wow? uy. “~~ PHYTOLOGIA An international journal to expedite plant systematic, phytogeographical and ecological publication Vol. 71 August 1991 HMONG ENON D CONTENTNWEW YORK BOTANICAL GARDEN Be aH R., A new combination in Panicum Poaceae subgenus Dichanthelium Ve Np ES ON a EAE OMA ASN ALON ALIEN Wo SST OBES 85 y KARTESZ, J.T. & K.N. GANDHI, Nomenclatural notes for the North ITI CHI SIGMA PAAR NOP ALA LD Ofer seth ey ah catvcle td bi de valoncalie bug 87 /SHEVOCK, J.R. & G.A. ALLEN, A new variety of Erythronium citrinum | (Liliaceae) from the Scott Mountains of northwest California ....... 101 SPENCER, M.A. & L.B. SMITH, A new species of Pitcairnia (BEORTCUBCEAG EWU MUEIMONACAC). 2b 0.1. vA. kev talel da lap bovercbbaeboenlaetduates 104 YNESOM, G.L., Union of Bradburia with Chrysopsis (Asteraceae: Astereae), with a phylogenetic hypothesis for Chrysopsis ........... 109 ANESOM, G.L., Transfer of Aster kingii to Tonestus (Asteraceae: PUOCE ECA Ere Ne ee nace day e Vi t ea VGMUUN RE A Albans dd § 122 NESOM, G.L., Tomentaurum (Asteraceae: Pra a new genus of goldenaster fomninahua? Mexico! cee OE MCW a 128 NESOM, G.L., Transfer of Heterotheca bartlettii to Osbertia PASTOR CEIET ARPETERE SUNN GMa ha ibid iid ects atouh dole da adn kop hectyn occa yaa: 132 -NESOM, G.L., A phylogenetic hypothesis for the goldenasters (MAREPACHOGVABIELCAG) Ae Ulaiiy Chios RON SS (AL. ciddcas igs vubiigse Yomwasd dagen’ 136 __-» Contents continued on the inside cover. Published by Michael J. Warnock 185 Westridge Drive Huntsville, Texas 77340 U.S.A. PHYTOLOGIA is printed on acid free paper. . NESOM, G.L., Two new species of Archibaccharis (Asteraceae: Astereae) from México, with a reevaluation of sectional groupings in the BOMUGE il are ee adedeed sah sen due cleanse halgt be Qala e «a #4 pelt ait a ean el ae 152 -JONES, G.D. & S.D. JONES, Sarcostemma clausum, series Clausa (Asclepiadaceae), new to! Texas... ...)).0).5,..1.4 Mee ea ang eeeen cee 160 Books received PHYTOLOGIA (ISSN 00319430) is published monthly with two volumes per year by Michael J. Warnock, 185 Westridge Drive, Huntsville, TX 77340. Second Class postage at Huntsville, TX. Copyright ©1991 by PHYTOLOGIA. Annual domestic individual subscription (12 issues): $36.00. Annual domestic institutional subscription (12 issues): $40.00. Foreign and/or airmail postage extra. Single copy sales: Current issue and back issues volume 67 to present, $3.50; Back issues (previous to volume 67), $3.00 (add $.50 per copy postage and handling US [$1.00 per copy foreign]). Back issue sales by volume: $17.00 per volume 42-66 (not all available as complete volumes); $21.00 per volume 67-present; add $2.00 per volume postage US ($4.00 per volume foreign). POSTMASTER: Send address changes to Phytologia, 185 Westridge Drive, Huntsville, TX 77340. Phytologia (August 1991) 71(2):85-86. A NEW COMBINATION IN PANICUM (POACEAE) SUBGENUS DICHANTHELIUM Ray Angelo New England Botanical Club, 22 Divinity Avenue, Cambridge, Massachusetts 02138 U.S.A. ABSTRACT The new combination Panicum sphaerocarpon Elliott var. iso- phyllum (Scribner) R. Angelo is proposed. KEY WORDS: Panicum, Poaceae, Dichanthelium. In Gould & Clark (1978) the taxon that was generally known as Panicum polyanthes J.A. Schultes was treated as a variety under the name Dichanthe- lium sphaerocarpon (Elliott) Gould var. isophyllum (Scribner) Gould & Clark. According to' Zuloaga (1987) and Lelong (1984) there is insufficient evidence at present to warrant elevating subgenus Dichanthelium to the generic level. If Panicum polyanthes is still to be considered as a variety of Panicum sphae- rocarpon, then a new combination will be required. The new combination Panicum sphaerocarpon Elliott var. polyanthes (Schultes) A.S. Sherif was pro- posed (Sherif 1983). However, under Article 60 of the International Code of Botanical Nomenclature (1988), a name does not have priority outside of its own rank. According to Gould & Clark (1978), “The name isophyllum Scribner is the earliest varietal name to be applied to this taxon and must be retained even though the species basionym, microcarpon, is illegitimate.” In light of this the correct new combination should be: Panicum sphaerocarpon Elliott var. isophyllum (Scribner) R. Angelo, comb. nov. BASIONYM: Panicum microcarpon Muhlenberg var. tso- phyllum Scribner, Bulletin of the Agricultural Experiment Station of the University of Tennessee. 7:51. Figure 54. 1894. Dichanthelium sphaero- carpon Elliott var. isophyllum (Scribner) Gould & Clark, Ann. Missouri Bot. Gard. 65:1105. 1978. TYPE: UNITED STATES. Tennessee: Blount County, Alleghany Springs, E.E. Gayle, August 1890 (HOLO- TYPE: US, Sheet #743578, not seen). 85 86 PHYTOL 0 GIA volume 71(2):85-86 August 1991 Panicum multiflorum Elliott, not Poiret, Sketch Bot. S. Carolina. 1:122. 1816. TYPE: UNITED STATES. South Carolina or Georgia: “Hab. in umbrosis. Flor. Mai. June” (HOLOTYPE: CHARL, not seen, photo in GH). Panicum microcarpon Muhlenberg, not Muhlenberg ez Elliott, Descr. Gram. 111. 1817. TYPE: UNITED STATES. “40, c Jul. 12. e Cherokee.” (HOLOTYPE: PH, not seen). Panicum polyanthes J.A. Schultes, Mantissa 2:257. 1824. Dichanthe- lium sphaerocarpon Elliott var. polyanthes (J.A. Schultes) Gould, Brittonia 26:60. 1974. Panicum sphaerocarpon Elliott var. polyan- thes (J.A. Schultes) A.S. Sherif, Sida 10:191. 1983. Dichanthe- lium polyanthes (J.A. Schultes) Mohlenbrock, Erigenia 6:26. 1985. TYPE: Based upon P. multiflorum Elliott (December 1816), not Poiret (June 1816). Panicum firmandum Steudel, Syn. Pl. Glumac. 1:418. 1855. TYPE: UNITED STATES. North Carolina: M.A. Curtis (HOLOTYPE: P, not seen). ACKNOWLEDGMENTS I thank Dr. Elizabeth A. Shaw and Dr. David E. Boufford for their assis- tance and advice. I thank Deborah Bell for assisting in the location of a type specimen. LITERATURE CITED Gould, F.W. & C.A. Clark. 1978. Dichanthelium (Poaceae) in the United States and Canada. Ann. Missouri Bot. Gard. 65:1088-1132. Lelong, M.G. 1984. New combinations for Panicum subgenus Panicum and subgenus Dichanthelium (Poaceae) of the southeastern United States. Brittonia 36:262-273. Sherif, A.S. 1983. Panicum sphaerocarpon Ell. var. polyanthes (Schultes) A.S. Sherif (Poaceae) comb. nov. Sida 10:191. Zuloaga, F.O. 1987. Systematics of New World species of Panicum (Poaceae: Paniceae). pp. 287-306 in T.R. Soderstrom, et. al. (editors), Grass Sys- tematics and Evolution. Smithsonian Press, Washington, D.C. Phytologia (August 1991) 71(2):87-100. NOMENCLATURAL NOTES FOR THE NORTH AMERICAN FLORA. VII John T. Kartesz & Kancheepuram N. Gandhi The North Carolina Botanical Garden, Department of Biology, Coker Hall, University of North Carolina, Chapel Hill, North Carolina 27599-3280, U.S.A. ABSTRACT The authorship of the following names is discussed: Horkelia fusca Lindl. var. parviflora (Nutt. ez Torr. & Gray) Wawra, Lotus purshianus Clements & Clements, Madia ezigua (Smith) A. Gray, Sagittaria latifo- lia Willd. var. obtusa (Engelm.) Wiegand, Sagittaria longiloba Engelm. ex J.G. Sm., and Thunbergia grandiflora Roxb. The parenthetical scien- tific names for poison oak ( Tozicodendron pubescens P. Mill.), for Span- ish clover (Lotus unifoliolatus [Hook.| Benth.), and for staghorn sumac (Rhus hirta [L.] Sudworth) are viewed to be correct. Three new com- binations are proposed: Chamaecrista nictitans (L.) Moench var. patellaria (DC. ez Colladon) Kartesz & Gandhi; Lotus unifoliola- tus (Hook.) Benth. var. helleri (Britt.) Kartesz & Gandhi; and Ruel- | lia caroliniensis (Walt.) Steud. var. cinerascens (Fern.) Kartesz & Gandhi. KEY WORDS: Floristics, nomenclature, Acanthaceae, Alismat- aceae, Anacardiaceae, Asteraceae, Fabaceae, Rosaceae, Spargani- aceae, Chamaecrista, Horkelia, Lotus, Madia, Rhus, Ruellia, Sagit- taria, Thunbergia, and Tozicodendron. INTRODUCTION Continuing with the “NOMENCLATURAL NOTES FOR THE NORTH AMERICAN FLORA” (Kartesz & Gandhi 1989, 1990a, 1990b, 1990c, 1991a, 1991b), a seventh note in the series is presented here towards advancing our understanding of North American plant names. 87 88 PHY TOLOGIA volume 71(2):87-100 August 1991 ACANTHACEAE Ruellia caroliniensis var. cinerascens Under the binomial Ruellia ciliosa Pursh, Fernald described var. cineras- cens. Since we treat R. ciliosa as a synonym of R. caroliniensis (Walt.) Steud., but still recognize Fernald’s variety, the following new combination is proposed. Ruellia caroliniensis (Walt.) Steud. var. cinerascens (Fern.) Kartesz & Gandhi, comb. nov. BASIONYM: Ruellia ciliosa Pursh var. cinerascens Fern., Rhodora 47:48. 1945. TYPE: U.S.A. Florida: Walton Co., Crestview, 22 Jul 1899, Curtis 6489 (US). Thunbergia grandiflora Thunbergia grandiflora, a native of India, has become naturalized in Florida (pers. comm. from Dr. R.C. Wunderlin). The authorship of T. grandiflora has been attributed to “Roxb.” (Jackson 1895; Wasshausen in Nicolson 1991), or to “(Roxb. ez Rottl.) Roxb.” (Mathew, Fl. Carnatic, Tamil Nadu, vol. 3. 1983, fide Nicolson (US), pers. comm.; Barker 1986), or to “(Roxb. ez Rottl.) Lodd.” (Howard 1989). Mathew, Barker, and Howard believed that this species was based on Flemingia grandiflora Roxb. ez Rottl., the type species for the genus Flemingia Roxb. ez Rottl. According to the ICBN (Greuter et af 1988:219), the legume genus name Flemingia Roxb. ez Ait. f. (Hort. Kew., ed. 2, 4:349. 1812) has been conserved over the homonym Flemingia Roxb. ez Rottler (Ges. Naturf. Freunde Berlin Neue Schriften 4:202. 1803). Furthermore, the ICBN indicated that F. grandiflora is an illegitimate name. In the protologue of F. grandiflora, we found the name T. fragrans Roxb., cited “as a synonym” in a footnote, which infers that F. grandiflora was a renaming of T. fragrans; i.e., the name F’. grand:florais superfluous and illegitimate. However, Dr. Nicolson informed us that footnotes in Rottler’s article were not by Rottler, but rather by Willdenow (editor of Rottler’s article). Hence, Nicolson concluded that the name F. grandiflora is legitimate and that ICBN erred in its treatment of Rottler’s binomial. With this established, we discuss the authorship of the binomial T. grandiflora. Roxburgh’s 1814 work (Hort. Beng.), in which the binomial Thunbergia grandiflora was proposed, does not contain valid descriptions. For T. grandi- flora, Roxburgh did not refer to Rottler’s published description of F. grandi- flora; hence, the binomial T. grandiflora was invalid. Loddiges (Bot. Cab., vol. 4(3): t. 324. Jan 1820) independently used the name T. grandiflora and pro- vided an illustration and generalized remarks. Except for his comment that T. grandiflora was odorless, Loddiges’ protologue did not provide an illustration with analysis and did not meet the requirements of ICBN Art. 44.1. He neither referred to Roxburgh nor cited any other reference. Since Loddiges stated that the plant was native to India, it is possible that he was aware of Roxburgh’s Kartesz & Gandhi: Nomenclatural notes on North American flora VII 89 1814 usage of the name, but this speculation can not be verified. Subsequently, Roxburgh, again without referencing a botanical work, provided a description (in Ker-Gawler, Bot. Reg. 6: t. 495. Nov 1820) and thus validated the name. Although Wasshausen (in Nicolson 1991:18) attributed the name T. gran- diflora to Roxburgh, his treatment was followed by Nicolson’s (editor and chief author of the publication) remarks that the nomenclature of this binomial re- mains unsettled. Until or unless additional nomenclatural evidence can be provided, we are compelled to accept Roxburgh alone as the author of the binomial dating from Nov 1820. Thunbergia grandiflora Roxb. {Hort. Beng. 45. 1814, nom. nud.] in Ker- Gawler, Bot. Reg. 6: t. 495. Nov 1820. TYPE: Bot. Reg. 6: t. 495. Flemingia grandiflora Roxb. ex Rottl., Ges. Naturf. Freunde Berlin Neue Schriften 4:202. 1803. TYPE: INDIA. Tamil Nadu, Madras, Marmelon(g), 28 Nov 1799, Berry s.n. (not traced; fide Nicolson). ALISMATACEAE Sagittarza latifolia var. obtusa For his new varietal combination Sagittaria latifolia Willd. var. obtusa, Wiegand gave a direct and full reference to S. obtusa Muhl. ez Willd. and also cited a reference to S. variabilis Engelm. var. obtusa Engelm. in A. Gray (1856) as asynonym. Engelmann, who provided the treatment for the suborder Alismeae (including Alisma, Echinodorus, and Sagittaria) in the second edition of Gray’s Manual (fide Gray’s preface; p. xii), also based his variety on S. obtusa. Unfortunately, Sagittaria obtusa Muhl. ez Willd. is a later homonym of S. obtusa Thunb. Although the name S. obtusa Thunb. is superfluous, and thus illegitimate (Thunberg cited S. obtusifolia L. as a synonym), S. obtusa Muhl. ez Willd. must still be rejected as a later homonym (ICBN Art. 64.1). Therefore, “Muhl. ez Willd.” must not be cited as a parenthetical author for either S. latifolia var. obtusa or for S. variabilis var. obtusa. Furthermore, Engelmann’s new combination S. variabilis var. obtusa must be treated as a nomen novum, with its priority from 1856 (ICBN Art. 72.2, Note 1), and Engelmann must be placed as the parenthetical author for the name S. latzfolia var. obtusa. Sagittaria latifolia Willd. var. obtusa (Engelm.) Wiegand, Rhodora 27:186. 1925. BASIONYM: Sagittaria variabilis Engelm. var. obtusa Engelm. in A. Gray, Man. Bot., ed. 2. 439. 1856. Sagittaria obtusa Muhl. ez Willd., Sp. Pl. 4:409. 1805, non Thunb., 1784. TYPE: U.S.A. Muhlenberg s.n. (PH; fiche !). 90 PAY TeOLO GLA volume 71(2):87-100 August 1991 Sagittaria longiloba The name Sagittaria longiloba has been generally attributed to Engelm. ez Torrey (Smith 1895; Bogin 1955; Soil Conservation Service 1982) or to Engelmann (Kaul 1986). Torrey (1859:212), in a footnote, indicated that he had a Sagittarza specimen (sent by Bigelow) and considered it to be the same as “S. longiloba Engelm.” According to Torrey, Engelmann had provisionally proposed (in manuscript) the name S. longiloba for a west Texas Sagittarza and regarded S. longiloba to be closely related to S. sumpler Torr. Although Torrey briefly described S. longiloba (“The leaves, however, are sagittate, with very long, narrow, and widely diverging lobes, a state in which we have never seen S. sumplez”), he remarked that these characters were insufficient for separating it from S. simplez. Since Torrey did not accept the name S. longiloba, he did not validly publish it, and the name must not be attributed to him (ICBN Art. 34.1a). Subsequently, Smith (1895) accepted the name S. longiloba, attributed it to Engelmann ez Torrey, and provided an adequate description for it. He thereby validated the name, with its priority beginning from 1895. Although Bogin (1955) made a similar analysis, he failed to correct the authorship. The proper authorship should be as follows. Sagittaria longiloba Engelm. ez J.G. Sm., Annual Rep. Missouri Bot. Gard. 6:42. 1895. ANACARDIACEAE Rhus hirta Although Britton (Bull. Torrey Bot. Club 18:269. 1891) indicated that Datisca hirta L. (published in 1753) was the earliest name for the staghorn sumac, he believed that transfer of D. hirta to the genus Rhus would create a later homonym of Harvey’s R. hirta. We quote from Britton’s article: “Al- though hirtais thus the oldest specific name associated with the plant, we are, I think, debarred from using it by the publication of Rhus hirta Harv., as a synonym by Engler (in C. DC., Monogr. Phan. 4:425. 1883), where this is referred to R. tridentata, Sond.” Consequently, he accepted the name R. ty- phina L. (published in 1756), for the staghorn sumac. Britton’s nomenclatural understanding of the sumac in question was probably appropriate during his time, but inappropriate under the present Code (Greuter et al. 1988). Har- vey’s manuscript’s name R. hirta, which first appeared in Engler’s treatment (as a synonym of R. tridentata, an African sumac), was indeed effectively pub- lished, but never validly published. Hence, Harvey’s manuscript name has no nomenclatural standing (ICBN Art. 34.1). Therefore, contrary to Britton’s Kartesz & Gandhi: Nomenclatural notes on North American flora VII 91 belief, the transfer of D. hirta to the genus Rhus would not have created a later homonym of R. hirta Harvey ez Engler, pro syn. Sudworth (Bull. Torrey Bot. Club 19:80-81. 1892) argued against Brit- ton’s assertion and concluded that Datisca hirta could be transferred to the genus Rhus. Accordingly, he made the new combination: Rhus hirta (L.) Sud- worth. In a rejoinder to Sudworth’s note [published on the same page that the new combination (R. hirta) was published], Britton rejected Sudworth’s new combination, but subsequently (in Britton & Brown 1913) accepted the new combination. Sudworth (1927:180) used the name 2. hirta and remarked about the usage of the binomial R. typhina by others. Consistent with Sud- worth’s remark, many subsequent workers, such as Barkley (1937), Gleason (1952), Radford et al. (1968), Little (1979), McGregor (1986), Voss (1985), and Wofford (1989) accepted the name R. typhina. In addition to the name R. hirta (L.) Sudworth, Barkley (1937:326) also mentioned R. hirta “L. ex Small” as a synonym of R. typhina. On verification, we found that Small (1903) had indeed used the name R. hirta and attributed it to Linnaeus. We believe that Small certainly should have been aware of Britton’s and Sudworth’s notes on the sumac name under consideration and thereafter, should have chosen to fol- low Sudworth. However, he erred on the authorship in attributing the name to Linnaeus. Likewise, Barkley also erred by attributing the name R. hirta to “L. ‘ex omall.” In order to legitimately use the name Rhus typhina, either it must be con- served over D. hirta, or the latter name must be rejected by the Nomenclatural Committee of the ICBN. Dr. Jim Reveal (MARY) has been pursuing the re- tention of the name R. typhina. Until or unless the committee decides in favor of Reveal’s proposal (a decision we would not endorse), we accept the name R. hirta for the North American flora, and provide the following nomenclatural details. Rhus hirta (L.) Sudworth, Bull. Torrey Bot. Club 19:81. 1892; Small, Fl. S.E. U.S. 1334. 1903. BASIONYM: Datisca hirta L., Sp. Pl. 1037. 1753. Rhus typhina L., Cent. Pl. 2:14. 1756. Tozicodendron pubescens Historically, the following names have been associated with the poison-oak of eastern North America: Rhus tozicarium Salisb., R. toricodendron L., Toz- icodendron pubescens P. Mill., T. quercifoltum (Michx.) E. Greene, T. tozicar- ium (Salisb.) Gillis, T. tozicodendron (L.) Britt., and T. vulgare P. Mill. After reviewing the nomenclature for this species, it is apparent that the earliest name for this complex is R. tozicodendron L. Unfortunately, this Linnaean epithet can not be transferred to Tozicodendron, since it would create a tau- tonym [T. tozicodendron (L.) Britt., an illegitimate name (ICBN Art. 23.4)]. 92 PHYTOLOGIA volume 71(2):87-100 August 1991 Gillis (1971) presented an informative analysis on the nomenclature of the poison-oak. With reference to the name Tozicodendron vulgare, Gillis stated (p. 413): “Tozicodendron vulgare is too inaccurately described to be Eastern poison-oak, even though indirectly linked with that species in the literature ..., this binomial must be rejected under (ICBN) Art. 69 of the code.” Fur- thermore, he also rejected the name T. pubescens remarking: “His (Miller’s) description fits T. tozicarium which has pubescent leaves and fruits, but not to the exclusion of all other taxa.” (Both rejections are not permissible under the current code.) Consequently, Gillis chose the next earliest name (Rhus tozi- carium Salisb.), transferred it to Toricodendron, and made the combination: T. tozicarium (Salisb.) Gillis. Unfortunately, the basionym R. tozicarium is superfluous, since Salisbury cited R. tozicodendron in synonymy. With reference to Gillis’ new combination, we analyzed his treatment and concluded the following: Gillis did not exclude the Linnaean type of Rhus tozi- codendron L. Gillis presumed that R. torzicarium had priority and was unaware that it was an illegitimate name and that ICBN Art. 45, Note 2 (pertaining to priority of names) and Art. 49 (pertaining to parenthetical authorship) applied to his new combination (Lanjouw et al. 1966). We apply ICBN Art. 72.2, Note 1, and recognize Gillis’ “new combination” as a nomen novum (i.e., based on the same type as R. tozicodendron L. and R. tozicarium Salisb., nomen super- fluum). In other words, Tozicodendron tozicarium Gillis is a legitimate name, with priority from 1971. Barkley (1937) recognized the name Tozicodendron quercifolium for the plant in question. (The name 7. quercifolium was based on Rhus toricodendron var. quercifolum Michx.). He rejected the name T. pubescens and placed it (in parts) as synonyms of both T. quercifolium and T. radicans (L.) Kuntze. Barkley’s rejection of the name T. pubescens was probably appropriate for the Code of his time, but inappropriate under today’s Code. For the legitimate use of T. pubescens, Reveal (Taxon 40:334. 1991) designated a neotype. In his article, Reveal was correct in his assessment that T. tozicarium Gillis is a “new name” but erred in considering it to be “nom. illeg. superfl.” Tozicodendron pubescens P. Mill., Gard. Dict., ed. 8, Art. Tozricodendron, no. 2. 1768. NEOTYPE (vide Reveal, Taxon 40:334. 1991): Uvedale collection, Sloane herbarium, H.S. 315:86 (BM-SL). Rhus tozicodendron L., Sp. Pl. 1:266. 1753. Rhus tozicarium Salisb., Prodr. 170. 1796. Tozicodendron tozicodendron (L.) Britt. in Britt. & Brown, Ill. Fl. N. U.S., ed. 2. 2:484. 1913, nom. illegit. Tozico- dendron tozicarium Gillis, Rhodora 73:402. 1971. Tozicodendron quercifolium (Michx.) E. Greene, Leafl. Bot. Observ. Crit. 12127. 1905. Kartesz & Gandhi: Nomenclatural notes on North American flora VII 93 ASTERACEAE Madia ezigua In the second note of his article, Gray (Proc. Amer. Acad. Arts 8:372-412. 1872) included a total of 701 numbers, of which the nos. 284-288 pertained to the genus Madia. For no. 288 (p. 391), Gray stated: “Madia ( Harpecar- pus) filipes. Harpecarpus madrioides Nutt. This and the related M. ezigua (Sclerocarpus exiguus Smith) form a marked section of the genus expanded.” Cronquist (1955), Kartesz & Kartesz (1980), and Dorn (1988) accepted Gray as the combining author of the name Madia ezigua, but the Soil Conser- vation Service (1982) recognized Greene (Erythea 1:90. 