cornell Univer- & A. G. Kluge. 1996. Parsimony jackknifing out- sity’s chapter of Sigma Xi, the Cornell performs neighbor-joining. Cladistics 12: 99–124. Goloboff, P. 1999. Analyzing large data sets in reason- University Graduate School, and Cornell able times: Solutions for composite optima. Cladistics University’s Latin American Studies Pro- 15: 415–428. gram. Without Shannon C. K. Straub, María ———, J. S. Farris & K. C. Nixon. 2000. TNT (Tree Hilda Flores Olvera, Helga Ochoterena, analysis using New Technology) Ver. 1. Published by authors, Tucuman, Argentina. Socorro Gonzalez, Paty Ledesma Hernandez, Hall, T. A. 1999. BioEdit: a user-friendly biological Fernando Alzate, Lucía Vázquez, Janelle M. sequence alignment editor and analysis program for Burke, and Caroline D. Kellogg, it would Windows 95/98/NT. Nucleic Acids Symposium Se- have been much more difficult to locate and ries. 41:95–98. collect these plants. In addition, the National Johnston, I. M. 1935. Studies in the Boraginaceae, XI. Journal of the Arnold Arboretum. 16:145–205. Botanic Garden of Belgium, Missouri Botan- ———. 1952. Studies in the Boraginaceae, XXIII. A ical Garden, and Cornell Plantations provided survey of the genus Lithospemum. With three plates. leaf material, and the Royal Botanic Garden, Journal of the Arnold Arboretum 33:299–366. Kew supplied one DNA isolation. Thanks to ———. 1953a. Studies in the Boraginaceae, XXIV. A. Three genera segregated from Lithospermum. B. Caroline D. Kellogg for her support and Supplementary notes on Lithospermum. Journal of comments on this manuscript. James L. the Arnold Arboretum 34:1–16. Reveal gave invaluable advice concerning ———. 1953b. Studies in the Boraginaceae, XXV. A the nomenclature of Lithospermum as well revaluation of some genera of the Lithospermeae. as helpful comments on the manuscript. Journal of the Arnold Arboretum 34: 258–300. ———. 1954a. Studies in the Boraginaceae, XXVI. James S. Miller’s specific comments and Further revaluations of the genera of the Lithosper- advice allowed this manuscript to be ready meae. Journal of the Arnold Arboretum 35:1–81. for submission in a timely manner. Shannon ———. 1954b. Studies in the Boraginaceae, XXVII. C. K. Straub gave helpful comments on an Some general observations concerning the Lithosper- meae. Journal of the Arnold Arboretum 35: 158–166. early draft, and thanks to Monica Geber, M. Langstrom, E. & M. W. Chase. 2002. Tribes of Alejandra Gandolfo, Michael Simpson, and Boraginoideae (Boraginaceae) and placement of Lawrence Kelly for their detailed reading and Antiphytum, Echiochilon, Ogastemma and Sericos- comments and corrections. toma: A phylogenetic analysis based on atpB plastid DNA sequence data. Plant Systematics and Evolution 234: 137–153. Literature Cited Lehmann, J. G. C. 1818. Plantae e Familia Asperifo- Barfuss, H. J., R. Samuel, W. Till & T. F. Stuessy. liarum Nuciferae. Dümmler, Berlin. 