1893) as the combining author. From Gray’s treatment of the name, it is clear that he validly and effectively made the new combination. Any rejection of Gray’s usage of the name, as an incidental mention, and thus invalid (Voss et al. 1983; ICBN Art. 34.1c), is now incorrect, since the preceding Article was dropped in the Berlin Congress (Greuter et al. 1988). Hence, we concur with Cronquist (1955), Kartesz & Kartesz (1980), and Dorn (1988) that Gray is the combining author of the combination. Madia exigua (Smith) A. Gray, Proc. Amer. Acad. Arts 8:391. 1872. BA- SIONYM: Sclerocarpus eziguus Smith in Rees, Cycl. 31. 1815. FABACEAE Chamaecrista nictitans Irwin & Barneby (1982) proposed (among several others) a new combi- nation: Chamaecrista nictitans (L.) Moench var. ramosa (Vogel) Irwin & Barneby, which was based on Cassia patellaria DC. ez Colladon var. ramosa Vogel. The authors also cited Cassia patellaria as a synonym. When Vogel proposed his var. ramosa, he automatically created the au- tonym: Cassia patellaria DC. ex Colladon var. patellaria. It was most likely that Irwin & Barneby followed the Leningrad Congress, which stated that au- tonyms were not to be taken into consideration for purposes of priority (Stafleu 1978; ICBN Art. 26.2), whereas in the Sydney Congress that principle was re- versed, with autonyms having priority over the names that established them (Voss et al. 1983; ICBN Art. 57.3). A new combination is therefore needed and is proposed below: Chamaecrista nictitans (L.) Moench var. patellaria (DC. ez Colladon) Kartesz & Gandhi, comb. nov. BASIONYM: Cassia patellaria DC. ez Colladon var. patellaria, automatically established by Cassia patellaria var. ramosa Vogel, 1837. Cassia patellaria var. ramosa Vogel, Syn. Gen. Cass. 66. 1837. Chamaecrista nictitans (L.) Moench var. ramosa ( Vogel) Irwin & Barneby, Mem. New York Bot. Garden 35:817. 1982. 94 PRY PO O'GTA volume 71(2):87-100 August 1991 Lotus unifoliolatus Lotus purshianus has historically been the scientific name used for the Spanish clover (Hitchcock in Hitchcock & Cronquist 1961; Isely 1981; Barneby 1989); however, Dorn (1988:171) used the name L. unifoliolatus Benth. Hitch- cock as well as Isely attributed the name L. purshianus to “(Benth.) Clements & Clements” and cited the names L. sericeus Pursh (a later homonym), Trigonella americana Nutt., Hosackia purshiana Benth., H. unifoliolata Benth., and L. americanus (Nutt.) Bisch. (a later homonym) in synonymy. Barneby attributed the name L. purshianus to “(Benth. ez Lindl.) F. & E. Clements ex Ottley” but did not include H. unzfoliolata in synonymy. Under the name L. untfoliolatus, Dorn (1988:303) cited L. purshianus as a synonym and men- tioned the following: “Lotus purshianus - Hosackia purshiana, on which name was based, is illegitimate. The epithet americanus should have been taken up by Bentham.” Since Dorn’s remarks were brief and since the most recent work of Barneby differed from that of Dorn, we decided to investigate the name L. purshianus. The earliest binomial in this complex appears to be Lotus sericeus Pursh, and hitherto, this binomial has been considered to be a later homonym of L. sericeus DC. In our study, we found that prior to de Candolle’s usage of the name L. sericeus, Moench used the binomial L. sericeus. Although Moench’s binomial was superfluous (he cited L. creticus L. as a synonym), and thus illegitimate, it still rendered both de Candolle’s and Pursh’s binomials to be illegitimate (ICBN Art. 64.1, Note 1). Regarding the authorship of Hosackia purshiana, the name was validly published in Lindley’s (Bot. Reg. 15: t. 1257. 1829) work. Possibly for this reason, Barneby (1989) considered Lindley to be the validating author. However, Lindley clearly indicated that Bentham contributed the description of this species; hence, Bentham is the author of the binomial H. purshiana. The next question is whether to consider the name L. purshianus as a new combination (as did Hitchcock, Isely, and Barneby), as a nomen novum, or as a species novum. Clements & Clements (1914:183) used the name Lotus purshianus and pro- vided key characters, but did not provide a citation. In their preface, they referenced Britton & Brown (among others), but their key characters were not copied from Britton & Brown (1897, 1913). Neither Hitchcock (in Hitchcock & Cronquist 1961) nor Barneby (1989) discussed the nomenclature of Lotus purshianus, but Isely (1981:243-244) did provide a discussion. Isely contended that Clements & Clements’ preface ref- erence to Britton & Brown (who included Hosackia purshiana as a synonym of L. americanus) should be considered as an indirect reference to Bentham’s H. purshiana. With this assertion, Isely concluded that Clements & Clements had validly made the combination: L. purshianus (Benth.) Clements & Clements. Kartesz & Gandhi: Nomenclatural notes on North American flora VII 95 Our analysis, contrary to Isely’s assertion, follows. Nuttall transferred Lotus sericeus to the genus Trigonella L. and provided anew name: T. americana. Bentham (in Lindley 1829) proposed the name Hosackia purshiana for L. sericeus and cited T. americana as a synonym. Later, Bentham (Trans. Linn. Soc. London. 17:368. 1837) rejected the name H. purshiana, accepted the name L. sericeus, and cited T. americana and H. purshiana as synonyms. Since both H. purshiana and T. americana were based on the type of L. sericeus, Bentham should have accepted the epithet amerz- cana in lieu of the epithet purshiana. However, he did not. This rendered the name H. purshiana to be superfluous, and thus illegitimate (ICBN Art. 63.1). Moreover, the name H. purshiana must not be considered for purpose of prior- ity (ICBN Art. 45.3). Any resultant new combination, based on H. purshiana, must be considered as a nomen novum, without a parenthetical author (ICBN Art. 72.2, Note 1). Hence, even if Hitchcock, Isely, and Barneby. were correct in assuming that the name L. purshianus was based on H. purshiana, they erred in treating L. purshtanus as a new combination. Ottley’s (1944) usage of the name L. purshianus suggested that he considered it to be a nomen novum for L. americanus. However, we believe that Clements & Clements’ usage of the name L. purshianus should be considered a species novum. Clements & Clements’ preface reference to Britton & Brown (1897) was a generalized statement and was not specific enough to provide even an indirect reference to either Nuttall or Bentham, i.e., the requirements of ICBN Art. 32.4 were never met for an indirect reference to make a nomen novum or combinatio novum. Regarding the usage of the epithet purshianus, perhaps Clements & Clements were aware of Bentham’s treatment; however, due to a lack of citation, this speculation can not be verified. The use of the epithet purshiana by Bentham as well as by Clements & Clements should be considered coincidental. Therefore, we conclude that Clements & Clements inadvertently, but validly and effectively proposed a new species, with priority from 1914. With the nomenclature of the name Lotus purshianus resolved, additional discussion on Barneby’s nomenclature is not made here. We conclude that L. unifoliolatus is the correct name for Spanish clover, as indicated by Dorn (1988). Lotus unifoliolatus (Hook.) Benth., Trans. Linn. Soc. 17:368. 1837. BA- SIONYM: Hosackia unifoliolata Hook., Fl. Bor. Amer. 1:135. 1833. Lotus sericeus Pursh, Fl. Amer. Sept. 2:489. 1814, non Moench, 1802, nec DC., 1813. Trigonella americana Nutt., Gen. Pl. 2:120. 1818. Lotus americanus (Nutt.) Bisch., Linnaea 14(App.):132. 1840, non Vell., 1825. Hosackia purshiana Benth. in Lindl., Bot. Reg. 15: t. 1257. 1829, nom. illegit. Lotus purshianus Clements & Clements, Rocky Mt. Fls. 183. 1914. 96 Po Y DOE O'GT A volume 71(2):87-100 August 1991 The above treatment necessitates the transfer of L. purshianus var. helleri (Britt.) Isely to L. untfoliolatus, for which the following new combination is proposed. Lotus unifoliolatus (Hook.) Benth. var. helleri (Britt.) Kartesz & Gandhi, comb. nov. BASIONYM: Lotus helleri Britt., Bull. Torrey Bot. Club 17:312. 1890. Lotus purshianus Clements & Clements var. helleri (Britt.) Isely, Brittonia 30:468. 1928. LECTOTYPE (vide Isely, Ic.): U.S.A. North Carolina: Mecklenberg Co., 1835, Curtis s.n. (NY). ROSACEAE Horkelia fusca var. parviflora Dorn (1988) attributed the authorship of Horkelia fusca Lindl. var. parvi- flora to (Nutt. ex Hook. & Arn.) Peck, whereas the Soil Conservation Service (1982, vol. 2) attributed it to (Nutt. ez Torr. & Gray) Wawra. Hence, we decided to research the nomenclature. Hooker & Arnott (1839) numbered each taxon that they accepted. In the protologue of Horkelia cuneata Lindl. (p. 338, no. 2), they mentioned the Nuttalean manuscript name H. parviflora and provided a brief description, but did not include this name in the index. Since the name H. parviflora was not separately numbered and not indexed, its inclusion in the protologue of H. cuneata could be interpreted as either a provisional name (ICBN Art. 34.1b), a described name in synonymy (ICBN Art. 34.1c), or both, none of which can be considered to be legitimate. Torrey & Gray (1840) independently described Horkelia parviflora (at- tributing the name to Nuttall), and thus validated the name; hence, the correct authorship of H. parviflora is: Nutt. ex Torr. & Gray. Wawra transferred H. parviflora to H. fusca at varietal status. Perhaps unaware of Wawra’s earlier combination, Peck independently transferred H. parviflorato H. fusca at vari- etal status, and attributed the basionym to Hooker & Arnott. Since the type specimen for both Hooker & Arnott and for Torrey & Gray was a Nuttalean specimen, Peck’s new combination must be considered as an isonym (Nicolson 1975). Horkelia fusca Lindl. var. parviflora (Nutt. er Torr. & Gray) Wawra, Itin. Princ. §. Cobungi 1:17. 1883; Peck, Madrono 6:134. 1941. BASIONYM: Horkelia parviflora Nutt. [ex Hook. & Arn., Fl. Bor. Amer. 338. 1839, nom. invalid] ex Torr. & Gray, Fl. N. Amer. 1:435. 1840. Kartesz & Gandhi: Nomenclatural notes on North American flora VII 97 ACKNOWLEDGMENTS We are grateful: to Dr. Paul A. Fryxell (TAES) and to Dr. Larry E. Brown (SBSC) for their valuable comments on the manuscript; to Dr. Jim Reveal (MARY) for a discussion on the nomenclature of Rhus hirta; and to Dr. Dan H. Nicolson (US) for the suggestion on the application of ICBN Art. 72.2, Note 1 to the name Tozicodendron toricarium. We also thank Dr. Thomas Duncan (UC), Dr. Jim Reveal, Dr. David E. Boufford (GH), Dr. James W. Hardin (NCSC), Dr. J.E. Simonis (U), Ms. Ruth F. Schallert (Botanical Li- brarian, US), and librarians at the Biology-Forestry Library (DUKE) and of the Botanical Library (H) for copies of the literature used in this study. REFERENCES Barker, R.M. 1986. A taxonomic revision of Australian Acanthaceae. J. Adelaide Bot. Gard. 9:1-128. Barkley, F.A. 1937. A monographic study of Rhus and its immediate allies in North and Central America, including the West Indies. Ann. Missouri Bot. Gard. 24:265-498. Barneby, R.C. 1989. Intermountain Fl., vol. 3(B): Fabales. The New York Botanical Garden, Bronx, New York. Bogin, C. 1955. Revision of the genus Sagittaria (Alismataceae). Mem. New York Bot. Gard. 9(2):179-233. Britton, N.L. & A. Brown. 1897. Ill. Fl. N. U.S., vol. 2. Charles Scribner’s Sons, New York, New York. ee BA NG135 NL PLN. U.S.,,e0./2, vols, 24 -Dover Publications, Inc-, New York. Clements, F.E. & E.S. Clements. 1914. Rocky Mt. Fls. The H.W. Wilson Company, White Plains, New York. Cronquist, A..1955. Compositae. In: C.L. Hitchcock et al., Vasc. Pl. Pacific NW., part 5. Univ. Wash. Publ. Biol. 17(5). Dorn, R.D. 1988. Vasc. Pl. Wyoming. Mountain West Publishing, Box 1471, Cheyenne, Wyoming. 98 PHYTOLOGIA volume 71(2):87-100 August 1991 Gillis, W.T. 1971. The systematics and ecology of poison-ivy and the poison- oaks ( Tozicodendron, Anacardiaceae). Rhodora 73:72-159, 161-237, 370- 443, 465-540. Gleason, H.A. 1952. New Britt. & Brown Ill. Fl. NE. Sts., vol. 2. The New York Botanical Garden, Bronx, New York. Gray, A. 1856. Man. Bot., ed. 2. George P. Putnam, New York, New York. Greuter, W., H.M. Burdet, W.G. Chaloner, V. Demoulin, R. Grolle, D.L. Hawksworth, D.H. Nicolson, P.C. Silva, F.A. Stafleu, E.G. Voss, & J. Mc- Neill. 1988. International Code of Botanical Nomenclature. Adopted by the Fourteenth International Botanical Congress, Berlin, Jul-Aug 1987. Regnum Veg. 118. Hitchcock, C.L. & A. Cronquist. 1961. Vasc. Pl. Pac. NW., part 3. Univ. Wash. Publ. Biol. 17(3). Hooker, W.J. & G.A.W. Arnott. 1839. Bot. Beechey Voy., part 8. London. Reprinted in 1965 by J. Cramer, Germany. Howard, R.A. 1989. Fl. Lesser Antilles, vol. 6: Dicotyledoneae - part 3. Arnold Arboretum, Harvard University. Irwin, H.S. & R.C. Barneby. 1982. The American Cassiinae. Mem. New York Bot. Gard. vol. 35, part 1. Isely, D. 1981. Leguminosae of the United States. III subfamily Papil- ionoideae: tribes Sophoreae, Podalyrieae, Loteae. Mem. New York Bot. Gard. 25(3):1-264. Jackson, B.D. 1895. Ind. Kew., vol. 2. Clarendon Press, Oxford. Kartesz, J.T. 1991. Syn. Checkl. Vasc. Fl. U.S. Canad. Greenl., ed. 2 in press. Timber Press, Portland, Oregon. & K.N. Gandhi. 1989. Nomenclatural notes for the North American flora - I. Phytologia 67:461-468. . 1990a. Nomenclatural notes for the North American flora - II. Phytologia 68:421-427. . 1990b. Nomenclatural notes for the North American flora - III. Phytologia 69:129-137. . 1990c. Nomenclatural notes for the North American flora - IV. Phytologia 69:301-312. Kartesz & Gandhi: Nomenclatural notes on North American flora VII 99 . 1991a. Nomenclatural notes for the North American flora - V. Phytologia 70:194-208. . 1991b. Nomenclatural notes for the North American flora - VI. Phytologia 71:58-65. Kartesz, J.T. & R. Kartesz. 1980. Syn. Checkl. Vasc. Fl. U.S. Canad. Greenl. University of North Carolina Press, Chapel Hill, North Carolina. Kaul, R.B. 1986. Alismataceae. Pp. 1022-1028 in: Great Plains Flora Association, Fil. Great Plains. University Press of Kansas, Lawrence, Kansas. Lanjouw J., S.H. Mamay, R. McVaugh, W. Robyns, R.C. Collins, R. Ross, J. Rousseau, G.M. Schulze, R. de Vilmorin, & F.A. Stafleu (ed.). 1966. International Code of Botanical Nomenclature. Adopted by the Tenth International Botanical Congress, Edinburgh, August 1964. Regnum Veg., vol. 46. Little, E.L., Jr. 1979. Checkl. U.S. Trees. U.S.D.A., Agr. Handb. no. 541. Govt. Printing Office, Washington, DC 20402. McGregor, R.L. 1986. Anacardiaceae. Pp. 571-575 in: Great Plains Flora Association, Fl. Great Plains. University Press of Kansas, Lawrence, Kansas. Nicolson, D.H. 1975. Isonyms and pseudo-isonyms: identical combination with the same type. Taxon 24:461-466. . 1991. Fl. Dominica, part 2: Dicotyledoneae. Smithsonian Contr. Bot. no. 77. Ottley, A.M. 1944. The American loti with special consideration of a pro- posed new section, Simpeteria. Brittonia 5:81-123. Radford, A.E., H.A. Ahles, & C.R. Bell. 1968. Man. Vasc. Fl. Carolinas. The University of North Carolina Press, Chapel Hill, North Carolina. Small, J.K. 1903. Fl. SE. U.S. Published by the author, New York, New York. Smith, J.G. 1895. A revision of the North American species of Sagittaria and Lophotocarpus. Annual Rep. Missouri Bot. Gard. 6:27-64. Soil Conservation Service. 1982. Nat. List Sci. Pl. Nam., vols. 1 & 2. U.S.D.A., SCS-TP 159. Govt. Printing Office, Washington DC. 100 PHY TOLO:GIA volume 71(2):87-100 August 1991 Stafleu, F.A., V. Demoulin, W. Greuter, P. Hiepko, 1.A. Linczevski, R. Mc- Vaugh, R.D. Meikle, R.C. Rollins, R. Ross, J.M. Schopf, & E.G. Voss (ed.). 1978. International Code of Botanical Nomenclature. Adopted by the Twelfth International Botanical Congress, Leningrad, July 1975. Regnum Veg., vol. 97. Sudworth, G.B. 1927. Checkl. Forest Trees U.S. U.S.D.A., Misc. Cir. no. 92. U.S. Govt. Printing Office, Washington, DC. Torrey, J. 1859. Botany of the Boundary. Pp. 29-270 in: W.H. Emory, United States and Mexico Boundary Survey, Department of the Interior, Washington, DC. & A. Gray. 1840. Fl. N. Amer., vol. 1. Wiley & Putnam, New York, New York. Voss, E.G. 1985. Michigan Fl., part II: Dicots (Saururaceae - Cornaceae). Bull. Cranebrook Inst. Sci., vol. 59. Voss, E.G., H.M. Burdet, W.G. Chaloner, V. Demoulin, P. Hiepko, J. Mc- Neill, R.D. Meikle, D.H. Nicolson, R.C. Rollins, P.C. Silva, & W. Greuter. 1983. International Code of Botanical Nomenclature. Adopted by the Thirteenth International Botanical Congress, Sydney, August 1981. Reg- num Veg., vol. 111. Wofford, B.E. 1989. Guide Vasc. Pl. Blue Ridge. The University of Georgia Press, Athens, Georgia. Phytologia (August 1991) 71(2):101-103. A NEW VARIETY OF ERYTHRONIUM CITRINUM (LILIACEAE) FROM THE SCOTT MOUNTAINS OF NORTHWEST CALIFORNIA James R. Shevock Department of Botany, California Academy of Sciences, San Francisco, California 94118-4599 U.S.A. & Geraldine A. Allen Department of Biology, University of Victoria, Victoria, British Columbia, V8W 2Y2, CANADA ABSTRACT A new variety of Erythronium citrinum S. Watson, the var. roder- ickii, is described from ultrabasic soils in the Scott Mountains, Trinity . River drainage, northwest California. It differs from typical FE. citrinum in having pink to purplish anthers and a brighter yellow zone at the base of the tepals. KEY WORDS: Liliaceae, taxonomy, Erythronium, California INTRODUCTION In the course of our study of Erythronium for the Jepson Manual Project, a number of interesting collections and field observations dealing with range extensions and potential new taxa have been brought to our attention (She- vock, Bartel, & Allen 1990). The new variety described herein has been known to bulb specialist Wayne Roderick since 1961. In the spring of 1991, Roder- ick collected herbarium specimens and sent them to the senior author for a more detailed study. We have concluded that this entity is indeed worthy of recognition and we formally describe it here. 101 102 PHY DOL OIGiA volume 71(2):101-103 August 1991 TAXONOMY Erythronium citrinum S. Watson var. roderickii Shevock & Allen var. nov. TYPE: UNITED STATES. California: Trinity Co., along Cal- ifornia Highway 3, adjacent to Scott Mountain Creek, Trinity River drainage, NE facing slopes on ultrabasic (serpentine) soils, mixed conifer forest, Shasta-Trinity National Forest, T39N, R7W, sect. 15 NW1/4. MDB&M., 4000 ft. (1220 m), 3 May 1991, Wayne Roderick s.n. (HOLO- TYPE: CAS; Isotypes: JEPS,|K,MO,NY,OSC,RSA,UVIC,US). Tepalis albis, differt a Erythronium citrinum S. Wats. var. cit- rinum basibus aureoribus; antheris roseis vel purpureis. Bulb 3-6 cm long, slender. Leaves 2, strongly mottled, 8-30 mm wide, 5.5-16 cm long, lanceolate to narrowly ovate. Scapes 7.5-16.5 cm tall; flowers 1(-2). Perianth segments recurved, lanceolate, acute to acuminate, 20-35 mm long, 5-8(-10) mm wide, white, fading pinkish after anthesis, with a 5-7 mm long golden yellow zone at base (darker yellow than the typical variety): inner perianth segments with saclike folds at base. Stamens unequal in two series, filaments 8-10 and 10-12 mm long, slender, anthers pink to purple, drying reddish brown, 3-5 mm long. Style clavate, 6-9 mm long, white to cream colored, occasionally tinged pinkish, stigma entire to short lobed, the lobes < 0.5 mm long. Capsules obovoid, 1.5-3.0 cm long. Specimens examined. UNITED STATES. California: Trinity Co.: from the type locality, (specimen in fruit), 28 May 1991, Linda Barker 2012 (CAS); Bear Creek Trail, approximately 1 mile W of California Highway 3 near bound- ary of Trinity Alps Wilderness, T39N, R7W, SE1/4 NW1/4 section 34, (spec- imen in fruit), 9 June 1991, Wayne Steffes s.n. (CAS). This Erythronium first came to our attention in May 1991, although it has been in very limited cultivation since the mid 1970’s. We are pleased to name this variety for its discoverer, the noted horticulturist Wayne Roderick, who specializes in California bulbs (Wolf 1986). We propose the vernacular name be Scott Mountains fawn lily. The Klamath Mountains of northwest California and southwest Oregon contain at least six taxa of Erythronium, and are a center of diversity for the genus (Applegate 1935). Erythronium citrinum var. roderickit is known at present from only three localities, and appears to be a relatively rare localized endemic of this region. There is a high likelihood that continued field work on serpentine soils in the Scott Mountains will yield additional occurrences. The origins of Erythronium citrinum var. roderickii are problematic. In most aspects it closely resembles typical E. citrinum. However, the anther color suggests the possibility of past hybridization between the cream anthered E. citrinum and a purple anthered species, the most likely candidate being E. Shevock & Allen: New variety of Erythronium citrinum 103 hendersoniu S. Watson. Both typical E. citrinum and E. hendersonw occur farther north in the Klamath Mountains (Klamath River drainage), and no other Erythronium species is currently found in close proximity to FE. citrinum var. roderickit. ACKNOWLEDGMENTS We wish to thank Dr. Barbara Ertter and Dr. James C. Hickman of the University and Jepson Herbaria respectively at the University of California in Berkeley, for reviewing this paper. Special thanks to Linda Barker (former forest botanist. Klamath National Forest) and Julie Nelson, forest botanist, Shasta-Trinity National Forest for field observations and specimens; and to Wayne Steffes who provided the specimen from Bear Creek’and also color slides. We also thank Wayne Roderick for sharing his knowledge of this Ery- thronium with us. LITERATURE CITED Applegate. E.I. 1935. The genus Erythronium: a taxonomic and distribu- tional study of western North American species. Madrono 3:58-113. Shevock, J.R., Bartel, J.A., & G.A. Allen. 1990. Distribution, ecology, and taxonomy of Erythronium (Liliaceae) in the Sierra Nevada of California. Madrono 37:261-273. Wolf, M.R. 1986. Wayne Roderick, horticulturist [New California Native Plant Society Fellow. 1985]. Fremontia 14:29. Phytologia (August 1991) 71(2):104-108. A NEW SPECIES OF PITCAIRNIA (BROMELIACEAE: PITCAIRNIOIDEAE) Michael A. Spencer & Lyman B. Smith Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560 U.S.A. ABSTRACT Pitcairnia ramosii, a new species from Colombia, is described and discussed. KEY WORDS: Bromeliaceae, Colombia, Pitcairnia Pitcairnia ramosii M.A. Spencer & L.B. Smith, sp. nov. Fig. 1. TYPE: COLOMBIA. Chocé: Municipio San Jose del Palmar, Cerro del Torra, vertiente oriental del Rio Negro, abajo del helipuerto, vereda del Rio Negro, 1630 m, 8 Aug 1988, J.E. Ramos, P.A. Silverstone, & L.H. Ramos 1052 (HOLOTYPE: CUVC!; Isotypes: MO!, NY!, US!). A Pitcairnia lepidopetalone L.B. Smith affinis, a qua differt, inflorescentia subdense multiflora, pedicelis brevibus, sepalis bre- vibus margine undulatis et apice rotundatis, petalis apice acutis et luteo-viridis, ovario semisupero. PLANT caulescent, epiphytic or terrestrial, flowering to 8 dm high. LEAVES isomorphic, subdistichous, 10-13 dm long. SHEATHS inconspicuous, narrowly oblong, 9-12 cm long and 1.5-2.0 cm wide, entire, densely and minutely brown lepidote. BLADES linear-lanceolate, long attenuate, 9-12 dm long and 1.0-2.5 cm wide, plicate, entire, glabrous above, sparsely and minutely brown lepidote below, slightly narrowed at the base but not petiolate. SCAPE erect, slender, 6.5-7.0 dm long and 5-7 mm in diameter, glabrous. SCAPE BRACTS strict, the lower subfoliaceous and exceeding the internodes, the upper broadly ovate with an acuminate apex, 20-25 mm long and 15-20 mm wide, slightly shorter than to equaling the internodes, entire, membranaceous, light yellow-green. INFLORESCENCE simple, subdensely 15-25 flowered, racemose, 10-15 cm long and 5-7 cm in diameter, glabrous. FLORAL BRACTS spreading, ovate 104 Spencer & Smith: ITCAIRwWIA Rao Figure 1. Holotype dt A a é ti Speman 9 od ~*~ New Pitcairnia 20002 : Sn, d0eé del Palmar 2t,: 1630 Aierbe sobre €i piso, ca OC c=; infiorescencia 75 cm, ve de exerilierts; @pice de los s@palos rogos, couspicuoc, bese tlancay base de ios pétalos biente y hecia ts apac peiletinerente verde azari:lentoe; filamertos emarilier bese ¢el estiic blanca, facia el Qpice exariliexto; ant res ¥ polen amirillos, Cerro del Torrd, vertiente Oriental ds. rio Negrc, ubasc Relipuertc, Vereda de Rio hecre, Col.: J,E Ramos, F, A, Silverstoze : LH. Ramos et el, Nor 1052 Forks. & ten GRR of Pitcairnia ramosii, Chocé, Colombia. 105 106 PHYVTOLOGTA volume 71(2):104-108 August 1991 11cm : fi Q Figure 2. Pitcairnia ramosii: A, pedicel; H, sepal; R, petal. Pitcairnia lepi- dopetalon: B, pedicel; G, sepal; Q, petal. Spencer & Smith: New Pitcatrnia 107 with an acuminate apex, 20-25 mm long and 15-20 mm wide, inflated, undu- late, entire, membranaceous, much exceeding the pedicels, light yellow-green. FLOWERS spreading. PEDICELS slender, 10 mm long and 1 mm in diam- eter, glabrous. SEPALS slightly asymmetric, oblong, rounded and apiculate, 32 mm long and 15 mm wide, alate-carinate, membranaceous, entire, undu- late, sparsely and minutely stellate lepidote, light green with a bright red apex before anthesis, turning light. green with only a small, dark purple spot at the apex by anthesis. COROLLA zygomorphic. PETALS lingulate, acute, 70 mm long and 10 mm wide, sparsely white lepidote, bearing a single, erose-undulate ligule at the base, white becoming light green at apex. STAMENS included, 65 mm long. FILAMENTS free, linear, 54 mm long and 0.5 mm wide. AN- THERS linear, 11 mm long and 1 mm wide, basally dorsifixed, yellow. PISTIL included. STYLE terete, 55 mm long and 1 mm in diameter. STIGMA 5 mm long, lobes unequal, loosely spiraled. OVARY spheroidal, 5 mm long and 5 mm in diameter, 1/2 superior. OVULES long caudate. OTHER SPECIMENS EXAMINED: COLOMBIA. Choco: Municipio San Jose del Palmar, Cerro del Torra, vertiente nordeste, bosque de neblina, 1850- 1930 m, 13 Aug 1982, Felipe Silverstone-Sopkin 1341 (CUVC, MO!). DISCUSSION: The presence of caudate ovules places this new species in Pitcairnia subgenus Pitcairnia (Smith 1974). Pitcairnia ramosii is most closely related to P. lepidopetalon L.B. Smith but can be distinguished from the latter by its subdensely many flowered inflorescence, shorter pedicels, slightly asymmetric and shorter sepals that have undulate margins and a rounded apex, yellow-green petals that have an acute apex, and semisuperior ovary. In P. lepidopetalon, the inflorescence is lax and few flowered, the pedicels are twice as long, the sepals are longer, symmetric, the margins are even and the apex is truncate, the petals are red with an attenuate apex, and the ovary is 2/3 superior. Both Pitcairnia ramosii and P. lepidopetalon are known only from the moist, montane forests of eastern Colombia. ETYMOLOGY: This new species is named in honor of the collector, Dr. Jorge E. Ramos, Universidad De Valle, Cali, Colombia. ACKNOWLEDGMENTS We would like to thank John Kress and Harold Robinson for reviewing the manuscript. 108 PHY TOLOGdiIA volume 71(2):104-108 August 1991 LITERATURE CITED Smith, Lyman B. 1974. Pitcairnioideae. Flora Neotropica, Monograph No. 14, part 1. Hafner Press, New York, New York. Phytologia (August 1991) 71(2):109-121. UNION OF BRADBURIA WITH CHRYSOPSIS (ASTERACEAE: ASTEREAE), WITH A PHYLOGENETIC HYPOTHESIS FOR CHRYSOPSIS Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT The monotypic genus Bradburia (B. hirtella Torrey & Gray) is merged with Chrysopsis (as C. texana nom. nov.), enlarging the latter to eleven species. Within Chrysopsis, C. terana and C. pilosa are sis- ter species, based on a set of distinctive features common to both taxa. Both species are the westernmost of the genus and (as Chrysopsis sect. Bradburia comb. et stat. nov.) they form a group phyletically coor- dinate with the other species, which are centered in the southeastern United States. A phylogenetic hypothesis is presented for all species of Chrysopsis; the genera most closely related to it are Osbertia and Noticastrum. KEY WORDS: Bradburia, Chrysopsis, Osbertia, Noticastrum, Aster- aceae, Astereae The taxonomic history of the monotypic genus Bradburia Torrey & Gray has been summarized by Semple & Chinnappa (1984). Essentially, it has been recognized as closely related to the genera Heterotheca Cass. and Chrysopsis (Nutt.) Ell., but it differs prominently from both in its sterile disc flowers. The dimorphic achenes of Heterotheca (sensu Semple 1977; Semple et al. 1980) have been noted by some authors as similar, and perhaps homologous, to those of Bradburia, but Semple (1981) has observed that a closer relationship exists between Bradburia and Chrysopsis, in his more restricted view of the latter. Semple (1981, p. 339), indeed, placed the relationship so close that he hypothesized that some isolated, immediate precursors of C. pilosa Nutt. “ultimately evolved into Bradburia hirtella.” Semple & Chinnappa (1984, p. 95) noted that “While Bradburia hirtella could be transferred to Chrysopsis, the distinctive staminate florets of the for- mer are considered sufficient to separate the two at the generic level.” Later in 109 110 PHOT OL OG LA volume 71(2):109-121 August 1991 the same paper (p. 100), they commented that “The distinctive achene mor- phology, somewhat different flavonoid pattern, and karyotype support mainte- nance of Bradburia as a genus separate from Chrysopsis.” The data from the flavonoid studies, however, were not presented in their study, and its authors noted in the same paper (p. 95) that “Flavonoid patterns in the two genera were similar but did not confirm or refute a close relationship.” CHROMOSOMES Detailed studies of meiotic pairing in Chrysopsis hybrids (Semple & Chin- nappa 1980a) and of chromosome numbers and karyotypes (Semple & Chin- nappa 1980b) showed that four basic karyotypes exist within Chrysopsis: the zt = 4 of C. pilosa, the z = 4 of C. mariana (L.) Ell., the z = 5 of the ma- jority of the species, and the z =9 of C. gossypina (Michx.) Ell. The latter comprises “little more than a combination of the z = 5 karyotype and the z = 4 mariana karyotype and was therefore of allopolyploid origin.” (1980b, p. 164). These two studies provided convincing evidence of strong homology between the z = 4 mariana and the z = 5 karyotypes, the former probably derived by loss of chromosome “V” of the z = 5 set, with transfer of portions of that chromosome to both chromosomes “IJ” and “III” of the z = 4 mariana set. Semple & Chinnappa (1980b, p. 170) hypothesized that the “evolution of the z = 4 [pilosa] karyotype from the z = 4 [mariana] karyotype resulted in a considerable increase in asymmetry and an increase in the number of acro- centric homologues” but noted (p. 168) that “no clear homologies [between these two z = 4 karyotypes] are suggested by chromosome size and centromere position.” Plants of Bradburia have a chromosome number of n = 3 over most of the geographic range of the species, but Semple & Chinnappa (1984) reported the significant discovery of an individual with n = 4 from the southwesternmost portion of the range. They presented a detailed study of the karyotypes of the two Bradburia cytotypes and an artificially synthesized hybrid between the two. In comparing the n = 4 chromosome complement of Bradburia with that of Chrysopsis pilosa, they observed (p. 100) that “if the large pair of the C. pilosa karyotype was more metacentric, then the karyotype would be superficially very similar to that of the n = 4 karyotype of Bradburia.” Based on their very clear photographs and diagrams, the phrase “superficially very similar” could be replaced by “virtually identical.” Semple & Chinnappa (p. 99) further noted that comparison of their karyotypes “suggests a possible origin of Bradburia from an early ancestor of C. pilosa.” They did not offer a precise hypothesis of how the n = 3 complement originated from the n = 4, but clearly the two may be interpreted as homologous. Nesom: Union of Bradburia with Chrysopsis 111 MORPHOLOGY Phylogenetic analyses of the goldenaster lineage (Nesom 1991a) and of the genus Chrysopsis (see details below) indicate that Bradburia is closely related to Chrysopsis, based on their thin walled and vitreous upper cells of the Type A trichomes, and distinctive achene morphology (obovate, with thick, buried ribs). Among the species of Chrysopsis, it is similar to C. pilosa in its chro- mosome number, karyotype (see discussion above), annual duration, spring flowering, long flowering branches, Type A trichomes with the distal cells thin walled but not elongated into a flagelliform terminal portion, phyllaries with broad scarious margins, lack of characteristically large, elongated crystals in the tissues of the disc corolla throat, sharp pointed sweeping hairs on the disc style branches, and ray corollas remaining straight rather than coiling. Together these similarities provide evidence that the two taxa are related as sister species. Despite its close relationship to Chrysopsis pilosa, Bradburia differs from it in a number of features. The difference that has been heavily emphasized is the sterility of disc ovaries and concomitant reduction of the disc pappus to two awns. Additionally, the ray achenes of Bradburia are 3 sided, the outer face distinctly broadened (vs. merely 2 sided in C. pilosa), with an outer pappus composed of a whorl of relatively inconspicuous, short bristles and setae (vs. a whorl of broad scales); the pappus bristles of Bradburia are distinctly flattened and characteristically reddish near the base (vs. terete and tawny); the leaves of Bradburia are narrow at the base and strictly sessile (vs. basally broadened and subclasping); the heads of the glandular (Type C) trichomes are broader (5-8 cells across vs. 3-5 cells); and the phyllaries are essentially glabrous (vs. distinctly hairy glandular). Attempts to cross Bradburza with C. pilosa have been unsuccessful (Semple & Chinnappa 1984), and populations of the two species are sometimes intimately intermixed in nature (Nesom personal observ.). While the differences that separate Bradburia from Chrysopsis pilosa are at least as great as those that distinguish most other species of Chrysopsis, Semple (1981, p. 338) noted that C. pilosa itself is “the most divergent species [of the genus] in terms of morphology, chemistry, and cytology.” Further, C. pilosa and Bradburia are so distinct as a pair that C. pilosa might justifiably be transferred to Bradburia, but the similarity in achene morphology between them and the other species of Chrysopsis strongly suggests that the group as a whole is monophyletic and should stand as a single genus. Artificially produced hybrids between C. pilosaand C. gossypina (Semple 1981) emphasize a significant degree of genetic similarity between the two segments of the genus. No one has suggested that C. pilosa be formally recognized apart from the other species of the genus; if it is taxonomically treated as Chrysopsis, then so should Bradburia. The close relationship between these two species, apart 112 PHY TO 2:0:6 LA volume 71(2):109-121 August 1991 from the rest of the genus, is formalized in the nomenclature below. Chrysopsis sect. Bradburia (Torrey & Gray) Nesom, comb. et stat. nov. BASIONYM: Bradburia Torrey & Gray, Fl. N. Amer. 2:250. 1842. Species included: C. pilosa Nutt. and the type species: Chrysopsis texana Nesom, nom. nov. BASIONYM: Bradburia hirtella Torrey & Gray, Fl. N. Amer. 2:250. 1842; not Chrysopsis hirtella DC., 1836. TYPE: UNITED STATES. Texas: [no other data], Drummond 134 (HOLOTYPE: not seen; Probable isotype: LL!). CLADISTIC ANALYSIS In a broader cladistic study that includes all of the goldenasters (Nesom 1991a), Osbertia (x = 5) E. Greene of México and Central America and Not- icastrum DC. (z = 9) of South America are hypothesized to be the closest relatives of Chrysopsis (including Bradburia), based on their relatively long, smooth, thin walled, and vitreous Type A trichomes. In order to determine the phylogenetic position of Bradburia, it has been included in a phylogenetic analysis of the species of Chrysopsis, with the assumption that such a group is holophyletic only with the inclusion of Bradburia. The closely related gold- enaster genus Heterotheca has been added as part of the outgroup to clarify some of the character state polarities. Studies of microcharacters were made with a compound microscope using epidermal “skims” with a razor from stems and slides of dissected flowers mounted in Hoyer’s Medium. A list of specimens from which slides have been made is deposited at TEX; all voucher specimens are deposited in TEX. All specimens examined and scored of Bradburia presumably have a chromosome number of n = 3. Characters and character states used in the analysis are given in Table 1; coding for the individual taxa are in Table 2; discussion of variability in the characters used is provided in the Appendix. The data were analyzed using PAUP (Swofford 1985) with Wagner parsimony. Chrysopsis gossypina was not included in the analysis, since it apparently is of hybrid origin (Semple 1980b); its position relative to the other species is shown on the cladogram following Semple (1981). A single shortest tree of 42 steps (Fig. 1) accounts for the distribution of character states among the taxa included in the analysis. Chrysopsis terana (Bradburia) and C. pilosa(sect. Bradburia) are geographically the westernmost of the genus and form a clade phyletically coordinate with the other species (sect. Chrysopsis), which are centered in the southeastern United States. In contrast to Semple (1981), who placed the ancestor of the whole genus in Nesom: Union of Bradburia with Chrysopsis 113 TABLE 1. Characters and character states a 9. 10. im 12: 13. 14. 15. 16. Ei. Chromosome number, (0) n = 4 or 3, the “pilosa” karyotype, or n = 9 (1) n = 5 or 4, the “Chrysopsis” karyotype . Chromosome number, (0) z = 5 or 4, the “mariana” karyotype or its n = ) 5 homologue, or n = 9 (1) n = 4 or n = 3, the “pilosa” karyotype ) . Chromosome number, (0) n = 4 or n = 3, the “pilosa” karyotype, or n = 9 orn = 5 (1) z= 4, the “mariana” karyotype . Duration, (0) perennial or biennial (1) annual . Flowering phenology, (0) primarily in late summer and fall, rarely in the spring (1) typically beginning in the spring, often continuing into the fall . Rhizomes/roots, (0) rhizomes (1) taproots . Capitulescence, (0) loosely to compactly corymbose (1) monocephalous . Flowering branch length, (0) relatively short, arising from the upper third of the stem, or the stems monocephalous (1) relatively long, commonly arising from the lower half of the stem Leaf insertion, (0) clasping or subclasping (1) not at all clasping Vestiture of peduncles and phyllaries, (0) stipitate glandular (1) glandu- larity absent or highly reduced Type A trichomes, (0) terete with thick, papillate walls (1) terete to flattened with thin, smooth walls Type A trichome length, (0) relatively short, not at all flagelliform (1) all flagelliform, with the distal cells greatly elongated, at least in some species Phyllary midvein, (0) mostly included within lamina (1) prominently raised-swollen, orange glandular Phyllary margins, (0) scarious rim narrow or absent (1) scarious rim very broad Phyllary apex, (0) attenuate or linear-attenuate (1) blunt-acute Phyllary apex, (0) attenuate or blunt-acute (1) linear-attenuate Behavior of buds, (0) erect (1) nodding 114 PHY TOLOGIA volume 71(2):109-121 August 1991 TABLE 1. (continued). 18. 19. 20. au. 22: 23. 24. 20. 26. PAG 28. 29. 30. 31. 32. Ray color, (0) yellow, drying yellow (1) yellow or creamy white, drying yellow tinged with red Ray corolla behavior after stigma receptivity, (0) coiling (1) remaining straight Sweeping hair shape, (0) narrowly oblong or slightly broadened apically (1) distinctly lanceolate, at least those near the apex Sweeping hair apex, (0) narrowly oblong to lanceolate, without a sharp point, or narrowly acute with a sharp point (1) broadly acute with a sharp point Sweeping hair apex, (0) narrowly oblong to lanceolate, without a sharp point, or broadly acute with a sharp point (1) narrowly acute with a sharp point Elongate crystals in the tissues of the disc corolla throat, (0) present (1) absent or greatly reduced in size Resin ducts in disc corolla lobes, (0) continuous or absent (1) discontin- uous Radial walls of cells of upper throat of disc corollas, (0) straight (1) broadly sinuate © Achene walls, (0) eglandular, or with thin, barely raised glandular ridges (1) with thick, raised, glandular ridges Achene shape, (0) narrowly elliptic to oblanceolate in side view (1) obo- vate Achene compression, (0) distinctly compressed (1) terete or subterete Achene nervation, (0) thin, superficial (1) thick, beneath the surface Pappus insertion, (0) at shoulder rim of achene apex (1) inset from shoulder rim of achene apex Pappus series, (0) double (1) single Outer pappus, (0) setae or absent (1) scales Nesom: Union of Bradburia with Chrysopsis 115 TABLE 2. Data matrix for Osbertia, Bradburia, and the diploid species of Chrysopsis. Character State a Heterotheca (HETERO) Noticastrum (NOTICA) Osbertia (OSBERT) Bradburia (BRADBU) Chrysopsis pilosa (PILOSA) C. mariana (MARIAN) C. lanuginosa (LANUGI) C. godfreyi (GODFRE) C. floridana (FLORID) C. scabrella (SCABRE) C. latisquamea (LATISQ) C. linearifolia (LINEAR) C. subulata (SUBULA) Florida, the present analysis suggests that such an ancestor, derived from Latin American precursors, probably migrated northward into south central and southeastern North America. The initial divergence, which occurred in the south central region, separated the two major lineages. The species of sect. Chrysopsis form four main clades. The relationship among them has not been resolved despite a detailed search for potentially useful characters, but on the basis of their karyotype and long flagelliform hairs, these species are united as a group distinct from sect. Bradburia. The morphological data suggest that the ancestor to sect. Chrysopsis may have been more or less simultaneously fragmented into four lineages. The base chromosome number of ancestral Chrysopsis is most parsimo- niously z = 9 in the present hypothesis, but the lack of any primitively n = 9 species of Chrysopsis and the geographic proximity of the closely related Osbertia (Nesom 1991b), with its base chromosome number of z = 5, sug- gests that it, rather than Noticastrum, may be phyletically coordinate with Chrysopsis and that the primitive number for Chrysopsis is x = 5. If Osber- tia and Chrysopsis are positioned as sister genera, with Noticastrum primitive to both, only two steps are added to the cladogram in Fig. 1, both as easily conceivable parallelisms in characters 7 and 18. A study of the karyotype of Osbertia will be critical in further assessing its relationship to Chrysopsis. Whatever may prove to be true regarding the base chromosome number for the immediate ancestor of Chrysopsis, within the genus the evolutionary attainment of four pairs of chromosomes in C. mariana appears to have oc- 116 PHY TOLOGIA volume 71(2):109-121 August 1991 LINEAR SUBULA LANUGI| GODFRE ‘6 5 SCABRE FLORID LATISQ 9 2 \ yy \ N N Be 7 a GOSSYP 7 BRADBU PILOSA a 26 7 2 <9 / 13 7 ay ‘; Pe 2 / MARIAN w “ a Vo, So ® © b&b Vv /2 / NOTICA OSBERT J Seo ae a 3o x) 2 2> 2 2 ‘i (0) HETERO 4/ Figure 1. Phylogenetic reconstruction of the Chrysopsis lineage and the species of Chrysopsis, including Bradburia. Abbreviations for the tax follow those in Table 2; “GOSSYbP” is C. gossypina (see text); “BRADBU” is C. tezana. Sin- gle slash = apomorphy or autapomorphy; double slash = parallelism; down- ward arrow = reversal. Nesom: Union of Bradburia with Chrysopsis 117 curred independently of that in the ancestor of C. pilosa and C. terana. Even if characters related to chromosomes (characters 1, 2, and 3) are deleted from the analysis, the topology of the resulting cladogram is no different from that in Fig. 1. This supports the hypothesis that the z = 4 pilosa and the z = 4 mariana karyotypes are not homologous. The phylogenetic hypothesis presented here for Chrysopsis is similar to the one developed more intuitively by Semple (1981, Figure 3, p. 331). Apart from the general methodology of its construction, without the use of an outgroup for the polarization of character states and without the inclusion of Brad- buria, Semple’s diagrammed hypothesis differs primarily in its indication that C. pilosa and C. mariana are sister species (based on their reduced chromo- some number) and the placement of C. scabrella-floridana and C. lanuginosa- godfreyt in a single clade (apparently based on geography and what are inter- preted here as shared plesiomorphies). The narrative evolutionary scenario for the genus related by Semple (1981) differs from his diagram. In the narrative, he postulated that the divergence of the latisquamea lineage and the scabrella- lanugznosa lineage from the ancestor of the genus was the first cladistic event, followed by the divergence of the z = 4 lineage (including C. pilosa and C. mariana) from the scabrella-lanuginosa lineage. In the narrative, also, he hy- pothesized that some populations of the immediate precursor of C. pilosa, as it migrated from the Florida region toward the Ozark Plateau, became isolated and evolved into Bradburia. ACKNOWLEDGMENTS I thank Dr. B.L. Turner and Dr. T.P. Ramamoorthy for their review and comments on the manuscript. LITERATURE CITED Nesom, G.L. 199la. A phylogenetic hypothesis for the goldenasters (Aster- aceae: Astereae). Phytologia 71:136-151. . 