2005. Phylogenetic relationships in subfamily Till- Muhlenberg, H. 1813. Catalogus Plantarum Americae andsioideae (Bromeliaceae) based on DNA sequence Septentrionalis. William Hamilton, Lancaster. data from seven plastid regions. American Journal of Nixon, K. C. 1999. Parsimony ratchet, a new method for Botany 92: 337–351. rapid parsimony analysis. Cladistics 15: 407–414. Boyd, A. E. 2003. Phylogenetic relationships and corolla Ochoterena, H. In press. Homology in coding and non- size evolution among Macromeria. Systematic Bota- coding DNA sequences: a parsimony perspective. ny 28: 118–129. Plant Systematics and Evolution. Candolle, A. P. de. 1846. Boraginaceaede Candolle. In: Ralston, B. 1993. Unpublished doctoral dissertation. Uni- A. P. de Candolle and A. L. P. P. de Candolle (eds.), versity of Northern Arizona Library, Flagstaff, Arizona. Prodromus Systematis Naturalis Regni Vegetabilis. Rechinger, K. H. 1967. Boraginaceae. In: H. Riedl (ed.), 10: 1–178. Fortin, Masson et sociorum, Paris, France. Flora Iranica. 48: 1–281. Akademische Druk- u. Cohen, J. 2007. Lithospermum L. systematics: Utilizing Verlagsanstalt, Graz. multiple cpDNA regions and morphological charac- Shaw, J., E. B. Lickey, J. T. Beck, S. B. Farmer, W. ters. Botany 2007. [Abstract] Liu, J. Miller, K. C. Siripun, C. T. Winder, E. E. Cuénoud, P., V. Savolainen, L. W. Chatrou, M. Powell, Schilling & R. L. Small. 2005. The tortoise and the R. J. Grayer & M. W. Chase. 2002. Molecular hare II: Relative utility of 21 noncoding chloroplast phylogenetics of Caryophyllales based on nuclear DNA sequences for phylogenetic analysis. American 18S rDNA and plastid rbcL, atpB, and matK DNA Journal of Botany 92: 142–166. sequences. American Journal of Botany 89: 132–144. ———, ———, E. E. Schilling & R. L. Small. 2007. Doyle, J. J. & J. L. Doyle. 1990. Isolation of plant DNA Comparison of whole chloroplast genome sequences from fresh tissue. Focus. 12: 13–15. to choose noncoding regions for phylogenetic studies 2009] COHEN & DAVIS: LITHOSPERUMUM NOMENCLATURE 111

in angiosperms: the tortoise and the hare III. 3 48–58° for 1 minute American Journal of Botany 94: 275–288. 4 72° for 2 minutes Simmons, M. P. & H. Ochoterena. 2000. Gaps as 5 repeat steps 2–4 35X characters in sequence-based phylogenetic analyses. Systematic Biology 49: 369–381. 6 72° for 10 minutes Thomas, D. C., M. Weigand & H. H. Hilger. 2008. 7 hold 4° Phylogeny and systematics of Lithodora (Boragina- ceae-Lithospermeae) and its afﬁnities to the monotypic genera Mairetis, Halacsya and Paramoltkia Program for the rpl16 intron based on ITS1 and trnLUAA - sequence data and morphology. Taxon 57: 79–97. 1 80° for 5 minutes Turner, B. 1994a. Synoptical study of the genus Macro- 2 94° for 30 seconds meria (Boraginaceae). Phytologia 77: 393–407. 3 59°– .5° per replication for 30 seconds ———. 1994b. Revisionary study of Lasiarrhenum (Boraginaceae). Phytologia 77: 38–44. 4 65° for 2 minutes ———. 1995. Synopsis of the genus Onosmodium 5 repeat steps 2–4 10X (Boraginaceae). Phytologia 78: 39–60. 6 94° for 25 seconds Zhu, G., H. Riedl & R. V. Kamelin. 1995. Boragina- 7 54° for 25 seconds ceae. Pp. 329–427. In: Wu, Z. Y. & P. H. Raven (eds.), Flora of China. Vol. 16 (Gentianaceae through 8 65° for 2 minutes Boraginaceae). Science Press, Beijing, and Missouri 9 repeat steps 6–8 20X Botanical Garden Press, St. Louis. 10 hold 4°

Appendix 2. TNT analysis commands Appendix 1. PCR ampliﬁcation programs rs 0; hold 10000; ratchet: iter 1000 upfactor Program for all regions except the rpl16 10 downfact 10; ratchet:; drift: iter 100; drift:; intron sectsch:; xmult:; tfuse: rounds 100; tfuse:; 1 80° for 5 minutes xmult: hits 10 ratchet 1000 drift 1000 css rss 2 94° for 30 seconds fuse 100; xmult:; mult 1; xmult; quit;