1991b. Transfer of Heterotheca bartletta to Osbertia (Asteraceae: Astereae). Phytologia 71:132-135. Semple, J.C. 1977. Cytotaxonomy of Chrysopsis and Heterotheca (Compositae- Astereae): a new interpretation of phylogeny. Canad. J. Bot. 55:2503- 2513. 118 PHY TOL 0 GPA volume 71(2):109-121 August 1991 . 1981. A revision of the goldenaster genus Chrysopsis (Nutt.) Ell. nom. cons. (Compositae-Astereae). Rhodora 83:323-384. Semple, J.C. & C.C. Chinnappa. 1980a. Phylogenetic implications of meiosis in wild and cultivated interspecific hybrids in Chrysopsis (Compositae- Astereae): C. godfreyt (n = 5) X gossypina ssp. cruiseana (n = 9) and C. godfreyi (n = 5) X lineartfolia (n = 5). Canad. J. Bot. 58:172-181. 1980b. Karyotype evolution and chromosome numbers in Chrysopsis (Nutt.) Ell. sensu Semple (Compositae-Astereae). Canad. J. Bot. 58:164-171. . 1984. Observations on the cytology, morphology, and ecology of Bradburza hirtella (Compositae-Astereae). Syst. Bot. 9:95-101. Semple, J.C., V.C. Blok, & P. Heiman. 1980. Morphological, anatomical, habit, and habitat differences among the goldenaster genera Chrysop- sis, Heterotheca, and Pityopsis (Compositae-Astereae). Canad. J. Bot. 58:147-163. Swofford, D.L. 1985. PAUP: Phylogenetic analysis using parsimony, Vers. 2.4. Illinois Nat. Hist. Survey, Champaign, Illinois. APPENDIX 1. Notes on character variation. 1, 2, and 3. Chromosomes: see discussion in text. The scoring relies on data furnished by Semple & Chinnappa (1980b, 1984). 4. Duration: Semple & Chinnappa (1984) observed that the n = 4 plants of Bradburia lived for 2 1/2 years, as opposed to the strictly annual n = 3 ones. This species is scored here as annual, but it would not change the results if it were scored as perennial. 5. Flowering phenology: data from Semple (1981). 6. Rhizomes/roots: some species of Noticastrum are rhizomatous, oth- ers taprooted; the genus is scored ambiguously, but even if it were not, the topology of the cladogram would not change. The primitive condition of the goldenaster root system is not clear (Nesom 1991la), but Heterotheca, which is used as the outgroup in the present analysis, is apparently primitively tap- rooted. 7. Capitulescence: in Osbertia and nearly all species of Noticastrum the stems are monocephalous, but some plants of N. marginatum (Kunth) Cuatr. produce loose cymes, with several head-bearing branches arising on the upper third of the initial stem. The same is true for O. bartletti (S.F. Blake) Nesom. Nesom: Union of Bradburia with Chrysopsis 119 Such variants probably reflect the ancestral condition for these two genera, displayed with a release of the suppression of lateral branches. 8. Head bearing branches: in most of Chrysopsis, as well as in the genus Heterotheca, the capitula are produced on relatively short peduncles originat- ing from a branch point on the upper part of the main stem. In C. pilosa and C. tezana, the peduncles are usually longer and originate from lower portions of the stem. 9. Leaf insertion: in Bradburia as well as the three species of the C. latisquamea Pollard lineage, the leaves are clearly sessile; in C. scabrella Torr. & Gray, C. floridana Small, C. lanuginosa Small, and C. godfreyi Semple, they are clearly clasping or subclasping. Semple (1981) described leaf insertion in C. pilosa, C. mariana, and C. scabrella as sessile, but these species are more similar in leaf base to the clasping leaved species than to the sessile leaved ones. 11 and 12. The smooth, thin walled Type A trichomes that are often lengthened into a tomentose vestiture are the primary feature that distin- guishes the Chrysopsis lineage from all of the other goldenasters (Semple et al. 1980; Nesom 199la). Type A trichomes in Noticastrum are variable in length from species to species, but in some they are minutely flagelliform and form a dense tomentum identical to that of Chrysopsis species. The comparatively high degree of variability in vestiture, as well as habit, among the species of Noticastrum is perhaps reflective of the ancestral stock of the whole Chrysopsis clade. In the cladistic hypothesis offered here, the relatively short hairs of C. pilosa and C. terana are more primitive than the long flagelliform, tomentum forming hairs characteristic of the other species of Chrysopsis. 13. Phyllary midvein: the phyllaries of C. subulata Small and C. linearifolia Semple are distinctive in their prominently raised-swollen, orange glandular midveins. The midveins of the other species are somewhat variable but none approach the distinctive morphology of these two species. 15 and 16. Phyllary apex: these features are difficult to characterize with precision, but there appear to be two pairs of species, each pair with phyllary apices distinctive from all other species. 17. Behavior of buds: data for Chrysopsis from Semple (1981). Compar- ative data for most of the rest of the goldenaster lineage is lacking, but the buds of the species of Heterotheca sect. Heterotheca are nodding. 18. Ray color: among the 12-13 species of Noticastrum, only N. acumina- tum (DC.) Cuatr. has rays that are distinctly yellow when fresh. The other species are noted by collectors as having white to dark lavender rays. Upon drying, however, the rays of seven of the eight species I have studied are a densely opaque, dark orangish yellow, sometimes retaining a purplish tinge. The “dry” color of the rays, with a reddish purple tinge, is similar among the other goldenasters only to that in Osbertia. Based on these observations and the distribution of other character states in the present analysis, the white 120 PHY TOLOGIA volume 71(2):109-121 August 1991 ray color in Noticastrum is interpreted here as derived from the primitive yel- low rays of the ancestral goldenaster; the yellow rays of N. acumznatum have retained the primitive color. 19. Ray corolla behavior after stigma receptivity: the corollas of post- receptive ray flowers in nearly all genera of the goldenaster lineage tightly coil inward, with the adaxial surface exposed. Only in Noticastrum, Osbertia, Chrysopsis pilosa, and C. tezana, do they remain straight. In the hypothesis here, the noncoiling behavior of first two apparently has evolved in parallel with that of the second two, perhaps reflecting an inherited tendency. In other species of Chrysopsis the tendency to coil may be more weakly expressed than in other goldenasters, but it is clear nevertheless. This behavior can be easily observed even on herbarium specimens; I have also observed C. pilosa and C. tezana in the field and in cultivation. Comparative observations for other species are not available, but in C. pilosa and C. tezana, the ray corollas close upwards in parallel at night, forming a narrow pyramid, spreading the next morning. 20, 21, and 22. Sweeping hair shape and apex: in contrast to the rest of the goldenasters, there is a definite tendency in Chrysopsis for the sweeping hairs of the disc style appendages to be lanceolate rather than narrowly oblong, at least near the appendage apex. In two pairs of species this tendency is markedly more pronounced than the other species: in C. terana and C. pilosa, the sweeping hairs are broadly lanceolate to nearly triangular with sharp points; in C. scabrella and C. floridana, they are linear-lanceolate with sharp points. These two specialized morphologies do not appear to be strictly homologous. Outside of the goldenasters, lanceolate hairs occur in some species of the Aster L. lineage, and the sweeping hair apices of some genera of the Machaeranthera Nees lineage show the same tendency to be sharp pointed (Nesom in prep.). 23. Elongate crystals in the tissues of the disc corolla throat: relatively large, elongate, straight sided crystals are found in the throat cells of all species of goldenaster except Chrysopsis terana and C. pilosa (see Nesom 1991a for further details). The corolla throats of C. pilosa produce a mixture of stellate “sand” crystals and elongate crystals but the latter are markedly smaller, ap- parently through reduction, than those in other goldenaster species. Chrysop- sts terana produces only “sand” crystals. 24. Resin ducts in disc corolla lobes: in the three species of the lattsquamea lineage, the yellow-orange resin ducts associated with the veins are sharply discontinuous. This is much less pronounced in C. latisquamea than in the other two species. 25. Radial cell walls: the difference between the sinuate walls and straight walls can be easily seen in the cells of the upper throat of the disc corollas. Outside of the goldenasters, sinuate walls are particularly characteristic of the disc corolla throats in some groups of Aster (Nesom in prep.) as well as related genera. Nesom: Union of Bradburia with Chrysopsis 121 26, 27, 28, 28, and 30. Achene morphology: the flattened, obovate achenes of Chrysopsis (including C. pilosa and Bradburia) with subepidermal ribbing, truncate apices and inset pappus insertion are distinctive. Raised, orange, glandular ridges occur in the three species of the latzsqguamea lineage and in C. gossypina; similar but much thinner and barely raised ridges occur in Het- erotheca sect. Heterotheca, where they appear to have arisen independently as a specialization within that genus. Relatively long achenes with numer- ous, superficial nerves appear to be primitive within the goldenasters, but the degree of compression is more evolutionarily plastic and less useful in recon- structing phylogeny. The achenes of Heterotheca sect. Ammodia, for example, are distinctly flattened; those of some species of Noticastrum also are strongly compressed, while others are intermediate in compression between Chrysopsis and Osbertia. In the present analysis, only Osbertia is scored as having terete achenes within the Chrysopsis lineage. The achenes of Bradburia are strongly 3 sided, at first sight very different from other taxa of Chrysopsis, but ray achenes that are slightly but distinctly 3 sided, in exactly the same way as Bradburia, are produced by C. mariana, C. scabrella, and C. lanuginosa, although apparently not by C. pilosa. 31 and 32. Pappus: C. pilosa has an outer series of broad and conspicuous scales. The outer series in C. mariana is also more scalelike than the other species of Chrysopsis but less pronounced than in C. pilosa. The analysis here indicates that the broad wings of the nearly scalelike pair of bristles of the sterile disc achenes in C. terana may be homologous with the scales of C. pilosa. The scaly outer pappus in some taxa of Heterotheca sect. Heterotheca is interpreted to have arisen as a specialization within that genus and in parallel with that in Chrysopsis. Among the goldenasters, a uniseriate pappus occurs only in Osbertia, where it is clearly derived. Phytologia (August 1991) 71(2):122-127. TRANSFER OF ASTER KINGII TO TONESTUS (ASTERACEAE: ASTEREAE) Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT Aster kingit is not accommodated in Machaeranthera or Xylorhiza, the only genera outside of Aster with which it has been formally aligned. Studies have returned it to Aster, its original position, but it has not been accepted even there in recent taxonomic overviews. Placements alternative to Aster have not been suggested by previous workers. Al- though it is white rayed, the species is transferred here to the primarily yellow rayed genus Tonestus, where it is hypothesized to be most closely related to the rayless T. aberrans. The new combinations required are T. kingii (D.C. Eaton) Nesom and T. kingii var. barnebyana (Welsh & Goodrich) Nesom. KEY WORDS: Tonestus, Aster, Machaeranthera, Asteraceae, Aster- eae Aster kingu D.C. Eaton, a species restricted to the Wasatch and Canyon Mountains of central to north central Utah, has been unsettled in its system- atic placement in the last 35 years. It was accepted by earlier North American taxonomists within the bounds of the large and variable genus Aster L., but Cronquist & Keck (1957) excluded the species from Aster and placed it in their expanded version of Machaeranthera Nees (within sect. Xylorhiza [Nutt.| Cronq. & Keck series Integrifoliae Cronq. & Keck). These authors noted (p. 233) that “The monocephalous montane true perennial ‘Aster’ king D.C. Eaton is suggestive of Xylorhiza glabriuscula, but the involucre is that of Machaeranthera, and the plant blooms in late summer like other Machaer- antherae.” The species of Machaeranthera sect. Xylorhiza are now treated as the genus Xylorhiza Nutt., mostly following the monograph by Watson (1977). Aster kingi, however, was noted by Watson (1978, p. 209) to be “phenologically, ecologically, morphologically, and chromosomally anomalous ...” within Xy- lorhiza, while he observed resemblances in various features between A. kingit 122 Nesom: Transfer of Aster kingw to Tonestus 123 and A. alpigenus (Torr. & Gray) A. Gray (in habit), A. conspicuus Lindl. (in phyllary morphology), and A. integrifolius Nutt. (in vestiture), although in 1977 he had suggested that it might prove to belong to Machaeranthera sect. Machaeranthera. Welsh (1983) accepted Watson’s later point of view and treated the species as Aster, transferring a varietal taxon (var. barnebyana, see below) named in Machaeranthera to a position in Aster. In a taxonomic conspectus of the genus Machaeranthera, Hartman (1990) also excluded A. kingit, leaving it in Aster, although he did not provide a commentary regard- ing its possible affinities there. In contrast, recent taxonomic and phyletic overviews of the genus Aster (Jones 1980; Semple & Brouillet 1980; Jones & Young 1983) did not provide a position for A. kingzt or even a mention of it in any context, although Dr. Jones (pers. comm.) now accepts it as a species of Aster sect. Oreostemma (E. Greene) Peck. The chromosome number of Aster kingii (n = 9 pairs; Watson 1978, for var. kingii; Sanderson et al. 1984, for var. barnebyana) contrasts with all species of Xylorhiza (c = 6) as well as those of Machaeranthera (z = 4, 5). Species in groups of Aster hypothesized by Jones & Young (1983) to be primitive within the genus have a base chromosome number of z = 9, and this is presumably the area where A. kingzi would find its closest phyletic affinities were it established that it lies within that lineage. Aster kingiw might be left in a broadly conceived Aster but its placement there would remain anomalous, as even in that heterogeneous group there are no species to which it can be unequivocally related. The three species of sect. Oreostemma appear to be relatively homogeneous, the plants all with nine pairs of chromosomes, producing a taproot and short branched caudex, strictly monocephalous stems, stems and leaves eglandular or with few, short stipitate glands, entire, linear leaves with three parallel veins, and flattened pappus bristles. The phyllaries in subg. Oreostemma are slightly keeled but the midvein near the phyllary base is usually sunken rather than raised. None of these features except the chromosome number and taproot are matched in A. kangit. On the other hand, Aster kingzi clearly is similar in a suite of features to the group of species recently consolidated as the genus Tonestus A. Nels. (Nesom & Morgan 1990): habitats rocky and at high altitudes, plants rhizomatous or taprooted with ascending, woody caudex branches, leaves obovate, reticulate veined, and coarsely toothed, the basal persistent, the cauline continuing to immediately below the heads, outer phyllaries foliaceous, at least the inner with a raised keel beginning at the phyllary base and with distinctively ex- tended, foliaceous apices, and achenes strigose, long, narrow, and multinerved, with a nearly symmetrical carpopodium and with a single series of pappus bris- tles terete to the very base and more or less even in length. The vestiture, particularly of the stems, is composed of very long (up to 0.7 mm), vitreous, biseriate, gland tipped trichomes (Type C trichomes, see Nesom 1976). Type 124 PAY TOL OGL A volume 71(2):122-127 August 1991 C trichomes are ubiquitous throughout the tribe, but such distinctively long ones are rare among American Astereae, where to my knowledge, outside of Tonestus they occur only in Aster integrifolius and A. modestus Lindl., Xy- lorhiza wrighti (A. Gray) E. Greene, and a few species of Solidago L. In other features, A. integrifolius is securely positioned within Aster but A. modestus is anomalous and more difficult to place; neither could be considered to be closely related to Tonestus. Xylorhiza is more similar to Machaeranthera Nees, and Solidago has been shown by DNA studies to be closely related to Tonestus (see Nesom et al. 1990 for a summary). Two species within Tonestus are atypical in their vestiture —- T. microcephalus (Cronq.) Nesom & Morgan is glabrous to glabrate and T. pygmaeus (Torr. & Gray) A. Nels. usually is eglandular — but they belong in the genus on the basis of overall morphology. The type species, T. lyallia (A. Gray) A. Nels., which is apparently one of the most specialized in the genus with its calvous achenes and (at least in some plants) somewhat flattened pappus bristles, also produces the very long glandular trichomes. Among the species of Tonestus, T. aberrans (A. Nels.) Nesom & Morgan is most similar to Aster kingu. Indeed, as a pair the two are somewhat set apart from the rest of the genus in their toothed leaves often with spinulose teeth, narrowly lanceolate-attenuate and apically spreading or reflexed phyllar- ies, style appendages with more widely arranged sweeping hairs, and purplish disc corolla lobes. Tonestus aberrans, however, has leaves more predominately cauline and consistently produces multi-headed capitulescences in contrast to the normally monocephalous stems of A. kingiz, although the latter occasion- ally may produce up to five heads in a loose corymb (fide Welsh 1983). Further, the leaves of T. aberrans are glandular pubescent while those of A. kingi are glabrous or glabrate, and the phyllaries of T. aberrans are distinctly graduated in length, while those of A. kingit are somewhat variable in this respect but also tend to be graduated. A conspicuous feature that has deterred consideration of Aster kingti as a member of any yellow rayed genus is its white ray color. The complete ab- sence of rays, however, in three species of Tonestus (T. aberrans, T. graniticus [Tiehm & Shultz] Nesom & Morgan, and T. alpinus [Anderson & Goodrich] Nesom & Morgan) has not previously prevented botanists from recognizing their close relationship with the yellow rayed species (Anderson 1980; Tiehm & Shultz 1985). If the hypothesis of close relationship regarding A. kingit and T. aberrans is correct, the purplish pigments in the disc corolla lobes of T. aberrans are perhaps unmasked by a reduced amount of yellow pigment. Tonestus graniticus has strongly graduated phyllaries and is probably closely related to T. aberrans. On the basis of its broader phyllaries of nearly even length, 7. alpinus is probably most closely related to T. eximius (H.M. Hall) A. Nels. & Macbr. and T. peirsoni (Keck) Nesom & Morgan. A second disparity, at first sight, between Aster kingiz and the species of Tonestus might be perceived in the relatively well defined taproot of the for- Nesom: Transfer of Aster kingu to Tonestus 125 mer. All but two species of Tonestus, however, produce woody, ascending caudex branches apparently arising from a single axis, although the latter is usually broken off of herbarium specimens, probably because of the difficulty in extracting them from their crevice habitats. Tonestus pygmaeus, however, is clearly taprooted with a multicipital crown and foreshortened caudex branches and is nearly identical in habit to A. king; T. microcephalus, and T. petrsoniu are taprooted with longer caudex branches. The relatively slender, more dif- fuse rhizomes of T. lyallii and T. ezimius do not arise from a central axis and in this respect are unusual in the genus. Within Tonestus, taproots have perhaps developed in the manner hypothesized by Welsh (1983), but a more detailed analysis of the generic phylogeny would be required to establish this with cer- tainty. In any case, other genera of Astereae accommodate both taprooted and rhizomatous species (e.g., Chrysopsis [Nutt.| Ell., Erigeron L., Grindelia Willd., Gutierrezia Lag., Machaeranthera Nees, Noticastrum DC. and numer- ous others). Within Aster, A. subulatus Michx. of subg. Ozytripolium (DC.) Torr. & Gray is a taprooted annual, its closest relatives rhizomatous. The species of Aster subg. Janthe (Torr. & Gray) A. Gray (or the genus Jonactis E. Greene) are variable in their production of short, woody rhizomes, caudex branches, and taproots (Nesom & Leary, submitted). The remaining species of Aster sensu lato are somewhat more uniform in their rhizomatous bases, al- though the cormoid rhizomes of subg. Virgulus (Raf.) A. Jones are distinctive. In summary, Aster kingii shares with species of Tonestus a significant num- ber of features, particularly including the distinctive vestiture of glandular tri- chomes with extremely long stipes. And in spite of its white rays, it fits more securely in that genus than in Aster, particularly when it is placed next to T. aberrans. This view of the taxonomic placement of A. kingzz is reflected in the following two nomenclatural combinations: Tonestus kingii (D.C. Eaton) Nesom, comb. nov. BASIONYM: Aster kingiz D.C. Eaton, Bot. Fortieth Parallel 5:141. 1871. Machaeranthera kingi (D.C. Eaton) Cronquist & Keck, Brittonia 9:239. 1957. Tonestus kingii (D.C. Eaton) Nesom var. barnebyana (Welsh & Goodrich) Nesom, comb. nov. BASIONYM: Machaeranthera kingi (D.C. Eaton) Cronquist & Keck var. barnebyana Welsh & Goodrich, Brittonia 33:299. 1981. Aster king D.C. Eaton var. barnebyana (Welsh & Goodrich) Welsh, Great Basin Naturalist 43:221. 1983. Var. barnebyana is weakly distinguished, but it is restricted to the south- western corner of the range of the species, and as noted in its original de- scription, its leaves are characteristically toothed and the biseriate glands are longer than in the typical variety. In these features, var. barnebyana is more similar than var. kingzi to related species within Tonestus. 126 PHY TOLOGIA volume 71(2):122-127 August 1991 ACKNOWLEDGMENTS I thank Dr. B.L. Turner for his review. Dr. Almut Jones also reviewed the manuscript, and although we do not agree on the placement of Aster kingit, her detailed and insightful comments have been helpful and are greatly appreciated. LITERATURE CITED Anderson, L.C. 1980. Haplopappus alpinus (Asteraceae): a new species from Nevada. Great Basin Naturalist 40:73-77. Cronquist, A. & D.D. Keck. 1957. A reconstitution of the genus Machaer- anthera. Brittonia 9:231-239. Hartman, R.L. 1990. A conspectus of Machaeranthera (Asteraceae: Aster- eae). Phytologia 68:439-465. Jones, A.G. 1980. A classification of the New World species of Aster (Aster- aceae). Brittonia 32:230-239. Jones, A.G. & D. Young. 1983. Generic concepts of Aster (Asteraceae): A comparison of cladistic, phenetic, and cytological approaches. Syst. Bot. 8:71-84. Nesom, G.L. 1976. A new species of Erigeron (Asteraceae) and its relatives in southwestern Utah. Brittonia 28:263-272. Nesom, G.L. & D. R. Morgan. 1990. Reinstatement of Tonestus (Astereae: Asteraceae). Phytologia 68:174-180. Nesom, G.L., D.R. Morgan, Y. Suh, & B.B. Simpson. 1990. Xylothamia (Asteraceae: Astereae), a new genus related to Futhamia. Sida 14:101- 116. Nesom, G.L. & T.J. Leary. Submitted. A new species of Jonactis (Asteraceae: Astereae) from southern Nevada. Brittonia Sanderson, S.C., S. Goodrich, & E.D. McArthur. 1984. Chromosome number reports LXXXV. Taxon 33:756-760. Semple, J.C. & L. Brouillet. 1980. A synopsis of North American asters: the subgenera, sections and subsections of Aster and Lasallea. Amer. J. Bot. 67:1010-1026. Nesom: Transfer of Aster kingi to Tonestus 127 Tiehm, A. & L.M. Shultz. 1985. A new Haplopappus (Asteraceae: Astereae) from Nevada. Brittonia 37:165-168. Watson, T.J., Jr. 1977. The taxonomy of Xylorhiza (Asteraceae-Astereae). Brittonia 29:199-216. ____. 1978. Chromosome numbers in Xylorhiza Nuttall (Asteraceae - Astereae). Madrono 25:205-210. Welsh, S.L. 1983. Utah flora: Compositae (Asteraceae). Great Basin Natu- ralist 43:179-357. Phytologia (August 1991) 71(2):128-131. TOMENTAURUM (ASTERACEAE: ASTEREAE), A NEW GENUS OF GOLDENASTER FROM CHIHUAHUA, MEXICO Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT ? Tomentaurum, a new monotypic genus of the goldenaster lineage, is based on Heterotheca vandevenderorum, with the new combination Tomentaurum vandevenderorum. These plants are endemic to west central Chihuahua, México and are distinguished by the follow- ing combination of features: rhizomatous, leaves mostly basal, obovate- oblanceolate, dense and closely white tomentose, heads large, solitary, on long scapes, rays yellow and coiling, disc corollas very long, achenes narrowly oblong and strongly compressed, with numerous, whitish, su- perficial nerves, and pappus double. In its dense vestiture of minutely filiform hairs arising from similarly thin bases, Tomentaurum is hypoth- esized to be most closely related to the North American genera Pityopsis and Croptilon. KEY WORDS: Tomentaurum, Heterotheca, Pityopsis, Astereae, Asteraceae, México In a review and phylogenetic analysis (Nesom 1991a) of the genera closely related to the goldenasters Heterotheca, Chrysopsis, and Pityopsis (sensu Sem- ple et al. 1980), it has become evident that a recently described species known by only a few collections from northwest México cannot be placed within any of the known goldenaster genera, although it is clearly a member of that phylad. Turner (1987) noted that the species was “very distinctive” but apart from placing it in Heterotheca, he did not offer an hypothesis regarding its possible closest relatives. The distinctive plants from Chihuahua are described here as a monotypic genus. 128 Nesom: Tomentaurum new genus of goldenaster 129 Tomentaurum Nesom, gen. nov. Type species: Heterotheca vandevendero- rum B. Turner Plantae rhizomatosae, folia praecipue basalia obovati-oblanceo- lata dense arcte albo tomentosa, capitula grandia solitaria in scapis longis, corollae radii flavae circinnatae, corollae disci longae, ache- nia longae valde compressae nervis numerosis albidis superficia- ribusque, et pappus duplex. Tomentaurum vandevenderorum (B. Turner) Nesom, comb. nov. BA- SIONYM: Heterotheca vandevenderorum B. Turner, Phytologia 63:127. 1987. TYPE: MEXICO. Chihuahua: Rio Mayo Region, El Capitan, 28° 13’ 30” N, 108° 07’ 30” W, 2000 m, 26 Jun 1986, P.S. Martin 56 (HOLOTYPE: TEX!; Isotype: ARIZ). Perennial herbs from slender, slightly woody rhizomes, these producing long, caudexlike branches with scale leaves; stems, leaves, and phyllaries dense- ly invested with short, biseriate trichomes (Type C trichomes, see Nesom 1991a), each with a few celled, orange resinous head, the glands sometimes nearly sessile on the phyllaries, stems and leaves tomentose with very long, uniseriate, few celled, minutely filiform white hairs from equally small bases (Type B trichomes), completely lacking coarser uniseriate trichomes (“oste- olate” or Type A trichomes). Stems unbranched, erect, stipitate glandular, closely tomentose. Leaves in a basal rosette or on the basal 3-10 cm of the stems, oblanceolate-obovate, 2-4 cm long, 3-9 mm wide, entire, the cauline subclasping, densely woolly tomentose, tardily glabrescent above. Heads 16- 20 mm wide, solitary on scapose peduncles 12-25 cm long; receptacles very shallow convex, barely foveolate; phyllaries narrowly ovate-lanceolate, strongly graduated in 5-6 series, the inner 10-13 mm long, the margins purplish, with a prominent, scarious flange, the outer herbaceous, the middle strongly keeled, herbaceous only at the apex. Ray flowers 16-21, pistillate, fertile, yellow, 10-13 mm long, the ligules 2.0-2.5 mm wide, tightly coiling after receptivity of the ray stigmas. Disc corollas perfect, tubular, 8-9 mm long, the lobes with unis- eriate, acicular trichomes, with elongate, straight sided crystals in the tissues of the lower throat; style branches 1.4-1.7 mm long, with linear-lanceolate ap- pendages 1/3-1/2 the length of the branches. Achenes densely white-sericeous, 4.5-5.5 mm long, 0.6-0.7 mm wide, strongly compressed, with 7-9 thin, closely adjacent, superficial, white nerves on each side, one or two of these often slightly thicker than the others; carpopodium strongly asymmetric; pappus of 45-60 white, barbellate bristles in several series, with a few, inconspicuous setae or very slightly widened bristles 0.5-1.5 mm long. Chromosome number unknown. Endemic to the plateau in the area of Basaseachic, Chihuahua; rocky stream beds or sandy-cobbly soil, area of pine woodlands; 2000-2100 m; flow- ering July-September. 130 PHY TOLO GIA volume 71(2):128-131 August 1991 Additional collections examined: MEXICO. Chihuahua: Rancho El Capitan, 28° 13’ 30” N, 108° 07’ 30” W, 22 Jun 1987, T.R. & R.K. Van Devender 87-134 with P. S. Martin (TEX); N edge of town of Basaseachic on Rio de Basaseachic, 28° 12’ 40” N, 108° 12’ 50” W, 24 Jun 1987, T.R. & R.K. Van Devender 87-165 with P. S. Martin (TEX). Tomentaurum is a member of the goldenaster lineage (Nesom 1991a), as evidenced by its stipitate glandular herbage, keeled phyllaries, ray flowers with yellow, coiling ligules, disc corolla throats with elongate crystals, linear- lanceolate disc style appendages, achenes with asymmetric carpopodia, and the pappus with an outer series of members much shorter than the inner. The narrow achenes with numerous, superficial nerves appear to be of relatively primitive morphology (Nesom 1991a). Plants of Tomentaurum are habitally distinctive: they are rhizomatous, the rosettes arising from somewhat herbaceous, slender caudex branches with scale leaves, the basal leaves persistent, obovate-oblanceolate, densely and closely white tomentose, and the scapose stems long and unbranched, bearing large, solitary heads. The only goldenaster species that are somewhat similar in habit (ie., rhizomatous, basal leaves persistent, monocephalous) are Heterotheca chthuahuana (Turner & Sundberg) B. Turner and those of Osbertza E. Greene and Noticastrum DC. (Nesom 1991b), but all of these differ strongly from Tomentaurum in significant features (Nesom 1991a). In its dense tomentum of highly elaborated, minutely filiform hairs arising from similarly thin bases (Type B trichomes; see Nesom 1991a for trichome terminology), Tomentaurum is hypothesized to be closely related to the North American genera Pityopsis Nutt. and Croptilon Rafin. (Nesom 1991la). To- mentaurum differs from both of the latter genera, particularly in its less spe- cialized leaves (shape, venation, and anatomy), glandular trichomes with much less highly elaborated heads, large, solitary capitula, and strongly compressed achenes. All species of Heterotheca (except H. chthuahuana and H. mucronata Harms ez Turner) are taprooted and produce relatively short, thick walled Type A trichomes mixed with an understory of minute, inconspicuous Type B trichomes. ACKNOWLEDGMENTS I thank Dr. B.L. Turner and Dr. Andrew McDonald for their review and comments on the manuscript. Nesom: Tomentaurum new genus of goldenaster 131 LITERATURE CITED Nesom, G.L. 1991a. An phylogenetic hypothesis for the goldenasters (Aster- aceae: Astereae). Phytologia 71:136-151. Nesom, G.L. 1991b. Transfer of Heterotheca bartlettii to Osbertia (Aster- aceae: Astereae). Phytologia 71:132-135. Semple, J.C., V.C. Blok, & P. Heiman. 1980. Morphological, anatomical, habit and habitat differences among the goldenaster genera Chrysop- sis, Heterotheca, and Pityopsis (Compositae-Astereae). Canad. J. Bot. 58:147-163. Turner, B.L. 1987. New species and combinations in Mexican Heterotheca (Asteraceae-Astereae). Phytologia 63:127-128. Phytologie (August 1991) 71(2):132-135. TRANSFER OF HETEROTHECA BARTLETTII TO OSBERTIA (ASTERACEAE: ASTEREAE) Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT Heterotheca bartletti1, endemic to Tamaulipas and Nuevo Leon, Méx- ico, is transferred to the genus Osbertia. The latter has been monotypic for most of its taxonomic history, with the exception of one species re- cently added to it but subsequently transferred to Heterotheca. With the addition of O. bartlettii, comb. nov., Osbertia again becomes di- typic. KEY WORDS: Osbertia, Heterotheca, Haplopappus, Asteraceae, Astereae, México In the process of preparing a taxonomic treatment of the Mexican genera of goldenasters for the “Asteraceae of México” (Turner & Nesom, in prep.) as well as a phylogenetic analysis of the whole goldenaster lineage (Nesom 1991), it has become apparent that Heterotheca bartletti (S.F. Blake) M.C. Johnston (= Aplopappus bartletti S.F. Blake) is in need of taxonomic realignment. In the original description of (H)Aplopappus bartlettu, Blake noted that the uniseriate pappus was the single feature that argued for the placement of the species in Haplopappus DC. rather than Chrysopsis (Nutt.) Ell. He referred the species to Haplopappus sect. Isopappus (Torr. & Gray) Hall (= Croptilon Rafin.), but it is not clear why he did not ally it instead with Haplopappus sect. Osbertia(DC.) Hall, because he recognized the large difference between H. bartlettz and the species of sect. Jsopappus. Johnston transferred the species to Heterotheca Cass., positioning it with phylogenetically close relatives, includ- ing some species in which the outer pappus is obscure. Smith (1965) agreed in excluding Haplopappus bartlettii from sect. Jsopappus, accepting its placement in Heterotheca pending further study. Turner & Sundberg (1986) did not mention Heterotheca bartletti in their formal treatment of Osbertia DC. They did, however, add a second species, O. chihuahuana Turner & Sundberg, to the single original one, O. stolonifera 132 Nesom: Transfer of Heterotheca bartlettzi to Osbertia 133 (DC.) E. Greene. In a re-examination of O. chthuahuana, Turner (1987) has transferred it to Heterotheca, where although it is still somewhat anomalous in habit, it is more naturally situated in its overall combination of characters, particularly in its vestiture. On the other hand, Heterotheca bartlettz differs from Heterotheca (including sects. Heterotheca, Phyllotheca [Nutt.] Harms, and Ammodia [Nutt.] Harms) and all of the other genera of the goldenaster lineage in a combination of char- acters that in turn match those in Osbertia stolonifera: uniseriate, long, thin walled, and vitreous trichomes (Type A trichomes, see Nesom 1991a for termi- nology), herbaceous phyllaries without an indurated keel, reddish tinged ray flowers with noncoiling ligules, subterete, nearly cylindric achenes with numer- ous, superficial nerves, and a simple pappus. The achenes of Osbertza, with numerous, superficial nerves, are primitive within the goldenaster lineage. The simple pappus and the herbaceous, nearly unkeeled phyllaries are found only in Osbertia, but several features of O. stolonifera and H. bartletti are shared with species of two closely related genera (Nesom 1991b): the distinctive Type A trichomes occur in the South American genus Noticastrum DC. and in all species of Chrysopsis; noncoiling ray ligules are found in Noticastrum as well as a pair of species of Chrysopsis; reddish pigments in the ray corollas occur in Osbertia and Noticastrum. An additional similarity between Osbertia stolonifera and Heterotheca bart- lettii is found in their disc corollas, which in both species produce elongated crystals from the very base of the tube to the lower throat; stellate “sand” crystals are not produced at all. The disc corollas of all other genera of the goldenaster lineage typically (and diagnostically) produce elongated crystals (Nesom 1991a), but only in Osbertia are “sand” crystals completely lacking. The chromosome number of Osbertia stolonifera has been reported as n = 5 pairs (commonly) and n = 4 pairs (rarely) (see Turner & Sundberg 1986, for numerous records); no chromosome count is available for Heterotheca bartletti, but this information will be significant in further substantiating the compari- son between the two species. In summary, Heterotheca bartlettii is such a close morphological match for Osbertia stolonifera that the two are best placed in the same genus. Osbertia bartlettii (S.F. Blake) Nesom, comb. nov. BASIONYM: Aplopap- pus bartlettii S.F. Blake, J. Washington Acad. Sci. 22:328. 1932. TYPE: MEXICO. Tamaulipas: Above La Vegonia near San Jose, 1100 m, 3 Jul 1930, Bartlett 10046 (HOLOTYPE: MICH; Isotype: US!). Heterotheca bartletti: (S.F. Blake) M.C. Johnston, Southw. Nat. 2:172. 1958. Perennial herbs from slender, woody rhizomes, the stems, leaves, and phyl- laries prominently invested with long, stipitate glandular hairs, also sparsely and closely villous-puberulent with minute, nonglandular hairs, sparsely long 134 PHY POLO GTA volume 71(2):132-135 August 1991 pilose with vitreous, sometimes flattened and twisted trichomes (Type A) aris- ing from thin bases. Stems 8-25 cm tall, often scapose when the cauline leaves strongly reduced. Basal leaves oblanceolate, persistent, the cauline with sub- clasping bases, when present restricted to the lower third of the stem. Heads 7-10 mm wide, usually solitary on long, naked peduncles 3-10 cm long; recep- tacles smooth or very slightly foveolate; phyllaries linear-lanceolate, 1 nerved, the nerve sometimes orangish, not keeled, with a herbaceous midregion and indurated margins from base to tip, strongly graduated. Ray flowers 23-33, the corollas 8-10 mm long, usually drying with an abaxial, purplish midstripe, not coiling. Disc corollas 4.0-4.5 mm long, the lobes minutely viscid glandu- lar; style appendages linear-lanceolate, 0.6-0.8 mm long, 1/3-1/2 the length of the style branches. Achenes sparsely short strigose, cylindric, terete to slightly compressed, with ca. 16, slender, evenly spaced, superficial nerves; carpopodium asymmetric; pappus of ray and disc achenes uniseriate, of 15-20 barbellate bristles, without an outer series. Nuevo Leon and Tamaulipas, México; meadows or openings, oak, oak-pine, or pine woodlands, 900-1800 m; flowering July to September. Additional collections examined: MEXICO. Nuevo Leén: Dulces Nombres, and just E of border into Tamaulipas, 20 Jul 1948, Meyer & Rogers 2827(MO); Mpio. Villa Santiago, Canon Denuncio, Rancho La Bolla, upper canyon, 22 Jun 1935, Mueller 2008 (MO); Mpio. Villa Santiago, trail between Potrero Redondo and Laguna Sanchez, 16 Aug 1939, Mueller 2721 (MO); Mpio. Montemorelos, La Trinidad, 19 Aug 1939, Muller 2854 (LL). Tamaulipas: Sierra de San Carlos, Cerro El Diente, 7 Jun 1985, Jiménez 268 (TEX); road from Vicente Guerrero toward El Molino, 30 km W of Victoria, 23 Sep 1985, Yanez 547 (TEX); Sierra de San Carlos, 5 mi S of San Carlos, N side of Bufa El Diente, 15 Apr 1988, Nesom 6296 (MEXU, TEX). Standley marked collections of Mueller 2008 (cited above) as the type of a species of Chrysopsis, using the epithet “longipes,” but this name apparently was never published. Osbertia is most closely related to the genera Chrysopsis of the eastern and south central United States and Noticastrum of South America (Nesom 1991a, 1991b). The two species of Osbertia are restricted primarily to eastern México and northern Central America, where they have allopatric, nearly contiguous geographic ranges. They can be distinguished by the following contrasts: 1. Plants with short, thick rhizomes and leafy runners producing terminal plantlets; heads 16-30 mm wide; ray flowers 34-150, 15-25 mm long; southern Nuevo Leén to Guatemala. ...............00008 O. stolonifera 1. Plants with slender, woody rhizomes, without runners; heads 7-10 mm wide; ray flowers 23-33, 8-10 mm long; west-central Tamaulipas and adjacent Nuevo Leon? 2.00055 220.034. 2 aia eee O. bartlettz Nesom: Transfer of Heterotheca bartlettu to Osbertia 135 ACKNOWLEDGMENTS I thank Dr. B.L. Turner and Dr. Andrew McDonald for their review and comments on the manuscript. LITERATURE CITED Johnston, M.C. 1957. Heterotheca bartlettii (Blake) M.C. Johnston, new combination. Southw. Nat. 2:172. Nesom, G.L. 1991a. A phylogenetic hypothesis for the goldenasters (Aster- aceae: Astereae). Phytologia 71:136-151. ______. 1991b. Union of Bradburia with Chrysopsis (Asteraceae: Astereae), with a phylogenetic hypothesis for Chrysopsis. Phytologia 71:109-121. Smith, E.B. 1965. Taxonomy of Haplopappus, section /sopappus (Composi- tae). Rhodora 67:217-238. Turner, B.L. 1987. New species and combinations in Mexican Heterotheca (Asteraceae-Astereae). Phytologia 63:127-128. & S.D. Sundberg. 1986. Systematic study of Osbertia (Asteraceae- Astereae). Pl. Syst. Evol. 151:229-239. Phytologia (August 1991) 71(2):136-151. A PHYLOGENETIC HYPOTHESIS FOR THE GOLDENASTERS (ASTERACEAE: ASTEREAE) Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT The seven genera of the goldenaster lineage are included in a phy- logenetic analysis based primarily on morphological data. They are hypothesized to form three separate clades: (1) Chrysopsis (includ- ing Bradburia), Osbertia, and Noticastrum, (2) Croptilon and Pityop- sis, which are sister genera, and Tomentaurum, and (3) Heterotheca (including sects. Ammodia, Heterotheca, and Phyllotheca). The genus most closely related to the goldenaster lineage is taken to be Jonactis (= Aster sect. Janthe). KEY WORDS: Heterotheca, Ionactis, Aster, Asteraceae, Astereae Recent studies of restriction site variation in chloroplast DNA of North American Astereae (Suh 1989; Morgan 1990) have produced a relatively clear delimitation of a phyletically coherent group of genera termed the “golde- nasters.” Suh’s sample included Croptilon Rafin., Pityopsis Nutt., Heterotheca Cass. (sect. Heterotheca), Bradburia Torr. & Gray, and Chrysopsis (Nutt.) Ell.; Morgan’s sample included Croptilon and Osbertia E. Greene. Semple et al. (1980) postulated that Chrysopsis and Bradburia are part of an evo- lutionary lineage completely apart from the others, but for the most part, recent treatments have recognized the goldenaster genera as a natural group, at least implicitly, since many authors have viewed Chrysopsis, Pityopsis, and Heterotheca (including sects. Heterotheca, Phyllotheca [Nutt.] Harms, and Am- modia [Nutt.] Harms) in various combinations as only one or two genera (see Semple 1977 for a review). The studies of Semple and colleagues have been in- strumental in illuminating the distinctions among these taxa (see Semple et al. 1980 for a review and overview) and have provided separate taxonomic studies of Chrysopsis (Semple 1981), Pityopsis (Semple & Bowers 1985), Heterotheca sect. Ammodza (Semple 1988), and Bradburia (Semple & Chinnappa 1984). Other of the genera have also been the subjects of relatively recent taxonomic 136 Nesom: Phylogenetic hypothesis for goldenasters 137 treatments (Croptilon, Smith 1966, 1981; Osbertia, Turner & Sundberg 1986, Nesom 1991a; Heterotheca sect. Heterotheca, Nesom 1990), and Nesom (1991c) has presented evidence for uniting the monotypic Bradburia with Chrysopsis in a phylogenetic analysis of the latter. In addition to the genera above, Turner & Sundberg (1986) noted that the South American genus Noticastrum DC. (see Cuatrecasas 1973; Zardini 1978) probably should be included as a member of the goldenaster lineage, and evidence from the present study supports its inclusion. Tomentaurum Nesom, a newly described monotypic genus from western Mexico has been added to the goldenasters (Nesom 1991b). There is further indication, based on the cpDNA studies of Suh (1989) and Morgan (1990), that the white rayed Boltonia L’Herit. and Chloracantha Nesom et al. are more closely related to the goldenasters than to any other species in their studies, although they are highly divergent in morphology and appear to be weakly associated somewhere at the very base of the phylad (Nesom et al. 1991). They are not included in the present analysis. The molecular studies of Suh & Morgan placed the goldenaster lineage most closely to the Machaeranthera Nees and Aster L. lineages (see Nesom et al. 1989, fora summary). Neither of their studies, however, included a wide range of species traditionally accommodated in Aster. A morphologically based phy- logenetic overview of Aster and related genera (Nesom in prep.) hypothesizes that the taxa most similar to the goldenasters lie among the groups of Aster not included in the recent molecular analyses, those with carinate phyllaries, a double pappus, and a base chromosome number of z = 9. Among these, Aster subg. Janthe (Torr. & Gray] A. Gray, with strongly asymmetric carpopodia, is hypothesized to be the sister group to the goldenasters. Additional comments are provided by Nesom & Leary (submitted), who treat the group at generic rank as Jonactis E. Greene and add a fourth species to it. The studies noted above have offered various ideas regarding relationships of genera within the goldenaster group, but none have included all of the taxa involved. The present study confirms the evolutionary unity of the seven genera of goldenasters on a morphological basis and provides a phylogenetic hypothesis for all of them. CLADISTIC ANALYSIS - METHODS Studies of microcharacters were made with a compound microscope. Glands, trichomes, foliar veins were studied from epidermal “skims” and cross sections made with a razor from stems and leaves; these and dissected flowers have been mounted on slides in Hoyer’s Medium. A list of specimens from which slides have been made is deposited at TEX; all voucher specimens are deposited in TEX. Characters and character states used in the analysis are given in Table 138 PHY DOL OG FA volume 71(2):136-151 August 1991 1; coding for the individual taxa are in Table 2; discussion of variability in the characters is provided in the Appendix. The data were analyzed using PAUP (Swofford 1985) with Wagner parsimony. Some decisions regarding evolutionarily primitive character states in Chry- sopsis, which comprises two sharply disparate sections, have been made in a separate analysis (Nesom 1991c). Section Phyllotheca of Heterotheca is a variable group and is represented in the present analysis by two species, H. mezicana Harms ez B. Turner and H. stenophylla (A. Gray) Shinners, these chosen to broaden the variability included in the character scoring. The genus Ionactis serves as the outgroup in the present analysis, since it appears to be the group most closely similar and probably most closely related to the goldenasters. RESULTS and DISCUSSION The goldenaster lineage is characterized by a number of morphological fea- tures, some shared with related groups and some that distinguish it from the Aster and Machaeranthera lineages: stipitate glandular herbage, carinate phy]- laries, yellow ray flowers, throat tissues of disc corollas with numerous, straight sided, elongate crystals, disc style branches have linear-lanceolate appendages, multinerved achenes, asymmetric carpopodia, and pappus of slender, basally terete bristles with a much shorter outer series. In Heterotheca and Chrysop- sis, the heads are borne in a corymboid cyme arising from the distal portion of the main branches, with a tendency to be somewhat more loosely arranged in Pityopsis and Croptilon. Further, it seems clear that the base chromosome number for the group is z = 9. Of particular significance in this study has been the discovery that the goldenaster genera appear to stand apart from almost all other North American Astereae in their production of long, straight sided crystals in the tissues of the disc corolla throats. Similar crystals have been observed elsewhere only in Xylorhiza Nutt., Grindelia Willd., and Prionopsis Nutt. (Nesom et al. submitted), which are related to Machaeranthera rather than the goldenasters (Morgan 1990). Further details regarding these crystals are provided in the Appendix. A single tree of 36 steps (Fig. 1) accounts for the distribution of character states scored in the present analysis. Three primary lineages are defined, referred to here as the Chrysopsis, Heterotheca, and Pityopsis lineages, since the present study fully corroborates the earlier conclusion by Semple et al. (1980) that each of these three genera is phyletically distinct from the other two. The single character (no. 18, vestiture of the disc corolla lobes) that unites the Heterotheca and Pityopsis lineages is the most homoplasious of all those in the present study, and it must be viewed as nearly weightless evidence in resolving what otherwise would be a basal trichotomy. If it were assumed Nesom: Phylogenetic hypothesis for goldenasters 139 TABLE 1. Characters and character states. 8. o: 10. ll. 12. 13. 14. 15. . Chromosome number, (0) z = 9 or 7 (1) z= 5 or 4 . Taproot or rhizomes, (0) rhizomes (1) taproot . Leaf shape and venation pattern, (0) obovate to oblanceolate, net veined 1) linear, parallel veined or with a strong tendency for parallel veins & Leaf venation/anatomy, (0) veins not strongly sclerenchymatous, not raised above the lamina (1) veins with associated massive sclerenchyma, prominently superficial and forming ridges, Biseriate (Type C) trichomes, (0) without a differentiated head or the head merely 2 celled, the only form of Type C trichomes present (1) at least some with a highly elaborated head 4-8 cells wide at the apex, on a distinct stalk . Type A trichomes on herbage, (0) abundant (1) few in number or com- pletely absent . Type A trichomes, (0) thick-walled, terete, or absent (1) thin walled, often flattened and twisted Type A trichomes - outer walls, (0) smooth or absent (1) papillate Type B trichomes, (0) minute, less than 0.5 mm long (1) filamentous, highly elaborated in length, sometimes forming a dense, close, tomentum Head arrangement, (0) numerous in a corymboid cyme arising from branches on the upper third of the primary stem (1) solitary on mono- cephalous stems Head width, (0) mostly (8-)10-20 mm (1) mostly 4-7 mm Phyllary morphology, (0) indurated at least basally, with a narrow, raised, white indurated keel (1) herbaceous throughout, or at least the medial nerve not at all keeled Ray corolla color, (0) bluish (1) yellow Ray corolla color, (0) bluish or yellow (1) yellow, tinged with reddish purple Ray corolla behavior after stigma receptivity, (0) coiling (1) remaining straight 140 PAH YTOLOCTA volume 71(2):136-151 August 1991 TABLE 1. (continued). 16 Pie 18. 1: 20. PALE 22. 23. 24. 20. 26. Cells of disc corolla throat, (0) with stellate “sand” crystals or without crystals (1) with elongate, straight sided crystals Radial wall of cells of disc corolla throat, (0) broadly sinuate (1) straight Disc corolla lobes, vestiture (0) glabrous or with biseriate, glandular hairs (1) with uniseriate, acicular hairs, commonly also with obscure glandular hairs Vascular bundles in collecting appendages of disc flower style branches, (0) unsclerified (1) sclerified Disc achene surface, (0) with nerves subepidermal or raised but not resinous (1) with numerous, thin, slightly resinous nerves flush with sur- face Disc achene surface, (0) with (2-)8-26 thin, superficial, nerves, or the nerves not thick and rounded (1) with thick, rounded ridges, the nerves completely below the epidermal surface Disc achene nervation, (0) (6-)8-26 nerved (1) 2-(3-6) nerved Achene shape, side view, (0) narrowly oblong to oblong elliptic, symmet- tic (1) obovate, asymmetric Disc achene shape, (0) distinctly compressed (1) terete or subterete Pappus series, (0) double, the outer much shorter than the inner and varying from setae to linear scales (1) single, of bristles relatively even in length Pappus bristle shape in cross section, (0) terete (1) flattened Nesom: Phylogenetic hypothesis for goldenasters 141 TABLE 2. Data matrix for the goldenaster taxa analyzed. Character State | 00000 00000 00000 00000 00000 01001 00100 00100 11110 01000 01001 00100 00100 11111 01000 01001 00100 00100 11011 01100 01001 00100 00100 11011 01000 70001 10011 00100 11110 01000 00111 10010 10100 11110 01010 01111 10010 10100 11010 01011 07001 01001 00111 11110 01000 10001 01001 01111 10010 01011 Tonactis (IONACT) Sect. Ammodia (AMMODI) Heterotheca mezicana (HMEXIC) Sect. Heterotheca (HETERO) Heterotheca stenophylla (HSTENO) Tomentaurum (TOMENT) Pityopsis (PITYOP) Croptilon (CROPTI) Noticastrum (NOTICA) Osbertia (OSBERT) Chrysopsis (CHRYSO) eo qo © © oOo Hk = — © that the Type A trichomes were present in the goldenaster ancestor and lost twice in the Chrysopsis lineage, the trichotomy would remain unresolved with the addition of a single additional step in the same character. Each of the three goldenaster lineages is distinguished primarily by the.na- ture of its vestiture (see Nesom 1976 and further comments in the Appendix). All of the goldenaster taxa usually have stipitate glandular hairs, but those of the Chrysopsis lineage are characterized by their relatively long Type A trichomes with thin, smooth walls, those of the Heterotheca lineage by Type A trichomes with papillate outer walls, and those of the Pityopsis lineage by the the absence or near absence of Type A trichomes and the presence instead of greatly lengthened and prominent Type B trichomes. Apart from the relationship of the three primary lineages, the goldenaster phylogeny appears to be relatively well resolved, based on the data available, but several other problematic and unsatisfying points remain. First, the place- ment of Osbertia and Chrysopsis as sister species, rather than Osbertza and Noticastrum, would be more consistent with a well known geographic pattern of diversification, where closely related taxa occur both in the southwestern United States and in eastern Mexico, and it would not require a parallel re- duction in chromosome number from z = 9 to z= 5. A study of the z= 5 Chrysopsis and z = 5 Osbertia karyotypes might prove to be critical in this interpretation. Second, Tomentaurum is so very different from Pityopsis and Croptilon in habit, leaf morphology, and other features, as well as different from the probable ancestor to the whole goldenaster phylad, that it would not be surprising if it were placed, on the basis of additional evidence, closer to any of the other genera. Finally, Heterotheca (sensu Semple et al. 1980 and the 142 PH Y¥T.O.L.O.G 1A volume 71(2):136-151 HSTENO HETERO HMEXIC = CROPTI PITYOP lg io AMMODI 2 L 20 NOTICA OSBERT pha SS ot ® 1” 4. > ¥ ° TOMENT So © a 26 “0 9 8 © 2 @ > IONACT is? ‘> “6 ee, August 1991 7 CHRYSO Figure 1. Phylogenetic reconstruction of the goldenaster lineage. Abbrevi- ations for the taxa follow those in Table 2. Single slash = apomorphy or autapomorphy; double slash = parallelism; downward arrow = reversal. Nesom: Phylogenetic hypothesis for goldenasters 143 present analysis) can be maintained as a single genus, but sect. Phyllotheca probably is paraphyletic and better treated as several separate sections (the data of this analysis do not show this as clearly as possible). Alternatively, the monotypic Heterotheca sect. Ammodia (H. oregona [Nutt.] Shinners) might justifiably be segregated as a monotypic genus, as did Nuttall originally. The cladistic topology postulated here, based on morphology, differs slightly from a relatively unresolved cladogram based on molecular data (Suh 1989), although the latter analysis included only four of the taxa treated here. Suh positioned Heterotheca sect. Heterotheca, Pityopsis, and Croptilon as a tri- chotomy that is phyletically coordinate with the species pair Chrysopsis pilosa Nutt. and C. terana Nesom (= Bradburia). Both of the latter two are treated by Nesom (1991c) as Chrysopsis sect. Bradburia (Torr. & Gray) Nesom. Mor- gan’s analysis (1990) included only two goldenaster genera and placed Osbertia coordinate with several species of Croptilon. ACKNOWLEDGMENTS I thank Dr. B.L. Turner and Dr. Andrew McDonald for their review and comments on the manuscript, and I am grateful to Dr. P.O. Karis for pointing out the variability and potential utility in the crystals within corolla tissues. LITERATURE CITED Carlquist, S. 1961. Comparative Plant Anatomy. Holt, Rinehart, & Winston, New York. Cuatrecasas, J. 1973. Miscellaneous notes on Neotropical Flora, III. Phy- tologia 25:249-256. Drury, D.G. & L. Watson. 1966. Taxonomic implications of a comparative anatomical study of Inuloideae-Compositae. Amer. J. Bot. 53:828-833. Hartman, R.L. 1990. A conspectus of Machaeranthera (Asteraceae: Aster- eae). Phytologia 68:439-465. Jones, A.G. 1980. A classification of the New World species of Aster (Aster- aceae). Brittonia 32:230-239. Jones, A.G. & D. Young. 1983. Generic concepts of Aster (Asteraceae): A comparison of cladistic, phenetic, and cytological approaches. Syst. Bot. 8:71-84. 144 PHY FOLOCGIA volume 71(2):136-151 August 1991 Morgan, D.R. 1990. A systematic study of Machaeranthera (Asteraceae) and related groups using restriction analysis of chloroplast DNA and a taxo- nomic revision of Machaeranthera section Psilactis. Ph.D. dissertation, University of Texas, Austin. Nesom, G.L. 1976. A new species of Erigeron (Asteraceae) and its relatives in southwestern Utah. Brittonia 28:263-272. . 1989. Infrageneric taxonomy of New World Erigeron (Compositae: Astereae). Phytologia 67:67-93. 1990. Taxonomy of Heterotheca sect. Heterotheca (Asteraceae: Astereae) in México, with comments on the taxa of the United States. Phytologia 69:282-294. . 199la. Transfer of Heterotheca bartletti: to Osbertia (Asteraceae: Astereae). Phytologia 71:132-135. 1991b. Tomentaurum (Asteraceae: Astereae), a new genus of goldenaster from Mexico. Phytologia 71:128-131. . 1991c. Union of Bradburia with Chrysopsis (Asteraceae: Astereae), with a phylogenetic hypothesis for Chrysopsis. Phytologia 71:109-121. . 1991d. Transfer of Aster kingii to Tonestus (Asteraceae: Astereae). Phytolagia 71:122-127. . In prep. An overview of the phylogeny of Aster sensu lato (Aster- aceae: Astereae). Nesom, G.L. & D.R. Morgan. 1990. Reinstatement of Tonestus (Astereae: Asteraceae). Phytologia 68:174-180. Nesom, G.L., Y. Suh, D.R. Morgan, & B.B. Simpson. 1989. Xylothamia (Astereae: Asteraceae), a new genus related to Euthamia. Sida 14:101- 116. Nesom, G.L., Y. Suh, & B.B. Simpson. Submitted. The phylogenetic position of Stephanodoria (Asteraceae: Astereae). Brittonia Nesom, G.L., Y. Suh, D-R. Morgan, S.D. Sundberg, & B.B. Simpson. 1991. Chloracantha, a new genus of North American Astereae (Asteraceae). Phytologia 70:371-381. Nesom, G.L. & P.J. Leary. Submitted. A new species of Jonactis (Asteraceae: Astereae) from southern Nevada. Brittonia. Nesom: Phylogenetic hypothesis for goldenasters 145 Semple, J.C. 1977. Cytotaxonomy of Chrysopsis and Heterotheca (Compositae- Astereae): a new interpretation of phylogeny. Canad. J. Bot. 55:2503- 2513. . 1981. A revision of the goldenaster genus Chrysopsis (Nutt.) Ell. nom. cons. (Compositae-Astereae). Rhodora 83:323-384. . 1984. Observations on the cytology, morphology, and ecology of Bradburia hirtella (Compositae-Astereae). Syst. Bot. 9:95-101. . 1988. Heterotheca sect. Ammodia (Compositae: Astereae): A mul- tivariate study of H. oregona and specimens of Brewer’s (golden)aster. Syst. Bot. 13:547-558. Semple, J.C., V.C. Blok, & P. Heiman. 1980. Morphological, anatomical, habit, and habitat differences among the goldenaster genera Chrysop- sis, Heterotheca, and Pityopsis (Compositae-Astereae). Canad. J. Bot. 58:147-163. Semple, J.C. & F.D. Bowers. 1985. A revision of the goidenaster genus Pityopsis Nutt. (Compositae: Astereae). Univ. Waterloo Biol. Ser. 29:1- 34. Semple, J.C. & C.C. Chinnappa. 1980. Karyotype evolution and chromosome numbers in Chrysopsis (Nutt.) Ell. sensu Semple (Compositae-Astereae). Canad. J. Bot. 58:164-171. Smith, E.B. 1966. Cytogenetics and phylogeny of Haplopappus section Isopappus (Compositae). Canad. J. Genet. Cytol. 8:14-36. . 1981. New combinations in Croptilon (Compositae-Astereae). Sida 9:59-63. Stirton, C.H. 1983. Nocturnal petal movements in the Asteraceae. Bothalia 14:1003-1006. Suh, Y. 1989. Phylogenetic studies of North American Astereae (Asteraceae) based on chloroplast DNA. Ph.D. dissertation, University of Texas, Austin. Swofford, D.L. 1985. PAUP: Phylogenetic analysis using parsimony, Vers. 2.4. Illinois Nat. Hist. Survey, Champaign, Illinois. Turner, B.L. & S.D. Sundberg. 1986. Systematic study of Osbertia ( Asteraceae- Astereae). Pl. Syst. Evol. 151:229-239. Zardini, E.M. 1978. Novedades en el genéro Noticastrum (Compositae - Astereae). Hickenia 1:134-138. 146 PHYTOLOGIA volume 71(2):136-151 August 1991 APPENDIX 1. Notes on character variation 1. The ancestral chromosome number for the Heterotheca lineage appears to be z = 9, based on the hypothesis that Noticastrum is closely related to Osbertia and Chrysopsis (Nesom 1991c), as well as on the taxa presumably most closely related to the goldenasters, including Jonactis and others (Nesom in prep.). The species of Croptilon have an aneuploid series of n = 7, n = 6, n= 5, and n = 4; the genus is scored with the z = 9 taxa, based on evidence presented by Smith (1966), who showed that the lower numbers within Crop- tilon are derived sequentially from the higher. Although no populations with n = 8 or n = 9 have been discovered, it is probable and assumed here that n = 7 in Croptilon has also been derived from a base of z = 9. 2. Noticastrum comprises both rhizomatous and taprooted species and it has been scored ambiguously. All species of Heterotheca are taprooted, ex- cept the rhizomatous H. chthuahuana (Turner & Sundberg) B. Turner and C. mucronata Harms ez Turner, which apparently are derived from taprooted ancestors. Plants of a previously undescribed species of Jonactis produce a tap- root, but this is also apparently evolutionarily derived from the short, woody rhizomes of the other species (Nesom & Leary submitted). 3 and 4. Pityopsis and Croptilon stand alone among the other genera in their linear or narrowly oblong leaves with acute apices. The major foliar veins of Pityopsis are strictly parallel; the veins of Croptilon show a strong tendency to run parallel, usually with a single pair of laterals continuous from base to apex. And in contrast to the other genera, the foliar veins of both Pityopsis and Croptilon have large masses of associated sclerenchyma. Another distinctive feature of Pityopsis illustrated by Semple et al. (1980), the “sunken” abaxial epidermis, is present and even more pronounced in Croptilon, where it could be described as “lacunate.” 5-9. Trichome terminology follows the designations in Nesom (1976), based on a study by Drury & Watson (1966). Type A trichomes have also referred to as “osteolate” (Semple et al. 1980) or as “helianthoid;” these are uniseriate and have thick, often roughened walls and the bases commonly are multicellu- lar. Type B trichomes are also uniseriate but have thin, single celled bases and much thinner walls; in most taxa they are much smaller than the Type A tri- chomes, although in some of the goldenaster genera, they have supplanted the Type A trichomes and are the largest and most prominent. Type C trichomes are biseriate, usually with thin walled cells, and they typically are glandular in appearance. The simplest form has no head, with apical cells similar to the stalk cells - these are rarely found on herbage in the goldenaster group but elongated ones typically occur on the tubes of both ray and disc corollas. In the goldenasters, the apical cells are usually numerous and massed into a distinct, glandular head (3-8 cells wide at the apex) at the end of a stalk, Nesom: Phylogenetic hypothesis for goldenasters 147 which is often long but variable in length among species. In these capitate tri- chomes, a droplet of transparent, yellow-orange resin may be formed at the tip of the head, and a membrane sometimes seems to enclose the droplet. Stages in the ontogeny of Type C trichomes have been documented and illustrated by Carlquist (1961). Within Noticastrum, N. hatschbachi Zardini produces large glands; in the other species, Type C trichomes are present but unelaborated from the primi- tive biseriate, noncapitate form. Similar variation is present within Heterotheca sect. Phyllotheca. Various species within Croptilon, Pityopsis, and Chrysop- sis usually lack Type C trichomes, but this is clearly a derived condition. In Heterotheca sect. Ammodia, there appear to be two distinct size classes of Type C trichomes, undifferentiated ones and large headed ones on thick stalks, although it is possible that these may simply represent ontogenetic stages. In the tribe Astereae, Type A, B, and C or modifications of these appear to be the only kinds of trichomes present on the herbage. Homologies among trichome types can be established by examining the array of types present on a single plant. If only one of the uniseriate types is present, wall thickness and the nature of the trichome base are usually sufficient to distinguish it. The achene surfaces typically produce a second type of biseriate hair, the Zwillingshaare, which are ubiquitous throughout the tribe. Zwillingshaare rarely are produced on corollas, but they are primarily restricted to the achenes and have not been given an equivalent “Type” designation. ; Type A trichomes are present in most species of all genera involved in the present analysis except Pityopsis and Tomentaurum, where they are com- pletely absent. They are greatly reduced in number in Croptilon, where they are found primarily as spreading cilia along the petiolar margins. In Osberiza, Chrysopsis, and Noticastrum, Type A trichomes have thin, smooth, and vit- reous walls, and they are commonly flattened and longitudinally twisted. In many taxa of Chrysopsis and Noticastrum they are distally elongated into fine, flexuous, flagelliform hairs. In the remainder of the goldenaster genera, the Type A trichomes are terete and have thicker walls. Semple et al. (1980) have shown that such trichomes in Heterotheca (including sects. Heterotheca, Phyl- lotheca, and Ammodia) have minutely papillate cell surfaces. Such surfaces apparently do not occur in other taxa of the goldenaster lineage, judging from studies with the compound microscope. Heterotheca inuloides Cass. produces vitreous trichomes similar to those in Chrysopsis, but they still have slightly papillate surfaces and certainly have developed their characteristics indepen- dently of those in the Chrysopsis lineage. In plants of Heterotheca the Type A trichomes often appear to be highly variable in length, even perhaps grading into Type B trichomes. Comparative data in greater detail (SEM) would be helpful for these taxa. Type B trichomes are present in the goldenasters in their primitive form as thin walled, uniseriate hairs at least an order of magnitude smaller than 148 PHYTOLOGIA volume 71(2):136-151 August 1991 the Type A trichomes, except in the Pityopsis lineage, where they appear in several modifications. In Pityopsis and Tomentaurum, they are minutely and evenly filiform, forming a dense, close tomentum; they are parallel and sometimes anastamosing in Pityopsis but irregularly oriented and matted in Tomentaurum. In Croptilon, conspicuously lengthened Type B trichomes are apparent only on the abaxial surface of some species, where they originate in the “lacunae.” On some plants of C. divaricatum (Nutt.) Rafin., these are dense enough to form a thin but conspicuous tomentum over the leaf surface. Some taxa of Croptilon and Pityopsis are glabrous or nearly so, lacking all trichomes. 10. Plants of Tomentaurum produce strictly monocephalous stems; those of Osbertia are mostly monocephalous, but the primary stems of O. bartletta (S.F. Blake) Nesom produce a few lateral branches. Most species of Noticastrum are monocephalous, and Heterotheca chihuahuana is the only monocephalous species of its genus. It seems clear that monocephaly is a derived feature in the goldenasters, having arisen independently in several lineages. 11. The heads of Croptilon and Pityopsis are distinctly smaller than those of the other genera. Further, they are longer than wide, appearing cylindric, in contrast to those in the rest of the goldenasters, which are about as long as wide. The heads of Chrysopsis terana (= Bradburia) are exceptionally small, but this certainly has resulted from the sterility of the disc flowers. Among. the species of Jonactis, the heads of J. cieloatica are also small (Nesom & Leary submitted) and are reduced independently of the small ones among the goldenasters. 12. Keeled phyllaries are one of the conspicuous features of the goldenaster group, although they are also characteristic of related genera (Nesom in prep.). At least on the basal portion of the phyllary, the medial nerve is indurated, not at all resinous, and sharply and narrowly raised. The medial nerve in phyllaries of Croptilon, and in some Pityopsis, is sometimes resinous from base to tip, lacking an indurated covering, but some taxa have a clearly discernible keel with a morphology that is similar to the other genera. The orange nerves in Osbertia also are sometimes exposed, and among the goldenasters, only its two species have completely herbaceous phyllaries without a distinct keel, although they are approached by some in Chrysopsis. The medial nerve in the phyllaries of Noticastrum is keeled, but at least in the inner phyllaries there are usually an additional 1 or 2 pairs of smaller, unraised nerves conspicuously present on either side of the medial one. 13 and 14. In a phylogenetic overview of Aster and related groups (Nesom in prep.), the blue rayed genus Jonactis is the sister group of the yellow rayed genera of goldenasters. Ray color variation in Noticastrum, where white rays apparently are derived from yellow ones, is discussed in more detail in another paper (Nesom 1991c). The primarily yellow rayed Machaeranthera alliance also appears to be Nesom: Phylogenetic hypothesis for goldenasters 149 closely related the Aster group, based on molecular data of Suh (1989) and Morgan (1990). Only two genera within it have white or blue rayed taxa: Machaeranthera (sensu Hartman 1990), which comprises both yellow and white or blue rayed species, and Xylorhiza, where all taxa are white or blue rayed. Morgan’s molecular data show Xylorhiza to be the most primitive genus in the Machaeranthera lineage, its position there analogous to that hypothesized for Ionactis in the goldenaster lineage. Both yellow and white rayed species are hypothesized to occur within the genus Tonestus A. Nels. (Nesom 1991d) as well as among genera related to Gutierrezia Lag. (Nesom in prep.). 15. The corollas of postreceptive ray flowers in most genera of the golden- aster lineage coil tightly inward, with the adaxial surface exposed. However, they remain straight in Noticastrum and Osbertia, as well as in two species of Chrysopsis, where that behavior is hypothesized to have arisen in parallel with the former two genera (Nesom 1991c). Coiling ray corollas occur in most of North American Astereae, including the Aster and the Machaeranthera lineages. Ray corollas in Erigeron L., however, display a number of different behaviors, although for the most part the behavior appears to be consistent among species of a single section (Nesom 1989). Further aspects of variation in ligule behavior in the family are noted by Stirton (1983). 16. The difference between sinuate and straight walls can be easily seen in the cells of the upper throat of the disc corollas. Sinuate walls are particularly characteristic of the disc corolla throats in some groups of Aster as well. as related genera but not in the goldenasters. 17. Relatively large, elongate, straight sided crystals are found in the throat cells of all species of species of goldenaster except Chrysopsis terana. Straight sided crystals much reduced in size occur in C. pilosa. The occurrence of these crystals appears to be one of the most significant diagnostic features of the goldenaster lineage, and their absence or reduction in two taxa clearly within the lineage on the basis of other characters is interpreted as apomorphic. The disc corolla tubes usually produce stellate “sand” crystals, and the transition between the large crystals of the throat and the much smaller “sand” crystals of the tube is abrupt. The disc corollas of C. terana produce only “sand” crystals. Both species of Osbertia produce only straight sided crystals in the disc corollas, from the base of the tube into the throat. In a survey of disc corolla morphology of other genera of American Aster- eae, crystals similar to those in the goldenasters have been observed (Nesom et al. submitted) only in the genus Xylorhiza (8 species examined), Grindelia (8 species examined), Prionopsis Nutt., and Hazardia E. Greene (4 species examined), where they are highly reduced in size. The following genera have been examined and found to produce only “sand” crystals (number of species studied in parenthesis, if more than one): Aphanostephus DC., Aster sensu lato (18), Baccharis L., Boltonia L’Herit., Chaetopappa DC., Chloracantha Nesom et al., Corethrogyne DC., Ericameria Nutt., Erigeron (11), Euthamza Nutt., 150 PH YGT Ou. OsG A volume 71(2):136-151 August 1991 Gutierrezia, Gymnosperma Less., Haplopappus DC. (3 South American), Hys- teritonica Willd., Jsocoma Nutt., Machaeranthera(8), Monoptilon Torr. & Gray, Olivaea Sch.-Bip. ez Benth., Oonopsis E. Greene (2), Oritrophium (Kunth) Cabr., Petradoria E. Greene, Pyrrocoma Hook., Solidago L. (3), Stenotus Nutt., Stephanodoria E. Greene, Tonestus A. Nels. (6), Townsendia Hook., Vanclevea E. Greene, Xanthisma DC., Xanthocephalum Willd. (5), and Xy- lothamia Nesom et al. In these genera, the crystals may be restricted to the corolla tube or they may extend into the throat. 18. The disc corolla lobes are either glabrous, invested with Type C tri- chomes noncapitate or only slightly capitate, or else they have uniseriate, acicular hairs clearly homologous with Type A trichomes found elsewhere on the plant. The occurrence of the latter on disc corolla lobes is rare in the Astereae (Nesom pers. obs.), and this feature probably has had a single origin in the goldenaster lineage, with losses in a number of the taxa; the distribu- tion of character states could with equal parsimony be regarded as a number of parallel gains. In Pityopsis, only P. falcata (Pursh) Nutt. and P. ruthw (Small) Small have disc corolla lobes with Type A trichomes. Lobes of the other species of Pityopsis usually have Type C trichomes or else they are glabrous, although some individuals of P. microcephala (Small) Semple produce Zwillingshaare on the disc corolla lobes. Pityopsis falcata and P. ruthit are among the four species hypothesized by Semple & Bowers (1985), apparently based on their geographic distribution, to be the most primitive in the genus. The genus is scored here as primitively possessing Type A trichomes on the disc corol- las. Similar variation occurs in Noticastrum, where only N. hatschbachu, N. acuminatum (DC.) Cuatr., and N. macrocephalum (Baker) Cuatr. have lobes hairy with Type A trichomes. 19. The vascular traces, usually of 2-4 veins extending to nearly the tip of the style appendages, are densely sclerified’in all of the goldenaster genera. Sclerification of such traces is variable in other genera of Astereae (Nesom pers. obs.). 20-24. Possible homologies among the different achene morphologies are among the most difficult to interpret. The achenes of Noticastrum, Heterotheca sect. Ammodia, Tomentaurum, Pityopsis, Croptilon, and Osbertia produce 6-26 thin, whitish, superficial and slightly raised, evenly spaced nerves; C. rigidifolium (Smith) Smith has four resinous, larger nerves interspersed be- tween smaller ones. In cross section, the achenes are fusiform and terete to slightly compressed (in Pityopsis, Osbertia, and Croptilon), but among the goldenasters compressed achenes appear to be primitive. In Heterotheca sect. Ammodia and Tomentaurum the achenes are very strongly compressed; in Noticastrum they are plump but distinctly compressed. The ray achenes of Heterotheca sect. Heterotheca are 3 angled, most in sect. Phyllotheca are 2 angled, but those of H. stenophylla are slightly but distinctly 3 angled. Nesom: Phylogenetic hypothesis for goldenasters 151 In Heterotheca sects. Heterotheca and Phyllotheca, one to several of the fa- cial nerves of the disc achenes are resinous and clearly evident but not raised above the achenial surface. No other achenes of the group are like these, al- though three specialized species of Chrysopsis and one of Croptilon produce thick, distinctly raised, resinous veins. The achenes of most Chrysopsis (includ- ing Bradburia) are distinctive in their walls with broad and rounded ridges, the nerves completely below the surface. This feature separates Chrysopsis from Osbertia and Noticastrum, but it is not found in any of the other goldenaster genera. The achenes of H. oregona, with numerous superficial nerves, are most like those of Tomentaurum, although none of the goldenaster achenes have as many nerves (10-14) as in H. oregona. The 2-3(-6) nerved achenes of Jonactis are probably themselves evolutionarily derived from immediate ancestors with a greater number of achenial nerves, since these are common among the genera with keeled phyllaries and a base chromosome number of z = 9. Among the taxa of Heterotheca sect. Heterotheca, achene shape is rela- tively constant. In Heterotheca sect. Phyllotheca, the achenes vary from long and nearly straight sided, much like H. oregona, to shorter and obovate as in sect. Heterotheca. The achenes of Noticastrum, while somewhat obovate, are relatively long. 25. In the goldenaster lineage, the pappus typically comprises numerous, basally terete, antrorsely barbellate bristles and an additional, very short, “outer” series of setae, very slender bristles, or linear scales. In Osbertia and Croptilon there is no outer pappus and the bristles are somewhat reduced in number; in some species of Chrysopsis as well as in Heterotheca sect. Ammodia, the outer pappus is represented only by a few inconspicuous setae. Both Chrysopsis and sect. Ammodia have been scored as having an outer pappus, and in the phylogenetic hypothesis presented here, its apparent total absence in the other two genera is interpreted as apomorphic. A double pappus also occurs in Jonactis and apparently is primitive for the goldenasters as well as related groups (Nesom in prep.). 26. The pappus bristles in most of the goldenasters are terete from base to tip, with the divergent, “ciliate” hairs emerging spirally from all around the bristle. In contrast, the bristles of Heterotheca can be seen to be flattened, with the cilia arranged in two lateral rows, even though they are very slender. In some cases, this is most apparent immediately above the bristle base. There is variation among the species of Jonactis, but it has been scored as having terete bristles (see comments in Nesom & Leary submitted). Phytologia (August 1991) 71(2):152-159. TWO NEW SPECIES OF ARCHIBACCHARIS (ASTERACEAE: ASTEREAE) FROM MEXICO WITH A REEVALUATION OF SECTIONAL GROUPINGS IN THE GENUS Guy L. Nesom Department of Botany, University of Texas, Austin, Texas 78713 U.S.A. ABSTRACT Six sections are recognized within the genus Archibaccharis, com- pared to the two previously recognized by Jackson. The four newly described sections, which are primarily segregated from sect. Archibac- charts sensu Jackson, are sect. Tomentosa, sect. Stricta, sect. Glan- dulicarpa, and sect. Stylosa. In this modified view, the difference between “erect” and “scandent” habit is not as absolute as conceived by Jackson, since five of the six sections include species with scandent tendencies. The sectional placement of all 32 species of the genus is proposed, including two species that are described as new. Archibac- charis tuxtlensis sp. nov. (sect. Glandulicarpa) is described from the peak of Volcan San Martin Tuxtla in Veracruz, México. Archibaccha- ris vesticaulis sp. nov. (sect. Stylosa) from southern Chiapas is appar- ently most closely related to A. salmeoides (S.F. Blake) S.F. Blake. KEY WORDS: Archibaccharis, Asteraceae, Astereae, México, tax- onomy I. Sectional classification of Archibaccharis The taxonomic revision of the genus Archibaccharis Heering by J. D. Jack- son (1974, 1975) provided a detailed and coherent taxonomic interpretation of the genus, but since his study there have been changes in the nomenclature and delimitation of numerous taxa (Nesom 1988a), and a number of additional species have been described (Turner 1984; Sundberg 1984; Nesom 1988b, 1989, 1990). Jackson divided the genus into two sections, based on the habit of the plants (erect - sect. Archibaccharis vs. scandent - sect. Hirtella J.D. Jackson). In a reevaluation of the relationships of all the species, the present study di- vides sect. Archibaccharis sensu Jackson into four separate sections, three of 152 Nesom: New species and sections in Archibaccharis 153 them previously undescribed, and two species of sect. Hirtella are transferred to a fourth new section, where they join a third species described in the present paper. Archibaccharis Heering, Jahr. Hamb. Wissensch. Anst. 21, Beiheft 3:40. 1904. Type species: A. hieracitfoha Heering (= A. auriculata [Hemsl.] Nesom) 1. Sect. Archibaccharis Type species: A. auriculata (Hemsl.) Nesom, Phytologia 65:123. 1988. Plants erect to subscandent, 0.5-2.0(-3.0) m tall, stems greenish, herba- ceous, straight to weakly zig-zag, stems and leaves stipitate glandular, eglan- dular in some, leaves slightly thickened but not coriaceous, capitulescences corymboid and terminal (rounded paniculate and axillary in Archibaccharis blakeana Stand]. & Steyerm. and A. pringlet [Greenm.| S.F. Blake), pistillate heads 6-10(-15) mm high (4-6 mm in A. pringlei and A. blakeana), on rela- tively long pedicels, staminate flower styles with linear collecting appendages (0.6-)0.8-1.2 mm long, achenes strigose to glabrous, eglandular. Species included: Archibaccharis auriculata, A. campit S.F. Blake, A. hi- eracioides (S.F. Blake) S.F. Blake, A. macdonaldii Nesom, A. simplez (S.F. Blake) S.F. Blake, A. blakeana, and A. pringlez. Only the first two species listed have all of the features noted above, but the section appears to be monophyletic. The first four species produce a stipitate glandular vestiture, and the first five have larger heads than any others in the genus. Archibaccharis simplez is distinct in its glabrate, nonglandular vestiture but has an erect habit, sessile to slightly clasping leaves, and produces a corymboid capitulescence of large heads on long pedicels. Archibaccharis hieracioides and A. macdonaldii have petiolate leaves, but the petioles are broad, and in the latter, they have a flaring, foliar base. Jackson (1975) placed A. blakeana and A. campziin sect. Hirtella because of their subscandent tendencies (he characterized them as “weak-stemmed scramblers”), fractiflex stems, but in their auriculate leaves and linear collecting appendages, they appear to belong in sect. Archibaccharis. 2. Sect. Glandulicarpa Nesom, sect. nov. Type species: Archibaccharis standley: S.F. Blake, J. Washington Acad. Sci. 19:271. 1929. Herbae erectae vel subscandentes, folia plerumque coriacea, ap- pendices styli lineares 0.8-1.0 mm long, achenia glandulosa. Plants 0.4-3.0 m tall, erect but usually with marked subscandent tenden- cies (sprawling, leaning, or climbing), stems greenish, herbaceous, straight to weakly zig-zag, glabrous or glabrate (except for Archibaccharis corymbosa), leaves thick, lanceolate to ovate, glabrous to glabrate with a shiny upper sur- face, capitulescence corymboid to rounded or convex paniculate, terminal or 154 PHY TE.0O L/0.GI A volume 71(2):152-159 August 1991 terminal and axillary, pistillate heads 4-7(-8) mm high, on relatively short pedicels, staminate flower styles with linear to linear-lanceolate collecting ap- pendages 0.8-1.0 mm long, achenes glandular and usually strigose as well. Species included: Archibaccharis standleyi (including A. aequivenia [S.F. Blake] D. Nash), A. nicaraguensis Nesom, A. venturana Nesom, A. tuztlensis Nesom, A. veracruzana Nesom, A. subsessilis S.F. Blake, A. linearilobis J.D. Jackson, and A. corymbosa (J.D. Smith) S.F. Blake. Archibaccharis subsessilis, A. linearilobis, and A. corymbosa produce ovate leaves with truncate to cordate bases and relatively long ligules on the pistillate flowers and appear to form a closely related group; the leaves of the first two are epetiolate or nearly so. Archibaccharis veracruzana, A. venturana, and A. tuztlensis have ovate leaves with tapering bases, while the rest have lanceolate leaves. 3. Sect. Stricta Nesom, sect. nov. Type species: Archibaccharis caloneura S.F. Blake, Proc. Biol. Soc. Washington 55:117. 1942. Caules stricti lignescentes porphyreisque et appendices styli lin- eares 0.8-1.0 mm longi. Plants 0.6-3.0(-6.5) m tall, strictly erect, stems usually reddish brown, lig- nescent, straight, glabrous or glabrate, leaves thickened to coriaceous, lanceo- late (or ovate to ovate-lanceolate in Archibaccharis asperifolia), capitulescence corymboid or broadly paniculate, terminal, rarely also from the upper ax- ils, pistillate heads (3.5-)4.0-7.5(-9.0) mm high, on relatively short pedicels, staminate flower style appendages linear to linear-lanceolate, 0.8-1.0 mm long, achenes sparsely strigose, eglandular. Species included: Archibaccharis caloneura, A. androgyna (Brandeg.) S.F. Blake, A. panamensis S.F. Blake, A. irazuensis (S.F. Blake) S.F. Blake, A. jacksoni Sundberg, and A. aspertfolia (Benth.) S.F. Blake (including A. ses- centiceps [S.F. Blake] S.F. Blake). Archibaccharis caloneura and A. androgyna have coriaceous leaves and ap- pear to be closely related. Archibaccharis irazuensis, A. panamensis, and A. jacksoniz have thinner leaves, and as pointed out by Sundberg (1984), the last two have more achenial nerves than any other species of the genus. Archibac- charis asperifolia is unusual in its broader leaves and the strongly developed indument on its leaves, but in diagnostic features it belongs with sect. Stricta. 4. Sect. Tomentosa Nesom, sect. nov. Type species: Archibaccharis serrat- ifolia (Kunth) S.F. Blake, Contr. U. S. Natl. Herb. 26:236. 1930. Herbae erectae vel subscandentes, caules ac folia tomentosa vel villosa, appendices styli deltati vel triangulares 0.2-0.5 mm long. Nesom: New species and sections in Archibaccharis 155 Plants 0.5-3.0 m tall, erect to leaning or sprawling, stems greenish, herba- ceous, straight to weakly zig-zag, stems and leaves sparsely to densely tomen- tose or villous, the leaves thickened, lanceolate-ovate, capitulescence corym- boid or rounded paniculate, terminal, sometimes with axillary branches, pistil- late heads most 2.0-5.0 mm high, on relatively short pedicels, staminate flower styles with deltate to triangular appendages 0.2-0.5 mm long, achenes sparsely strigose, eglandular. Species included: Archibaccharis serratifolia, A. nephocephala Nesom, and A. peninsularis §.F. Blake. These three species are similar in their primarily erect habit, vestiture, and style branch morphology. Archibaccharis corymbosa of sect. Glandulicarpa also produces tomentose vestiture, but the similarity clearly is parallel. The only other species in which the style collecting appendage are so short are those of sect. Hirtella, and the two groups may ultimately prove to be closely related, despite their disparity in habit. 5. Sect. Stylosa Nesom, sect. nov. Type species: Archibaccharis lucentifolia L.O. Wms., Fieldiana, Bot. 29:388. 1962. Herbae scandentes vel subscandentes, appendices styli ovati- deltati 0.7-0.8 mm long. Plants subscandent herbs (up to 5 m tall) or true vines, stems greenish, herbaceous, straight to weakly zig-zag, stems and leaves glabrous to glabrate, the leaves coriaceous to slightly thickened, ovate, capitulescence rounded pan- iculate, terminal and axillary, pistillate heads 3.0-8.0 mm high, on relatively short pedicels, staminate flower styles with ovate-deltate appendages 0.7-0.8 mm long, achenes sparsely strigose, eglandular. Species included: Archibaccharis lucentifolia, A. salmeoides (S.F. Blake) S.F. Blake, and A. vesticaulis Nesom. These species are separated primarily on the basis of their very thick stami- nate style branch appendages with thick, spreading, sweeping hairs. No other species of the genus are similar. The first two species are strongly scandent, while Archibaccharis vesticaulis is more shrublike. All have thick leaves, and their relationship may prove to be with sect. Glandulicarpa, where their posi- tion would be analogous to the two of sect. Archibaccharis that are hypothe- sized to have evolved a scandent habit from primarily erect ancestors. 6. Sect. Hirtella J.D. Jackson, Phytologia 32:158. 1975. Type species: Archibaccharis hirtella (DC.) Heering, Jahr. Hamb. Wissensch. Anst. 21, Beiheft 3:41. 1904. Scandent herbs, sometimes leaning, stems greenish, sharply zig-zag to sin- uous, stems and leaves hirsute to sparsely hirsutulous, the leaves thin and not 156 Pea YT OLsOVGH A volume 71(2):152-159 August 1991 shiny, ovate, capitulescence mostly rounded to convex panicles, terminal and axillary, heads 3.5-5.5(-7.0) mm high, on relatively short pedicels, of the stami- nate flower styles with minute, ovate-deltate collecting appendages 0.2-0.5 mm long, achenes strigose, eglandular. Species included: Archibaccharis hirtella*, A. intermedia* (S.F. Blake) B. Turner, A. albescens* (J.D. Jackson) Nesom, A. taeniotricha* (S.F. Blake) Nesom, A. flezilis (S.F. Blake) S.F. Blake, and A. schiedeana (Benth.) J.D. Jackson. From sect. Hirtella as delimited by Jackson (1975), Archibaccharis salmeor- des and A. lucentifolia have been transferred to sect. Stylosa, and A. pringlei and A. blakeana have been transferred to sect. Archibaccharis sensu stricto. Jackson (1975) pointed out the similarity in collecting appendages among the taxa he regarded as four varieties of Archibaccharis hirtella (marked with asterisk above). These are the core of sect. Hirtella and are here considered four separate species. Archibaccharis flezilis shares with A. taeniotricha and A. albescens a truly scandent habit and densely hispid stems with coarse, erect, long, sharp pointed trichomes. The collecting appendages of both A. flezilis and A. schiedeana, however, are much more linear than the core species, and the evolutionary affinities of both of these may ultimately prove to belong elsewhere. As noted by Jackson (1975), the plants of Archibaccharis schiedeana are initially decumbent or procumbent to sprawling or arching herbs, but at ma- turity the stems elongate greatly and they become true vines. Plants of A. hartella and its three closest relatives, however, as well as A. flezilis, have been collected only as vines. The stems of most of the scandent plants are zig-zag (“fractiflex” in the terminology of Jackson 1975) and produce a terminal capit- ulescence as well as similar sized axillary ones at least on the upper portions, while those of erect plants are usually relatively straight and produce only a single, terminal capitulescence. The stems of many of the erect species in other sections with scandent tendencies, however, tend to be slightly zig-zag, and the capitulescences are often axillary as well as terminal. Further, the stems even of the scandent species are variable in orientation. For example, the stems of A. schiedeana are usually nearly straight, and A. flezilis has sinuous rather than zig-zag stems. In any case, the definite and consistent tendency for plants of almost all species of Archibaccharis to become at least somewhat vinelike argues for the monophyletic status of the genus. It is assumed, however, that primarily erect stems, rather than true vines, are the primitive condition for the genus. II. Two new species Continued study of Archibaccharis has revealed the existence of two pre- viously undescribed species. The first apparently is restricted to Veracruz, Nesom: New species and sections in Archibaccharis 157 México, where it is known only from Volcan San Martin Tuxtla. Archibaccharis tuxtlensis Nesom, sp. nov. TYPE: MEXICO. Veracruz: Mpio. San Andres Tuxtla, near summit (upper 300 ft) of Volcan San Martin, ca. 6000 ft, 28 Dec 1964, A.C. Faberge s.n. (HOLOTYPE: TEX!). Differt a A. venturana Nesom phyllariis interioribus longioribus, corollis staminalibus longioribus, et acheniis glandes non nisi fer- entibus. Subscandent shrubs ca. 0.5 m tall, stems noticeably but not strongly zig- zag, eglandular, sparsely to moderately invested with thick, vitreous hairs with conspicuous brownish crosswalls. Leaves thick, the upper surface shiny, with a raised reticulum of veins, mostly glabrous except along the midvein, lower sur- face with a duller texture, blades ovate with acuminate apices and rounded to obtuse bases, 5-10 cm long, 2-4 cm wide, on petioles 5-8 mm long, the margins minutely and widely mucronulate with 2-9 pairs of mucros. Capitulescences axillary and terminal on the upper portion of the stems, in rounded, ebracteate panicles. Staminate heads not seen. Pistillate heads with 7-12 outer, pistillate flowers and 1 central, staminate flower; phyllaries lanceolate, greenish, with fringed- ciliate distal margins, otherwise glabrous, in 3-4 graduated series, the inner 4.5-5.5 mm long, the outer 1/5 as long. Pistillate flowers fertile, the corollas eligulate or with an extension 0.2 mm long, the tube 2.5-3.5 mm long, with thick, viscid trichomes, the style 2.5-3.5 mm long, with branches 1 mm long; achenes 1.2-1.5 mm long, gland dotted and viscid, without other hairs; pappus of numerous bristles. Staminate flowers with sterile ovaries, the corol- las 3.5-4.0 mm long, the lobes purplish, lanceolate, 1.5 mm long, cut 2/3 to the tube, tube densely hairy with viscid trichomes. Additional collection examined: MEXICO. Veracruz: Mpio. San Andres Tuxtla, top of Volcan San Martin Tuxtla, ca. 1730 m, 14 Feb 1972, Beaman & Castillo 5686 (TEX). Archibaccharis tuztlensis is most similar in habit, leaf morphology, and fruit morphology to A. venturanaand A. subsessilis of sect. Glandulicarpa. The new species is distinguished from its closest relatives by the following contrasts. 1. Leaves gland dotted above and beneath, not distinctly thick or shiny, basally rounded to slightly cordate, sessile or on petioles 1 mm long; heads mostly in loose, broad, terminal corymbs; pistillate corollas ligu- late; Guatemala, Chiapas, and Oaxaca. ............-..-- A. subsessilis 1. Leaves eglandular, thick, the upper surface shiny, basally obtuse to acute, on petioles 5-8 mm long; heads in relatively dense, terminal and ax- illary, corymboid panicles; pistillate corollas eligulate or essentially so; ISG CIT sp ene opis, Otisl oycho iviues sala ge ape a aittens ciate.» cgay ha ais hodeaage = (2) 158 PHYTOLOGIA volume 71(2):152-159 August 1991 2. Inner phyllaries 2.0-3.0 mm long; pistillate heads with 2 staminate flowers, the corollas 2.0-2.8 mm long; achenes both sparsely strigose and glandular; vicinity of Xalapa to Perote. ........ A. venturana 2. Inner phyllaries 4.5-5.5 mm long; pistillate heads with 1 stami- nate flower, the corolla 3.5-4.0 mm long; achenes glandular, without other vestiture; Volcan San Martin Tuxtla. ......... A. tuztlensis Archibaccharis vesticaulis Nesom, sp. nov. TYPE: MEXICO. Chiapas: Mpio. Las Margaritas, E of Comitan Dominguez, 2.5 mi E of Ejido Tzis- cao turnoff, along a short spur road toward lake, disturbed ground, 7 Jan 1984, S. Sundberg 2423 (HOLOTYPE: TEX!). Differt a A. salmeoide (S.F. Blake) S.F. Blake habitu subscan- denti, foliis glandulosis, caulibus dense pubescentibus, capitulis brevioribus, et corollis staminalibus brevioribus. Subscandent shrubs 5 m tall, stems slightly zig-zag, invested with thick, reddish brown trichomes completely obscuring the stem surface, eglandular. Leaves coriaceous, dark green, the upper surface shiny, with a slightly raised reticulum of veins, sparsely and minutely puberulous along the veins, eglan- dular, lower surface minutely but definitely gland-dotted, the blades elliptic- obovate with short attenuate apices and acute bases, 3-9 cm long, 2-4 cm wide, on petioles 5-15 mm long minutely but densely viscid puberulous, the mar- gins entire or usually mucronulate with 2-7 pairs of mucros. Capitulescence of rounded, ebracteate panicles, terminal and on short axillary branches. Stami- nate heads not seen. Pistillate heads with 11-12 outer, pistillate flowers and 1 central, staminate flower; phyllaries narrowly oblong-lanceolate, greenish, with fringed ciliate distal margins, otherwise glabrous, in 3-4 graduated series, the inner 3.0-4.0 mm long, the outer 1/5 as long. Pistillate flowers fertile, the corollas tubular-filiform, 2.0-2.4 mm long, with a purple apex, eligulate; achenes 1.2-1.4 mm long, flattened, with 3-4 thick nerves, sparsely strigose, eg- landular; pappus of numerous bristles. Staminate flowers with sterile ovaries, the corollas 3.0-3.2 mm long, the tube 1.8-2.0 mm long, densely hairy with viscid trichomes, the lobes purple, lanceolate, 1.2 mm long, cut nearly to the tube, style branches 0.8 mm long, with broad, elliptic-ovate appendages 0.5 mm long. Additional collection examined: MEXICO. Chiapas: Mt. Pasitar, Jan 1937, Matuda 1544 (MEXU). Jackson (1975) cited a MICH duplicate of Matuda 1544 as a specimen of Archibaccharis schiedeana, but the MEXU specimen is clearly the same as the type. Archibaccharis schiedeana has neither coriaceous leaves nor densely invested stems, and its style branch morphology is very different from that of A. vesticaulis. The new species appears to be most closely related to A. Nesom: New species and sections in Archibaccharis 159 salmeoides and A. lucentifolia and is placed with them in sect. Stylosa. It is distinguished from A. salmeozdes by the following contrasts. 1. Stems densely invested with reddish brown trichomes, obscuring the sur- face; lower surface of leaves sessile-glandular; pistillate heads 3.0-4.0 mm high; staminate corollas 3.0-3.2 mm long. ............... A. vesticaulis 1. Stems sparsely puberulous with relatively thin trichomes; leaves eglan- dular; pistillate heads 4.0-6.4 mm high; staminate corollas 3.9-4.4 mm NON F ois ora odo tides xisvessieas nent ai Aa Sete sence mim ows 9 wma A. salmeoides ACKNOWLEDGMENTS I thank Dr. B.L. Turner and Dr. A. McDonald for their review and com- ments. LITERATURE CITED Jackson, J.D. 1974. Notes on Archibaccharis (Compositae - Astereae). Phy- tologia 28:296-302. . 1975. A revision of the genus Archibaccharis Heering (Compositae - Astereae). Phytologia 32:81-194. Nesom, G.L. 1988a. Studies in Mexican Archibaccharis (Compositae: Aster- eae). Phytologia 65:122-128. . 1988b. A new species of Archibaccharis (Compositae: Astereae) from Central America. Phytologia 65:158-159. 1989. A new species of Mexican Archibaccharis (Compositae: Astereae). Phytologia 67:377-378. . 1990. Two new species of Archibaccharis (Compositae: Astereae) from Oaxaca, México. Phytologia 68:122-124. Sundberg, S.D. 1984. Archibaccharis jacksonii (Compositae: Astereae): a new species from Costa Rica. Syst. Bot. 9:295-296. Turner, B.L. 1984. Archibaccharis intermedia (Blake) Turner, comb. nov. Phytologia 56:377-379. Phytologia (August 1991) 71(2):160-162. SARCOSTEMMA CLAUSUM, SERIES CLAUSA (ASCLEPIADACEAE), NEW TO TEXAS 'Gretchen D. Jones & ?Stanley D. Jones ‘Department of Biology, Texas A&M University, College Station, Texas 77843-3258 U.S.A. ?§.M. Tracy Herbarium, Department of Rangeland Ecology and Management, Texas A&M University, College Station, Texas 77843-2126 U.S.A. ABSTRACT Sarcostemma clausum (Jacq.) Roemer & Schultes, subgenus Cera- manthus, series Clausa (Asclepiadaceae) previously unreported in Texas has been found in Hidalgo County. A key is provided for the Texas species of Sarcostemma. KEY WORDS: Sarcostemma, Sarcostemma clausum, subgenus Ceramanthus, series Clausa, Asclepiadaceae, Texas Sarcostemma clausum (Jacq.) Roemer & Schultes, subgenus Ceramanthus Kunze, series Clausa Holm (Asclepiadaceae) was first described by Jacquin (1763) as Cynanchum clausum. Roemer & Schultes (1820) moved it to Sar- costemma. Sarcostemma clausum is a southern species, reported from south- ern Florida, México (including Baja California Sur), Central America, South America, and the Caribbean Islands (Holm 1950). Members of this genus are twining or trailing vines which climb by turning to the right. They have milky sap; conspicuous flowers with a 5 lobed calyx, corolla rotate to almost cam- panulate, 5 lobed, and 5 filaments forming a column. The fruits are follicles, fusiform or obclavate in shape. Cory & Parks (1937), Holm (1950), Correll & Johnston (1970), Gould (1975), nor Hatch et al. (1990) listed Sarcostemma clausum as occurring in Texas. Based on previously mapped distributions (Holm 1950), it appears that S. clausum has had a natural migration up the eastern coastline of México from the state of Tamaulipas to its present location in Hidalgo Co., Texas. There are now four (4) species of Sarcostemma with five (5) taxa represented in Texas. 160 Jones & Jones: Sarcostemma clausum, new to Texas 161 The following key; modified from Correll & Johnston (1970), will differentiate the Sarcostemma found in Texas. KEY TO TEXAS SARCOSTEMMA 1. Peduncles as thick as, or thicker than the adjacent internode. .......... SSR aiaiarsicio Maine iaiee Sarcostemma clausum (Jacq.) Roemer & Schultes 1’ Peduncles thinner than the adjacent internode. 2. Margins of leaves crisped, occasionally purple tinged, sepals mostly greater than 3 times as long as wide. .......... S. crispum Benth. 2' Margins of leaves not crisped, green, sepals mostly less than 3 times as long as wide. 3. Sepals 4-6 mm long, pubescent on both surfaces. ............ ee eit eas ei ais weaeosteve Bae teralG oss S. torreyi (A. Gray) Woodson 3’ Se ialk 2-3 mm long, pubescent on back side (dorsal) only. 4. Blades cordate at base, mostly less than 3 times as long as WEEE. 5 .eaenhese S. cynanchoides Dene. var. cynanchoides 4’ Blades hastate to rounded cuneate at base, greater than 3 unten aalongias! WAGES o iAicick Usk aopioete soccer anes Soh ACES ae S. cynanchoides var. hartwegii (Vail) Shinners Specimen collected: UNITED STATES. Texas: Hidalgo Co., west side of the westernmost resaca in Bentsen-Rio Grande Valley State Park, south of Mission, 27 Dec 1988, S. & G. Jones 2229 (TAES). It was frequent, climbing on Celtis along the wooded resaca adjacent to an open fallow field. Associated genera: Celtis, Saliz, Pennisetum, and Phragmites. The soils are of the Rio Grande-Matamoros association (RT) and are deep, moderately and slowly permeable, loamy soils of floodplains and low terraces. More specifically they are typically light brownish gray or grayish brown silt loam or have a silty clay surface layer. The geology of the site is of Alluvium formation (Qos) (Recent). During follow up trips in December 1989 and 1990, vines of Sarcostemma clausum were found, but in neither year did we find it in flower. ACKNOWLEDGMENTS We thank Warren Stevens (MO) for verifying our collection. We thank Robert Lonard (PAUH), Guy Nesom (TEX), and J.K. Wipff (TAES) for re- viewing this manuscript. We also thank Barney Lipscomb (SMU) for his tech- nical directions, and we thank the curators and staff at SMU and TEX for looking for specimens in their herbaria. 162 PH Yat OE .OxG TA volume 71(2):160-162 August 1991 LITERATURE CITED Correll, D.S. & M.C. Johnston. 1970. Manual of the Vascular Plants of Tezas. Texas Research Foundation, Renner, Texas. Cory, V.L. & H.B. Parks. 1937. Catalogue of the Flora of Tezas. Tex. Agric. Exp. Sta. Bull. 550. Gould, F.W. 1975. Tezas Plants-A Checklist and Ecological Summary. Tex. Agric. Exp. Sta. Bull. MP-585. Hatch, $.L., K.N. Gandhi, & L.E. Brown. 1990. Checklist of the Vascular Plants of Texas. Tex. Agric. Exp. Stat. Bull. MP-1655. Holm, R.W. 1950. American species of Sarcostemma. Ann. Missouri Bot. Gard. 37:480-590. Jacquin, N.J. 1763. Selectarum Stirpium Americanarum Historia. 1:87, t. 60. Facsimile ed. (1971): Hafner Publishing Co., Inc., New York, New York. Roemer, J.J. & J.A. Schultes. 1820. Systema Vegetabilium. 6:116. Phytologia (August 1991) 71(2):163-164. BOOKS RECEIVED Memoria II Simposio Latinoamericano de Briologia. Claudio Delgadillo M. (ed.). Instituto de Biologia, Universidad Nacional Autonoma de México, México, D.F. México. 1991. 48 pp. Price unknown. No ISBN. The book comprises papers from a symposium on bryophytes held in conjunction with the fifth Latin American botanical Congress which was held in Havana, Cuba in June 1990. Eight papers are included, authored by nine individuals. Most of the presentations deal with mosses or liverworts of Cuba. One paper treats Frul- laniaceae in Brasil, and another examines distribution patterns of neotropical mosses. Phytochemical Induction by Herbivores. Douglas W. Tallamy & Michael J. Raupp (eds.). John Wiley & Sons, Inc., 1 Wiley Dr., Somerset, New Jersey 08875-1272. 1991. 431 pp. xx. $98.00 (hardcover). ISBN 0471- 63241-4. ; A total of 33 contributors participated in writing this book. As stressed by the editors in their preface, the concept of phyto- chemical induction in response to herbivory is extremely difficult to study and quantify. Authors of papers in the present volume have examined a number of different plant taxa (aspen, birch, gourds, grasses, tobacco, etc.) subjected to several different types of actual or simulated herbivore stress. Plant responses have been measured by examination of a number of different plant products from alka- loids to phenolics. Effects of plant response on the herbivores was also examined. Most of the papers are of a review nature. The book as a whole provides a broad spectrum of activity in this field. 163 164 PA YT OO GEA volume 71(2):163-164 August 1991 Plant Genetic Resources of Ethiopia. J.M.M. Engels, J.G. Hawkes, & Melaku Worede (eds.). Cambridge University Press, 40 W. 20th St., New York, New York 10011. 1991. 383 pp. xv. $85.00 (hardcover). ISBN 0-521- 38456-7. The plant genetic resources of Ethiopia are described by a total of 29 contributors to thirty papers in this book. The contributors are participants in an Ethiopian Plant Genetic Resources Center, and are familiar with the genetic resources of Ethiopia and how they may relate to similar resources in other parts of the world. Four sections in the book give a general introduction, describe Ethiopia as a center of plant genetic diversity, describe germplasm collection and conservation efforts in Ethiopia, and describe meth- ods of evaluation and utilization of Ethiopian plant genetic re- sources. Vascular Plants of Minnesota A Checklist and Atlas. Gerald B. Ownbey & Thomas Morley. University of Minnesota Press, 2037 University Avenue SE, Minneapolis, Minnesota 55414. 1991. 307 pp. xi. $39.95 (cloth with dust jacket). ISBN 0-8166-1915-8. The introduction to this work provides a brief but useful his- tory of botanical activity in Minnesota, with references to previous floristic works. Also included in the preferatory material is a veg- etation map for the state. The bulk of the work consists of two parts. First, the checklist (by Ownbey), wherein each plant known to occur in the state is listed along with synonyms (those used in past treatments for the state). Families and genera are listed alphabetically within the major categories of Pteridophytes, Gym- nosperms, and Angiosperms. Each genus citation is accompanied by a reference to the maps in the atlas where distributions for the included species may be found, and for most, references to recent studies of the genus. The atlas (by Morley) follows the same order as the checklist. Localities for each species are reported on indi- vidual maps showing the outlines of counties in the state. Dots representing localities are placed so as to reasonably accurately re- flect the localities, not merely to indicate the presence of the plant in a given county. Thus, multiple localities may be indicated in a single county for any given species. ~ Pe eee +, ——_ Information for Authors 4 Articles from botanical systematics and ecology, including \ biographical sketches, critical reviews, and summaries of literature will _ ‘ be considered for publication in PHYTOLOGIA. Manuscripts may be } submitted either on computer diskette, or as typescript. Diskettes will be ? returned to authors after action has been taken on the manuscript. 7 Diskettes may be 5.25 inches or 3.5 inches and may be written in any IBM or MacIntosh compatible format. Typescript manuscripts should be ‘ single spaced and will be read into the computer using a page scanner. ‘ The scanner will read standard typewriter fonts but will not read dot x matrix print. Manuscripts submitted in dot matrix print cannot be ac- y cepted. Use underscore (not italics) for scientific names. Corrections made on typescript manuscripts must be complete and neat as the scan- 1 ner will not read them otherwise. Language of manuscripts may be either English or Spanish. Figures will be reduced to fit within limits of text pages and therefore, should be submitted with an internal scale and have dimensions proportional to those for text pages. Legends for figures should be included in figures whenever possible. Each manuscript should have an abstract and key word list. Specimen citations should be consistent throughout the manuscript. Serial titles should be cited with abbreviations used in Botanico Periodicum Huntianum. References cited only as part of nomenclatural summaries should not appear in Literature Cited. Nomenclatural work should include one paragraph per basionym and must provide proper (as defined by the current /nterna- tional Code of Botanical Nomenclature) citation of sources of epithets and combinations. Authors should arrange for two workers in the appropriate field to review the manuscript before submission. Copies of reviews should be forwarded to the editor with the manuscript. Manuscripts will not be — published without review. q Cost of publication is currently $13.00 US per page for publication without reprints. Publication with 100 reprints is provided for $18.00US =~ per page, 200 reprints for $21.50 US per page. Page charges are due with | manuscript and no paper will be published before payment is receivedin ~~ full. Reprints must be ordered and paid for in advance. Page charges will © be determined on the basis of a typescript page (single spaced, 10 points, =~ blank line between paragraphs) with all type inside a rectangle 143 mm} (horizontal) by 219 mm (vertical), not including running head and page number. ‘Title page should include title, author(s) name(s), and me address(es). Two blank lines should appear above and below section | headings (Abstract, Discussion, Literature Cited, etc.) in the manuscript. — y No extra charge is made for line drawings provided they conform to — 4 limitations of size and proportion for normal text. Halftones require an BY ; extra charge of $10.00 US per page.